Applied Evidence

Caring for a newborn can be a source of joy for family physicians (FPs). In this article, we examine care provided in the first month of life, including a thorough physical examination, safe hospital discharge procedures, assessment of neonatal feeding, evaluation of jaundice and fever, and prevention of sudden infant death syndrome (SIDS). In addition, we describe how FPs can support women of childbearing age between pregnancies, with the goal of reducing the risk of adverse outcomes in future pregnancies. (See “Your role in risk assessment and interventions during the interconception period.”)

Before parents leave the hospital with their newborn, it is essential that they receive written and verbal counseling on important issues in neonatal care. A discharge checklist can help make sure all topics have been covered.¹ A hearing screen and pulse oximetry before discharge are required for all newborns in most states, in addition to important preventive counseling for parents. TABLE 1² and TABLE 2² summarize important newborn physical exam findings and common skin conditions. Parents should be given additional written information regarding prevention of SIDS and proper use of car seats.

Hospital physicians should assess maternal medical and psychosocial readiness for discharge. Through shared decision-making with the newborn’s parents, physicians should create a plan for outpatient follow-up. Assessment through a physician home visit can provide safe and effective care similar to what is provided at a visit to an office medical practice.^3-7 A follow-up appointment should be made 2 to 5 days before discharge, preferably connecting the newborn to a medical home where comprehensive health care services are offered.^1,5,6,8

Age, gestational age, risk factors for hyperbilirubinemia, and the timing and level of bilirubin testing should be considered when establishing a follow-up interval. Most newborns who are discharged before 72 hours of age should have a follow-up visit in 2 days; a newborn who has a recognized risk factor for a health problem should be seen sooner. Newborns in the “low-risk zone” (ie, no recognized risk factors) should be seen based on age at discharge or need for breastfeeding support.⁹

Tracking baby’s weight, ensuring proper feeding

A newborn who is discharged at 24 hours of life, or sooner, should be seen in the office within 2 days of discharge to 1) ensure that he (she) is getting proper nutrition and 2) monitor his weight^1,3,5 (TABLE 3^10-13). All newborns should be seen again at 2 weeks of life, with additional visits more frequently if there are concerns about nutrition.¹

Recording an accurate weight is critical; the newborn should be weighed completely undressed and without a diaper. Healthy newborns can safely lose up to 10% of birth weight within the first week of life; they should be back to their birth weight by approximately 2 weeks of life.^10,11 A healthy newborn loses approximately 0.5 to 1 oz a day;¹¹ greater than 10% loss of birth weight should trigger a thorough medical work-up and feeding assessment.

Breastfeeding. For breastfeeding mothers, physicians should recommend on-demand feeding or a feeding at least every 2 or 3 hours. Adequate intake in breastfed infants can be intimidating for new parents to monitor, but they can use a written chart or any of several available smartphone applications to document length and timing of feeds and frequency of urination and bowel movements. By the fifth day of life, a newborn should be having at least 6 voids and 3 or 4 stools a day.^10-12

In addition, physicians can counsel parents on what to look for—in the mother and the newborn—to confirm that breastfeeding is successful, with adequate nutritional intake (TABLE 3^10-13). Physicians should recommend against providing a pacifier to breastfeeding infants during the first several weeks of life—or until breastfeeding is well established (usually at 3 or 4 weeks of age). The World Health Organization (WHO) recommends against providing bottles, pacifiers, and artificial nipples to breastfeeding newborns.¹⁴ Liquids other than colostrum or breast milk should not be given unless there is a documented medical need, such as inadequate weight gain or feeding difficulty.¹⁵ If the newborn experiences early latch difficulties, supplementation with expressed breast milk is preferable to supplementation with formula. Assistance from a trained lactation consultant is a key element in the support of the breastfeeding dyad.^11,12,16

Breastfeeding optimizes development of the newborn’s immune system, thus bolstering disease prevention; it also assists with maternal postpartum weight loss and psychological well-being. Exclusively or primarily formula-fed newborns are at increased risk of gastrointestinal, ear, and respiratory infections throughout infancy and childhood; type 1 diabetes mellitus; asthma; childhood and adult obesity; and leukemia.^17,18 Mothers who feed their newborn primarily formula increase their own risk of obesity, type 2 diabetes mellitus, ovarian and breast cancer, and depression.^17-22

Infant feeding is a personal and family choice but, in the absence of medical contraindications—such as maternal human immunodeficiency virus infection and galactosemia—exclusive breastfeeding should be recommended.^17,18 FPs are well suited to support the mother–infant breastfeeding dyad in the neonatal period, based on expert recommendations. Specifically, the American Academy of Family Physicians (AAFP) and American Academy of Pediatrics (AAP) recommend that all infants be exclusively breastfed for the first 6 months of life and continue some breastfeeding through the first year or longer.^17,18 WHO recommends breastfeeding until 24 months of age—longer if mother and infant want to, unless breastfeeding is contraindicated.^14,17,18

Physicians should provide up-to-date information to parents regarding the risks and benefits of feeding choices. Support for breastfeeding mothers postnatally has been shown to be helpful in lengthening the time of exclusive breastfeeding.¹² Certain medications pass through breast milk, and updated guides to medication cautions can be found at the National Institutes of Health’s LACTMED Web site (https://toxnet.nlm.nih.gov/newtoxnet/lactmed.htm).¹³ In many cases, when a maternal medication is incompatible with breastfeeding, the family physician can consider substituting another appropriate medication that is compatible.

Physician recommendation and support improves the rate of breastfeeding, but many mother–infant dyads require additional support to maintain breastfeeding for the recommended duration; such support can take the form of a certified lactation consultant or counselor, doula, or peer counselor.^23-25 Although structured breastfeeding education in the antenatal period has been demonstrated to be effective in improving breastfeeding initiation and duration, recent research shows that support groups and assistance from the professionals previously mentioned also improve the breastfeeding rate.^26-28

Adequate intake in breastfed infants can be intimidating for new parents to monitor, but they can use a written chart or any of several available smartphone applications.

The AAFP recommends that FPs’ offices adopt specific, evidence-based practices that can have an impact on breastfeeding initiation and duration. Such practices include phone and in-person breastfeeding support from nursing staff and removing any formula advertisements from the office.¹⁷

Formula feeding. When parents choose formula feeding, most infants tolerate cow’s milk-based formula.²⁹ For healthy term infants, differences between brands of formula are generally insignificant. Soy-protein formulas are of value only if lactose intolerance is strongly suspected, such as after prolonged episodes of loose stools. Even then, intolerance is usually transient and cow’s milk-based formula can be tried again in 2 to 4 weeks.

Physicians should recommend 20 kcal/oz of iron-fortified formula for infants who are fed formula—except in special circumstances, such as premature newborns, who may require a more calorie-dense formula. Parents should pay special attention to the manufacturer’s instructions for mixing formula with water because overdilution can cause hyponatremia. Typical volume for newborns should be at least 15 to 30 mL/feed for the first few days; newborns should not go more than 4 hours between feedings. Within the first week, newborns will start taking 60 to 90 mL/feed and increase that gradually to approximately 120 mL/feed by the end of the first month of life. On average, infants need a little more than 100 kcal/kg of body weight a day; for a 3.5-kg infant, that is at least 500 mL of formula over the course of a day.^17,22

Because formula does not contain fluoride, physicians should recommend that parents mix formula that is provided as a powder with fluoridated water. Low-iron formula offers no advantage; feeding with it will cause iron-deficiency anemia in most infants.

When tongue-tie interferes with feeding

Tongue-tie—or ankyloglossia, an atypically short or thick lingual frenulum—is present in 3% to 16% of all births. The condition can make breastfeeding difficult; result in poor neonatal weight gain; and cause sore nipples in 25% to 44% of cases.³⁰ Once tongue-tie is noted, the physician should talk to the mother about the history of feeding success, including whether her nipples are sore and whether the newborn is having difficulty feeding (ie, transferring milk). The Hazelbaker Assessment Tool for Lingual Frenulum Function and the simpler Bristol Tongue Assessment Tool can be used to assess the severity of tongue-tie.^30-35

When tongue-tie interferes with feeding, a physician who is not trained in treatment can refer the mother and infant to a specialist in the community. Frenotomy has been used for many years as a treatment for tongue-tie; improvement in nipple pain and the mother-reported breastfeeding score have been reported postoperatively in several studies.^30-33

Ensure proper vitamin D intake through supplementation

Newborns should consume 400 IU/d of supplemental vitamin D to prevent deficiency and its clinical manifestation, rickets, or other associated abnormalities of calcium metabolism. Deficiency of vitamin D has also been linked to a number of other conditions, including developmental delay and, possibly, type 1 diabetes mellitus in childhood and cardiovascular disease later in life.³⁶

Newborns should consume 400 IU/d of supplemental vitamin D to prevent deficiency and its clinical manifestation, rickets, or other associated abnormalities of calcium metabolism.

In the first months of life, few infants who are solely formula-fed will consume a full liter daily; for them, supplementation of vitamin D for at least one month should be prescribed.³⁵ For breastfed infants, high-dosage maternal vitamin D supplementation may be effective, precluding infant oral vitamin D supplementation³⁶; however, neither the AAFP nor the AAP has issued guidance promoting maternal supplementation in lieu of direct oral infant supplementation.³⁷

Jaundice prevention—and recognition

An elevated bilirubin level is seen in most newborns in the first days of life because of increased production and decreased clearance of bilirubin—a condition known as physiologic jaundice. Conditions that aggravate physiologic hyperbilirubinemia include inborn errors of metabolism, ABO blood-group incompatibility, hemoglobin variants, and inflammatory states such as sepsis. It is important to distinguish physiologic jaundice from exaggerated physiologic and pathologic forms of hyperbilirubinemia; the latter is a medical emergency. Before we get to that, a word about prevention.

Prevention. Because poor caloric intake and dehydration are associated with hyperbilirubinemia, physicians should advise breastfeeding mothers to feed their newborn at least 8 to 12 times daily during the first week of life. However, routine supplementation of liquids other than breast milk should be discouraged in newborns who are not dehydrated.³⁸

The total serum bilirubin level should be tested in every newborn who has clinical jaundice in the first 24 hours of life.

All pregnant women should be tested for ABO and Rh (D) blood types and undergo serum screening for isoimmune antibodies. Randomized trials have demonstrated that the incidence of significant hyperbilirubinemia can be reduced if, for Rh-negative mothers and those who did not undergo prenatal blood-group testing, infant cord blood is tested for 1) ABO and Rh (D) types and 2) direct antibody (Coombs’ test).^38,39

Screening and assessment. It is recommended that all newborns be screened for jaundice before discharge by 1) assessment of clinical risk factors or 2) testing of transcutaneous bilirubin (TcB) or total serum bilirubin (TSB). Furthermore, because evidence shows that treating clinical jaundice can improve outcomes and rehospitalization, TSB should be measured in every newborn who has clinical jaundice in the first 24 hours of life. Measurement of TcB or TSB should also be performed on all infants in whom there appears to be clinical jaundice that is excessive for age.^38,39

During routine clinical care, TcB measurement provides a reasonable estimate of the TSB level in healthy newborns at levels less than 15 mg/dL,⁴⁰ although TcB testing might not be available in the outpatient office. An AAP management algorithm can help determine when a newborn should be seen for outpatient follow-up based on risk of hyperbilirubinemia; higher-risk newborns should be reevaluated in 24 hours.⁹ Outpatient visual assessment of jaundice for cephalocaudal progression—in a well-lit room, with a fully undressed newborn—correlates well with TSB test results. However, visual assessment should not be used alone to screen for hyperbilirubinemia; recent studies have demonstrated that such assessment lacks clinical reliability.⁴⁰

Laboratory assessment. All bilirubin levels should be interpreted based on the newborn’s age in hours. The need for phototherapy should be based on the zone (low, low-intermediate, high-intermediate, or high, as categorized in the AAP nomogram³⁸ in which the TSB level falls. TABLE 4^38-40 provides recommendations for laboratory studies based on risk factors. Standard curves for risk stratification have been developed by the AAP.^37,38

Treatment. Decisions to initiate treatment should be based on the AAP algorithm.³⁸ When initiating phototherapy, precautions include ensuring adequate fluid intake, patching eyes, and monitoring temperature. Phototherapy can generally be stopped when the TSB level falls by 5 mg/dL or below 14 mg/dL. Home phototherapy, using a fiberoptic blanket, for uncomplicated jaundice (in carefully selected newborns with reliable parents) allows continued breastfeeding and bonding with the family, and can significantly decrease the rate of rehospitalization for infants older than 34 weeks.⁴¹

Breastfeeding is often associated with a higher bilirubin level than is seen in infants fed formula exclusively; increasing the frequency of feeding usually reduces the bilirubin level. So-called breast-milk jaundice is a delayed, but common, form of jaundice that is usually diagnosed in the second week of life and peaks by the end of the second week, resolving gradually over one to 4 months. If evaluation reveals no pathologic source, breastfeeding can generally be continued. Temporary discontinuation of breastfeeding to consider a diagnosis of breast-milk jaundice or other reasons for an elevated bilirubin level increases the risk of breastfeeding failure and is usually unnecessary.^12,37,39

Fever—a full work-up, thorough history are key

Concern about serious bacterial illness (SBI) makes the evaluation of fever critical for those who care for newborns. Many studies have attempted to identify which newborns might be able to be cared for safely as outpatients to prevent unnecessary testing and antibiotics.^5,42 Regrettably, SBI in infants remains difficult to predict, and protocols that have been developed may miss as many as 1 of every 10 newborns who has SBI.⁴³ Initial management of all infants 28 days old or younger with fever must therefore include a full work-up, including lumbar puncture and empiric antibiotics.⁴⁴

Evaluation. When an infant younger than 28 days has a fever, the physician should first verify that the temperature was taken rectally and how it was documented. In an infant who has a history of prematurity, it is crucial to correct for chronological age when deciding on proper evaluation.

Additional important findings in the history include a significant change in behavior, associated symptoms, and exposure to sick contacts. The maternal and birth history, including prolonged rupture of membranes, colonization with group B Streptococcus, administration of antibiotics at delivery, and genital herpes simplex virus (HSV) infection may suggest a cause for fever.⁴⁵

The evaluation of fever might include the white blood cell (WBC) count, blood culture, measurement of markers of inflammation, urine studies, lumbar puncture, stool culture, and chest radiograph. Traditionally, the WBC count has been utilized as a standard marker for sepsis, although it has a low sensitivity and specificity for SBI, especially in newborns.⁴⁶ Blood cultures should be obtained routinely in the newborn with fever, and before antibiotics are administered in older infants.

Procalcitonin (PCT; a calcitonin precursor) and the inflammatory marker C-reactive protein (CRP) have been shown, in several large studies, to have relatively high sensitivity and specificity for SBI; measurement of these constituents may enhance detection of serious illness.^46-49 In a large study of 2047 febrile infants older than 30 months, the PCT level was determined to be more accurate than the CRP level, the WBC count, and the absolute neutrophil count in predicting SBI.^48,49 PCT shows the most promise for preventing a full fever work-up and empiric antibiotics. It has not yet been widely translated into practice, however, because of a lack of clear guidance on how to combine PCT levels with other laboratory markers and clinical decision-making.^48-50

Urinalysis (UA) should be obtained for all newborns who present with fever. Traditionally, it was recommended that urine should be cultured for all newborns with fever; however, more recent data show that the initial urinalysis is much more sensitive than once thought. In a study, UA was positive (defined as pyuria or a positive leukocyte esterase test, or both) in all but 1 of 203 infants who had bacteremic UTI (sensitivity, 99.5%).⁵¹

The procalcitonin level was determined to be more accurate than the C-reactive protein level, the white blood cell count, and the absolute neutrophil count in predicting serious bacterial illness.

Stool culture is necessary in newborns only when they present with blood or mucus in diarrhea. Lumbar puncture should be performed in all febrile newborns and all newborns for whom empiric antibiotics have been prescribed.^43,44 A chest radiograph may be useful in diagnosis when a newborn has any other sign of pulmonary disease: respiratory rate >50/min, retractions, wheezing, grunting, stridor, nasal flaring, cough, and positive findings on lung examination.^43,44

Treatment. Management for all newborns who have a rectal temperature ≥38° C includes admission to the hospital and empiric antibiotics; guidance is based primarily on expert consensus. Common pathogens for SBI include group B Strep, Escherichia coli, Enterococcus spp., and Listeria monocytogenes.^43,44 Empiric antibiotics, including ampicillin (to cover L monocytogenes) and cefotaxime or gentamicin should be started immediately after sending for blood, urine, and cerebrospinal fluid (CSF) cultures.^43-45

Management for all newborns who have rectal temperature ≥38° C includes admission to the hospital and empiric antibiotics.

All infants who are ill-appearing or have vesicles, seizures, or a maternal history of genital HSV infection should also be started on empiric acyclovir. Vesicles should be cultured and CSF should be sent for HSV DNA polymerase chain reaction before acyclovir is administered.^43-45

Sudden infant death syndrome: Steps to take to minimize risk

SIDS is defined as the sudden death of a child younger than 1 year that remains unexplained after a thorough case investigation and comprehensive review of the clinical history. The risk of SIDS in the United States is less than 1 for every 1000 live births; incidence peaks between 2 and 4 months of age.⁵² In the United States, SIDS and other sleep-related infant deaths, such as strangulation in bed or accidental suffocation, account for more than 4000 deaths a year.⁵³ The incidence of SIDS declined markedly after the “Back to Sleep” campaign was launched in 2003, but has leveled off since 2005.^53-55

Numerous risk factors for SIDS have been identified, including maternal factors (young maternal age, maternal smoking during pregnancy, late or no prenatal care) and infant and environmental factors (prematurity, low birth weight, male gender, prone sleeping position, sleeping on a soft surface or with bedding accessories, bed-sharing (ie, sleeping in the parents’ bed), and overheating. In many cases, the risk factors are modifiable; sleeping in the prone position is the most meaningful modifiable risk factor.

Home monitors have not been proven to reduce the incidence of SIDS and are not recommended for that purpose.

To minimize the risk for SIDS, parents should be educated on the risk factors—prenatally as well as at each infant well visit. Home monitors have not been proven to reduce the incidence of SIDS and are not recommended for that purpose.^54-57 Although evidence is strongest for supine positioning as a preventive intervention for SIDS, other evidence-based recommendations include use of a firm sleep surface; breastfeeding; use of a pacifier; room-sharing with parents without bed-sharing; routine immunization; avoidance of overheating; avoiding falling asleep with the infant on a chair or couch; and avoiding exposure to tobacco smoke, alcohol, and drugs of abuse.^55,56 A recent systematic review showed that large-scale community interventions and education campaigns can play a significant role in parental and community adoption of safe sleep recommendations; however, families and communities rarely exhibit complete adherence to safe sleep practices.⁵⁷

Other concerns in the first month of life and immediately beyond

In TABLE 5,² we list additional common newborn problems not reviewed in the text of this article and summarize evidence-based treatment strategies.

CORRESPONDENCE
Scott Hartman, MD, Associate Professor, Department of Family Medicine, University of Rochester Medical Center, 777 South Clinton Avenue, Rochester, NY 14620; [email protected].

Acknowledgement
We thank Nancy Phillips for her assistance in the preparation of this article.

Caring for a newborn can be a source of joy for family physicians (FPs). In this article, we examine care provided in the first month of life, including a thorough physical examination, safe hospital discharge procedures, assessment of neonatal feeding, evaluation of jaundice and fever, and prevention of sudden infant death syndrome (SIDS). In addition, we describe how FPs can support women of childbearing age between pregnancies, with the goal of reducing the risk of adverse outcomes in future pregnancies. (See “Your role in risk assessment and interventions during the interconception period.”)

Before parents leave the hospital with their newborn, it is essential that they receive written and verbal counseling on important issues in neonatal care. A discharge checklist can help make sure all topics have been covered.¹ A hearing screen and pulse oximetry before discharge are required for all newborns in most states, in addition to important preventive counseling for parents. TABLE 1² and TABLE 2² summarize important newborn physical exam findings and common skin conditions. Parents should be given additional written information regarding prevention of SIDS and proper use of car seats.

Hospital physicians should assess maternal medical and psychosocial readiness for discharge. Through shared decision-making with the newborn’s parents, physicians should create a plan for outpatient follow-up. Assessment through a physician home visit can provide safe and effective care similar to what is provided at a visit to an office medical practice.^3-7 A follow-up appointment should be made 2 to 5 days before discharge, preferably connecting the newborn to a medical home where comprehensive health care services are offered.^1,5,6,8

Age, gestational age, risk factors for hyperbilirubinemia, and the timing and level of bilirubin testing should be considered when establishing a follow-up interval. Most newborns who are discharged before 72 hours of age should have a follow-up visit in 2 days; a newborn who has a recognized risk factor for a health problem should be seen sooner. Newborns in the “low-risk zone” (ie, no recognized risk factors) should be seen based on age at discharge or need for breastfeeding support.⁹

Tracking baby’s weight, ensuring proper feeding

A newborn who is discharged at 24 hours of life, or sooner, should be seen in the office within 2 days of discharge to 1) ensure that he (she) is getting proper nutrition and 2) monitor his weight^1,3,5 (TABLE 3^10-13). All newborns should be seen again at 2 weeks of life, with additional visits more frequently if there are concerns about nutrition.¹

Recording an accurate weight is critical; the newborn should be weighed completely undressed and without a diaper. Healthy newborns can safely lose up to 10% of birth weight within the first week of life; they should be back to their birth weight by approximately 2 weeks of life.^10,11 A healthy newborn loses approximately 0.5 to 1 oz a day;¹¹ greater than 10% loss of birth weight should trigger a thorough medical work-up and feeding assessment.

Breastfeeding. For breastfeeding mothers, physicians should recommend on-demand feeding or a feeding at least every 2 or 3 hours. Adequate intake in breastfed infants can be intimidating for new parents to monitor, but they can use a written chart or any of several available smartphone applications to document length and timing of feeds and frequency of urination and bowel movements. By the fifth day of life, a newborn should be having at least 6 voids and 3 or 4 stools a day.^10-12

In addition, physicians can counsel parents on what to look for—in the mother and the newborn—to confirm that breastfeeding is successful, with adequate nutritional intake (TABLE 3^10-13). Physicians should recommend against providing a pacifier to breastfeeding infants during the first several weeks of life—or until breastfeeding is well established (usually at 3 or 4 weeks of age). The World Health Organization (WHO) recommends against providing bottles, pacifiers, and artificial nipples to breastfeeding newborns.¹⁴ Liquids other than colostrum or breast milk should not be given unless there is a documented medical need, such as inadequate weight gain or feeding difficulty.¹⁵ If the newborn experiences early latch difficulties, supplementation with expressed breast milk is preferable to supplementation with formula. Assistance from a trained lactation consultant is a key element in the support of the breastfeeding dyad.^11,12,16

Breastfeeding optimizes development of the newborn’s immune system, thus bolstering disease prevention; it also assists with maternal postpartum weight loss and psychological well-being. Exclusively or primarily formula-fed newborns are at increased risk of gastrointestinal, ear, and respiratory infections throughout infancy and childhood; type 1 diabetes mellitus; asthma; childhood and adult obesity; and leukemia.^17,18 Mothers who feed their newborn primarily formula increase their own risk of obesity, type 2 diabetes mellitus, ovarian and breast cancer, and depression.^17-22

Infant feeding is a personal and family choice but, in the absence of medical contraindications—such as maternal human immunodeficiency virus infection and galactosemia—exclusive breastfeeding should be recommended.^17,18 FPs are well suited to support the mother–infant breastfeeding dyad in the neonatal period, based on expert recommendations. Specifically, the American Academy of Family Physicians (AAFP) and American Academy of Pediatrics (AAP) recommend that all infants be exclusively breastfed for the first 6 months of life and continue some breastfeeding through the first year or longer.^17,18 WHO recommends breastfeeding until 24 months of age—longer if mother and infant want to, unless breastfeeding is contraindicated.^14,17,18

Physicians should provide up-to-date information to parents regarding the risks and benefits of feeding choices. Support for breastfeeding mothers postnatally has been shown to be helpful in lengthening the time of exclusive breastfeeding.¹² Certain medications pass through breast milk, and updated guides to medication cautions can be found at the National Institutes of Health’s LACTMED Web site (https://toxnet.nlm.nih.gov/newtoxnet/lactmed.htm).¹³ In many cases, when a maternal medication is incompatible with breastfeeding, the family physician can consider substituting another appropriate medication that is compatible.

Physician recommendation and support improves the rate of breastfeeding, but many mother–infant dyads require additional support to maintain breastfeeding for the recommended duration; such support can take the form of a certified lactation consultant or counselor, doula, or peer counselor.^23-25 Although structured breastfeeding education in the antenatal period has been demonstrated to be effective in improving breastfeeding initiation and duration, recent research shows that support groups and assistance from the professionals previously mentioned also improve the breastfeeding rate.^26-28

Adequate intake in breastfed infants can be intimidating for new parents to monitor, but they can use a written chart or any of several available smartphone applications.

The AAFP recommends that FPs’ offices adopt specific, evidence-based practices that can have an impact on breastfeeding initiation and duration. Such practices include phone and in-person breastfeeding support from nursing staff and removing any formula advertisements from the office.¹⁷

Formula feeding. When parents choose formula feeding, most infants tolerate cow’s milk-based formula.²⁹ For healthy term infants, differences between brands of formula are generally insignificant. Soy-protein formulas are of value only if lactose intolerance is strongly suspected, such as after prolonged episodes of loose stools. Even then, intolerance is usually transient and cow’s milk-based formula can be tried again in 2 to 4 weeks.

Physicians should recommend 20 kcal/oz of iron-fortified formula for infants who are fed formula—except in special circumstances, such as premature newborns, who may require a more calorie-dense formula. Parents should pay special attention to the manufacturer’s instructions for mixing formula with water because overdilution can cause hyponatremia. Typical volume for newborns should be at least 15 to 30 mL/feed for the first few days; newborns should not go more than 4 hours between feedings. Within the first week, newborns will start taking 60 to 90 mL/feed and increase that gradually to approximately 120 mL/feed by the end of the first month of life. On average, infants need a little more than 100 kcal/kg of body weight a day; for a 3.5-kg infant, that is at least 500 mL of formula over the course of a day.^17,22

Because formula does not contain fluoride, physicians should recommend that parents mix formula that is provided as a powder with fluoridated water. Low-iron formula offers no advantage; feeding with it will cause iron-deficiency anemia in most infants.

When tongue-tie interferes with feeding

Tongue-tie—or ankyloglossia, an atypically short or thick lingual frenulum—is present in 3% to 16% of all births. The condition can make breastfeeding difficult; result in poor neonatal weight gain; and cause sore nipples in 25% to 44% of cases.³⁰ Once tongue-tie is noted, the physician should talk to the mother about the history of feeding success, including whether her nipples are sore and whether the newborn is having difficulty feeding (ie, transferring milk). The Hazelbaker Assessment Tool for Lingual Frenulum Function and the simpler Bristol Tongue Assessment Tool can be used to assess the severity of tongue-tie.^30-35

When tongue-tie interferes with feeding, a physician who is not trained in treatment can refer the mother and infant to a specialist in the community. Frenotomy has been used for many years as a treatment for tongue-tie; improvement in nipple pain and the mother-reported breastfeeding score have been reported postoperatively in several studies.^30-33

Ensure proper vitamin D intake through supplementation

Newborns should consume 400 IU/d of supplemental vitamin D to prevent deficiency and its clinical manifestation, rickets, or other associated abnormalities of calcium metabolism. Deficiency of vitamin D has also been linked to a number of other conditions, including developmental delay and, possibly, type 1 diabetes mellitus in childhood and cardiovascular disease later in life.³⁶

Newborns should consume 400 IU/d of supplemental vitamin D to prevent deficiency and its clinical manifestation, rickets, or other associated abnormalities of calcium metabolism.

In the first months of life, few infants who are solely formula-fed will consume a full liter daily; for them, supplementation of vitamin D for at least one month should be prescribed.³⁵ For breastfed infants, high-dosage maternal vitamin D supplementation may be effective, precluding infant oral vitamin D supplementation³⁶; however, neither the AAFP nor the AAP has issued guidance promoting maternal supplementation in lieu of direct oral infant supplementation.³⁷

Jaundice prevention—and recognition

An elevated bilirubin level is seen in most newborns in the first days of life because of increased production and decreased clearance of bilirubin—a condition known as physiologic jaundice. Conditions that aggravate physiologic hyperbilirubinemia include inborn errors of metabolism, ABO blood-group incompatibility, hemoglobin variants, and inflammatory states such as sepsis. It is important to distinguish physiologic jaundice from exaggerated physiologic and pathologic forms of hyperbilirubinemia; the latter is a medical emergency. Before we get to that, a word about prevention.

Prevention. Because poor caloric intake and dehydration are associated with hyperbilirubinemia, physicians should advise breastfeeding mothers to feed their newborn at least 8 to 12 times daily during the first week of life. However, routine supplementation of liquids other than breast milk should be discouraged in newborns who are not dehydrated.³⁸

The total serum bilirubin level should be tested in every newborn who has clinical jaundice in the first 24 hours of life.

All pregnant women should be tested for ABO and Rh (D) blood types and undergo serum screening for isoimmune antibodies. Randomized trials have demonstrated that the incidence of significant hyperbilirubinemia can be reduced if, for Rh-negative mothers and those who did not undergo prenatal blood-group testing, infant cord blood is tested for 1) ABO and Rh (D) types and 2) direct antibody (Coombs’ test).^38,39

Screening and assessment. It is recommended that all newborns be screened for jaundice before discharge by 1) assessment of clinical risk factors or 2) testing of transcutaneous bilirubin (TcB) or total serum bilirubin (TSB). Furthermore, because evidence shows that treating clinical jaundice can improve outcomes and rehospitalization, TSB should be measured in every newborn who has clinical jaundice in the first 24 hours of life. Measurement of TcB or TSB should also be performed on all infants in whom there appears to be clinical jaundice that is excessive for age.^38,39

During routine clinical care, TcB measurement provides a reasonable estimate of the TSB level in healthy newborns at levels less than 15 mg/dL,⁴⁰ although TcB testing might not be available in the outpatient office. An AAP management algorithm can help determine when a newborn should be seen for outpatient follow-up based on risk of hyperbilirubinemia; higher-risk newborns should be reevaluated in 24 hours.⁹ Outpatient visual assessment of jaundice for cephalocaudal progression—in a well-lit room, with a fully undressed newborn—correlates well with TSB test results. However, visual assessment should not be used alone to screen for hyperbilirubinemia; recent studies have demonstrated that such assessment lacks clinical reliability.⁴⁰

Laboratory assessment. All bilirubin levels should be interpreted based on the newborn’s age in hours. The need for phototherapy should be based on the zone (low, low-intermediate, high-intermediate, or high, as categorized in the AAP nomogram³⁸ in which the TSB level falls. TABLE 4^38-40 provides recommendations for laboratory studies based on risk factors. Standard curves for risk stratification have been developed by the AAP.^37,38

Treatment. Decisions to initiate treatment should be based on the AAP algorithm.³⁸ When initiating phototherapy, precautions include ensuring adequate fluid intake, patching eyes, and monitoring temperature. Phototherapy can generally be stopped when the TSB level falls by 5 mg/dL or below 14 mg/dL. Home phototherapy, using a fiberoptic blanket, for uncomplicated jaundice (in carefully selected newborns with reliable parents) allows continued breastfeeding and bonding with the family, and can significantly decrease the rate of rehospitalization for infants older than 34 weeks.⁴¹

Breastfeeding is often associated with a higher bilirubin level than is seen in infants fed formula exclusively; increasing the frequency of feeding usually reduces the bilirubin level. So-called breast-milk jaundice is a delayed, but common, form of jaundice that is usually diagnosed in the second week of life and peaks by the end of the second week, resolving gradually over one to 4 months. If evaluation reveals no pathologic source, breastfeeding can generally be continued. Temporary discontinuation of breastfeeding to consider a diagnosis of breast-milk jaundice or other reasons for an elevated bilirubin level increases the risk of breastfeeding failure and is usually unnecessary.^12,37,39

Fever—a full work-up, thorough history are key

Concern about serious bacterial illness (SBI) makes the evaluation of fever critical for those who care for newborns. Many studies have attempted to identify which newborns might be able to be cared for safely as outpatients to prevent unnecessary testing and antibiotics.^5,42 Regrettably, SBI in infants remains difficult to predict, and protocols that have been developed may miss as many as 1 of every 10 newborns who has SBI.⁴³ Initial management of all infants 28 days old or younger with fever must therefore include a full work-up, including lumbar puncture and empiric antibiotics.⁴⁴

Evaluation. When an infant younger than 28 days has a fever, the physician should first verify that the temperature was taken rectally and how it was documented. In an infant who has a history of prematurity, it is crucial to correct for chronological age when deciding on proper evaluation.

Additional important findings in the history include a significant change in behavior, associated symptoms, and exposure to sick contacts. The maternal and birth history, including prolonged rupture of membranes, colonization with group B Streptococcus, administration of antibiotics at delivery, and genital herpes simplex virus (HSV) infection may suggest a cause for fever.⁴⁵

The evaluation of fever might include the white blood cell (WBC) count, blood culture, measurement of markers of inflammation, urine studies, lumbar puncture, stool culture, and chest radiograph. Traditionally, the WBC count has been utilized as a standard marker for sepsis, although it has a low sensitivity and specificity for SBI, especially in newborns.⁴⁶ Blood cultures should be obtained routinely in the newborn with fever, and before antibiotics are administered in older infants.

Procalcitonin (PCT; a calcitonin precursor) and the inflammatory marker C-reactive protein (CRP) have been shown, in several large studies, to have relatively high sensitivity and specificity for SBI; measurement of these constituents may enhance detection of serious illness.^46-49 In a large study of 2047 febrile infants older than 30 months, the PCT level was determined to be more accurate than the CRP level, the WBC count, and the absolute neutrophil count in predicting SBI.^48,49 PCT shows the most promise for preventing a full fever work-up and empiric antibiotics. It has not yet been widely translated into practice, however, because of a lack of clear guidance on how to combine PCT levels with other laboratory markers and clinical decision-making.^48-50

Urinalysis (UA) should be obtained for all newborns who present with fever. Traditionally, it was recommended that urine should be cultured for all newborns with fever; however, more recent data show that the initial urinalysis is much more sensitive than once thought. In a study, UA was positive (defined as pyuria or a positive leukocyte esterase test, or both) in all but 1 of 203 infants who had bacteremic UTI (sensitivity, 99.5%).⁵¹

The procalcitonin level was determined to be more accurate than the C-reactive protein level, the white blood cell count, and the absolute neutrophil count in predicting serious bacterial illness.

Stool culture is necessary in newborns only when they present with blood or mucus in diarrhea. Lumbar puncture should be performed in all febrile newborns and all newborns for whom empiric antibiotics have been prescribed.^43,44 A chest radiograph may be useful in diagnosis when a newborn has any other sign of pulmonary disease: respiratory rate >50/min, retractions, wheezing, grunting, stridor, nasal flaring, cough, and positive findings on lung examination.^43,44

Treatment. Management for all newborns who have a rectal temperature ≥38° C includes admission to the hospital and empiric antibiotics; guidance is based primarily on expert consensus. Common pathogens for SBI include group B Strep, Escherichia coli, Enterococcus spp., and Listeria monocytogenes.^43,44 Empiric antibiotics, including ampicillin (to cover L monocytogenes) and cefotaxime or gentamicin should be started immediately after sending for blood, urine, and cerebrospinal fluid (CSF) cultures.^43-45

Management for all newborns who have rectal temperature ≥38° C includes admission to the hospital and empiric antibiotics.

All infants who are ill-appearing or have vesicles, seizures, or a maternal history of genital HSV infection should also be started on empiric acyclovir. Vesicles should be cultured and CSF should be sent for HSV DNA polymerase chain reaction before acyclovir is administered.^43-45

Sudden infant death syndrome: Steps to take to minimize risk

SIDS is defined as the sudden death of a child younger than 1 year that remains unexplained after a thorough case investigation and comprehensive review of the clinical history. The risk of SIDS in the United States is less than 1 for every 1000 live births; incidence peaks between 2 and 4 months of age.⁵² In the United States, SIDS and other sleep-related infant deaths, such as strangulation in bed or accidental suffocation, account for more than 4000 deaths a year.⁵³ The incidence of SIDS declined markedly after the “Back to Sleep” campaign was launched in 2003, but has leveled off since 2005.^53-55

Numerous risk factors for SIDS have been identified, including maternal factors (young maternal age, maternal smoking during pregnancy, late or no prenatal care) and infant and environmental factors (prematurity, low birth weight, male gender, prone sleeping position, sleeping on a soft surface or with bedding accessories, bed-sharing (ie, sleeping in the parents’ bed), and overheating. In many cases, the risk factors are modifiable; sleeping in the prone position is the most meaningful modifiable risk factor.

Home monitors have not been proven to reduce the incidence of SIDS and are not recommended for that purpose.

To minimize the risk for SIDS, parents should be educated on the risk factors—prenatally as well as at each infant well visit. Home monitors have not been proven to reduce the incidence of SIDS and are not recommended for that purpose.^54-57 Although evidence is strongest for supine positioning as a preventive intervention for SIDS, other evidence-based recommendations include use of a firm sleep surface; breastfeeding; use of a pacifier; room-sharing with parents without bed-sharing; routine immunization; avoidance of overheating; avoiding falling asleep with the infant on a chair or couch; and avoiding exposure to tobacco smoke, alcohol, and drugs of abuse.^55,56 A recent systematic review showed that large-scale community interventions and education campaigns can play a significant role in parental and community adoption of safe sleep recommendations; however, families and communities rarely exhibit complete adherence to safe sleep practices.⁵⁷

Other concerns in the first month of life and immediately beyond

In TABLE 5,² we list additional common newborn problems not reviewed in the text of this article and summarize evidence-based treatment strategies.

CORRESPONDENCE
Scott Hartman, MD, Associate Professor, Department of Family Medicine, University of Rochester Medical Center, 777 South Clinton Avenue, Rochester, NY 14620; [email protected].

Acknowledgement
We thank Nancy Phillips for her assistance in the preparation of this article.

Caring for a newborn can be a source of joy for family physicians (FPs). In this article, we examine care provided in the first month of life, including a thorough physical examination, safe hospital discharge procedures, assessment of neonatal feeding, evaluation of jaundice and fever, and prevention of sudden infant death syndrome (SIDS). In addition, we describe how FPs can support women of childbearing age between pregnancies, with the goal of reducing the risk of adverse outcomes in future pregnancies. (See “Your role in risk assessment and interventions during the interconception period.”)

Before parents leave the hospital with their newborn, it is essential that they receive written and verbal counseling on important issues in neonatal care. A discharge checklist can help make sure all topics have been covered.¹ A hearing screen and pulse oximetry before discharge are required for all newborns in most states, in addition to important preventive counseling for parents. TABLE 1² and TABLE 2² summarize important newborn physical exam findings and common skin conditions. Parents should be given additional written information regarding prevention of SIDS and proper use of car seats.

Hospital physicians should assess maternal medical and psychosocial readiness for discharge. Through shared decision-making with the newborn’s parents, physicians should create a plan for outpatient follow-up. Assessment through a physician home visit can provide safe and effective care similar to what is provided at a visit to an office medical practice.^3-7 A follow-up appointment should be made 2 to 5 days before discharge, preferably connecting the newborn to a medical home where comprehensive health care services are offered.^1,5,6,8

Age, gestational age, risk factors for hyperbilirubinemia, and the timing and level of bilirubin testing should be considered when establishing a follow-up interval. Most newborns who are discharged before 72 hours of age should have a follow-up visit in 2 days; a newborn who has a recognized risk factor for a health problem should be seen sooner. Newborns in the “low-risk zone” (ie, no recognized risk factors) should be seen based on age at discharge or need for breastfeeding support.⁹

Tracking baby’s weight, ensuring proper feeding

A newborn who is discharged at 24 hours of life, or sooner, should be seen in the office within 2 days of discharge to 1) ensure that he (she) is getting proper nutrition and 2) monitor his weight^1,3,5 (TABLE 3^10-13). All newborns should be seen again at 2 weeks of life, with additional visits more frequently if there are concerns about nutrition.¹

Recording an accurate weight is critical; the newborn should be weighed completely undressed and without a diaper. Healthy newborns can safely lose up to 10% of birth weight within the first week of life; they should be back to their birth weight by approximately 2 weeks of life.^10,11 A healthy newborn loses approximately 0.5 to 1 oz a day;¹¹ greater than 10% loss of birth weight should trigger a thorough medical work-up and feeding assessment.

Breastfeeding. For breastfeeding mothers, physicians should recommend on-demand feeding or a feeding at least every 2 or 3 hours. Adequate intake in breastfed infants can be intimidating for new parents to monitor, but they can use a written chart or any of several available smartphone applications to document length and timing of feeds and frequency of urination and bowel movements. By the fifth day of life, a newborn should be having at least 6 voids and 3 or 4 stools a day.^10-12

In addition, physicians can counsel parents on what to look for—in the mother and the newborn—to confirm that breastfeeding is successful, with adequate nutritional intake (TABLE 3^10-13). Physicians should recommend against providing a pacifier to breastfeeding infants during the first several weeks of life—or until breastfeeding is well established (usually at 3 or 4 weeks of age). The World Health Organization (WHO) recommends against providing bottles, pacifiers, and artificial nipples to breastfeeding newborns.¹⁴ Liquids other than colostrum or breast milk should not be given unless there is a documented medical need, such as inadequate weight gain or feeding difficulty.¹⁵ If the newborn experiences early latch difficulties, supplementation with expressed breast milk is preferable to supplementation with formula. Assistance from a trained lactation consultant is a key element in the support of the breastfeeding dyad.^11,12,16

Breastfeeding optimizes development of the newborn’s immune system, thus bolstering disease prevention; it also assists with maternal postpartum weight loss and psychological well-being. Exclusively or primarily formula-fed newborns are at increased risk of gastrointestinal, ear, and respiratory infections throughout infancy and childhood; type 1 diabetes mellitus; asthma; childhood and adult obesity; and leukemia.^17,18 Mothers who feed their newborn primarily formula increase their own risk of obesity, type 2 diabetes mellitus, ovarian and breast cancer, and depression.^17-22

Infant feeding is a personal and family choice but, in the absence of medical contraindications—such as maternal human immunodeficiency virus infection and galactosemia—exclusive breastfeeding should be recommended.^17,18 FPs are well suited to support the mother–infant breastfeeding dyad in the neonatal period, based on expert recommendations. Specifically, the American Academy of Family Physicians (AAFP) and American Academy of Pediatrics (AAP) recommend that all infants be exclusively breastfed for the first 6 months of life and continue some breastfeeding through the first year or longer.^17,18 WHO recommends breastfeeding until 24 months of age—longer if mother and infant want to, unless breastfeeding is contraindicated.^14,17,18

Physicians should provide up-to-date information to parents regarding the risks and benefits of feeding choices. Support for breastfeeding mothers postnatally has been shown to be helpful in lengthening the time of exclusive breastfeeding.¹² Certain medications pass through breast milk, and updated guides to medication cautions can be found at the National Institutes of Health’s LACTMED Web site (https://toxnet.nlm.nih.gov/newtoxnet/lactmed.htm).¹³ In many cases, when a maternal medication is incompatible with breastfeeding, the family physician can consider substituting another appropriate medication that is compatible.

Physician recommendation and support improves the rate of breastfeeding, but many mother–infant dyads require additional support to maintain breastfeeding for the recommended duration; such support can take the form of a certified lactation consultant or counselor, doula, or peer counselor.^23-25 Although structured breastfeeding education in the antenatal period has been demonstrated to be effective in improving breastfeeding initiation and duration, recent research shows that support groups and assistance from the professionals previously mentioned also improve the breastfeeding rate.^26-28

Adequate intake in breastfed infants can be intimidating for new parents to monitor, but they can use a written chart or any of several available smartphone applications.

The AAFP recommends that FPs’ offices adopt specific, evidence-based practices that can have an impact on breastfeeding initiation and duration. Such practices include phone and in-person breastfeeding support from nursing staff and removing any formula advertisements from the office.¹⁷

Formula feeding. When parents choose formula feeding, most infants tolerate cow’s milk-based formula.²⁹ For healthy term infants, differences between brands of formula are generally insignificant. Soy-protein formulas are of value only if lactose intolerance is strongly suspected, such as after prolonged episodes of loose stools. Even then, intolerance is usually transient and cow’s milk-based formula can be tried again in 2 to 4 weeks.

Physicians should recommend 20 kcal/oz of iron-fortified formula for infants who are fed formula—except in special circumstances, such as premature newborns, who may require a more calorie-dense formula. Parents should pay special attention to the manufacturer’s instructions for mixing formula with water because overdilution can cause hyponatremia. Typical volume for newborns should be at least 15 to 30 mL/feed for the first few days; newborns should not go more than 4 hours between feedings. Within the first week, newborns will start taking 60 to 90 mL/feed and increase that gradually to approximately 120 mL/feed by the end of the first month of life. On average, infants need a little more than 100 kcal/kg of body weight a day; for a 3.5-kg infant, that is at least 500 mL of formula over the course of a day.^17,22

Because formula does not contain fluoride, physicians should recommend that parents mix formula that is provided as a powder with fluoridated water. Low-iron formula offers no advantage; feeding with it will cause iron-deficiency anemia in most infants.

When tongue-tie interferes with feeding

Tongue-tie—or ankyloglossia, an atypically short or thick lingual frenulum—is present in 3% to 16% of all births. The condition can make breastfeeding difficult; result in poor neonatal weight gain; and cause sore nipples in 25% to 44% of cases.³⁰ Once tongue-tie is noted, the physician should talk to the mother about the history of feeding success, including whether her nipples are sore and whether the newborn is having difficulty feeding (ie, transferring milk). The Hazelbaker Assessment Tool for Lingual Frenulum Function and the simpler Bristol Tongue Assessment Tool can be used to assess the severity of tongue-tie.^30-35

When tongue-tie interferes with feeding, a physician who is not trained in treatment can refer the mother and infant to a specialist in the community. Frenotomy has been used for many years as a treatment for tongue-tie; improvement in nipple pain and the mother-reported breastfeeding score have been reported postoperatively in several studies.^30-33

Ensure proper vitamin D intake through supplementation

Newborns should consume 400 IU/d of supplemental vitamin D to prevent deficiency and its clinical manifestation, rickets, or other associated abnormalities of calcium metabolism. Deficiency of vitamin D has also been linked to a number of other conditions, including developmental delay and, possibly, type 1 diabetes mellitus in childhood and cardiovascular disease later in life.³⁶

Newborns should consume 400 IU/d of supplemental vitamin D to prevent deficiency and its clinical manifestation, rickets, or other associated abnormalities of calcium metabolism.

In the first months of life, few infants who are solely formula-fed will consume a full liter daily; for them, supplementation of vitamin D for at least one month should be prescribed.³⁵ For breastfed infants, high-dosage maternal vitamin D supplementation may be effective, precluding infant oral vitamin D supplementation³⁶; however, neither the AAFP nor the AAP has issued guidance promoting maternal supplementation in lieu of direct oral infant supplementation.³⁷

Jaundice prevention—and recognition

An elevated bilirubin level is seen in most newborns in the first days of life because of increased production and decreased clearance of bilirubin—a condition known as physiologic jaundice. Conditions that aggravate physiologic hyperbilirubinemia include inborn errors of metabolism, ABO blood-group incompatibility, hemoglobin variants, and inflammatory states such as sepsis. It is important to distinguish physiologic jaundice from exaggerated physiologic and pathologic forms of hyperbilirubinemia; the latter is a medical emergency. Before we get to that, a word about prevention.

Prevention. Because poor caloric intake and dehydration are associated with hyperbilirubinemia, physicians should advise breastfeeding mothers to feed their newborn at least 8 to 12 times daily during the first week of life. However, routine supplementation of liquids other than breast milk should be discouraged in newborns who are not dehydrated.³⁸

The total serum bilirubin level should be tested in every newborn who has clinical jaundice in the first 24 hours of life.

All pregnant women should be tested for ABO and Rh (D) blood types and undergo serum screening for isoimmune antibodies. Randomized trials have demonstrated that the incidence of significant hyperbilirubinemia can be reduced if, for Rh-negative mothers and those who did not undergo prenatal blood-group testing, infant cord blood is tested for 1) ABO and Rh (D) types and 2) direct antibody (Coombs’ test).^38,39

Screening and assessment. It is recommended that all newborns be screened for jaundice before discharge by 1) assessment of clinical risk factors or 2) testing of transcutaneous bilirubin (TcB) or total serum bilirubin (TSB). Furthermore, because evidence shows that treating clinical jaundice can improve outcomes and rehospitalization, TSB should be measured in every newborn who has clinical jaundice in the first 24 hours of life. Measurement of TcB or TSB should also be performed on all infants in whom there appears to be clinical jaundice that is excessive for age.^38,39

During routine clinical care, TcB measurement provides a reasonable estimate of the TSB level in healthy newborns at levels less than 15 mg/dL,⁴⁰ although TcB testing might not be available in the outpatient office. An AAP management algorithm can help determine when a newborn should be seen for outpatient follow-up based on risk of hyperbilirubinemia; higher-risk newborns should be reevaluated in 24 hours.⁹ Outpatient visual assessment of jaundice for cephalocaudal progression—in a well-lit room, with a fully undressed newborn—correlates well with TSB test results. However, visual assessment should not be used alone to screen for hyperbilirubinemia; recent studies have demonstrated that such assessment lacks clinical reliability.⁴⁰

Laboratory assessment. All bilirubin levels should be interpreted based on the newborn’s age in hours. The need for phototherapy should be based on the zone (low, low-intermediate, high-intermediate, or high, as categorized in the AAP nomogram³⁸ in which the TSB level falls. TABLE 4^38-40 provides recommendations for laboratory studies based on risk factors. Standard curves for risk stratification have been developed by the AAP.^37,38

Treatment. Decisions to initiate treatment should be based on the AAP algorithm.³⁸ When initiating phototherapy, precautions include ensuring adequate fluid intake, patching eyes, and monitoring temperature. Phototherapy can generally be stopped when the TSB level falls by 5 mg/dL or below 14 mg/dL. Home phototherapy, using a fiberoptic blanket, for uncomplicated jaundice (in carefully selected newborns with reliable parents) allows continued breastfeeding and bonding with the family, and can significantly decrease the rate of rehospitalization for infants older than 34 weeks.⁴¹

Breastfeeding is often associated with a higher bilirubin level than is seen in infants fed formula exclusively; increasing the frequency of feeding usually reduces the bilirubin level. So-called breast-milk jaundice is a delayed, but common, form of jaundice that is usually diagnosed in the second week of life and peaks by the end of the second week, resolving gradually over one to 4 months. If evaluation reveals no pathologic source, breastfeeding can generally be continued. Temporary discontinuation of breastfeeding to consider a diagnosis of breast-milk jaundice or other reasons for an elevated bilirubin level increases the risk of breastfeeding failure and is usually unnecessary.^12,37,39

Fever—a full work-up, thorough history are key

Concern about serious bacterial illness (SBI) makes the evaluation of fever critical for those who care for newborns. Many studies have attempted to identify which newborns might be able to be cared for safely as outpatients to prevent unnecessary testing and antibiotics.^5,42 Regrettably, SBI in infants remains difficult to predict, and protocols that have been developed may miss as many as 1 of every 10 newborns who has SBI.⁴³ Initial management of all infants 28 days old or younger with fever must therefore include a full work-up, including lumbar puncture and empiric antibiotics.⁴⁴

Evaluation. When an infant younger than 28 days has a fever, the physician should first verify that the temperature was taken rectally and how it was documented. In an infant who has a history of prematurity, it is crucial to correct for chronological age when deciding on proper evaluation.

Additional important findings in the history include a significant change in behavior, associated symptoms, and exposure to sick contacts. The maternal and birth history, including prolonged rupture of membranes, colonization with group B Streptococcus, administration of antibiotics at delivery, and genital herpes simplex virus (HSV) infection may suggest a cause for fever.⁴⁵

The evaluation of fever might include the white blood cell (WBC) count, blood culture, measurement of markers of inflammation, urine studies, lumbar puncture, stool culture, and chest radiograph. Traditionally, the WBC count has been utilized as a standard marker for sepsis, although it has a low sensitivity and specificity for SBI, especially in newborns.⁴⁶ Blood cultures should be obtained routinely in the newborn with fever, and before antibiotics are administered in older infants.

Procalcitonin (PCT; a calcitonin precursor) and the inflammatory marker C-reactive protein (CRP) have been shown, in several large studies, to have relatively high sensitivity and specificity for SBI; measurement of these constituents may enhance detection of serious illness.^46-49 In a large study of 2047 febrile infants older than 30 months, the PCT level was determined to be more accurate than the CRP level, the WBC count, and the absolute neutrophil count in predicting SBI.^48,49 PCT shows the most promise for preventing a full fever work-up and empiric antibiotics. It has not yet been widely translated into practice, however, because of a lack of clear guidance on how to combine PCT levels with other laboratory markers and clinical decision-making.^48-50

Urinalysis (UA) should be obtained for all newborns who present with fever. Traditionally, it was recommended that urine should be cultured for all newborns with fever; however, more recent data show that the initial urinalysis is much more sensitive than once thought. In a study, UA was positive (defined as pyuria or a positive leukocyte esterase test, or both) in all but 1 of 203 infants who had bacteremic UTI (sensitivity, 99.5%).⁵¹

The procalcitonin level was determined to be more accurate than the C-reactive protein level, the white blood cell count, and the absolute neutrophil count in predicting serious bacterial illness.

Stool culture is necessary in newborns only when they present with blood or mucus in diarrhea. Lumbar puncture should be performed in all febrile newborns and all newborns for whom empiric antibiotics have been prescribed.^43,44 A chest radiograph may be useful in diagnosis when a newborn has any other sign of pulmonary disease: respiratory rate >50/min, retractions, wheezing, grunting, stridor, nasal flaring, cough, and positive findings on lung examination.^43,44

Treatment. Management for all newborns who have a rectal temperature ≥38° C includes admission to the hospital and empiric antibiotics; guidance is based primarily on expert consensus. Common pathogens for SBI include group B Strep, Escherichia coli, Enterococcus spp., and Listeria monocytogenes.^43,44 Empiric antibiotics, including ampicillin (to cover L monocytogenes) and cefotaxime or gentamicin should be started immediately after sending for blood, urine, and cerebrospinal fluid (CSF) cultures.^43-45

Management for all newborns who have rectal temperature ≥38° C includes admission to the hospital and empiric antibiotics.

All infants who are ill-appearing or have vesicles, seizures, or a maternal history of genital HSV infection should also be started on empiric acyclovir. Vesicles should be cultured and CSF should be sent for HSV DNA polymerase chain reaction before acyclovir is administered.^43-45

Sudden infant death syndrome: Steps to take to minimize risk

SIDS is defined as the sudden death of a child younger than 1 year that remains unexplained after a thorough case investigation and comprehensive review of the clinical history. The risk of SIDS in the United States is less than 1 for every 1000 live births; incidence peaks between 2 and 4 months of age.⁵² In the United States, SIDS and other sleep-related infant deaths, such as strangulation in bed or accidental suffocation, account for more than 4000 deaths a year.⁵³ The incidence of SIDS declined markedly after the “Back to Sleep” campaign was launched in 2003, but has leveled off since 2005.^53-55

Numerous risk factors for SIDS have been identified, including maternal factors (young maternal age, maternal smoking during pregnancy, late or no prenatal care) and infant and environmental factors (prematurity, low birth weight, male gender, prone sleeping position, sleeping on a soft surface or with bedding accessories, bed-sharing (ie, sleeping in the parents’ bed), and overheating. In many cases, the risk factors are modifiable; sleeping in the prone position is the most meaningful modifiable risk factor.

Home monitors have not been proven to reduce the incidence of SIDS and are not recommended for that purpose.

To minimize the risk for SIDS, parents should be educated on the risk factors—prenatally as well as at each infant well visit. Home monitors have not been proven to reduce the incidence of SIDS and are not recommended for that purpose.^54-57 Although evidence is strongest for supine positioning as a preventive intervention for SIDS, other evidence-based recommendations include use of a firm sleep surface; breastfeeding; use of a pacifier; room-sharing with parents without bed-sharing; routine immunization; avoidance of overheating; avoiding falling asleep with the infant on a chair or couch; and avoiding exposure to tobacco smoke, alcohol, and drugs of abuse.^55,56 A recent systematic review showed that large-scale community interventions and education campaigns can play a significant role in parental and community adoption of safe sleep recommendations; however, families and communities rarely exhibit complete adherence to safe sleep practices.⁵⁷

Other concerns in the first month of life and immediately beyond

In TABLE 5,² we list additional common newborn problems not reviewed in the text of this article and summarize evidence-based treatment strategies.

CORRESPONDENCE
Scott Hartman, MD, Associate Professor, Department of Family Medicine, University of Rochester Medical Center, 777 South Clinton Avenue, Rochester, NY 14620; [email protected].

Acknowledgement
We thank Nancy Phillips for her assistance in the preparation of this article.

An estimated 39.4 million US adults suffer from persistent pain,¹ and the National Institutes of Health indicate that pain affects more Americans than diabetes, heart disease, and cancer combined.²

As physicians, we know that conventional options to manage chronic pain leave much to be desired and that more evidence-based treatment options are sorely needed. Patients know this, too, and turn to complementary therapies for pain more than for any other diagnosis.³

Case in point: The use of acupuncture is growing. Its use in the United States tripled between 1997 and 2007.⁴ In addition, the research base for acupuncture is rapidly expanding. From 1991 to 2009, nearly 4000 acupuncture research studies were published, with studies on pain accounting for 41% of the acupuncture literature.⁴

But acupuncture is not without controversy. This is due to a lack of a universally accepted biologic mechanism, theories of use and efficacy based in an alternative medical system (traditional Chinese medicine [TCM]), and conflicting views of the evidence.

This article will help make sense of this growing body of knowledge by summarizing the latest evidence and addressing 7 common questions about acupuncture for pain conditions. Applying this information will give you the confidence to counsel patients appropriately and decide if acupuncture fits within their pain management plan.

1. What is acupuncture and how does it work?

Acupuncture, which has a 2000-year history of use, involves inserting needles at various points throughout the body to promote healing and improve function. Although acupuncture represents one piece of TCM (which is a holistic system that also includes herbal medicine, nutrition, meditation, and movement), it is often offered as an independent therapy.

Acupuncture point locations are determined either by using an underlying theoretical framework, such as TCM, or by using anatomic structures, such as muscular trigger points. Providers today often employ a hybrid approach when delivering acupuncture treatment. That is, practitioners may choose point locations based on TCM, but they may combine the practice with local treatments that are based on current knowledge of anatomy. For example, a patient presenting with low back pain may be treated utilizing traditional points located near the ankle and knee, and also by needling active trigger points in the quadratus lumborum muscle.

The mechanism of action. One of the reasons for the continuing controversy surrounding acupuncture is the lack of a clear understanding of its underlying mechanism of action. For centuries the “how” of acupuncture has been explained in poetic terms such as yin, yang, and qi. Only in the past half-century have we begun investigating the potential biologic mechanisms responsible for the physiologic effects seen with acupuncture treatment.

While research has uncovered several interesting theories, how these mechanisms interact to produce therapeutic effects is still unclear. However, looking at various components of the nervous system helps to provide some insight.

Consider the nervous system. One way to conceptualize the mechanisms of acupuncture is to consider the various levels of the nervous system and how each level is affected. In the central nervous system, needling an acupuncture point stimulates the natural endorphin system, altering the pain sensation.⁵ This effect is reversible with naloxone in animal models, indicating that blocking the endorphin system interferes with the analgesic benefits of acupuncture.⁵

Serotonergic systems are also involved centrally. Functional magnetic resonance imaging studies have shown that needling specific acupuncture points modulates areas of the brain.

In the spinal cord, the gate control theory is believed to play a role. (The gate control theory puts forth that nonpainful input closes the “gates” to painful input, which prevents pain sensations from traveling to the central nervous system.) Modulation of sensory input occurs at the level of the dorsal horn of the spinal cord during an acupuncture treatment, which can affect the physiologic pain response.⁶ Opioid receptors are also affected at the spinal cord level.⁷

Acupuncture use in the United States tripled between 1997 and 2007.

Lastly, multiple chemicals released peripherally, including interleukins, substance P, and adenosine, appear to contribute to acupuncture’s analgesia.⁶ We know this because a local anesthetic injected around a peripheral nerve at an acupoint blocks the analgesic effect of acupuncture.⁸ Taken together, acupuncture treatment produces physiologic changes in the brain, spinal cord, and at the periphery, making it a truly unique therapeutic modality.

2. Is acupuncture an effective treatment for pain?

Yes, but before we look at the individual studies, it is important to mention some of the shortcomings of the research to date. First, acupuncture trials lack a standard sham control intervention. Some sham treatments involve skin penetration, while others do not. This has led to controversy regarding whether the sham interventions themselves are physiologically active, thus lessening the magnitude of effect for acupuncture. This is a point of contention in the acupuncture literature and a factor to consider when deciding if results have clinical significance.

In addition, the acupuncturist providing treatment in a trial is typically unblinded. This is also true of trials measuring other physical modalities, but it contributes to the debate surrounding the magnitude of placebo response in acupuncture studies.

Finally, many randomized trials involving acupuncture have had low methodologic quality. Fortunately, there are now several high-quality systematic reviews that have attempted to filter out the lower-quality research and provide a better representation of the evidence (TABLE^9-14). A discussion of them follows.

General chronic pain. A 2012 meta-analysis¹⁵ evaluated the effectiveness of acupuncture for the treatment of chronic pain with one of 4 etiologies: nonspecific back or neck pain, chronic headache, osteoarthritis, and shoulder pain. This analysis attempted to control for the high variability of study quality in the acupuncture literature by including only studies of high methodologic character. The final analysis included 29 randomized controlled trials (N=17,922). The authors concluded that acupuncture was superior to both no acupuncture and sham (placebo) acupuncture for all pain conditions in the study. The average effect size was 0.5 standard deviations on a 10-point scale. The authors considered this to be clinically relevant, although the magnitude of benefit was modest.¹⁵

Low back pain. A 2017 systematic review by Chou et al⁹ evaluated 32 trials (N=5931) reviewing acupuncture for the treatment of chronic low back pain. This review found acupuncture was associated with lower pain intensity and improved function in the short term when compared with no treatment. And while acupuncture was associated with lower pain intensity when compared with a sham control, there was no difference in function between the 2 groups. Of note, 3 of the included trials compared acupuncture to standard medications used in the treatment of low back pain and found acupuncture to be superior in terms of both pain reduction and improved function.⁹

Trials that compared acupuncture to another active therapy have found that it often has fewer adverse effects.

The authors of a 2008 systematic review that evaluated 23 trials (N=6359)¹⁰ similarly concluded that there is moderate evidence for the use of acupuncture (compared to no treatment) for the treatment of nonspecific low back pain, but did not find evidence that acupuncture was superior to sham controls.¹⁰ The 2017 American College of Physicians clinical practice guidelines support the use of acupuncture for the treatment of chronic low back pain.¹⁶

In addition to helping with chronic low back pain, acupuncture is also showing promise as a treatment for acute spinal pain. A 2013 systematic review (11 trials, N=1139) showed that acupuncture may be more effective than nonsteroidal anti-inflammatory drugs (NSAIDs) in treating acute low back pain and may cause fewer adverse effects.¹⁷

Headache pain. Evidence favoring acupuncture in the management of headache has been fairly consistent over the past decade. An updated Cochrane review on the prevention of migraine headaches was published in 2016.¹¹ Acupuncture was compared with no treatment in 4 trials (n=2199). The authors found moderate quality evidence that acupuncture reduces headache frequency (number needed to treat=4). Acupuncture achieved at least 50% headache reduction in 41% vs 17% in the groups that received no acupuncture. When compared with sham control groups (10 trials, n=1534), acupuncture demonstrated a small but statistically significant improvement in headache frequency. Three trials (n=744) compared acupuncture to medication prophylaxis for migraine headaches and found acupuncture had similar effectiveness with fewer adverse effects.¹¹

Osteoarthritis (OA). Most studies have focused on OA of the knee, and, thus far, have generated conflicting results. A Cochrane review in 2010 included 4 trials (n=884) that had a wait list control and 9 trials (n=1835) that compared acupuncture to a sham control.¹² When compared to a wait list control, acupuncture resulted in statistically significant and clinically relevant improvement in pain and function. However, when compared to sham treatment for OA, the review showed statistically significant improvement in pain and function for acupuncture that was unlikely to be clinically relevant.¹²

A more recent meta-analysis in 2016 evaluated 10 trials (N=2007) investigating acupuncture in the treatment of knee OA.¹³ The authors found acupuncture improved both pain and functional outcome measures when compared with either no treatment or a sham control.

Fibromyalgia. Systematic reviews in 2007 (5 trials, N=316)¹⁸ and 2010 (7 trials, N=385)¹⁹ showed that acupuncture did provide short-term pain relief in patients with fibromyalgia, but that the effect was not sustained at follow-up.These reviews were limited by a high risk of bias, which was noted in the studies. The authors of both reviews concluded that acupuncture could not be recommended for the treatment of fibromyalgia.

A more recent Cochrane review published in 2013 (9 trials, N=395) offered low- to moderate-level evidence of benefit for acupuncture compared with no treatment at one month follow-up.¹⁴ Of note, there was also evidence of benefit in improved sleep and global well-being, in addition to pain and stiffness measures in this review. The overall magnitude of benefit was small, but clinically significant. Acupuncture also has evidence of benefit in the treatment of conditions commonly seen in conjunction with fibromyalgia, including headaches and low back pain as described earlier.

3. What does a typical acupuncture treatment entail?

In a typical treatment, anywhere from about 5 to 20 needles are inserted into the body. Common areas of needling include the arms and legs, especially below the elbows and knees. Other frequently used areas are the scalp, ears, and structures related to the painful condition.

The needles used are very thin (typically smaller than a 30-gauge needle) and do not have a beveled tip like phlebotomy needles do. Most patients have minimal pain as the needles are inserted. During the treatment, the needles may be left alone or they may be heated or stimulated electrically. An average treatment lasts 30 to 40 minutes; many patients find the sessions relaxing.

4. Are there any adverse effects or complications of treatment?

Acupuncture is generally considered a safe therapy, with most patients experiencing no adverse effects at all. Minor adverse effects can include post-treatment fatigue, minor bruising, or vasovagal reactions from the insertion of the needles. Serious complications, such as pneumothorax, are possible, but are considered rare.²⁰ A 2004 study estimated the incidence of severe complications to be .05 per 10,000 acupuncture treatments.²¹

Infections are also possible, but most reported cases were due to practitioners reusing needles.²² The standard of care in the United States is to use only sterilized, single-use needles. With this practice, infections due to acupuncture are thought to be rare.

Of note, trials that compare acupuncture to another active therapy often find that acupuncture has fewer adverse effects. This has been the case when acupuncture was compared to NSAIDs for low back pain and to topiramate for headaches.^17,23

5. How does acupuncture fit into a patient’s treatment?

The simple answer is that it is often most effectively used as part of a comprehensive management plan for chronic pain.

As our understanding of the complexity of chronic pain deepens, our therapeutic armamentarium for the management of chronic pain needs to broaden. This was summed up well in a 2016 article on the multimodal management of chronic pain when the authors stated, “Many targets need more than one arrow.”²⁴ Effective management of chronic pain involves addressing psychosocial and lifestyle factors in a patient-centered way and finding a combination of treatments that most effectively leads to improved coping and function.

It’s important to note that like medications and injections, acupuncture is a passive therapy. Although there is evidence for efficacy of improved pain with acupuncture in certain conditions, it should be combined with treatments that actively engage patients, such as exercise, behavioral treatments, development of coping skills, sleep hygiene, and educational strategies.

6. To whom do I refer patients for acupuncture treatment?

In the United States, licensed acupuncturists and physicians most commonly perform acupuncture. There are more than 50 schools that train licensed acupuncturists in the United States, and it usually takes 3 years to meet the requirements.²⁵

Physicians are often trained through continuing medical education (CME) programs that take several months to complete. These programs often combine live lectures, distance learning, and hands-on training and are typically sponsored by a university. Most require 300 hours of CME to complete. Licensure varies by state, but in many states, having an MD or DO degree automatically allows physicians to practice acupuncture. (See “Online resources,” above for links to Web sites that can be useful in finding qualified acupuncturists in your area.)

7. Is acupuncture covered by insurance?

Patients can expect to pay $75 to $150 for an acupuncture session.

It depends. Insurance coverage of acupuncture is highly variable and based on region and insurance type. Medicare and Medicaid plans do not pay for acupuncture. There are some private insurance plans that do. If covered, there may be limitations regarding diagnosis, number of visits, or provider. It is best for patients to call their insurance plan directly to inquire about coverage and any limitations. If paying out of pocket, patients can expect to pay $75 to $150 per treatment session.

CORRESPONDENCE
Russell Lemmon, DO, 1100 Delaplaine Court, Madison, WI 53715; [email protected].

An estimated 39.4 million US adults suffer from persistent pain,¹ and the National Institutes of Health indicate that pain affects more Americans than diabetes, heart disease, and cancer combined.²

As physicians, we know that conventional options to manage chronic pain leave much to be desired and that more evidence-based treatment options are sorely needed. Patients know this, too, and turn to complementary therapies for pain more than for any other diagnosis.³

Case in point: The use of acupuncture is growing. Its use in the United States tripled between 1997 and 2007.⁴ In addition, the research base for acupuncture is rapidly expanding. From 1991 to 2009, nearly 4000 acupuncture research studies were published, with studies on pain accounting for 41% of the acupuncture literature.⁴

But acupuncture is not without controversy. This is due to a lack of a universally accepted biologic mechanism, theories of use and efficacy based in an alternative medical system (traditional Chinese medicine [TCM]), and conflicting views of the evidence.

This article will help make sense of this growing body of knowledge by summarizing the latest evidence and addressing 7 common questions about acupuncture for pain conditions. Applying this information will give you the confidence to counsel patients appropriately and decide if acupuncture fits within their pain management plan.

1. What is acupuncture and how does it work?

Acupuncture, which has a 2000-year history of use, involves inserting needles at various points throughout the body to promote healing and improve function. Although acupuncture represents one piece of TCM (which is a holistic system that also includes herbal medicine, nutrition, meditation, and movement), it is often offered as an independent therapy.

Acupuncture point locations are determined either by using an underlying theoretical framework, such as TCM, or by using anatomic structures, such as muscular trigger points. Providers today often employ a hybrid approach when delivering acupuncture treatment. That is, practitioners may choose point locations based on TCM, but they may combine the practice with local treatments that are based on current knowledge of anatomy. For example, a patient presenting with low back pain may be treated utilizing traditional points located near the ankle and knee, and also by needling active trigger points in the quadratus lumborum muscle.

The mechanism of action. One of the reasons for the continuing controversy surrounding acupuncture is the lack of a clear understanding of its underlying mechanism of action. For centuries the “how” of acupuncture has been explained in poetic terms such as yin, yang, and qi. Only in the past half-century have we begun investigating the potential biologic mechanisms responsible for the physiologic effects seen with acupuncture treatment.

While research has uncovered several interesting theories, how these mechanisms interact to produce therapeutic effects is still unclear. However, looking at various components of the nervous system helps to provide some insight.

Consider the nervous system. One way to conceptualize the mechanisms of acupuncture is to consider the various levels of the nervous system and how each level is affected. In the central nervous system, needling an acupuncture point stimulates the natural endorphin system, altering the pain sensation.⁵ This effect is reversible with naloxone in animal models, indicating that blocking the endorphin system interferes with the analgesic benefits of acupuncture.⁵

Serotonergic systems are also involved centrally. Functional magnetic resonance imaging studies have shown that needling specific acupuncture points modulates areas of the brain.

In the spinal cord, the gate control theory is believed to play a role. (The gate control theory puts forth that nonpainful input closes the “gates” to painful input, which prevents pain sensations from traveling to the central nervous system.) Modulation of sensory input occurs at the level of the dorsal horn of the spinal cord during an acupuncture treatment, which can affect the physiologic pain response.⁶ Opioid receptors are also affected at the spinal cord level.⁷

Acupuncture use in the United States tripled between 1997 and 2007.

Lastly, multiple chemicals released peripherally, including interleukins, substance P, and adenosine, appear to contribute to acupuncture’s analgesia.⁶ We know this because a local anesthetic injected around a peripheral nerve at an acupoint blocks the analgesic effect of acupuncture.⁸ Taken together, acupuncture treatment produces physiologic changes in the brain, spinal cord, and at the periphery, making it a truly unique therapeutic modality.

2. Is acupuncture an effective treatment for pain?

Yes, but before we look at the individual studies, it is important to mention some of the shortcomings of the research to date. First, acupuncture trials lack a standard sham control intervention. Some sham treatments involve skin penetration, while others do not. This has led to controversy regarding whether the sham interventions themselves are physiologically active, thus lessening the magnitude of effect for acupuncture. This is a point of contention in the acupuncture literature and a factor to consider when deciding if results have clinical significance.

In addition, the acupuncturist providing treatment in a trial is typically unblinded. This is also true of trials measuring other physical modalities, but it contributes to the debate surrounding the magnitude of placebo response in acupuncture studies.

Finally, many randomized trials involving acupuncture have had low methodologic quality. Fortunately, there are now several high-quality systematic reviews that have attempted to filter out the lower-quality research and provide a better representation of the evidence (TABLE^9-14). A discussion of them follows.

General chronic pain. A 2012 meta-analysis¹⁵ evaluated the effectiveness of acupuncture for the treatment of chronic pain with one of 4 etiologies: nonspecific back or neck pain, chronic headache, osteoarthritis, and shoulder pain. This analysis attempted to control for the high variability of study quality in the acupuncture literature by including only studies of high methodologic character. The final analysis included 29 randomized controlled trials (N=17,922). The authors concluded that acupuncture was superior to both no acupuncture and sham (placebo) acupuncture for all pain conditions in the study. The average effect size was 0.5 standard deviations on a 10-point scale. The authors considered this to be clinically relevant, although the magnitude of benefit was modest.¹⁵

Low back pain. A 2017 systematic review by Chou et al⁹ evaluated 32 trials (N=5931) reviewing acupuncture for the treatment of chronic low back pain. This review found acupuncture was associated with lower pain intensity and improved function in the short term when compared with no treatment. And while acupuncture was associated with lower pain intensity when compared with a sham control, there was no difference in function between the 2 groups. Of note, 3 of the included trials compared acupuncture to standard medications used in the treatment of low back pain and found acupuncture to be superior in terms of both pain reduction and improved function.⁹

Trials that compared acupuncture to another active therapy have found that it often has fewer adverse effects.

The authors of a 2008 systematic review that evaluated 23 trials (N=6359)¹⁰ similarly concluded that there is moderate evidence for the use of acupuncture (compared to no treatment) for the treatment of nonspecific low back pain, but did not find evidence that acupuncture was superior to sham controls.¹⁰ The 2017 American College of Physicians clinical practice guidelines support the use of acupuncture for the treatment of chronic low back pain.¹⁶

In addition to helping with chronic low back pain, acupuncture is also showing promise as a treatment for acute spinal pain. A 2013 systematic review (11 trials, N=1139) showed that acupuncture may be more effective than nonsteroidal anti-inflammatory drugs (NSAIDs) in treating acute low back pain and may cause fewer adverse effects.¹⁷

Headache pain. Evidence favoring acupuncture in the management of headache has been fairly consistent over the past decade. An updated Cochrane review on the prevention of migraine headaches was published in 2016.¹¹ Acupuncture was compared with no treatment in 4 trials (n=2199). The authors found moderate quality evidence that acupuncture reduces headache frequency (number needed to treat=4). Acupuncture achieved at least 50% headache reduction in 41% vs 17% in the groups that received no acupuncture. When compared with sham control groups (10 trials, n=1534), acupuncture demonstrated a small but statistically significant improvement in headache frequency. Three trials (n=744) compared acupuncture to medication prophylaxis for migraine headaches and found acupuncture had similar effectiveness with fewer adverse effects.¹¹

Osteoarthritis (OA). Most studies have focused on OA of the knee, and, thus far, have generated conflicting results. A Cochrane review in 2010 included 4 trials (n=884) that had a wait list control and 9 trials (n=1835) that compared acupuncture to a sham control.¹² When compared to a wait list control, acupuncture resulted in statistically significant and clinically relevant improvement in pain and function. However, when compared to sham treatment for OA, the review showed statistically significant improvement in pain and function for acupuncture that was unlikely to be clinically relevant.¹²

A more recent meta-analysis in 2016 evaluated 10 trials (N=2007) investigating acupuncture in the treatment of knee OA.¹³ The authors found acupuncture improved both pain and functional outcome measures when compared with either no treatment or a sham control.

Fibromyalgia. Systematic reviews in 2007 (5 trials, N=316)¹⁸ and 2010 (7 trials, N=385)¹⁹ showed that acupuncture did provide short-term pain relief in patients with fibromyalgia, but that the effect was not sustained at follow-up.These reviews were limited by a high risk of bias, which was noted in the studies. The authors of both reviews concluded that acupuncture could not be recommended for the treatment of fibromyalgia.

A more recent Cochrane review published in 2013 (9 trials, N=395) offered low- to moderate-level evidence of benefit for acupuncture compared with no treatment at one month follow-up.¹⁴ Of note, there was also evidence of benefit in improved sleep and global well-being, in addition to pain and stiffness measures in this review. The overall magnitude of benefit was small, but clinically significant. Acupuncture also has evidence of benefit in the treatment of conditions commonly seen in conjunction with fibromyalgia, including headaches and low back pain as described earlier.

3. What does a typical acupuncture treatment entail?

In a typical treatment, anywhere from about 5 to 20 needles are inserted into the body. Common areas of needling include the arms and legs, especially below the elbows and knees. Other frequently used areas are the scalp, ears, and structures related to the painful condition.

The needles used are very thin (typically smaller than a 30-gauge needle) and do not have a beveled tip like phlebotomy needles do. Most patients have minimal pain as the needles are inserted. During the treatment, the needles may be left alone or they may be heated or stimulated electrically. An average treatment lasts 30 to 40 minutes; many patients find the sessions relaxing.

4. Are there any adverse effects or complications of treatment?

Acupuncture is generally considered a safe therapy, with most patients experiencing no adverse effects at all. Minor adverse effects can include post-treatment fatigue, minor bruising, or vasovagal reactions from the insertion of the needles. Serious complications, such as pneumothorax, are possible, but are considered rare.²⁰ A 2004 study estimated the incidence of severe complications to be .05 per 10,000 acupuncture treatments.²¹

Infections are also possible, but most reported cases were due to practitioners reusing needles.²² The standard of care in the United States is to use only sterilized, single-use needles. With this practice, infections due to acupuncture are thought to be rare.

Of note, trials that compare acupuncture to another active therapy often find that acupuncture has fewer adverse effects. This has been the case when acupuncture was compared to NSAIDs for low back pain and to topiramate for headaches.^17,23

5. How does acupuncture fit into a patient’s treatment?

The simple answer is that it is often most effectively used as part of a comprehensive management plan for chronic pain.

As our understanding of the complexity of chronic pain deepens, our therapeutic armamentarium for the management of chronic pain needs to broaden. This was summed up well in a 2016 article on the multimodal management of chronic pain when the authors stated, “Many targets need more than one arrow.”²⁴ Effective management of chronic pain involves addressing psychosocial and lifestyle factors in a patient-centered way and finding a combination of treatments that most effectively leads to improved coping and function.

It’s important to note that like medications and injections, acupuncture is a passive therapy. Although there is evidence for efficacy of improved pain with acupuncture in certain conditions, it should be combined with treatments that actively engage patients, such as exercise, behavioral treatments, development of coping skills, sleep hygiene, and educational strategies.

6. To whom do I refer patients for acupuncture treatment?

In the United States, licensed acupuncturists and physicians most commonly perform acupuncture. There are more than 50 schools that train licensed acupuncturists in the United States, and it usually takes 3 years to meet the requirements.²⁵

Physicians are often trained through continuing medical education (CME) programs that take several months to complete. These programs often combine live lectures, distance learning, and hands-on training and are typically sponsored by a university. Most require 300 hours of CME to complete. Licensure varies by state, but in many states, having an MD or DO degree automatically allows physicians to practice acupuncture. (See “Online resources,” above for links to Web sites that can be useful in finding qualified acupuncturists in your area.)

7. Is acupuncture covered by insurance?

Patients can expect to pay $75 to $150 for an acupuncture session.

It depends. Insurance coverage of acupuncture is highly variable and based on region and insurance type. Medicare and Medicaid plans do not pay for acupuncture. There are some private insurance plans that do. If covered, there may be limitations regarding diagnosis, number of visits, or provider. It is best for patients to call their insurance plan directly to inquire about coverage and any limitations. If paying out of pocket, patients can expect to pay $75 to $150 per treatment session.

CORRESPONDENCE
Russell Lemmon, DO, 1100 Delaplaine Court, Madison, WI 53715; [email protected].

An estimated 39.4 million US adults suffer from persistent pain,¹ and the National Institutes of Health indicate that pain affects more Americans than diabetes, heart disease, and cancer combined.²

As physicians, we know that conventional options to manage chronic pain leave much to be desired and that more evidence-based treatment options are sorely needed. Patients know this, too, and turn to complementary therapies for pain more than for any other diagnosis.³

Case in point: The use of acupuncture is growing. Its use in the United States tripled between 1997 and 2007.⁴ In addition, the research base for acupuncture is rapidly expanding. From 1991 to 2009, nearly 4000 acupuncture research studies were published, with studies on pain accounting for 41% of the acupuncture literature.⁴

But acupuncture is not without controversy. This is due to a lack of a universally accepted biologic mechanism, theories of use and efficacy based in an alternative medical system (traditional Chinese medicine [TCM]), and conflicting views of the evidence.

This article will help make sense of this growing body of knowledge by summarizing the latest evidence and addressing 7 common questions about acupuncture for pain conditions. Applying this information will give you the confidence to counsel patients appropriately and decide if acupuncture fits within their pain management plan.

1. What is acupuncture and how does it work?

Acupuncture, which has a 2000-year history of use, involves inserting needles at various points throughout the body to promote healing and improve function. Although acupuncture represents one piece of TCM (which is a holistic system that also includes herbal medicine, nutrition, meditation, and movement), it is often offered as an independent therapy.

Acupuncture point locations are determined either by using an underlying theoretical framework, such as TCM, or by using anatomic structures, such as muscular trigger points. Providers today often employ a hybrid approach when delivering acupuncture treatment. That is, practitioners may choose point locations based on TCM, but they may combine the practice with local treatments that are based on current knowledge of anatomy. For example, a patient presenting with low back pain may be treated utilizing traditional points located near the ankle and knee, and also by needling active trigger points in the quadratus lumborum muscle.

The mechanism of action. One of the reasons for the continuing controversy surrounding acupuncture is the lack of a clear understanding of its underlying mechanism of action. For centuries the “how” of acupuncture has been explained in poetic terms such as yin, yang, and qi. Only in the past half-century have we begun investigating the potential biologic mechanisms responsible for the physiologic effects seen with acupuncture treatment.

While research has uncovered several interesting theories, how these mechanisms interact to produce therapeutic effects is still unclear. However, looking at various components of the nervous system helps to provide some insight.

Consider the nervous system. One way to conceptualize the mechanisms of acupuncture is to consider the various levels of the nervous system and how each level is affected. In the central nervous system, needling an acupuncture point stimulates the natural endorphin system, altering the pain sensation.⁵ This effect is reversible with naloxone in animal models, indicating that blocking the endorphin system interferes with the analgesic benefits of acupuncture.⁵

Serotonergic systems are also involved centrally. Functional magnetic resonance imaging studies have shown that needling specific acupuncture points modulates areas of the brain.

In the spinal cord, the gate control theory is believed to play a role. (The gate control theory puts forth that nonpainful input closes the “gates” to painful input, which prevents pain sensations from traveling to the central nervous system.) Modulation of sensory input occurs at the level of the dorsal horn of the spinal cord during an acupuncture treatment, which can affect the physiologic pain response.⁶ Opioid receptors are also affected at the spinal cord level.⁷

Acupuncture use in the United States tripled between 1997 and 2007.

Lastly, multiple chemicals released peripherally, including interleukins, substance P, and adenosine, appear to contribute to acupuncture’s analgesia.⁶ We know this because a local anesthetic injected around a peripheral nerve at an acupoint blocks the analgesic effect of acupuncture.⁸ Taken together, acupuncture treatment produces physiologic changes in the brain, spinal cord, and at the periphery, making it a truly unique therapeutic modality.

2. Is acupuncture an effective treatment for pain?

Yes, but before we look at the individual studies, it is important to mention some of the shortcomings of the research to date. First, acupuncture trials lack a standard sham control intervention. Some sham treatments involve skin penetration, while others do not. This has led to controversy regarding whether the sham interventions themselves are physiologically active, thus lessening the magnitude of effect for acupuncture. This is a point of contention in the acupuncture literature and a factor to consider when deciding if results have clinical significance.

In addition, the acupuncturist providing treatment in a trial is typically unblinded. This is also true of trials measuring other physical modalities, but it contributes to the debate surrounding the magnitude of placebo response in acupuncture studies.

Finally, many randomized trials involving acupuncture have had low methodologic quality. Fortunately, there are now several high-quality systematic reviews that have attempted to filter out the lower-quality research and provide a better representation of the evidence (TABLE^9-14). A discussion of them follows.

General chronic pain. A 2012 meta-analysis¹⁵ evaluated the effectiveness of acupuncture for the treatment of chronic pain with one of 4 etiologies: nonspecific back or neck pain, chronic headache, osteoarthritis, and shoulder pain. This analysis attempted to control for the high variability of study quality in the acupuncture literature by including only studies of high methodologic character. The final analysis included 29 randomized controlled trials (N=17,922). The authors concluded that acupuncture was superior to both no acupuncture and sham (placebo) acupuncture for all pain conditions in the study. The average effect size was 0.5 standard deviations on a 10-point scale. The authors considered this to be clinically relevant, although the magnitude of benefit was modest.¹⁵

Low back pain. A 2017 systematic review by Chou et al⁹ evaluated 32 trials (N=5931) reviewing acupuncture for the treatment of chronic low back pain. This review found acupuncture was associated with lower pain intensity and improved function in the short term when compared with no treatment. And while acupuncture was associated with lower pain intensity when compared with a sham control, there was no difference in function between the 2 groups. Of note, 3 of the included trials compared acupuncture to standard medications used in the treatment of low back pain and found acupuncture to be superior in terms of both pain reduction and improved function.⁹

Trials that compared acupuncture to another active therapy have found that it often has fewer adverse effects.

The authors of a 2008 systematic review that evaluated 23 trials (N=6359)¹⁰ similarly concluded that there is moderate evidence for the use of acupuncture (compared to no treatment) for the treatment of nonspecific low back pain, but did not find evidence that acupuncture was superior to sham controls.¹⁰ The 2017 American College of Physicians clinical practice guidelines support the use of acupuncture for the treatment of chronic low back pain.¹⁶

In addition to helping with chronic low back pain, acupuncture is also showing promise as a treatment for acute spinal pain. A 2013 systematic review (11 trials, N=1139) showed that acupuncture may be more effective than nonsteroidal anti-inflammatory drugs (NSAIDs) in treating acute low back pain and may cause fewer adverse effects.¹⁷

Headache pain. Evidence favoring acupuncture in the management of headache has been fairly consistent over the past decade. An updated Cochrane review on the prevention of migraine headaches was published in 2016.¹¹ Acupuncture was compared with no treatment in 4 trials (n=2199). The authors found moderate quality evidence that acupuncture reduces headache frequency (number needed to treat=4). Acupuncture achieved at least 50% headache reduction in 41% vs 17% in the groups that received no acupuncture. When compared with sham control groups (10 trials, n=1534), acupuncture demonstrated a small but statistically significant improvement in headache frequency. Three trials (n=744) compared acupuncture to medication prophylaxis for migraine headaches and found acupuncture had similar effectiveness with fewer adverse effects.¹¹

Osteoarthritis (OA). Most studies have focused on OA of the knee, and, thus far, have generated conflicting results. A Cochrane review in 2010 included 4 trials (n=884) that had a wait list control and 9 trials (n=1835) that compared acupuncture to a sham control.¹² When compared to a wait list control, acupuncture resulted in statistically significant and clinically relevant improvement in pain and function. However, when compared to sham treatment for OA, the review showed statistically significant improvement in pain and function for acupuncture that was unlikely to be clinically relevant.¹²

A more recent meta-analysis in 2016 evaluated 10 trials (N=2007) investigating acupuncture in the treatment of knee OA.¹³ The authors found acupuncture improved both pain and functional outcome measures when compared with either no treatment or a sham control.

Fibromyalgia. Systematic reviews in 2007 (5 trials, N=316)¹⁸ and 2010 (7 trials, N=385)¹⁹ showed that acupuncture did provide short-term pain relief in patients with fibromyalgia, but that the effect was not sustained at follow-up.These reviews were limited by a high risk of bias, which was noted in the studies. The authors of both reviews concluded that acupuncture could not be recommended for the treatment of fibromyalgia.

A more recent Cochrane review published in 2013 (9 trials, N=395) offered low- to moderate-level evidence of benefit for acupuncture compared with no treatment at one month follow-up.¹⁴ Of note, there was also evidence of benefit in improved sleep and global well-being, in addition to pain and stiffness measures in this review. The overall magnitude of benefit was small, but clinically significant. Acupuncture also has evidence of benefit in the treatment of conditions commonly seen in conjunction with fibromyalgia, including headaches and low back pain as described earlier.

3. What does a typical acupuncture treatment entail?

In a typical treatment, anywhere from about 5 to 20 needles are inserted into the body. Common areas of needling include the arms and legs, especially below the elbows and knees. Other frequently used areas are the scalp, ears, and structures related to the painful condition.

The needles used are very thin (typically smaller than a 30-gauge needle) and do not have a beveled tip like phlebotomy needles do. Most patients have minimal pain as the needles are inserted. During the treatment, the needles may be left alone or they may be heated or stimulated electrically. An average treatment lasts 30 to 40 minutes; many patients find the sessions relaxing.

4. Are there any adverse effects or complications of treatment?

Acupuncture is generally considered a safe therapy, with most patients experiencing no adverse effects at all. Minor adverse effects can include post-treatment fatigue, minor bruising, or vasovagal reactions from the insertion of the needles. Serious complications, such as pneumothorax, are possible, but are considered rare.²⁰ A 2004 study estimated the incidence of severe complications to be .05 per 10,000 acupuncture treatments.²¹

Infections are also possible, but most reported cases were due to practitioners reusing needles.²² The standard of care in the United States is to use only sterilized, single-use needles. With this practice, infections due to acupuncture are thought to be rare.

Of note, trials that compare acupuncture to another active therapy often find that acupuncture has fewer adverse effects. This has been the case when acupuncture was compared to NSAIDs for low back pain and to topiramate for headaches.^17,23

5. How does acupuncture fit into a patient’s treatment?

The simple answer is that it is often most effectively used as part of a comprehensive management plan for chronic pain.

As our understanding of the complexity of chronic pain deepens, our therapeutic armamentarium for the management of chronic pain needs to broaden. This was summed up well in a 2016 article on the multimodal management of chronic pain when the authors stated, “Many targets need more than one arrow.”²⁴ Effective management of chronic pain involves addressing psychosocial and lifestyle factors in a patient-centered way and finding a combination of treatments that most effectively leads to improved coping and function.

It’s important to note that like medications and injections, acupuncture is a passive therapy. Although there is evidence for efficacy of improved pain with acupuncture in certain conditions, it should be combined with treatments that actively engage patients, such as exercise, behavioral treatments, development of coping skills, sleep hygiene, and educational strategies.

6. To whom do I refer patients for acupuncture treatment?

In the United States, licensed acupuncturists and physicians most commonly perform acupuncture. There are more than 50 schools that train licensed acupuncturists in the United States, and it usually takes 3 years to meet the requirements.²⁵

Physicians are often trained through continuing medical education (CME) programs that take several months to complete. These programs often combine live lectures, distance learning, and hands-on training and are typically sponsored by a university. Most require 300 hours of CME to complete. Licensure varies by state, but in many states, having an MD or DO degree automatically allows physicians to practice acupuncture. (See “Online resources,” above for links to Web sites that can be useful in finding qualified acupuncturists in your area.)

7. Is acupuncture covered by insurance?

Patients can expect to pay $75 to $150 for an acupuncture session.

It depends. Insurance coverage of acupuncture is highly variable and based on region and insurance type. Medicare and Medicaid plans do not pay for acupuncture. There are some private insurance plans that do. If covered, there may be limitations regarding diagnosis, number of visits, or provider. It is best for patients to call their insurance plan directly to inquire about coverage and any limitations. If paying out of pocket, patients can expect to pay $75 to $150 per treatment session.

CORRESPONDENCE
Russell Lemmon, DO, 1100 Delaplaine Court, Madison, WI 53715; [email protected].

CASE A 40-year-old man came to our office slightly agitated. He had an acute illness that was minor in nature. However, he was not interested in answering my questions or undergoing a physical exam. The more I tried to proceed with the visit, the more agitated he became—pacing the room, muttering, avoiding eye contact. I was uncomfortable and knew that the situation could quickly escalate if it was not brought under control.

What steps would you take if this were your patient?

The scene described above occurred several years ago, but more recently, one of the institutions in my (TIM) area was affected by a shooter in the workplace. The apprehension felt by all of us who were on the periphery paled in comparison to what was experienced by those at the scene. The outcome was horrific. Communicating with those directly involved during, and immediately after, the event was heart-wrenching. The trauma that they continue to relive is unimaginable, and some are not yet able to return to work.

Situations involving agitated patients are not uncommon in health care settings, although ones that escalate to the level of a shooting are. And no matter where on the spectrum an incident involving an agitated patient falls, it can leave those involved with various levels of physical, emotional, and psychological harm. It can also leave everyone asking themselves: “How can I better prepare for such occurrences?”

This article offers some answers by providing tips and guidelines for handling agitated and/or violent patients in various settings.

[polldaddy:9948472]

Defining the problem, assessing its severity

Between 2011 and 2013, workplace assaults ranged from 23,540 and 25,630 annually, with 70% to 74% occurring in health care and social service settings.^1,2

Agitation is defined as a state that may include inattention, disinhibition, emotional lability, impulsivity, motor restlessness, and aggression.^3,4 Violence in a clinical setting may be seen as an extreme expression of agitation sufficient enough to cause harm to an individual or damage to an object.^5,6

The causes of agitation can be grouped into categories: those due to a general medical condition, those due to a psychiatric condition, and those due to drug intoxication and/or withdrawal.⁷ We have chosen to add a fourth category—iatrogenic (see TABLE 1^3,4,7-9). They are not distinct categories, as there is sometimes overlap among areas.

Determining the level of agitation. Various scales and approaches can help determine the level of agitation in a patient (eg, the Agitated Behavior Scale [ABS; FIGURE];⁵ the Behavioral Activity Rating Scale [BARS]¹⁰) and the risk for violence (eg, the ABC violence risk assessment, TABLE 2⁸).

Scales like the ABS should be employed as soon as a patient shows signs of agitation sufficient to warrant intervention. The idea is for the family physician (FP) to be familiar enough with the tool to be able to mentally check it off, fill it out when time permits, and keep it in the patient’s chart. The first version of the form serves as a baseline so that if care is handed off to another provider, that provider can monitor whether signs and symptoms are improving or worsening.

Setting often drives the solution

Much of the evidence-based research on managing patient agitation and violence stems from inpatient psychiatric and emergency department (ED) settings. To make other health care providers aware of the experience gained in those settings, the American Association for Emergency Psychiatry created Project BETA (Best Practices in Evaluation and Treatment of Agitation). This project is designed to help promote consistency across health care settings and specialties in the way clinicians respond to agitated patients and to emphasize for all health care providers the availability of more than just pharmacologic approaches.⁷

De-escalating the situation. General tenets of de-escalation apply across practice settings. Among them:

• Stay calm. Avoid aggressive postures and prolonged eye contact.
• Be nonconfrontational. Acknowledge the patient’s frustration/perceptions and ask open-ended questions.
• Assess available resources such as clinical team members, family members, and silent alarms.
• Manage the situation and the patient’s underlying issues/diagnoses. This includes mobilizing other patients to avoid collateral damage and exploring solutions with the patient.

For more on de-escalation tools, see (TABLE 3^4,6,9,11).

Your setting matters. It’s worth noting that the settings in which clinicians practice greatly influence the resources available to de-escalate a situation and ensure the safety of the patient and others.⁷ The review that follows provides some issues—and tips—that are unique to different practice settings.

Ambulatory settings

Sim and colleagues⁹ noted that aggressive behavior in the general practice setting may stem not only from factors related to the patient’s own physical or psychological discomfort, but from patients feeling that they are being treated unfairly, whether it be because of wait times, uncomfortable waiting conditions, or something else. A number of international studies have shown high rates of abuse toward FPs.^9,12 Of 831 primary care physicians surveyed in a German study, close to three-quarters indicated that within the last year, they had experienced aggression (ranging from verbal abuse and threats to physical violence and property damage) from a patient.¹² This statistic increased to 91% when it included the length of their career.

Between 2011 and 2013, 7 out of 10 workplace assaults occurred in health care and social service settings.

Bell¹³ suggests that physicians be aware that transference and countertransference issues are often at play when dealing with hostile or potentially violent patients. Suggestions to prevent aggression include some practice-level approaches (eg, providing waiting room distractions, making patients aware of potential delays), as well as being aware of nonverbal cues suggesting increased agitation (eg, clenched fists, crossed arms, chin thrusts, finger pointing).⁹

Group practice

An FP who practices with other health care providers and clinical staff has a built-in team that can assist with de-escalation. When meeting with a patient who has a history of violence or agitation in an exam room or office, try to ensure that you can get to an exit quickly if necessary. Also, alert staff to any concerns, and have a system for at least one staff member to check in periodically during the visit.

It is also helpful to develop an evacuation plan and create a “panic room” or “safe zone” for emergencies.^14,15 Such a space may be nothing more than an area or room for staff to gather. It should have access to the police or other emergency services via a land and/or cell phone line.

Solo practice

If you practice alone, institute safeguards whereby a colleague (at a different practice, building, or location) can be alerted if concerns arise. In addition, consider the following precautions: locking the door when alone after hours, screening potential patients, having a way to call for help (keep the number for the local police station and ED readily available), prohibiting potential weapons (as some states allow them to be carried), and learning some form of self-defense.¹⁵

Ensure that agitated patients are not positioned between you and the exit so that you can quickly escape if necessary.

Resources exist that offer guidelines for developing policies and procedures, checklists, and sample incident forms (eg, the International Association for Healthcare Security and Safety; iahss.org). Other organizations that can help with the development of a preparedness plan include the Occupational Safety and Health Administration (https://www.osha.gov/SLTC/workplaceviolence/evaluation.html), the Department of Homeland Security (https://www.dhs.gov/sites/default/files/publications/ISC%20Violence%20in%20%20the%20Federal%20Workplace%20Guide%20April%202013.pdf), and The Joint Commission (https://www.jointcommission.org/workplace_violence.aspx).

In long-term care settings, such as nursing homes, and shorter-term care settings, such as rehabilitation facilities, agitation may stem from causes related to a head injury or dementia or from living in an unfamiliar environment. Assessment can be accomplished using a formal scale (eg, the ABS), as well as by identifying potential underlying health-related factors that can lead to agitation, such as pain, an infection, bowel and bladder issues, seizures, wounds, endocrine anomalies, cardiac or pulmonary problems, gastrointestinal dysfunction, and metabolic abnormalities.³

Modify the environment. For this population, a primary approach involves modifying the environment to decrease the likelihood of agitation. This may involve decreasing noise or light or ensuring adequate levels of stimulation. Preventing disorientation can be addressed through verbal and visual reminders of the date, schedule, etc. If a particular situation or activity is identified as a source of agitation, attempts at modifications are called for.³

For patients with dementia, the American Psychiatric Association recommends using the lowest effective dose of an antipsychotic in conjunction with environmental and behavioral measures.¹⁶ A benzodiazepine (lorazepam, oxazepam) may be used for infrequent agitation. Trazodone or a selective serotonin reuptake inhibitor are alternatives for those without psychosis or who are intolerant to antipsychotics.¹⁶

For individuals in a rehabilitation setting, agitation can impede participation in therapy and has been associated with poorer functioning at the time of discharge.³ Agitation can also be disruptive and lead to distress for family members and caregivers, as well as for fellow patients. And because this environment has a greater likelihood of visitors unrelated to the patient being exposed to the aberrant behavior, it is especially important to have established policies and procedures for de-escalation in place.

Home care

More and more FPs and residents are conducting home visits. That’s because the Accreditation Council for Graduate Medical Education Program Requirements for Graduate Medical Education in Family Medicine now include integrating a patient’s care across settings—including the home.¹⁷ Those who do provide home care may find themselves in circumstances similar to those of domestic disputes.

The German study mentioned earlier of more than 800 primary care physicians found that while the vast majority of physicians felt safe in their offices, 66% of female doctors and 34% of male doctors did not feel safe making home visits.¹²

Know the neighborhood. There’s no doubt that working in the home health sector makes one vulnerable. More than 61% of home care workers report workplace violence annually.^18,19 An action plan, as well as established policies and procedures, are essential when making home visits. Prior to the visit, be aware of the community and the location of the nearest police department and hospital.

Unwin and Tatum²⁰ suggest not wearing a white coat or carrying a doctor’s bag so as not to stand out as a physician in neighborhoods where personal safety is an issue. Make sure that your cell phone is fully charged and that there is a GPS mechanism activated that allows others to locate you.²¹ Note the available exits in a patient’s home, and position yourself near them, if possible. Have someone call or text you at predetermined times so that the absence of a response from you will alert someone to send help.

In such situations, it is imperative to remain calm and to use the same verbal de-escalation techniques (TABLE 3^4,6,9,11) that would be used in any other health care setting. It is prudent to set expectations for the patient and family members prior to the home visit regarding the tools and services that will be provided in the home setting and the limitations in terms of scope of practice.

Emergency department

The ED is one of the most common settings for patient agitation and violence within the health care continuum.²² Providers must quickly determine the cause of the agitation while de-escalating the situation and ensuring that they do not miss a pertinent medical finding related to a time-sensitive issue, such as an intracerebral bleed or poisoning.⁷ In addition, the ED is usually heavily populated, providing an opportunity for tremendous collateral human damage should the violence escalate or weapons be deployed. The upside is that many EDs are now staffed with security personnel and, depending on the community, police officers may be on the premises or in the vicinity.²²

Avoid wearing a white coat or carrying a doctor's bag when doing a home visit so as not to stand out as a physician in neighborhoods where personal safety is an issue.

Etiologies for agitation in the ED can range from ingestion of unknown or unidentified substances to psychiatric or medical conditions. Knowledge of etiology is necessary prior to initiation of treatment.⁴

As in other settings, the safety of the patient and others present is of utmost importance. Key recommendations for managing agitated patients in the ED include: ⁴

1. Have an established plan for the management of agitated patients.
2. Identify signs of agitation early, and complete an agitation rating scale.
3. Attempt verbal de-escalation before using medication whenever possible.
4. Employ a “show of concern” rather than “a show of force” in response to escalating agitation/violence. Doing so can strengthen the perception that interventions are coming from a place of caring.
5. Use physical restraint as a last resort. When used, it should be with the intention of protecting the patient and those present, rather than as punishment.

Inpatient units

Unlike the ED, patients on units generally have a working diagnosis, and the provider has some background information with which to work, such as laboratory test results and radiology reports, facilitating more expedient and accurate situational assessment. However, the recommendations for assessment and early identification, as described for the ED, still apply.

If a provider finds him- or herself in an escalating situation, the call bells located in the rooms are of use. An alternative is to call out for help from someone in the hallway. One needs to be aware of the current policies and procedures for de-escalation, as some facilities have a specific “code” that is called for such occasions.¹⁹

Postop delirium is a common cause of agitation in the inpatient setting. Ng and colleagues¹¹ recommend a cognitive assessment before surgery to establish a baseline in order to determine the risk for delirium after surgery. Additionally, the FP must remain aware of preexisting conditions that may surface during a hospital stay, such as dehydration or unrecognized alcohol or medication withdrawal.

A "show of concern" rather than a "show of force" can strengthen the perception that interventions are coming from a place of caring.

Medication choice should be based on the type of delirium. Hyperactive delirium (restlessness, emotional lability, hallucinations) and mixed delirium (a combination of signs of hyperactive and hypoactive dementia) both hold the potential for agitation and even violence. The approach to hyperactive delirium includes consideration of an antipsychotic medication, although the efficacy of antipsychotics is considered controversial. In the case of mixed delirium, behavioral and environmental modifications are useful (eg, reducing noise and early ambulation).¹¹

No medications are registered with the US Food and Drug Administration for the management of delirium, and it is suggested that antipsychotics be considered only when other, less invasive, strategies have been attempted.²³

Regardless of the setting in which FPs work, witnessing or being directly involved in a traumatic event puts one at risk for symptoms—or a full diagnosis—of posttraumatic stress disorder (PTSD), acute stress disorder, or anxiety or mood disorders.^24,25 Although findings vary, studies have found that as many as 12% of ED personnel meet the criteria for PTSD^26,27 and 12% to 15% report having been threatened physically.^28,29 More than half of physicians in another study had witnessed a physical attack.³⁰

Physicians and other health care personnel who have experienced a traumatic incident, or offered help to another during an incident, may attempt to cope through avoidance, cutting down on work hours, leaving the work setting in which the event took place, or leaving the profession altogether.^29,31,32

There is a paucity of methodologically sound research with regard to prevention and treatment of PTSD symptoms in this population.²⁴ According to a 2002 Cochrane review, the effectiveness of individual, single-session debriefing does not have solid research support,³³ and there are concerns about potential harms due to reliving the traumatic event when sessions are led by poorly trained debriefing staff.^34-36

Critical incident stress debriefing (CISD), however, holds promise in terms of facilitating a return to pretrauma functioning based on studies of first responders.^34,35 This may be because CISD follows a specific protocol and that group sessions may capitalize on the social support/camaraderie within a group that has undergone a traumatic event.^34,35 It is important that those providing debriefing and support be well-trained.³⁵

Debriefing, however, is not always sufficient, and those who appear to be affected on an ongoing basis may require individual treatment for PTSD symptoms. Evidence-based treatments for PTSD, such as trauma-focused psychotherapy and/or pharmacotherapy, may be considered³⁷ (TABLE 4^24,34,38).

Ongoing support in the workplace. The Cleveland Clinic has developed a “Code Lavender” to combat stress in the workplace. Like a Code Blue for medical emergencies, a Code Lavender is called when a health care worker is in need of emotional or spiritual support.³⁸ A provider who initiates the call is met by a team of holistic nurses within 30 minutes. The team provides Reiki and massage, healthy snacks and water, and lavender arm bands to remind the individual to relax for the rest of the day. Further opportunities for spiritual support, mindfulness training, counseling, and yoga may also be made available.

CASE  Sensing that the situation with my patient might escalate, I lowered my voice, relaxed my shoulders, leaned casually against the desk, and asked him to tell me how I could best help him. As he spoke, I offered him a seat (by gesturing to the chair). I did this for 2 reasons: to move him away from blocking my exit from the room, and to put him at a lower level than me so that he was entirely in my view. I didn’t interrupt him as he spoke. I just nodded or tilted my head to show I was listening. In my mind, I played out the various scenarios that could ensue.

Like a Code Blue for medical emergencies, a Code Lavender is called when a health care worker is in need of emotional or spiritual support.

Fortunately, I was able to get him to relax enough for an assessment, which involved a more relevant history and the exam, which he agreed to once an aide had come into the room. He did not exhibit the concerning signs of flushed skin, dilated pupils, shallow rapid respirations, or perspiration. He did have a comorbid behavioral health issue, which we were able to address. His earlier behavioral indicators of agitation were controlled with verbal and physical cues on my part. Our conversation didn't reveal an intent to harm himself or others. In this case, physical restraints were not required. Throughout the encounter the door was left open, and the patient was reminded that we were there to help.

Once he left, I made the relevant notes in the chart regarding his agitated state at the start of the visit and his final state at the end of the visit so as to assist any other providers. We (TIM, MG) also held a quick debrief after the encounter with the office staff and decided that we needed to create a policy and protocol regarding how to handle such situations in the future.

CORRESPONDENCE
Tochi Iroku-Malize, MD, MPH, MBA, Family Medicine Department, Southside Hospital, 301 East Main Street, Bay Shore, NY 11706; [email protected].

CASE A 40-year-old man came to our office slightly agitated. He had an acute illness that was minor in nature. However, he was not interested in answering my questions or undergoing a physical exam. The more I tried to proceed with the visit, the more agitated he became—pacing the room, muttering, avoiding eye contact. I was uncomfortable and knew that the situation could quickly escalate if it was not brought under control.

What steps would you take if this were your patient?

The scene described above occurred several years ago, but more recently, one of the institutions in my (TIM) area was affected by a shooter in the workplace. The apprehension felt by all of us who were on the periphery paled in comparison to what was experienced by those at the scene. The outcome was horrific. Communicating with those directly involved during, and immediately after, the event was heart-wrenching. The trauma that they continue to relive is unimaginable, and some are not yet able to return to work.

Situations involving agitated patients are not uncommon in health care settings, although ones that escalate to the level of a shooting are. And no matter where on the spectrum an incident involving an agitated patient falls, it can leave those involved with various levels of physical, emotional, and psychological harm. It can also leave everyone asking themselves: “How can I better prepare for such occurrences?”

This article offers some answers by providing tips and guidelines for handling agitated and/or violent patients in various settings.

[polldaddy:9948472]

Defining the problem, assessing its severity

Between 2011 and 2013, workplace assaults ranged from 23,540 and 25,630 annually, with 70% to 74% occurring in health care and social service settings.^1,2

Agitation is defined as a state that may include inattention, disinhibition, emotional lability, impulsivity, motor restlessness, and aggression.^3,4 Violence in a clinical setting may be seen as an extreme expression of agitation sufficient enough to cause harm to an individual or damage to an object.^5,6

The causes of agitation can be grouped into categories: those due to a general medical condition, those due to a psychiatric condition, and those due to drug intoxication and/or withdrawal.⁷ We have chosen to add a fourth category—iatrogenic (see TABLE 1^3,4,7-9). They are not distinct categories, as there is sometimes overlap among areas.

Determining the level of agitation. Various scales and approaches can help determine the level of agitation in a patient (eg, the Agitated Behavior Scale [ABS; FIGURE];⁵ the Behavioral Activity Rating Scale [BARS]¹⁰) and the risk for violence (eg, the ABC violence risk assessment, TABLE 2⁸).

Scales like the ABS should be employed as soon as a patient shows signs of agitation sufficient to warrant intervention. The idea is for the family physician (FP) to be familiar enough with the tool to be able to mentally check it off, fill it out when time permits, and keep it in the patient’s chart. The first version of the form serves as a baseline so that if care is handed off to another provider, that provider can monitor whether signs and symptoms are improving or worsening.

Setting often drives the solution

Much of the evidence-based research on managing patient agitation and violence stems from inpatient psychiatric and emergency department (ED) settings. To make other health care providers aware of the experience gained in those settings, the American Association for Emergency Psychiatry created Project BETA (Best Practices in Evaluation and Treatment of Agitation). This project is designed to help promote consistency across health care settings and specialties in the way clinicians respond to agitated patients and to emphasize for all health care providers the availability of more than just pharmacologic approaches.⁷

De-escalating the situation. General tenets of de-escalation apply across practice settings. Among them:

• Stay calm. Avoid aggressive postures and prolonged eye contact.
• Be nonconfrontational. Acknowledge the patient’s frustration/perceptions and ask open-ended questions.
• Assess available resources such as clinical team members, family members, and silent alarms.
• Manage the situation and the patient’s underlying issues/diagnoses. This includes mobilizing other patients to avoid collateral damage and exploring solutions with the patient.

For more on de-escalation tools, see (TABLE 3^4,6,9,11).

Your setting matters. It’s worth noting that the settings in which clinicians practice greatly influence the resources available to de-escalate a situation and ensure the safety of the patient and others.⁷ The review that follows provides some issues—and tips—that are unique to different practice settings.

Ambulatory settings

Sim and colleagues⁹ noted that aggressive behavior in the general practice setting may stem not only from factors related to the patient’s own physical or psychological discomfort, but from patients feeling that they are being treated unfairly, whether it be because of wait times, uncomfortable waiting conditions, or something else. A number of international studies have shown high rates of abuse toward FPs.^9,12 Of 831 primary care physicians surveyed in a German study, close to three-quarters indicated that within the last year, they had experienced aggression (ranging from verbal abuse and threats to physical violence and property damage) from a patient.¹² This statistic increased to 91% when it included the length of their career.

Between 2011 and 2013, 7 out of 10 workplace assaults occurred in health care and social service settings.

Bell¹³ suggests that physicians be aware that transference and countertransference issues are often at play when dealing with hostile or potentially violent patients. Suggestions to prevent aggression include some practice-level approaches (eg, providing waiting room distractions, making patients aware of potential delays), as well as being aware of nonverbal cues suggesting increased agitation (eg, clenched fists, crossed arms, chin thrusts, finger pointing).⁹

Group practice

An FP who practices with other health care providers and clinical staff has a built-in team that can assist with de-escalation. When meeting with a patient who has a history of violence or agitation in an exam room or office, try to ensure that you can get to an exit quickly if necessary. Also, alert staff to any concerns, and have a system for at least one staff member to check in periodically during the visit.

It is also helpful to develop an evacuation plan and create a “panic room” or “safe zone” for emergencies.^14,15 Such a space may be nothing more than an area or room for staff to gather. It should have access to the police or other emergency services via a land and/or cell phone line.

Solo practice

If you practice alone, institute safeguards whereby a colleague (at a different practice, building, or location) can be alerted if concerns arise. In addition, consider the following precautions: locking the door when alone after hours, screening potential patients, having a way to call for help (keep the number for the local police station and ED readily available), prohibiting potential weapons (as some states allow them to be carried), and learning some form of self-defense.¹⁵

Ensure that agitated patients are not positioned between you and the exit so that you can quickly escape if necessary.

Resources exist that offer guidelines for developing policies and procedures, checklists, and sample incident forms (eg, the International Association for Healthcare Security and Safety; iahss.org). Other organizations that can help with the development of a preparedness plan include the Occupational Safety and Health Administration (https://www.osha.gov/SLTC/workplaceviolence/evaluation.html), the Department of Homeland Security (https://www.dhs.gov/sites/default/files/publications/ISC%20Violence%20in%20%20the%20Federal%20Workplace%20Guide%20April%202013.pdf), and The Joint Commission (https://www.jointcommission.org/workplace_violence.aspx).

In long-term care settings, such as nursing homes, and shorter-term care settings, such as rehabilitation facilities, agitation may stem from causes related to a head injury or dementia or from living in an unfamiliar environment. Assessment can be accomplished using a formal scale (eg, the ABS), as well as by identifying potential underlying health-related factors that can lead to agitation, such as pain, an infection, bowel and bladder issues, seizures, wounds, endocrine anomalies, cardiac or pulmonary problems, gastrointestinal dysfunction, and metabolic abnormalities.³

Modify the environment. For this population, a primary approach involves modifying the environment to decrease the likelihood of agitation. This may involve decreasing noise or light or ensuring adequate levels of stimulation. Preventing disorientation can be addressed through verbal and visual reminders of the date, schedule, etc. If a particular situation or activity is identified as a source of agitation, attempts at modifications are called for.³

For patients with dementia, the American Psychiatric Association recommends using the lowest effective dose of an antipsychotic in conjunction with environmental and behavioral measures.¹⁶ A benzodiazepine (lorazepam, oxazepam) may be used for infrequent agitation. Trazodone or a selective serotonin reuptake inhibitor are alternatives for those without psychosis or who are intolerant to antipsychotics.¹⁶

For individuals in a rehabilitation setting, agitation can impede participation in therapy and has been associated with poorer functioning at the time of discharge.³ Agitation can also be disruptive and lead to distress for family members and caregivers, as well as for fellow patients. And because this environment has a greater likelihood of visitors unrelated to the patient being exposed to the aberrant behavior, it is especially important to have established policies and procedures for de-escalation in place.

Home care

More and more FPs and residents are conducting home visits. That’s because the Accreditation Council for Graduate Medical Education Program Requirements for Graduate Medical Education in Family Medicine now include integrating a patient’s care across settings—including the home.¹⁷ Those who do provide home care may find themselves in circumstances similar to those of domestic disputes.

The German study mentioned earlier of more than 800 primary care physicians found that while the vast majority of physicians felt safe in their offices, 66% of female doctors and 34% of male doctors did not feel safe making home visits.¹²

Know the neighborhood. There’s no doubt that working in the home health sector makes one vulnerable. More than 61% of home care workers report workplace violence annually.^18,19 An action plan, as well as established policies and procedures, are essential when making home visits. Prior to the visit, be aware of the community and the location of the nearest police department and hospital.

Unwin and Tatum²⁰ suggest not wearing a white coat or carrying a doctor’s bag so as not to stand out as a physician in neighborhoods where personal safety is an issue. Make sure that your cell phone is fully charged and that there is a GPS mechanism activated that allows others to locate you.²¹ Note the available exits in a patient’s home, and position yourself near them, if possible. Have someone call or text you at predetermined times so that the absence of a response from you will alert someone to send help.

In such situations, it is imperative to remain calm and to use the same verbal de-escalation techniques (TABLE 3^4,6,9,11) that would be used in any other health care setting. It is prudent to set expectations for the patient and family members prior to the home visit regarding the tools and services that will be provided in the home setting and the limitations in terms of scope of practice.

Emergency department

The ED is one of the most common settings for patient agitation and violence within the health care continuum.²² Providers must quickly determine the cause of the agitation while de-escalating the situation and ensuring that they do not miss a pertinent medical finding related to a time-sensitive issue, such as an intracerebral bleed or poisoning.⁷ In addition, the ED is usually heavily populated, providing an opportunity for tremendous collateral human damage should the violence escalate or weapons be deployed. The upside is that many EDs are now staffed with security personnel and, depending on the community, police officers may be on the premises or in the vicinity.²²

Avoid wearing a white coat or carrying a doctor's bag when doing a home visit so as not to stand out as a physician in neighborhoods where personal safety is an issue.

Etiologies for agitation in the ED can range from ingestion of unknown or unidentified substances to psychiatric or medical conditions. Knowledge of etiology is necessary prior to initiation of treatment.⁴

As in other settings, the safety of the patient and others present is of utmost importance. Key recommendations for managing agitated patients in the ED include: ⁴

1. Have an established plan for the management of agitated patients.
2. Identify signs of agitation early, and complete an agitation rating scale.
3. Attempt verbal de-escalation before using medication whenever possible.
4. Employ a “show of concern” rather than “a show of force” in response to escalating agitation/violence. Doing so can strengthen the perception that interventions are coming from a place of caring.
5. Use physical restraint as a last resort. When used, it should be with the intention of protecting the patient and those present, rather than as punishment.

Inpatient units

Unlike the ED, patients on units generally have a working diagnosis, and the provider has some background information with which to work, such as laboratory test results and radiology reports, facilitating more expedient and accurate situational assessment. However, the recommendations for assessment and early identification, as described for the ED, still apply.

If a provider finds him- or herself in an escalating situation, the call bells located in the rooms are of use. An alternative is to call out for help from someone in the hallway. One needs to be aware of the current policies and procedures for de-escalation, as some facilities have a specific “code” that is called for such occasions.¹⁹

Postop delirium is a common cause of agitation in the inpatient setting. Ng and colleagues¹¹ recommend a cognitive assessment before surgery to establish a baseline in order to determine the risk for delirium after surgery. Additionally, the FP must remain aware of preexisting conditions that may surface during a hospital stay, such as dehydration or unrecognized alcohol or medication withdrawal.

A "show of concern" rather than a "show of force" can strengthen the perception that interventions are coming from a place of caring.

Medication choice should be based on the type of delirium. Hyperactive delirium (restlessness, emotional lability, hallucinations) and mixed delirium (a combination of signs of hyperactive and hypoactive dementia) both hold the potential for agitation and even violence. The approach to hyperactive delirium includes consideration of an antipsychotic medication, although the efficacy of antipsychotics is considered controversial. In the case of mixed delirium, behavioral and environmental modifications are useful (eg, reducing noise and early ambulation).¹¹

No medications are registered with the US Food and Drug Administration for the management of delirium, and it is suggested that antipsychotics be considered only when other, less invasive, strategies have been attempted.²³

Regardless of the setting in which FPs work, witnessing or being directly involved in a traumatic event puts one at risk for symptoms—or a full diagnosis—of posttraumatic stress disorder (PTSD), acute stress disorder, or anxiety or mood disorders.^24,25 Although findings vary, studies have found that as many as 12% of ED personnel meet the criteria for PTSD^26,27 and 12% to 15% report having been threatened physically.^28,29 More than half of physicians in another study had witnessed a physical attack.³⁰

Physicians and other health care personnel who have experienced a traumatic incident, or offered help to another during an incident, may attempt to cope through avoidance, cutting down on work hours, leaving the work setting in which the event took place, or leaving the profession altogether.^29,31,32

There is a paucity of methodologically sound research with regard to prevention and treatment of PTSD symptoms in this population.²⁴ According to a 2002 Cochrane review, the effectiveness of individual, single-session debriefing does not have solid research support,³³ and there are concerns about potential harms due to reliving the traumatic event when sessions are led by poorly trained debriefing staff.^34-36

Critical incident stress debriefing (CISD), however, holds promise in terms of facilitating a return to pretrauma functioning based on studies of first responders.^34,35 This may be because CISD follows a specific protocol and that group sessions may capitalize on the social support/camaraderie within a group that has undergone a traumatic event.^34,35 It is important that those providing debriefing and support be well-trained.³⁵

Debriefing, however, is not always sufficient, and those who appear to be affected on an ongoing basis may require individual treatment for PTSD symptoms. Evidence-based treatments for PTSD, such as trauma-focused psychotherapy and/or pharmacotherapy, may be considered³⁷ (TABLE 4^24,34,38).

Ongoing support in the workplace. The Cleveland Clinic has developed a “Code Lavender” to combat stress in the workplace. Like a Code Blue for medical emergencies, a Code Lavender is called when a health care worker is in need of emotional or spiritual support.³⁸ A provider who initiates the call is met by a team of holistic nurses within 30 minutes. The team provides Reiki and massage, healthy snacks and water, and lavender arm bands to remind the individual to relax for the rest of the day. Further opportunities for spiritual support, mindfulness training, counseling, and yoga may also be made available.

CASE  Sensing that the situation with my patient might escalate, I lowered my voice, relaxed my shoulders, leaned casually against the desk, and asked him to tell me how I could best help him. As he spoke, I offered him a seat (by gesturing to the chair). I did this for 2 reasons: to move him away from blocking my exit from the room, and to put him at a lower level than me so that he was entirely in my view. I didn’t interrupt him as he spoke. I just nodded or tilted my head to show I was listening. In my mind, I played out the various scenarios that could ensue.

Like a Code Blue for medical emergencies, a Code Lavender is called when a health care worker is in need of emotional or spiritual support.

Fortunately, I was able to get him to relax enough for an assessment, which involved a more relevant history and the exam, which he agreed to once an aide had come into the room. He did not exhibit the concerning signs of flushed skin, dilated pupils, shallow rapid respirations, or perspiration. He did have a comorbid behavioral health issue, which we were able to address. His earlier behavioral indicators of agitation were controlled with verbal and physical cues on my part. Our conversation didn't reveal an intent to harm himself or others. In this case, physical restraints were not required. Throughout the encounter the door was left open, and the patient was reminded that we were there to help.

Once he left, I made the relevant notes in the chart regarding his agitated state at the start of the visit and his final state at the end of the visit so as to assist any other providers. We (TIM, MG) also held a quick debrief after the encounter with the office staff and decided that we needed to create a policy and protocol regarding how to handle such situations in the future.

CORRESPONDENCE
Tochi Iroku-Malize, MD, MPH, MBA, Family Medicine Department, Southside Hospital, 301 East Main Street, Bay Shore, NY 11706; [email protected].

CASE A 40-year-old man came to our office slightly agitated. He had an acute illness that was minor in nature. However, he was not interested in answering my questions or undergoing a physical exam. The more I tried to proceed with the visit, the more agitated he became—pacing the room, muttering, avoiding eye contact. I was uncomfortable and knew that the situation could quickly escalate if it was not brought under control.

What steps would you take if this were your patient?

The scene described above occurred several years ago, but more recently, one of the institutions in my (TIM) area was affected by a shooter in the workplace. The apprehension felt by all of us who were on the periphery paled in comparison to what was experienced by those at the scene. The outcome was horrific. Communicating with those directly involved during, and immediately after, the event was heart-wrenching. The trauma that they continue to relive is unimaginable, and some are not yet able to return to work.

Situations involving agitated patients are not uncommon in health care settings, although ones that escalate to the level of a shooting are. And no matter where on the spectrum an incident involving an agitated patient falls, it can leave those involved with various levels of physical, emotional, and psychological harm. It can also leave everyone asking themselves: “How can I better prepare for such occurrences?”

This article offers some answers by providing tips and guidelines for handling agitated and/or violent patients in various settings.

[polldaddy:9948472]

Defining the problem, assessing its severity

Between 2011 and 2013, workplace assaults ranged from 23,540 and 25,630 annually, with 70% to 74% occurring in health care and social service settings.^1,2

Agitation is defined as a state that may include inattention, disinhibition, emotional lability, impulsivity, motor restlessness, and aggression.^3,4 Violence in a clinical setting may be seen as an extreme expression of agitation sufficient enough to cause harm to an individual or damage to an object.^5,6

The causes of agitation can be grouped into categories: those due to a general medical condition, those due to a psychiatric condition, and those due to drug intoxication and/or withdrawal.⁷ We have chosen to add a fourth category—iatrogenic (see TABLE 1^3,4,7-9). They are not distinct categories, as there is sometimes overlap among areas.

Determining the level of agitation. Various scales and approaches can help determine the level of agitation in a patient (eg, the Agitated Behavior Scale [ABS; FIGURE];⁵ the Behavioral Activity Rating Scale [BARS]¹⁰) and the risk for violence (eg, the ABC violence risk assessment, TABLE 2⁸).

Scales like the ABS should be employed as soon as a patient shows signs of agitation sufficient to warrant intervention. The idea is for the family physician (FP) to be familiar enough with the tool to be able to mentally check it off, fill it out when time permits, and keep it in the patient’s chart. The first version of the form serves as a baseline so that if care is handed off to another provider, that provider can monitor whether signs and symptoms are improving or worsening.

Setting often drives the solution

Much of the evidence-based research on managing patient agitation and violence stems from inpatient psychiatric and emergency department (ED) settings. To make other health care providers aware of the experience gained in those settings, the American Association for Emergency Psychiatry created Project BETA (Best Practices in Evaluation and Treatment of Agitation). This project is designed to help promote consistency across health care settings and specialties in the way clinicians respond to agitated patients and to emphasize for all health care providers the availability of more than just pharmacologic approaches.⁷

De-escalating the situation. General tenets of de-escalation apply across practice settings. Among them:

• Stay calm. Avoid aggressive postures and prolonged eye contact.
• Be nonconfrontational. Acknowledge the patient’s frustration/perceptions and ask open-ended questions.
• Assess available resources such as clinical team members, family members, and silent alarms.
• Manage the situation and the patient’s underlying issues/diagnoses. This includes mobilizing other patients to avoid collateral damage and exploring solutions with the patient.

For more on de-escalation tools, see (TABLE 3^4,6,9,11).

Your setting matters. It’s worth noting that the settings in which clinicians practice greatly influence the resources available to de-escalate a situation and ensure the safety of the patient and others.⁷ The review that follows provides some issues—and tips—that are unique to different practice settings.

Ambulatory settings

Sim and colleagues⁹ noted that aggressive behavior in the general practice setting may stem not only from factors related to the patient’s own physical or psychological discomfort, but from patients feeling that they are being treated unfairly, whether it be because of wait times, uncomfortable waiting conditions, or something else. A number of international studies have shown high rates of abuse toward FPs.^9,12 Of 831 primary care physicians surveyed in a German study, close to three-quarters indicated that within the last year, they had experienced aggression (ranging from verbal abuse and threats to physical violence and property damage) from a patient.¹² This statistic increased to 91% when it included the length of their career.

Between 2011 and 2013, 7 out of 10 workplace assaults occurred in health care and social service settings.

Bell¹³ suggests that physicians be aware that transference and countertransference issues are often at play when dealing with hostile or potentially violent patients. Suggestions to prevent aggression include some practice-level approaches (eg, providing waiting room distractions, making patients aware of potential delays), as well as being aware of nonverbal cues suggesting increased agitation (eg, clenched fists, crossed arms, chin thrusts, finger pointing).⁹

Group practice

An FP who practices with other health care providers and clinical staff has a built-in team that can assist with de-escalation. When meeting with a patient who has a history of violence or agitation in an exam room or office, try to ensure that you can get to an exit quickly if necessary. Also, alert staff to any concerns, and have a system for at least one staff member to check in periodically during the visit.

It is also helpful to develop an evacuation plan and create a “panic room” or “safe zone” for emergencies.^14,15 Such a space may be nothing more than an area or room for staff to gather. It should have access to the police or other emergency services via a land and/or cell phone line.

Solo practice

If you practice alone, institute safeguards whereby a colleague (at a different practice, building, or location) can be alerted if concerns arise. In addition, consider the following precautions: locking the door when alone after hours, screening potential patients, having a way to call for help (keep the number for the local police station and ED readily available), prohibiting potential weapons (as some states allow them to be carried), and learning some form of self-defense.¹⁵

Ensure that agitated patients are not positioned between you and the exit so that you can quickly escape if necessary.

Resources exist that offer guidelines for developing policies and procedures, checklists, and sample incident forms (eg, the International Association for Healthcare Security and Safety; iahss.org). Other organizations that can help with the development of a preparedness plan include the Occupational Safety and Health Administration (https://www.osha.gov/SLTC/workplaceviolence/evaluation.html), the Department of Homeland Security (https://www.dhs.gov/sites/default/files/publications/ISC%20Violence%20in%20%20the%20Federal%20Workplace%20Guide%20April%202013.pdf), and The Joint Commission (https://www.jointcommission.org/workplace_violence.aspx).

In long-term care settings, such as nursing homes, and shorter-term care settings, such as rehabilitation facilities, agitation may stem from causes related to a head injury or dementia or from living in an unfamiliar environment. Assessment can be accomplished using a formal scale (eg, the ABS), as well as by identifying potential underlying health-related factors that can lead to agitation, such as pain, an infection, bowel and bladder issues, seizures, wounds, endocrine anomalies, cardiac or pulmonary problems, gastrointestinal dysfunction, and metabolic abnormalities.³

Modify the environment. For this population, a primary approach involves modifying the environment to decrease the likelihood of agitation. This may involve decreasing noise or light or ensuring adequate levels of stimulation. Preventing disorientation can be addressed through verbal and visual reminders of the date, schedule, etc. If a particular situation or activity is identified as a source of agitation, attempts at modifications are called for.³

For patients with dementia, the American Psychiatric Association recommends using the lowest effective dose of an antipsychotic in conjunction with environmental and behavioral measures.¹⁶ A benzodiazepine (lorazepam, oxazepam) may be used for infrequent agitation. Trazodone or a selective serotonin reuptake inhibitor are alternatives for those without psychosis or who are intolerant to antipsychotics.¹⁶

For individuals in a rehabilitation setting, agitation can impede participation in therapy and has been associated with poorer functioning at the time of discharge.³ Agitation can also be disruptive and lead to distress for family members and caregivers, as well as for fellow patients. And because this environment has a greater likelihood of visitors unrelated to the patient being exposed to the aberrant behavior, it is especially important to have established policies and procedures for de-escalation in place.

Home care

More and more FPs and residents are conducting home visits. That’s because the Accreditation Council for Graduate Medical Education Program Requirements for Graduate Medical Education in Family Medicine now include integrating a patient’s care across settings—including the home.¹⁷ Those who do provide home care may find themselves in circumstances similar to those of domestic disputes.

The German study mentioned earlier of more than 800 primary care physicians found that while the vast majority of physicians felt safe in their offices, 66% of female doctors and 34% of male doctors did not feel safe making home visits.¹²

Know the neighborhood. There’s no doubt that working in the home health sector makes one vulnerable. More than 61% of home care workers report workplace violence annually.^18,19 An action plan, as well as established policies and procedures, are essential when making home visits. Prior to the visit, be aware of the community and the location of the nearest police department and hospital.

Unwin and Tatum²⁰ suggest not wearing a white coat or carrying a doctor’s bag so as not to stand out as a physician in neighborhoods where personal safety is an issue. Make sure that your cell phone is fully charged and that there is a GPS mechanism activated that allows others to locate you.²¹ Note the available exits in a patient’s home, and position yourself near them, if possible. Have someone call or text you at predetermined times so that the absence of a response from you will alert someone to send help.

In such situations, it is imperative to remain calm and to use the same verbal de-escalation techniques (TABLE 3^4,6,9,11) that would be used in any other health care setting. It is prudent to set expectations for the patient and family members prior to the home visit regarding the tools and services that will be provided in the home setting and the limitations in terms of scope of practice.

Emergency department

The ED is one of the most common settings for patient agitation and violence within the health care continuum.²² Providers must quickly determine the cause of the agitation while de-escalating the situation and ensuring that they do not miss a pertinent medical finding related to a time-sensitive issue, such as an intracerebral bleed or poisoning.⁷ In addition, the ED is usually heavily populated, providing an opportunity for tremendous collateral human damage should the violence escalate or weapons be deployed. The upside is that many EDs are now staffed with security personnel and, depending on the community, police officers may be on the premises or in the vicinity.²²

Avoid wearing a white coat or carrying a doctor's bag when doing a home visit so as not to stand out as a physician in neighborhoods where personal safety is an issue.

Etiologies for agitation in the ED can range from ingestion of unknown or unidentified substances to psychiatric or medical conditions. Knowledge of etiology is necessary prior to initiation of treatment.⁴

As in other settings, the safety of the patient and others present is of utmost importance. Key recommendations for managing agitated patients in the ED include: ⁴

1. Have an established plan for the management of agitated patients.
2. Identify signs of agitation early, and complete an agitation rating scale.
3. Attempt verbal de-escalation before using medication whenever possible.
4. Employ a “show of concern” rather than “a show of force” in response to escalating agitation/violence. Doing so can strengthen the perception that interventions are coming from a place of caring.
5. Use physical restraint as a last resort. When used, it should be with the intention of protecting the patient and those present, rather than as punishment.

Inpatient units

Unlike the ED, patients on units generally have a working diagnosis, and the provider has some background information with which to work, such as laboratory test results and radiology reports, facilitating more expedient and accurate situational assessment. However, the recommendations for assessment and early identification, as described for the ED, still apply.

If a provider finds him- or herself in an escalating situation, the call bells located in the rooms are of use. An alternative is to call out for help from someone in the hallway. One needs to be aware of the current policies and procedures for de-escalation, as some facilities have a specific “code” that is called for such occasions.¹⁹

Postop delirium is a common cause of agitation in the inpatient setting. Ng and colleagues¹¹ recommend a cognitive assessment before surgery to establish a baseline in order to determine the risk for delirium after surgery. Additionally, the FP must remain aware of preexisting conditions that may surface during a hospital stay, such as dehydration or unrecognized alcohol or medication withdrawal.

A "show of concern" rather than a "show of force" can strengthen the perception that interventions are coming from a place of caring.

Medication choice should be based on the type of delirium. Hyperactive delirium (restlessness, emotional lability, hallucinations) and mixed delirium (a combination of signs of hyperactive and hypoactive dementia) both hold the potential for agitation and even violence. The approach to hyperactive delirium includes consideration of an antipsychotic medication, although the efficacy of antipsychotics is considered controversial. In the case of mixed delirium, behavioral and environmental modifications are useful (eg, reducing noise and early ambulation).¹¹

No medications are registered with the US Food and Drug Administration for the management of delirium, and it is suggested that antipsychotics be considered only when other, less invasive, strategies have been attempted.²³

Regardless of the setting in which FPs work, witnessing or being directly involved in a traumatic event puts one at risk for symptoms—or a full diagnosis—of posttraumatic stress disorder (PTSD), acute stress disorder, or anxiety or mood disorders.^24,25 Although findings vary, studies have found that as many as 12% of ED personnel meet the criteria for PTSD^26,27 and 12% to 15% report having been threatened physically.^28,29 More than half of physicians in another study had witnessed a physical attack.³⁰

Physicians and other health care personnel who have experienced a traumatic incident, or offered help to another during an incident, may attempt to cope through avoidance, cutting down on work hours, leaving the work setting in which the event took place, or leaving the profession altogether.^29,31,32

There is a paucity of methodologically sound research with regard to prevention and treatment of PTSD symptoms in this population.²⁴ According to a 2002 Cochrane review, the effectiveness of individual, single-session debriefing does not have solid research support,³³ and there are concerns about potential harms due to reliving the traumatic event when sessions are led by poorly trained debriefing staff.^34-36

Critical incident stress debriefing (CISD), however, holds promise in terms of facilitating a return to pretrauma functioning based on studies of first responders.^34,35 This may be because CISD follows a specific protocol and that group sessions may capitalize on the social support/camaraderie within a group that has undergone a traumatic event.^34,35 It is important that those providing debriefing and support be well-trained.³⁵

Debriefing, however, is not always sufficient, and those who appear to be affected on an ongoing basis may require individual treatment for PTSD symptoms. Evidence-based treatments for PTSD, such as trauma-focused psychotherapy and/or pharmacotherapy, may be considered³⁷ (TABLE 4^24,34,38).

Ongoing support in the workplace. The Cleveland Clinic has developed a “Code Lavender” to combat stress in the workplace. Like a Code Blue for medical emergencies, a Code Lavender is called when a health care worker is in need of emotional or spiritual support.³⁸ A provider who initiates the call is met by a team of holistic nurses within 30 minutes. The team provides Reiki and massage, healthy snacks and water, and lavender arm bands to remind the individual to relax for the rest of the day. Further opportunities for spiritual support, mindfulness training, counseling, and yoga may also be made available.

CASE  Sensing that the situation with my patient might escalate, I lowered my voice, relaxed my shoulders, leaned casually against the desk, and asked him to tell me how I could best help him. As he spoke, I offered him a seat (by gesturing to the chair). I did this for 2 reasons: to move him away from blocking my exit from the room, and to put him at a lower level than me so that he was entirely in my view. I didn’t interrupt him as he spoke. I just nodded or tilted my head to show I was listening. In my mind, I played out the various scenarios that could ensue.

Like a Code Blue for medical emergencies, a Code Lavender is called when a health care worker is in need of emotional or spiritual support.

Fortunately, I was able to get him to relax enough for an assessment, which involved a more relevant history and the exam, which he agreed to once an aide had come into the room. He did not exhibit the concerning signs of flushed skin, dilated pupils, shallow rapid respirations, or perspiration. He did have a comorbid behavioral health issue, which we were able to address. His earlier behavioral indicators of agitation were controlled with verbal and physical cues on my part. Our conversation didn't reveal an intent to harm himself or others. In this case, physical restraints were not required. Throughout the encounter the door was left open, and the patient was reminded that we were there to help.

Once he left, I made the relevant notes in the chart regarding his agitated state at the start of the visit and his final state at the end of the visit so as to assist any other providers. We (TIM, MG) also held a quick debrief after the encounter with the office staff and decided that we needed to create a policy and protocol regarding how to handle such situations in the future.

CORRESPONDENCE
Tochi Iroku-Malize, MD, MPH, MBA, Family Medicine Department, Southside Hospital, 301 East Main Street, Bay Shore, NY 11706; [email protected].

CASE 1

Lisa G, 31 years old, gravida 0, complains of severe dysmenorrhea that began when she discontinued an oral contraceptive (OC) one year ago. Prior to stopping the OC, she had been taking an OC without interruption since she was 28, during which time she continued to have moderate symptoms of dysmenorrhea. Before taking an OC, the patient had a trial of an etonogestrel implant, which was removed because of irregular bleeding, and depot medroxyprogesterone acetate (MPA) injection, which she discontinued because of associated weight gain and fatigue.

Ms. G is not sexually active and doesn’t want to start a family at this time, but is interested in having a diagnosis. She has no other medical problems, no surgical history, and no history of sexually transmitted infection. She reports that her mother and sister had endometriosis, including pain that resolved after definitive treatment.

Ms. G reports menstrual cycles that are exquisitely painful and occur regularly (every 28 days for 4 or 5 days), with a moderate volume of bleeding that requires a regular-size tampon change every 4 to 6 hours. She reports crampy abdominal pain as 10, on a scale of one to 10; dyschezia (without hematochezia); and generalized achy abdominal pain that is continuous during menses. Pain is partially controlled by ibuprofen, 800 mg every 8 hours. Ms. G also describes gastrointestinal symptoms of bloating, constipation preceding her menstrual cycle, diarrhea during her menses, and occasionally nausea and vomiting with the severe pain.

On examination (which is not performed during menses), Ms. G appears well and is not in acute distress. Abdominal examination is benign. There is no tenderness to palpation or distension; bowel sounds are normal. Pelvic examination reveals mild tenderness upon palpation of a small and mobile uterus. Rectal examination is normal. She has no signs of hyperandrogenism (eg, male-pattern body hair, central obesity).

CASE 2

Rhonda M, 42 years old, gravida 3, para 3003, reports continuous pelvic pain for 7 years that is exacerbated by defecation, intercourse, and insertion of a tampon. She has a low level of dull baseline pain (3, on scale of one to 10) that occasionally spikes up to sharp, knifelike pain (10 on the pain scale), which, she says, brings her to tears. Ms. M describes the pain as “deep inside,” central in her pelvis, and radiating to the left and right, particularly during pain flares.

The patient’s 3 children were born by spontaneous vaginal delivery; however, she recalls that her youngest son was born via a traumatic vaginal delivery 8 years ago (he “got stuck coming out,” she reports). The only other component of Ms. M’s medical history is an anxiety disorder, for which she takes citalopram. She has a family history of cervical cancer.

Ms. M’s past diagnostic work-up for pelvic pain includes pelvic ultrasonography, endometrial biopsy, Pap smear, and diagnostic laparoscopy—all normal. She had a negative gastrointestinal work-up, including upper- and lower-tract endoscopy. Medical therapy, including opioids and nonsteroidal anti-inflammatory drugs (NSAIDs), did not provide significant relief of pain.

Despite the negative work-up, Ms. M is still concerned that the pain might be related to cancer. With her family history of cervical cancer, she says that she does not want to “miss anything.”

Ms. M is thin and appears anxious. The abdomen is mildly and diffusely tender to palpation with normal bowel sounds and no distension. Pelvic examination reveals some hyperesthesia upon single-digit palpation of the pelvic floor. Placement of the speculum is difficult because of discomfort.

How would you proceed with the care of these patients?

What is chronic pelvic pain? Why is management such a challenge?

Chronic pelvic pain (CPP) is defined as chronic or intermittent cyclic or noncyclic pelvic pain lasting longer than 6 months, localized to the pelvis, diminishing a woman’s quality of life, and requiring medical intervention.¹ It’s estimated that CPP affects as many as 15% of women of reproductive age in the United States each year, at a cost to the health care system of approximately $2 billion annually.^2,3

Chronic pelvic pain affects as many as 15% of women of reproductive age in the United States annually.

CPP can result from abnormal pain responses from multiple body systems, including gynecologic conditions such as endometriosis. Notably, a nongynecologic cause is more often the major pain generator, without significant identifiable pathology (TABLE 1). Like all chronic pain disorders, CPP can also result in central sensitization of the nervous system, altering how pain is processed at the level of the pain matrix in the brain.⁴

This article reviews the limited evidence for treating CPP and offers recommendations for the primary care physician on providing symptomatic relief in the absence of diagnosed pathology (TABLE 2^5-13).

Treatment

Analgesics

NSAIDs are frequently used as first-line treatment for any kind of pain, including CPP. There is some evidence of benefit from NSAIDs, compared to placebo, in cyclic CPP secondary to dysmenorrhea and endometriosis;^5,6 however, evidence of effectiveness in noncyclic CPP is absent. Because of the low cost and availability of NSAIDs, a trial is reasonable as a first-line intervention, particularly in CPP suspected to be endometriosis or of musculoskeletal origin. NSAIDs can cause adverse effects, including nausea, vomiting, headache, and drowsiness in 11% to 14% of women, although these agents are generally well-tolerated on a short-term basis.⁵

Opioids bind to opioid receptors in the central and peripheral nervous systems, resulting in an analgesic effect. Guidelines issued in 2016 by the Centers for Disease Control and Prevention recommend safer prescribing through careful evaluation of the risks and benefits of opioids for pain not caused by cancer and for palliation as part of end-of-life care.¹⁴

NSAIDs have inconclusive benefit over placebo in chronic pelvic pain secondary to endometriosis.

The risks of opioid use are well known in the medical community; they include tolerance, physical dependence, misuse, and death, in addition to common adverse effects such as nausea and vomiting, itching, constipation, and fatigue.^14,15 Because of those risks and limited long-term benefit in nonmalignant pain disorders, opioid therapy for CPP should be avoided.¹⁴ For patients already taking an opioid, discuss a strategy for weaning and, if possible, provide home naloxone therapy in the event of accidental overdose.¹⁴

Hormonal therapy

Hormonal therapies are the most common nonsurgical treatment of noncyclic CPP, with or without a definitive diagnosis of endometriosis, in reproductive-age women with CPP.

Combined OCs, despite a lack of quality evidence, are frequently the first hormonal treatment tried in both cyclic and noncyclic CPP. A low-dosage OC may decrease cyclic pain in endometriosis, although it can increase irregular bleeding and nausea.¹⁶ As many as 53% of women with CPP reported having undergone a trial of an OC for endometriosis, despite the absence of consistent evidence showing effectiveness in CPP.¹⁷

Depot MPA, in trials, decreased pain more than placebo. It can be tried as a treatment, but its use is often limited because of adverse effects, such as weight gain and bloating.⁸

A trial of a levonorgestrel-releasing intrauterine device (LNG-IUD) is supported by moderate-quality evidence for women whose CPP is thought to be a symptom of endometriosis or to have another uterine origin.⁷

Gonadotropin-releasing hormone agonists, such as depot leuprolide and goserelin acetate implant, may be considered in a woman with a diagnosis of endometriosis whose pelvic pain is not alleviated by MPA or an LNG-IUD.⁹

Nonhormonal therapies

CPP shares pain mechanisms with other pain syndromes, such as neuropathic pain. Antineuropathic medications, such as gabapentin and pregabalin, may, therefore, provide benefit. These medications also produce improvement in pain disorders of the musculoskeletal system, which may contribute to their analgesic effect.¹⁸

Gabapentin and amitriptyline have been studied in CPP; both were found successful in decreasing perceived pain. Of note, patients who received gabapentin, a gamma-aminobutyric acid analogue, with or without amitriptyline, had more pain relief than those treated with amitriptyline alone.¹⁰ Adverse effects of these medications may limit their use (TABLE 3^19-25).

Tricyclic antidepressants are well-supported, effective treatments for chronic pain through the central increase of norepinephrine. Beginning at a low dosage to diminish adverse effects (TABLE 3^19-25) and increasing the dosage slowly to an effective level may increase adherence. A trial of at least 6 to 8 weeks, at a moderate dosage, is recommended before discontinuing the medication. Although amitriptyline has the most evidence for value in the management of CPP disorders,¹⁰ second-generation tricyclic antidepressants nortriptyline and desipramine have also been used for pain control, and may be better tolerated.

Duloxetine and venlafaxine—serotonin–norepinephrine reuptake inhibitors—increase serotonin in addition to norepinephrine, which is believed to result in pain control. Although a systematic review of trials of duloxetine for chronic pain showed some improvement in diabetic peripheral neuropathy, fibromyalgia, chronic low back pain, and osteoarthritis, the review excluded CPP in its analysis.²⁶

Opioid therapy should be avoided in chronic pelvic pain because of significant risks presented by these agents and their limited long-term benefit in nonmalignant pain disorders.

In our opinion, a selective neurotransmitter reuptake inhibitor can be attempted to diminish the central pain sensitization of CPP. As with all drugs that increase the availability of serotonin, serotonin syndrome is a rare risk. Additionally, when stopping duloxetine, a prolonged taper may be required.

Pelvic floor dysfunction of the musculature within the bony pelvis may contribute to, or cause, CPP. The pelvic floor musculature may be hypertonic or hypotonic, and trigger points may exist. Despite the frequency of pelvic floor dysfunction, detailed examination of the pelvic floor is not routinely performed during a pelvic exam.

Because of the high prevalence of pelvic floor dysfunction in women with CPP, evaluation of the pelvic floor muscles is warranted.²⁷ (A protocol for this evaluation is detailed in TABLE 4.) Pelvic dynamometry may indicate muscle spasm or chronic tension; palpation of the pelvic floor during the exam can also identify a pain generator.

Although it might be difficult to distinguish pelvic floor myofascial pain as the primary or secondary cause of pain, pelvic floor physical therapy may clarify the role of the pelvic floor response (depending on the patient’s clinical exam and history). A low-quality retrospective case study on pelvic floor physical therapy reported significant improvement in pain that was proportional to the number of sessions completed.¹¹ Trigger-point injections and injections of botulinum toxin A have been used with reported improvement in the pelvic floor pain profile, and there is evidence to support the benefit of such injections in pelvic muscle dysfunction.¹²

Psychotherapy

Cognitive behavioral therapy (CBT) is well established as an option to manage a patient’s response to pain, including teaching coping skills for a chronic pain disorder and pain flares. Evidence supports using CBT or mindfulness techniques over usual care in reducing the intensity of pain in chronic low back pain,²⁸ and may be helpful in CPP. Patients with CPP who received 10 treatments of Mensendieck somatocognitive therapy (a mind–body therapy technique popular in Europe) over 90 days, compared with standard treatment alone, demonstrated improvement in pain, motor function, and psychological distress that persisted 9 months after treatment.¹³

Lifestyle changes, complementary and alternative therapies

Although medical and nonpharmacotherapeutic treatments are often important in the management of CPP, lifestyle modifications should be addressed initially and throughout treatment. Specifically, in patients with chronic, nonmalignant pain, diet modifications, exercise, complementary and alternative therapies, and sleep improvement can improve the patient’s ability to manage baseline pain and pain flares.

Diet modifications may relieve pain in some women with CPP. Although a systematic review in 2011 highlighted the lack of data available for the efficacy of dietary therapies for treating CPP, the authors did present data that a diet rich in antioxidants might alleviate pain sysmptoms.²⁹ Also, a gluten-free diet might reduce the symptoms of pain related to endometriosis and, thus, improve physical functioning, among other health domains.³⁰

Exercise can have important benefits as part of a treatment plan for chronic pelvic pain.

Exercise can be an important factor in the management of CPP, as with other chronic pain syndromes. In functional pain syndromes, the addition or maintenance of an exercise program has been shown to decrease the amount of pain medications required, improve depressive symptoms, increase energy, and decrease stress. Exercise also improves sleep quality and one’s ability to cope with pain.³¹

Yoga provides a good balance of aerobic and muscle-building activity and, in the authors’ experience, is tolerated by most women with CPP.

Acupuncture has limited evidence in the treatment of pelvic pain in women. Of the available studies, most are limited to pain related to endometriosis.³²

A recent meta-analysis reported mild-to-moderate immediate improvement in patients' pain after nonpharmacotherapeutic sleep interventions.

Sleep hygiene may be an important consideration in managing CPP. Sleep disturbances are reported in more than 80% of women with CPP,³³ including excessive time in bed and frequent napping, resulting in daytime fatigue and feeling generally unrested. A recent meta-analysis reported mild-to-moderate immediate improvement in patients’ pain after nonpharmacotherapeutic sleep interventions.³⁴ The National Sleep Foundation has produced a patient guide to assist in sleep hygiene.³⁵

Devising a management strategy despite sparse evidence

Because the cause of noncyclic CPP may be multifactorial, and because the literature on the etiology of CPP is limited (and, when there is research, it is inconclusive or of poor quality³⁶), there are few evidence-based recommendations for treating CPP. Given the paucity of quality evidence, physicians should treat patients empirically, based on their experience and their familiarity with the range of medical and nonpharmacotherapeutic options used to manage other chronic pain syndromes.

CASE 1

Ms. G’s cyclic pelvic pain was present only during menses. The dyschezia, severe pain that began only after she discontinued a combined OC, aching pain, and severe menstrual cramps are, taken together, suggestive of endometriosis, despite a normal physical exam.

Medical and surgical options were reviewed with Ms. G. She elected to undergo diagnostic laparoscopy. Several extrauterine foci of endometrial tissue were noted and excised; an LNG-IUD was inserted. Her pain improved significantly after surgery.

CASE 2

Ms. M was found to have significant pain on single-digit examination of the pelvic floor muscles, indicating likely pelvic floor muscle dysfunction. Pelvic dynamometry revealed significant tightness and spasm in the pelvic floor muscles—specifically, the levator ani complex.

Ms. M was started on gabapentin to reduce baseline pain and was referred for pelvic floor physical therapy. She felt reassured that her risk of cancer was low, considering her negative work-up, and that cancer was not the cause of her pain. Her symptoms improved greatly with a regimen of medical and physical therapy, although she continues to experience pain flares.

CORRESPONDENCE
Wendy S. Biggs, MD, Central Michigan University College of Medicine, 1632 Stone St., Saginaw, MI 48602; [email protected].

CASE 1

Lisa G, 31 years old, gravida 0, complains of severe dysmenorrhea that began when she discontinued an oral contraceptive (OC) one year ago. Prior to stopping the OC, she had been taking an OC without interruption since she was 28, during which time she continued to have moderate symptoms of dysmenorrhea. Before taking an OC, the patient had a trial of an etonogestrel implant, which was removed because of irregular bleeding, and depot medroxyprogesterone acetate (MPA) injection, which she discontinued because of associated weight gain and fatigue.

Ms. G is not sexually active and doesn’t want to start a family at this time, but is interested in having a diagnosis. She has no other medical problems, no surgical history, and no history of sexually transmitted infection. She reports that her mother and sister had endometriosis, including pain that resolved after definitive treatment.

Ms. G reports menstrual cycles that are exquisitely painful and occur regularly (every 28 days for 4 or 5 days), with a moderate volume of bleeding that requires a regular-size tampon change every 4 to 6 hours. She reports crampy abdominal pain as 10, on a scale of one to 10; dyschezia (without hematochezia); and generalized achy abdominal pain that is continuous during menses. Pain is partially controlled by ibuprofen, 800 mg every 8 hours. Ms. G also describes gastrointestinal symptoms of bloating, constipation preceding her menstrual cycle, diarrhea during her menses, and occasionally nausea and vomiting with the severe pain.

On examination (which is not performed during menses), Ms. G appears well and is not in acute distress. Abdominal examination is benign. There is no tenderness to palpation or distension; bowel sounds are normal. Pelvic examination reveals mild tenderness upon palpation of a small and mobile uterus. Rectal examination is normal. She has no signs of hyperandrogenism (eg, male-pattern body hair, central obesity).

CASE 2

Rhonda M, 42 years old, gravida 3, para 3003, reports continuous pelvic pain for 7 years that is exacerbated by defecation, intercourse, and insertion of a tampon. She has a low level of dull baseline pain (3, on scale of one to 10) that occasionally spikes up to sharp, knifelike pain (10 on the pain scale), which, she says, brings her to tears. Ms. M describes the pain as “deep inside,” central in her pelvis, and radiating to the left and right, particularly during pain flares.

The patient’s 3 children were born by spontaneous vaginal delivery; however, she recalls that her youngest son was born via a traumatic vaginal delivery 8 years ago (he “got stuck coming out,” she reports). The only other component of Ms. M’s medical history is an anxiety disorder, for which she takes citalopram. She has a family history of cervical cancer.

Ms. M’s past diagnostic work-up for pelvic pain includes pelvic ultrasonography, endometrial biopsy, Pap smear, and diagnostic laparoscopy—all normal. She had a negative gastrointestinal work-up, including upper- and lower-tract endoscopy. Medical therapy, including opioids and nonsteroidal anti-inflammatory drugs (NSAIDs), did not provide significant relief of pain.

Despite the negative work-up, Ms. M is still concerned that the pain might be related to cancer. With her family history of cervical cancer, she says that she does not want to “miss anything.”

Ms. M is thin and appears anxious. The abdomen is mildly and diffusely tender to palpation with normal bowel sounds and no distension. Pelvic examination reveals some hyperesthesia upon single-digit palpation of the pelvic floor. Placement of the speculum is difficult because of discomfort.

How would you proceed with the care of these patients?

What is chronic pelvic pain? Why is management such a challenge?

Chronic pelvic pain (CPP) is defined as chronic or intermittent cyclic or noncyclic pelvic pain lasting longer than 6 months, localized to the pelvis, diminishing a woman’s quality of life, and requiring medical intervention.¹ It’s estimated that CPP affects as many as 15% of women of reproductive age in the United States each year, at a cost to the health care system of approximately $2 billion annually.^2,3

Chronic pelvic pain affects as many as 15% of women of reproductive age in the United States annually.

CPP can result from abnormal pain responses from multiple body systems, including gynecologic conditions such as endometriosis. Notably, a nongynecologic cause is more often the major pain generator, without significant identifiable pathology (TABLE 1). Like all chronic pain disorders, CPP can also result in central sensitization of the nervous system, altering how pain is processed at the level of the pain matrix in the brain.⁴

This article reviews the limited evidence for treating CPP and offers recommendations for the primary care physician on providing symptomatic relief in the absence of diagnosed pathology (TABLE 2^5-13).

Treatment

Analgesics

NSAIDs are frequently used as first-line treatment for any kind of pain, including CPP. There is some evidence of benefit from NSAIDs, compared to placebo, in cyclic CPP secondary to dysmenorrhea and endometriosis;^5,6 however, evidence of effectiveness in noncyclic CPP is absent. Because of the low cost and availability of NSAIDs, a trial is reasonable as a first-line intervention, particularly in CPP suspected to be endometriosis or of musculoskeletal origin. NSAIDs can cause adverse effects, including nausea, vomiting, headache, and drowsiness in 11% to 14% of women, although these agents are generally well-tolerated on a short-term basis.⁵

Opioids bind to opioid receptors in the central and peripheral nervous systems, resulting in an analgesic effect. Guidelines issued in 2016 by the Centers for Disease Control and Prevention recommend safer prescribing through careful evaluation of the risks and benefits of opioids for pain not caused by cancer and for palliation as part of end-of-life care.¹⁴

NSAIDs have inconclusive benefit over placebo in chronic pelvic pain secondary to endometriosis.

The risks of opioid use are well known in the medical community; they include tolerance, physical dependence, misuse, and death, in addition to common adverse effects such as nausea and vomiting, itching, constipation, and fatigue.^14,15 Because of those risks and limited long-term benefit in nonmalignant pain disorders, opioid therapy for CPP should be avoided.¹⁴ For patients already taking an opioid, discuss a strategy for weaning and, if possible, provide home naloxone therapy in the event of accidental overdose.¹⁴

Hormonal therapy

Hormonal therapies are the most common nonsurgical treatment of noncyclic CPP, with or without a definitive diagnosis of endometriosis, in reproductive-age women with CPP.

Combined OCs, despite a lack of quality evidence, are frequently the first hormonal treatment tried in both cyclic and noncyclic CPP. A low-dosage OC may decrease cyclic pain in endometriosis, although it can increase irregular bleeding and nausea.¹⁶ As many as 53% of women with CPP reported having undergone a trial of an OC for endometriosis, despite the absence of consistent evidence showing effectiveness in CPP.¹⁷

Depot MPA, in trials, decreased pain more than placebo. It can be tried as a treatment, but its use is often limited because of adverse effects, such as weight gain and bloating.⁸

A trial of a levonorgestrel-releasing intrauterine device (LNG-IUD) is supported by moderate-quality evidence for women whose CPP is thought to be a symptom of endometriosis or to have another uterine origin.⁷

Gonadotropin-releasing hormone agonists, such as depot leuprolide and goserelin acetate implant, may be considered in a woman with a diagnosis of endometriosis whose pelvic pain is not alleviated by MPA or an LNG-IUD.⁹

Nonhormonal therapies

CPP shares pain mechanisms with other pain syndromes, such as neuropathic pain. Antineuropathic medications, such as gabapentin and pregabalin, may, therefore, provide benefit. These medications also produce improvement in pain disorders of the musculoskeletal system, which may contribute to their analgesic effect.¹⁸

Gabapentin and amitriptyline have been studied in CPP; both were found successful in decreasing perceived pain. Of note, patients who received gabapentin, a gamma-aminobutyric acid analogue, with or without amitriptyline, had more pain relief than those treated with amitriptyline alone.¹⁰ Adverse effects of these medications may limit their use (TABLE 3^19-25).

Tricyclic antidepressants are well-supported, effective treatments for chronic pain through the central increase of norepinephrine. Beginning at a low dosage to diminish adverse effects (TABLE 3^19-25) and increasing the dosage slowly to an effective level may increase adherence. A trial of at least 6 to 8 weeks, at a moderate dosage, is recommended before discontinuing the medication. Although amitriptyline has the most evidence for value in the management of CPP disorders,¹⁰ second-generation tricyclic antidepressants nortriptyline and desipramine have also been used for pain control, and may be better tolerated.

Duloxetine and venlafaxine—serotonin–norepinephrine reuptake inhibitors—increase serotonin in addition to norepinephrine, which is believed to result in pain control. Although a systematic review of trials of duloxetine for chronic pain showed some improvement in diabetic peripheral neuropathy, fibromyalgia, chronic low back pain, and osteoarthritis, the review excluded CPP in its analysis.²⁶

Opioid therapy should be avoided in chronic pelvic pain because of significant risks presented by these agents and their limited long-term benefit in nonmalignant pain disorders.

In our opinion, a selective neurotransmitter reuptake inhibitor can be attempted to diminish the central pain sensitization of CPP. As with all drugs that increase the availability of serotonin, serotonin syndrome is a rare risk. Additionally, when stopping duloxetine, a prolonged taper may be required.

Pelvic floor dysfunction of the musculature within the bony pelvis may contribute to, or cause, CPP. The pelvic floor musculature may be hypertonic or hypotonic, and trigger points may exist. Despite the frequency of pelvic floor dysfunction, detailed examination of the pelvic floor is not routinely performed during a pelvic exam.

Because of the high prevalence of pelvic floor dysfunction in women with CPP, evaluation of the pelvic floor muscles is warranted.²⁷ (A protocol for this evaluation is detailed in TABLE 4.) Pelvic dynamometry may indicate muscle spasm or chronic tension; palpation of the pelvic floor during the exam can also identify a pain generator.

Although it might be difficult to distinguish pelvic floor myofascial pain as the primary or secondary cause of pain, pelvic floor physical therapy may clarify the role of the pelvic floor response (depending on the patient’s clinical exam and history). A low-quality retrospective case study on pelvic floor physical therapy reported significant improvement in pain that was proportional to the number of sessions completed.¹¹ Trigger-point injections and injections of botulinum toxin A have been used with reported improvement in the pelvic floor pain profile, and there is evidence to support the benefit of such injections in pelvic muscle dysfunction.¹²

Psychotherapy

Cognitive behavioral therapy (CBT) is well established as an option to manage a patient’s response to pain, including teaching coping skills for a chronic pain disorder and pain flares. Evidence supports using CBT or mindfulness techniques over usual care in reducing the intensity of pain in chronic low back pain,²⁸ and may be helpful in CPP. Patients with CPP who received 10 treatments of Mensendieck somatocognitive therapy (a mind–body therapy technique popular in Europe) over 90 days, compared with standard treatment alone, demonstrated improvement in pain, motor function, and psychological distress that persisted 9 months after treatment.¹³

Lifestyle changes, complementary and alternative therapies

Although medical and nonpharmacotherapeutic treatments are often important in the management of CPP, lifestyle modifications should be addressed initially and throughout treatment. Specifically, in patients with chronic, nonmalignant pain, diet modifications, exercise, complementary and alternative therapies, and sleep improvement can improve the patient’s ability to manage baseline pain and pain flares.

Diet modifications may relieve pain in some women with CPP. Although a systematic review in 2011 highlighted the lack of data available for the efficacy of dietary therapies for treating CPP, the authors did present data that a diet rich in antioxidants might alleviate pain sysmptoms.²⁹ Also, a gluten-free diet might reduce the symptoms of pain related to endometriosis and, thus, improve physical functioning, among other health domains.³⁰

Exercise can have important benefits as part of a treatment plan for chronic pelvic pain.

Exercise can be an important factor in the management of CPP, as with other chronic pain syndromes. In functional pain syndromes, the addition or maintenance of an exercise program has been shown to decrease the amount of pain medications required, improve depressive symptoms, increase energy, and decrease stress. Exercise also improves sleep quality and one’s ability to cope with pain.³¹

Yoga provides a good balance of aerobic and muscle-building activity and, in the authors’ experience, is tolerated by most women with CPP.

Acupuncture has limited evidence in the treatment of pelvic pain in women. Of the available studies, most are limited to pain related to endometriosis.³²

A recent meta-analysis reported mild-to-moderate immediate improvement in patients' pain after nonpharmacotherapeutic sleep interventions.

Sleep hygiene may be an important consideration in managing CPP. Sleep disturbances are reported in more than 80% of women with CPP,³³ including excessive time in bed and frequent napping, resulting in daytime fatigue and feeling generally unrested. A recent meta-analysis reported mild-to-moderate immediate improvement in patients’ pain after nonpharmacotherapeutic sleep interventions.³⁴ The National Sleep Foundation has produced a patient guide to assist in sleep hygiene.³⁵

Devising a management strategy despite sparse evidence

Because the cause of noncyclic CPP may be multifactorial, and because the literature on the etiology of CPP is limited (and, when there is research, it is inconclusive or of poor quality³⁶), there are few evidence-based recommendations for treating CPP. Given the paucity of quality evidence, physicians should treat patients empirically, based on their experience and their familiarity with the range of medical and nonpharmacotherapeutic options used to manage other chronic pain syndromes.

CASE 1

Ms. G’s cyclic pelvic pain was present only during menses. The dyschezia, severe pain that began only after she discontinued a combined OC, aching pain, and severe menstrual cramps are, taken together, suggestive of endometriosis, despite a normal physical exam.

Medical and surgical options were reviewed with Ms. G. She elected to undergo diagnostic laparoscopy. Several extrauterine foci of endometrial tissue were noted and excised; an LNG-IUD was inserted. Her pain improved significantly after surgery.

CASE 2

Ms. M was found to have significant pain on single-digit examination of the pelvic floor muscles, indicating likely pelvic floor muscle dysfunction. Pelvic dynamometry revealed significant tightness and spasm in the pelvic floor muscles—specifically, the levator ani complex.

Ms. M was started on gabapentin to reduce baseline pain and was referred for pelvic floor physical therapy. She felt reassured that her risk of cancer was low, considering her negative work-up, and that cancer was not the cause of her pain. Her symptoms improved greatly with a regimen of medical and physical therapy, although she continues to experience pain flares.

CORRESPONDENCE
Wendy S. Biggs, MD, Central Michigan University College of Medicine, 1632 Stone St., Saginaw, MI 48602; [email protected].

CASE 1

Lisa G, 31 years old, gravida 0, complains of severe dysmenorrhea that began when she discontinued an oral contraceptive (OC) one year ago. Prior to stopping the OC, she had been taking an OC without interruption since she was 28, during which time she continued to have moderate symptoms of dysmenorrhea. Before taking an OC, the patient had a trial of an etonogestrel implant, which was removed because of irregular bleeding, and depot medroxyprogesterone acetate (MPA) injection, which she discontinued because of associated weight gain and fatigue.

Ms. G is not sexually active and doesn’t want to start a family at this time, but is interested in having a diagnosis. She has no other medical problems, no surgical history, and no history of sexually transmitted infection. She reports that her mother and sister had endometriosis, including pain that resolved after definitive treatment.

Ms. G reports menstrual cycles that are exquisitely painful and occur regularly (every 28 days for 4 or 5 days), with a moderate volume of bleeding that requires a regular-size tampon change every 4 to 6 hours. She reports crampy abdominal pain as 10, on a scale of one to 10; dyschezia (without hematochezia); and generalized achy abdominal pain that is continuous during menses. Pain is partially controlled by ibuprofen, 800 mg every 8 hours. Ms. G also describes gastrointestinal symptoms of bloating, constipation preceding her menstrual cycle, diarrhea during her menses, and occasionally nausea and vomiting with the severe pain.

On examination (which is not performed during menses), Ms. G appears well and is not in acute distress. Abdominal examination is benign. There is no tenderness to palpation or distension; bowel sounds are normal. Pelvic examination reveals mild tenderness upon palpation of a small and mobile uterus. Rectal examination is normal. She has no signs of hyperandrogenism (eg, male-pattern body hair, central obesity).

CASE 2

Rhonda M, 42 years old, gravida 3, para 3003, reports continuous pelvic pain for 7 years that is exacerbated by defecation, intercourse, and insertion of a tampon. She has a low level of dull baseline pain (3, on scale of one to 10) that occasionally spikes up to sharp, knifelike pain (10 on the pain scale), which, she says, brings her to tears. Ms. M describes the pain as “deep inside,” central in her pelvis, and radiating to the left and right, particularly during pain flares.

The patient’s 3 children were born by spontaneous vaginal delivery; however, she recalls that her youngest son was born via a traumatic vaginal delivery 8 years ago (he “got stuck coming out,” she reports). The only other component of Ms. M’s medical history is an anxiety disorder, for which she takes citalopram. She has a family history of cervical cancer.

Ms. M’s past diagnostic work-up for pelvic pain includes pelvic ultrasonography, endometrial biopsy, Pap smear, and diagnostic laparoscopy—all normal. She had a negative gastrointestinal work-up, including upper- and lower-tract endoscopy. Medical therapy, including opioids and nonsteroidal anti-inflammatory drugs (NSAIDs), did not provide significant relief of pain.

Despite the negative work-up, Ms. M is still concerned that the pain might be related to cancer. With her family history of cervical cancer, she says that she does not want to “miss anything.”

Ms. M is thin and appears anxious. The abdomen is mildly and diffusely tender to palpation with normal bowel sounds and no distension. Pelvic examination reveals some hyperesthesia upon single-digit palpation of the pelvic floor. Placement of the speculum is difficult because of discomfort.

How would you proceed with the care of these patients?

What is chronic pelvic pain? Why is management such a challenge?

Chronic pelvic pain (CPP) is defined as chronic or intermittent cyclic or noncyclic pelvic pain lasting longer than 6 months, localized to the pelvis, diminishing a woman’s quality of life, and requiring medical intervention.¹ It’s estimated that CPP affects as many as 15% of women of reproductive age in the United States each year, at a cost to the health care system of approximately $2 billion annually.^2,3

Chronic pelvic pain affects as many as 15% of women of reproductive age in the United States annually.

CPP can result from abnormal pain responses from multiple body systems, including gynecologic conditions such as endometriosis. Notably, a nongynecologic cause is more often the major pain generator, without significant identifiable pathology (TABLE 1). Like all chronic pain disorders, CPP can also result in central sensitization of the nervous system, altering how pain is processed at the level of the pain matrix in the brain.⁴

This article reviews the limited evidence for treating CPP and offers recommendations for the primary care physician on providing symptomatic relief in the absence of diagnosed pathology (TABLE 2^5-13).

Treatment

Analgesics

NSAIDs are frequently used as first-line treatment for any kind of pain, including CPP. There is some evidence of benefit from NSAIDs, compared to placebo, in cyclic CPP secondary to dysmenorrhea and endometriosis;^5,6 however, evidence of effectiveness in noncyclic CPP is absent. Because of the low cost and availability of NSAIDs, a trial is reasonable as a first-line intervention, particularly in CPP suspected to be endometriosis or of musculoskeletal origin. NSAIDs can cause adverse effects, including nausea, vomiting, headache, and drowsiness in 11% to 14% of women, although these agents are generally well-tolerated on a short-term basis.⁵

Opioids bind to opioid receptors in the central and peripheral nervous systems, resulting in an analgesic effect. Guidelines issued in 2016 by the Centers for Disease Control and Prevention recommend safer prescribing through careful evaluation of the risks and benefits of opioids for pain not caused by cancer and for palliation as part of end-of-life care.¹⁴

NSAIDs have inconclusive benefit over placebo in chronic pelvic pain secondary to endometriosis.

The risks of opioid use are well known in the medical community; they include tolerance, physical dependence, misuse, and death, in addition to common adverse effects such as nausea and vomiting, itching, constipation, and fatigue.^14,15 Because of those risks and limited long-term benefit in nonmalignant pain disorders, opioid therapy for CPP should be avoided.¹⁴ For patients already taking an opioid, discuss a strategy for weaning and, if possible, provide home naloxone therapy in the event of accidental overdose.¹⁴

Hormonal therapy

Hormonal therapies are the most common nonsurgical treatment of noncyclic CPP, with or without a definitive diagnosis of endometriosis, in reproductive-age women with CPP.

Combined OCs, despite a lack of quality evidence, are frequently the first hormonal treatment tried in both cyclic and noncyclic CPP. A low-dosage OC may decrease cyclic pain in endometriosis, although it can increase irregular bleeding and nausea.¹⁶ As many as 53% of women with CPP reported having undergone a trial of an OC for endometriosis, despite the absence of consistent evidence showing effectiveness in CPP.¹⁷

Depot MPA, in trials, decreased pain more than placebo. It can be tried as a treatment, but its use is often limited because of adverse effects, such as weight gain and bloating.⁸

A trial of a levonorgestrel-releasing intrauterine device (LNG-IUD) is supported by moderate-quality evidence for women whose CPP is thought to be a symptom of endometriosis or to have another uterine origin.⁷

Gonadotropin-releasing hormone agonists, such as depot leuprolide and goserelin acetate implant, may be considered in a woman with a diagnosis of endometriosis whose pelvic pain is not alleviated by MPA or an LNG-IUD.⁹

Nonhormonal therapies

CPP shares pain mechanisms with other pain syndromes, such as neuropathic pain. Antineuropathic medications, such as gabapentin and pregabalin, may, therefore, provide benefit. These medications also produce improvement in pain disorders of the musculoskeletal system, which may contribute to their analgesic effect.¹⁸

Gabapentin and amitriptyline have been studied in CPP; both were found successful in decreasing perceived pain. Of note, patients who received gabapentin, a gamma-aminobutyric acid analogue, with or without amitriptyline, had more pain relief than those treated with amitriptyline alone.¹⁰ Adverse effects of these medications may limit their use (TABLE 3^19-25).

Tricyclic antidepressants are well-supported, effective treatments for chronic pain through the central increase of norepinephrine. Beginning at a low dosage to diminish adverse effects (TABLE 3^19-25) and increasing the dosage slowly to an effective level may increase adherence. A trial of at least 6 to 8 weeks, at a moderate dosage, is recommended before discontinuing the medication. Although amitriptyline has the most evidence for value in the management of CPP disorders,¹⁰ second-generation tricyclic antidepressants nortriptyline and desipramine have also been used for pain control, and may be better tolerated.

Duloxetine and venlafaxine—serotonin–norepinephrine reuptake inhibitors—increase serotonin in addition to norepinephrine, which is believed to result in pain control. Although a systematic review of trials of duloxetine for chronic pain showed some improvement in diabetic peripheral neuropathy, fibromyalgia, chronic low back pain, and osteoarthritis, the review excluded CPP in its analysis.²⁶

Opioid therapy should be avoided in chronic pelvic pain because of significant risks presented by these agents and their limited long-term benefit in nonmalignant pain disorders.

In our opinion, a selective neurotransmitter reuptake inhibitor can be attempted to diminish the central pain sensitization of CPP. As with all drugs that increase the availability of serotonin, serotonin syndrome is a rare risk. Additionally, when stopping duloxetine, a prolonged taper may be required.

Pelvic floor dysfunction of the musculature within the bony pelvis may contribute to, or cause, CPP. The pelvic floor musculature may be hypertonic or hypotonic, and trigger points may exist. Despite the frequency of pelvic floor dysfunction, detailed examination of the pelvic floor is not routinely performed during a pelvic exam.

Because of the high prevalence of pelvic floor dysfunction in women with CPP, evaluation of the pelvic floor muscles is warranted.²⁷ (A protocol for this evaluation is detailed in TABLE 4.) Pelvic dynamometry may indicate muscle spasm or chronic tension; palpation of the pelvic floor during the exam can also identify a pain generator.

Although it might be difficult to distinguish pelvic floor myofascial pain as the primary or secondary cause of pain, pelvic floor physical therapy may clarify the role of the pelvic floor response (depending on the patient’s clinical exam and history). A low-quality retrospective case study on pelvic floor physical therapy reported significant improvement in pain that was proportional to the number of sessions completed.¹¹ Trigger-point injections and injections of botulinum toxin A have been used with reported improvement in the pelvic floor pain profile, and there is evidence to support the benefit of such injections in pelvic muscle dysfunction.¹²

Psychotherapy

Cognitive behavioral therapy (CBT) is well established as an option to manage a patient’s response to pain, including teaching coping skills for a chronic pain disorder and pain flares. Evidence supports using CBT or mindfulness techniques over usual care in reducing the intensity of pain in chronic low back pain,²⁸ and may be helpful in CPP. Patients with CPP who received 10 treatments of Mensendieck somatocognitive therapy (a mind–body therapy technique popular in Europe) over 90 days, compared with standard treatment alone, demonstrated improvement in pain, motor function, and psychological distress that persisted 9 months after treatment.¹³

Lifestyle changes, complementary and alternative therapies

Although medical and nonpharmacotherapeutic treatments are often important in the management of CPP, lifestyle modifications should be addressed initially and throughout treatment. Specifically, in patients with chronic, nonmalignant pain, diet modifications, exercise, complementary and alternative therapies, and sleep improvement can improve the patient’s ability to manage baseline pain and pain flares.

Diet modifications may relieve pain in some women with CPP. Although a systematic review in 2011 highlighted the lack of data available for the efficacy of dietary therapies for treating CPP, the authors did present data that a diet rich in antioxidants might alleviate pain sysmptoms.²⁹ Also, a gluten-free diet might reduce the symptoms of pain related to endometriosis and, thus, improve physical functioning, among other health domains.³⁰

Exercise can have important benefits as part of a treatment plan for chronic pelvic pain.

Exercise can be an important factor in the management of CPP, as with other chronic pain syndromes. In functional pain syndromes, the addition or maintenance of an exercise program has been shown to decrease the amount of pain medications required, improve depressive symptoms, increase energy, and decrease stress. Exercise also improves sleep quality and one’s ability to cope with pain.³¹

Yoga provides a good balance of aerobic and muscle-building activity and, in the authors’ experience, is tolerated by most women with CPP.

Acupuncture has limited evidence in the treatment of pelvic pain in women. Of the available studies, most are limited to pain related to endometriosis.³²

A recent meta-analysis reported mild-to-moderate immediate improvement in patients' pain after nonpharmacotherapeutic sleep interventions.

Sleep hygiene may be an important consideration in managing CPP. Sleep disturbances are reported in more than 80% of women with CPP,³³ including excessive time in bed and frequent napping, resulting in daytime fatigue and feeling generally unrested. A recent meta-analysis reported mild-to-moderate immediate improvement in patients’ pain after nonpharmacotherapeutic sleep interventions.³⁴ The National Sleep Foundation has produced a patient guide to assist in sleep hygiene.³⁵

Devising a management strategy despite sparse evidence

Because the cause of noncyclic CPP may be multifactorial, and because the literature on the etiology of CPP is limited (and, when there is research, it is inconclusive or of poor quality³⁶), there are few evidence-based recommendations for treating CPP. Given the paucity of quality evidence, physicians should treat patients empirically, based on their experience and their familiarity with the range of medical and nonpharmacotherapeutic options used to manage other chronic pain syndromes.

CASE 1

Ms. G’s cyclic pelvic pain was present only during menses. The dyschezia, severe pain that began only after she discontinued a combined OC, aching pain, and severe menstrual cramps are, taken together, suggestive of endometriosis, despite a normal physical exam.

Medical and surgical options were reviewed with Ms. G. She elected to undergo diagnostic laparoscopy. Several extrauterine foci of endometrial tissue were noted and excised; an LNG-IUD was inserted. Her pain improved significantly after surgery.

CASE 2

Ms. M was found to have significant pain on single-digit examination of the pelvic floor muscles, indicating likely pelvic floor muscle dysfunction. Pelvic dynamometry revealed significant tightness and spasm in the pelvic floor muscles—specifically, the levator ani complex.

Ms. M was started on gabapentin to reduce baseline pain and was referred for pelvic floor physical therapy. She felt reassured that her risk of cancer was low, considering her negative work-up, and that cancer was not the cause of her pain. Her symptoms improved greatly with a regimen of medical and physical therapy, although she continues to experience pain flares.

CORRESPONDENCE
Wendy S. Biggs, MD, Central Michigan University College of Medicine, 1632 Stone St., Saginaw, MI 48602; [email protected].

CASE An athletic 25-year-old woman presents to her family physician complaining of a painful and swollen knee. She says that she injured the knee the day before during a judo match. The injury occurred when her upper body suddenly changed direction while her foot remained planted and her knee rotated medially. A cruciate ligament injury immediately comes to mind, but other potential diagnoses include meniscal injury, collateral ligament injury, and patellar instability. The first step in determining an accurate diagnosis is to evaluate the stability of the knee by physical examination—often a difficult task immediately following an injury.

How would you proceed?

Rupture of the anterior cruciate ligament (ACL), partial or complete, is a common injury, especially in athletes who hurt their knee in a pivoting movement.¹ The number of patients who present with ACL injury is estimated at 252,000 per year.² Cruciate ligament injury may lead to complaints of instability with subsequent inability to engage in sports activities. Cruciate ligament injury is also associated with premature development of osteoarthritis later in life.³ Operative treatment seems to be superior to conservative treatment in improving both subjective and objective measures of knee instability and in helping athletes return to their former level of activity.⁴

Because early detection is key to achieving the best clinical outcome, it is essential that the most accurate physical examination tests are performed during the acute phase. Primary care physicians, emergency room doctors, physical therapists, and athletic trainers are the ones who most often see these patients immediately following the injury, and they often have only the physical examination with which to assess ACL injury. Their task is to identify the patient with potential ACL injury and to refer the patient swiftly.

Three physical examination tests are most commonly used to evaluate cruciate ligament injury. The best known and most frequently used technique is the anterior drawer test. The other 2 tests, the Lachman test and the pivot shift test, are more difficult to perform and are used less often, especially by physicians untrained in their use. In addition, there is a relatively new diagnostic test: the lever sign test. The aim of our article is to provide a short, clinically relevant overview of the literature and to assess the diagnostic value of physical examination for the primary care physician.

Anterior drawer test

How it’s done. In this test, the patient lies supine on the examination table with hips flexed to 45 degrees and knees flexed to 90 degrees (FIGURE 1).⁵ The examiner sits on the table with a leg resting on the patient's foot, grasps the tibia of the injured leg just below the knee, and draws the tibia forward. If the tibia, compared with the tibia of the uninjured leg, moves farther anteriorly, or if the endpoint feels softened or is absent, the result is positive for an ACL injury.

The literature. Nine systematic reviews conclude that the anterior drawer test is inferior to the Lachman test,^6-14 which we’ll describe in a moment. This is due, in part, to the anterior drawer test’s unacceptably low sensitivity and specificity in the clinical setting—especially during the acute phase.¹⁰ The most recent meta-analysis on the anterior drawer test reports a sensitivity of 38% and a specificity of 81%.⁹ In other words, out of 100 ruptured ligaments, only 38 will test positive with the anterior drawer test.

The literature offers possible explanations for findings on the test’s validity. First, rupture of the ACL is often accompanied by swelling of the knee caused by hemarthrosis and reactive synovitis that can prevent the patient from flexing the knee to 90 degrees. Second, the joint pain may induce a protective muscle action, also called guarding of the hamstrings, that creates a vector opposing the passive anterior translation.¹⁵

Apart from the matter of a test’s validity, it's also important to consider the test’s inter- and intra-rater reliability.¹⁶ Compared with the Lachman test, the anterior drawer test is inferior in reliability.⁷

Lachman test

How it’s done. The Lachman test is performed with the patient supine on the table and the injured knee flexed at 20 to 30 degrees (FIGURE 2).⁵ The examiner holds the patient’s thigh with one hand and places the other hand beneath the tibia with the thumb of that hand on the tibial joint line. As the tibia is pulled forward, firm resistance suggests an uninjured ACL. Free movement without a hard endpoint, compared with the uninjured knee, indicates ACL injury.

The literature. The Lachman test is the most accurate of the 3 diagnostic physical procedures. The most recent meta-analysis reports a sensitivity of 68% for partial ruptures and 96% for complete ACL ruptures.⁶ According to a recently published overview of systematic reviews, the Lachman test has high diagnostic value in confirming or ruling out an ACL injury.¹⁷

Two factors are important when assessing results of the Lachman test. The quantity of anterior translation of the tibia relative to the femur is as important as the quality of the endpoint of the anterior translation. Quantity of translation must always be compared with the unaffected knee. Quality of the endpoint in passive anterior translation should be assessed as “firm” or “sudden,” indicating an intact ACL, or as “absent, ill-defined, or softened,” indicating ACL pathology (TABLE).¹⁸

A drawback of the Lachman test is that it is challenging to perform correctly.¹⁹ The patient’s ability to relax the upper leg musculature is critically important. It is also essential to stabilize the distal femur, which can be problematic if the examiner has small hands relative to the size of the patient's leg musculature.¹⁰ These difficulties might be resolved by conducting the Lachman test with the patient in the prone position, known as the Prone Lachman.¹⁹ However, good evidence is not yet available to support this proposed solution. One systematic review, though, reports that the Prone Lachman test has the highest inter-rater reliability of all commonly used physical examination tests.⁷

The Lachman test is known as the test with highest validity on physical examination. When the outcome of a correctly performed Lachman test is negative, a rupture of the ACL is very unlikely.

Pivot shift test

How it’s done. With the patient lying supine on the table, the examiner uses one hand to hold the patient’s heel or ankle and the other hand to grasp the proximal portion of the lower leg (FIGURE 3).⁵ Lifting the leg to about 30 degrees from the table with the injured knee in full extension, the examiner rotates the foot or ankle medially, applies a valgus force to the knee, and slowly flexes it. During flexion, a ruptured ACL will cause the tibia to translate posteriorly to the femur. (Note that the starting position of the test has the tibia subluxed anteriorly. The posterior translation is the "pivot shift" back into the neutral position.)

The literature. The pivot shift test is technically more challenging to perform than the other 2 tests and is, therefore, less practical in the primary care setting. However, when this test is done correctly, a positive result is highly specific for ACL injury.^9,10 Reported sensitivity values are contradictory. The most recent meta-analysis reports a sensitivity of 85%.⁶ Two other studies cite much lower values: 24% and 28%.^9,10 These data suggest that the pivot shift test, when carried out correctly, can be of use in confirming a possible ACL rupture. However, the test should not be used alone in ruling out a possible ACL injury.

New diagnostic test: Lever sign test

How it’s done. The lever sign test (FIGURE 4),²⁰ introduced in the mid-2010s, is also performed with the patient lying in the supine position. The examiner stands at the side of the affected knee of the patient, places a closed fist just beneath the proximal third of the patient’s tibia, creating a slight flexion of the knee joint. With the other hand, the examiner applies a downward directed force to the distal third of the femur. With an intact ACL, the patient’s foot should rise from the table due to the induced lever mechanism. With a ruptured ACL, the lever effect is absent and the foot will not rise.

The literature. In the prospective clinical study that introduced the lever sign test, the sensitivity rate was reported at 100%—higher than that seen with the other commonly used tests.²⁰ Another study has reported that the lever sign test was easily adopted in clinical practice and showed higher sensitivity than the Lachman test (94% vs 80% in pre-anesthesia assessment).²¹ However, a more recent study has shown a sensitivity of 77% for the lever sign.²² The lever sign test is relatively easy to perform and requires less examiner strength than does the Lachman test. These factors enhance applicability of the lever sign test in the primary care office and in other settings such as physical therapy centers and emergency departments.

Applying this information in primary care

Given the importance of physical examination in diagnosing ACL injury, how can the current evidence best be applied in primary care practice? Based on its good test properties and feasibility, the Lachman test is preferred in primary care. The anterior drawer test can be used, but its low accuracy must be considered in making an assessment. The pivot shift test, given its difficulty of execution, should not be used by physicians unacquainted with it.

If future research supports early reports of the lever sign test’s accuracy, it could be very helpful in family practice. Going forward, research should aim at developing a constructive strategy for applying these physical examination tests in both primary care and specialty settings.

CORRESPONDENCE
Christiaan H. Koster, Department of Trauma Surgery, VU University Medical Centre, P.O. Box 7057, 1081 HV Amsterdam, The Netherlands; [email protected].

ACKNOWLEDGEMENTS
We thank Frits Oosterveld, PhD, for critically reviewing the manuscript and Ralph de Vries for his assistance in the literature search.

CASE An athletic 25-year-old woman presents to her family physician complaining of a painful and swollen knee. She says that she injured the knee the day before during a judo match. The injury occurred when her upper body suddenly changed direction while her foot remained planted and her knee rotated medially. A cruciate ligament injury immediately comes to mind, but other potential diagnoses include meniscal injury, collateral ligament injury, and patellar instability. The first step in determining an accurate diagnosis is to evaluate the stability of the knee by physical examination—often a difficult task immediately following an injury.

How would you proceed?

Rupture of the anterior cruciate ligament (ACL), partial or complete, is a common injury, especially in athletes who hurt their knee in a pivoting movement.¹ The number of patients who present with ACL injury is estimated at 252,000 per year.² Cruciate ligament injury may lead to complaints of instability with subsequent inability to engage in sports activities. Cruciate ligament injury is also associated with premature development of osteoarthritis later in life.³ Operative treatment seems to be superior to conservative treatment in improving both subjective and objective measures of knee instability and in helping athletes return to their former level of activity.⁴

Because early detection is key to achieving the best clinical outcome, it is essential that the most accurate physical examination tests are performed during the acute phase. Primary care physicians, emergency room doctors, physical therapists, and athletic trainers are the ones who most often see these patients immediately following the injury, and they often have only the physical examination with which to assess ACL injury. Their task is to identify the patient with potential ACL injury and to refer the patient swiftly.

Three physical examination tests are most commonly used to evaluate cruciate ligament injury. The best known and most frequently used technique is the anterior drawer test. The other 2 tests, the Lachman test and the pivot shift test, are more difficult to perform and are used less often, especially by physicians untrained in their use. In addition, there is a relatively new diagnostic test: the lever sign test. The aim of our article is to provide a short, clinically relevant overview of the literature and to assess the diagnostic value of physical examination for the primary care physician.

Anterior drawer test

How it’s done. In this test, the patient lies supine on the examination table with hips flexed to 45 degrees and knees flexed to 90 degrees (FIGURE 1).⁵ The examiner sits on the table with a leg resting on the patient's foot, grasps the tibia of the injured leg just below the knee, and draws the tibia forward. If the tibia, compared with the tibia of the uninjured leg, moves farther anteriorly, or if the endpoint feels softened or is absent, the result is positive for an ACL injury.

The literature. Nine systematic reviews conclude that the anterior drawer test is inferior to the Lachman test,^6-14 which we’ll describe in a moment. This is due, in part, to the anterior drawer test’s unacceptably low sensitivity and specificity in the clinical setting—especially during the acute phase.¹⁰ The most recent meta-analysis on the anterior drawer test reports a sensitivity of 38% and a specificity of 81%.⁹ In other words, out of 100 ruptured ligaments, only 38 will test positive with the anterior drawer test.

The literature offers possible explanations for findings on the test’s validity. First, rupture of the ACL is often accompanied by swelling of the knee caused by hemarthrosis and reactive synovitis that can prevent the patient from flexing the knee to 90 degrees. Second, the joint pain may induce a protective muscle action, also called guarding of the hamstrings, that creates a vector opposing the passive anterior translation.¹⁵

Apart from the matter of a test’s validity, it's also important to consider the test’s inter- and intra-rater reliability.¹⁶ Compared with the Lachman test, the anterior drawer test is inferior in reliability.⁷

Lachman test

How it’s done. The Lachman test is performed with the patient supine on the table and the injured knee flexed at 20 to 30 degrees (FIGURE 2).⁵ The examiner holds the patient’s thigh with one hand and places the other hand beneath the tibia with the thumb of that hand on the tibial joint line. As the tibia is pulled forward, firm resistance suggests an uninjured ACL. Free movement without a hard endpoint, compared with the uninjured knee, indicates ACL injury.

The literature. The Lachman test is the most accurate of the 3 diagnostic physical procedures. The most recent meta-analysis reports a sensitivity of 68% for partial ruptures and 96% for complete ACL ruptures.⁶ According to a recently published overview of systematic reviews, the Lachman test has high diagnostic value in confirming or ruling out an ACL injury.¹⁷

Two factors are important when assessing results of the Lachman test. The quantity of anterior translation of the tibia relative to the femur is as important as the quality of the endpoint of the anterior translation. Quantity of translation must always be compared with the unaffected knee. Quality of the endpoint in passive anterior translation should be assessed as “firm” or “sudden,” indicating an intact ACL, or as “absent, ill-defined, or softened,” indicating ACL pathology (TABLE).¹⁸

A drawback of the Lachman test is that it is challenging to perform correctly.¹⁹ The patient’s ability to relax the upper leg musculature is critically important. It is also essential to stabilize the distal femur, which can be problematic if the examiner has small hands relative to the size of the patient's leg musculature.¹⁰ These difficulties might be resolved by conducting the Lachman test with the patient in the prone position, known as the Prone Lachman.¹⁹ However, good evidence is not yet available to support this proposed solution. One systematic review, though, reports that the Prone Lachman test has the highest inter-rater reliability of all commonly used physical examination tests.⁷

The Lachman test is known as the test with highest validity on physical examination. When the outcome of a correctly performed Lachman test is negative, a rupture of the ACL is very unlikely.

Pivot shift test

How it’s done. With the patient lying supine on the table, the examiner uses one hand to hold the patient’s heel or ankle and the other hand to grasp the proximal portion of the lower leg (FIGURE 3).⁵ Lifting the leg to about 30 degrees from the table with the injured knee in full extension, the examiner rotates the foot or ankle medially, applies a valgus force to the knee, and slowly flexes it. During flexion, a ruptured ACL will cause the tibia to translate posteriorly to the femur. (Note that the starting position of the test has the tibia subluxed anteriorly. The posterior translation is the "pivot shift" back into the neutral position.)

The literature. The pivot shift test is technically more challenging to perform than the other 2 tests and is, therefore, less practical in the primary care setting. However, when this test is done correctly, a positive result is highly specific for ACL injury.^9,10 Reported sensitivity values are contradictory. The most recent meta-analysis reports a sensitivity of 85%.⁶ Two other studies cite much lower values: 24% and 28%.^9,10 These data suggest that the pivot shift test, when carried out correctly, can be of use in confirming a possible ACL rupture. However, the test should not be used alone in ruling out a possible ACL injury.

New diagnostic test: Lever sign test

How it’s done. The lever sign test (FIGURE 4),²⁰ introduced in the mid-2010s, is also performed with the patient lying in the supine position. The examiner stands at the side of the affected knee of the patient, places a closed fist just beneath the proximal third of the patient’s tibia, creating a slight flexion of the knee joint. With the other hand, the examiner applies a downward directed force to the distal third of the femur. With an intact ACL, the patient’s foot should rise from the table due to the induced lever mechanism. With a ruptured ACL, the lever effect is absent and the foot will not rise.

The literature. In the prospective clinical study that introduced the lever sign test, the sensitivity rate was reported at 100%—higher than that seen with the other commonly used tests.²⁰ Another study has reported that the lever sign test was easily adopted in clinical practice and showed higher sensitivity than the Lachman test (94% vs 80% in pre-anesthesia assessment).²¹ However, a more recent study has shown a sensitivity of 77% for the lever sign.²² The lever sign test is relatively easy to perform and requires less examiner strength than does the Lachman test. These factors enhance applicability of the lever sign test in the primary care office and in other settings such as physical therapy centers and emergency departments.

Applying this information in primary care

Given the importance of physical examination in diagnosing ACL injury, how can the current evidence best be applied in primary care practice? Based on its good test properties and feasibility, the Lachman test is preferred in primary care. The anterior drawer test can be used, but its low accuracy must be considered in making an assessment. The pivot shift test, given its difficulty of execution, should not be used by physicians unacquainted with it.

If future research supports early reports of the lever sign test’s accuracy, it could be very helpful in family practice. Going forward, research should aim at developing a constructive strategy for applying these physical examination tests in both primary care and specialty settings.

CORRESPONDENCE
Christiaan H. Koster, Department of Trauma Surgery, VU University Medical Centre, P.O. Box 7057, 1081 HV Amsterdam, The Netherlands; [email protected].

ACKNOWLEDGEMENTS
We thank Frits Oosterveld, PhD, for critically reviewing the manuscript and Ralph de Vries for his assistance in the literature search.

CASE An athletic 25-year-old woman presents to her family physician complaining of a painful and swollen knee. She says that she injured the knee the day before during a judo match. The injury occurred when her upper body suddenly changed direction while her foot remained planted and her knee rotated medially. A cruciate ligament injury immediately comes to mind, but other potential diagnoses include meniscal injury, collateral ligament injury, and patellar instability. The first step in determining an accurate diagnosis is to evaluate the stability of the knee by physical examination—often a difficult task immediately following an injury.

How would you proceed?

Rupture of the anterior cruciate ligament (ACL), partial or complete, is a common injury, especially in athletes who hurt their knee in a pivoting movement.¹ The number of patients who present with ACL injury is estimated at 252,000 per year.² Cruciate ligament injury may lead to complaints of instability with subsequent inability to engage in sports activities. Cruciate ligament injury is also associated with premature development of osteoarthritis later in life.³ Operative treatment seems to be superior to conservative treatment in improving both subjective and objective measures of knee instability and in helping athletes return to their former level of activity.⁴

Because early detection is key to achieving the best clinical outcome, it is essential that the most accurate physical examination tests are performed during the acute phase. Primary care physicians, emergency room doctors, physical therapists, and athletic trainers are the ones who most often see these patients immediately following the injury, and they often have only the physical examination with which to assess ACL injury. Their task is to identify the patient with potential ACL injury and to refer the patient swiftly.

Three physical examination tests are most commonly used to evaluate cruciate ligament injury. The best known and most frequently used technique is the anterior drawer test. The other 2 tests, the Lachman test and the pivot shift test, are more difficult to perform and are used less often, especially by physicians untrained in their use. In addition, there is a relatively new diagnostic test: the lever sign test. The aim of our article is to provide a short, clinically relevant overview of the literature and to assess the diagnostic value of physical examination for the primary care physician.

Anterior drawer test

How it’s done. In this test, the patient lies supine on the examination table with hips flexed to 45 degrees and knees flexed to 90 degrees (FIGURE 1).⁵ The examiner sits on the table with a leg resting on the patient's foot, grasps the tibia of the injured leg just below the knee, and draws the tibia forward. If the tibia, compared with the tibia of the uninjured leg, moves farther anteriorly, or if the endpoint feels softened or is absent, the result is positive for an ACL injury.

The literature. Nine systematic reviews conclude that the anterior drawer test is inferior to the Lachman test,^6-14 which we’ll describe in a moment. This is due, in part, to the anterior drawer test’s unacceptably low sensitivity and specificity in the clinical setting—especially during the acute phase.¹⁰ The most recent meta-analysis on the anterior drawer test reports a sensitivity of 38% and a specificity of 81%.⁹ In other words, out of 100 ruptured ligaments, only 38 will test positive with the anterior drawer test.

The literature offers possible explanations for findings on the test’s validity. First, rupture of the ACL is often accompanied by swelling of the knee caused by hemarthrosis and reactive synovitis that can prevent the patient from flexing the knee to 90 degrees. Second, the joint pain may induce a protective muscle action, also called guarding of the hamstrings, that creates a vector opposing the passive anterior translation.¹⁵

Apart from the matter of a test’s validity, it's also important to consider the test’s inter- and intra-rater reliability.¹⁶ Compared with the Lachman test, the anterior drawer test is inferior in reliability.⁷

Lachman test

How it’s done. The Lachman test is performed with the patient supine on the table and the injured knee flexed at 20 to 30 degrees (FIGURE 2).⁵ The examiner holds the patient’s thigh with one hand and places the other hand beneath the tibia with the thumb of that hand on the tibial joint line. As the tibia is pulled forward, firm resistance suggests an uninjured ACL. Free movement without a hard endpoint, compared with the uninjured knee, indicates ACL injury.

The literature. The Lachman test is the most accurate of the 3 diagnostic physical procedures. The most recent meta-analysis reports a sensitivity of 68% for partial ruptures and 96% for complete ACL ruptures.⁶ According to a recently published overview of systematic reviews, the Lachman test has high diagnostic value in confirming or ruling out an ACL injury.¹⁷

Two factors are important when assessing results of the Lachman test. The quantity of anterior translation of the tibia relative to the femur is as important as the quality of the endpoint of the anterior translation. Quantity of translation must always be compared with the unaffected knee. Quality of the endpoint in passive anterior translation should be assessed as “firm” or “sudden,” indicating an intact ACL, or as “absent, ill-defined, or softened,” indicating ACL pathology (TABLE).¹⁸

A drawback of the Lachman test is that it is challenging to perform correctly.¹⁹ The patient’s ability to relax the upper leg musculature is critically important. It is also essential to stabilize the distal femur, which can be problematic if the examiner has small hands relative to the size of the patient's leg musculature.¹⁰ These difficulties might be resolved by conducting the Lachman test with the patient in the prone position, known as the Prone Lachman.¹⁹ However, good evidence is not yet available to support this proposed solution. One systematic review, though, reports that the Prone Lachman test has the highest inter-rater reliability of all commonly used physical examination tests.⁷

The Lachman test is known as the test with highest validity on physical examination. When the outcome of a correctly performed Lachman test is negative, a rupture of the ACL is very unlikely.

Pivot shift test

How it’s done. With the patient lying supine on the table, the examiner uses one hand to hold the patient’s heel or ankle and the other hand to grasp the proximal portion of the lower leg (FIGURE 3).⁵ Lifting the leg to about 30 degrees from the table with the injured knee in full extension, the examiner rotates the foot or ankle medially, applies a valgus force to the knee, and slowly flexes it. During flexion, a ruptured ACL will cause the tibia to translate posteriorly to the femur. (Note that the starting position of the test has the tibia subluxed anteriorly. The posterior translation is the "pivot shift" back into the neutral position.)

The literature. The pivot shift test is technically more challenging to perform than the other 2 tests and is, therefore, less practical in the primary care setting. However, when this test is done correctly, a positive result is highly specific for ACL injury.^9,10 Reported sensitivity values are contradictory. The most recent meta-analysis reports a sensitivity of 85%.⁶ Two other studies cite much lower values: 24% and 28%.^9,10 These data suggest that the pivot shift test, when carried out correctly, can be of use in confirming a possible ACL rupture. However, the test should not be used alone in ruling out a possible ACL injury.

New diagnostic test: Lever sign test

How it’s done. The lever sign test (FIGURE 4),²⁰ introduced in the mid-2010s, is also performed with the patient lying in the supine position. The examiner stands at the side of the affected knee of the patient, places a closed fist just beneath the proximal third of the patient’s tibia, creating a slight flexion of the knee joint. With the other hand, the examiner applies a downward directed force to the distal third of the femur. With an intact ACL, the patient’s foot should rise from the table due to the induced lever mechanism. With a ruptured ACL, the lever effect is absent and the foot will not rise.

The literature. In the prospective clinical study that introduced the lever sign test, the sensitivity rate was reported at 100%—higher than that seen with the other commonly used tests.²⁰ Another study has reported that the lever sign test was easily adopted in clinical practice and showed higher sensitivity than the Lachman test (94% vs 80% in pre-anesthesia assessment).²¹ However, a more recent study has shown a sensitivity of 77% for the lever sign.²² The lever sign test is relatively easy to perform and requires less examiner strength than does the Lachman test. These factors enhance applicability of the lever sign test in the primary care office and in other settings such as physical therapy centers and emergency departments.

Applying this information in primary care

Given the importance of physical examination in diagnosing ACL injury, how can the current evidence best be applied in primary care practice? Based on its good test properties and feasibility, the Lachman test is preferred in primary care. The anterior drawer test can be used, but its low accuracy must be considered in making an assessment. The pivot shift test, given its difficulty of execution, should not be used by physicians unacquainted with it.

If future research supports early reports of the lever sign test’s accuracy, it could be very helpful in family practice. Going forward, research should aim at developing a constructive strategy for applying these physical examination tests in both primary care and specialty settings.

CORRESPONDENCE
Christiaan H. Koster, Department of Trauma Surgery, VU University Medical Centre, P.O. Box 7057, 1081 HV Amsterdam, The Netherlands; [email protected].

ACKNOWLEDGEMENTS
We thank Frits Oosterveld, PhD, for critically reviewing the manuscript and Ralph de Vries for his assistance in the literature search.

Point-of-care ultrasound (POCUS) has been gaining greater traction in recent years as a way to quickly (and cost-effectively) assess for conditions including systolic dysfunction, pleural effusion, abdominal aortic aneurysms (AAAs), and deep vein thrombosis (DVT). It involves limited and specific ultrasound protocols performed at the bedside by the health care provider who is trying to answer a specific question and, thus, help guide treatment of the patient.

POCUS was first widely used by emergency physicians starting in the early 1990s with the widespread adoption of the Focused Assessment with Sonography in Trauma (FAST) scan.^1,2 Since that time, POCUS has expanded beyond trauma applications and into family medicine.

One study assessed physicians’ perceptions of POCUS after its integration into a military family medicine clinic. The study showed that physicians perceived POCUS to be relatively easy to use, not overly time consuming, and of high value to the practice.³ In fact, the literature tells us that POCUS can help decrease the cost of health care and improve outcomes,^4-7 while requiring a relatively brief training period.

If residencies are any indication, POCUS may be headed your way

Ultrasound units are becoming smaller and more affordable, and medical schools are increasingly incorporating ultrasound curricula into medical student training.⁸ As of 2016, only 6% of practicing FPs reported using non-obstetric POCUS in their practices.⁹ Similarly, a survey from 2015 reported that only 2% of family medicine residency programs had established POCUS curricula.¹⁰ However, 50% of respondents in the 2015 survey reported early-stage development or interest in developing a POCUS curriculum.

Since then a validated family medicine residency curriculum has been published,¹¹ and the American Academy of Family Physicians (AAFP) recently released a POCUS Curriculum Guideline for residencies (https://www.aafp.org/dam/AAFP/documents/medical_education_residency/program_directors/Reprint290D_POCUS.pdf).

[polldaddy:9928416]

The potential applications of POCUS in family medicine are numerous and have been reviewed in several recent publications.^12,13 In this article, we will review the evidence for the use of POCUS in 4 areas: the cardiovascular exam (FIGURES 1 and 2), the lung exam (FIGURES 3-6), the screening exam for AAAs (FIGURE 7), and the evaluation for DVT (FIGURES 8 and 9). (Obstetric and musculoskeletal applications have been sufficiently covered elsewhere.^14-17) For all of these applications, POCUS is safe, accurate, and beneficial and can be performed with a relatively small amount of training by non-radiology specialists, including FPs (TABLEs 1 and 2).

Just 2 hours of cardio POCUS training enhanced Dx accuracy

The American Society of Echocardiography (ASE) issued an expert consensus statement for focused cardiac ultrasound in 2013.¹⁸ The guideline supports non-cardiologists utilizing POCUS to assess for pericardial effusion and right and left ventricular enlargement, as well as to review global cardiac systolic function and intravascular volume status. Cardiovascular POCUS protocols are relatively easy to learn; even small amounts of training and practice can yield competency.

Point-of-care ultrasound is safe, accurate, and beneficial and can be performed with a relatively small amount of training by family physicians.

For example, a 2013 study showed that after 2 hours of training with a pocket ultrasound device, medical students and junior physicians inexperienced with POCUS were able to improve their diagnostic accuracy for heart failure from 50% to 75%.¹⁹ In another study, internal medicine residents with limited cardiac ultrasound training (ie, 20 practice exams) were able to detect decreased left ventricular ejection fraction using a handheld ultrasound device with 94% sensitivity and specificity in patients admitted to the hospital with acute decompensated heart failure.²⁰ Similarly, after only 8 hours of training, a group of Norwegian general practitioners were able to obtain measurements of systolic function with a pocket ultrasound device that were not statistically different from a cardiologist’s measurements.²¹

In another study, rural FPs attended a 4-day course and then performed focused cardiac ultrasounds on primary care patients with a clinical indication for an echocardiogram.²² The scans were uploaded to a Web-based program for remote interpretation by a cardiologist. There was high concordance between the FPs’ interpretations of the focused cardiac ultrasounds and the cardiologist’s interpretations. Only 32% of the patients in the study group required a formal follow-up echocardiogram.

Kimura et al published a POCUS protocol for the rapid assessment of patients with heart failure, called the Cardiopulmonary Limited Ultrasound Exam (CLUE).²³ The CLUE protocol utilizes 4 views to assess left ventricular systolic and diastolic function along with signs of pulmonary edema or systemic volume overload (TABLE 3²³). The presence of pulmonary edema or a plethoric inferior vena cava (IVC) was highly prognostic of in-hospital mortality. The CLUE protocol has been successfully used by novices including internal medicine residents after brief training (ie, up to 60 supervised scans) and can be performed in less than 5 minutes.^24,25

More sensitive, specific than x-rays for pulmonary diagnoses

The chest x-ray has traditionally been the imaging modality of choice to evaluate primary care pulmonary complaints. However, POCUS can be more sensitive and specific than a chest x-ray for evaluating several pulmonary diagnoses including pleural effusion, pneumonia, and pulmonary edema.

Pleural effusion can be difficult to detect with a physical exam alone. A systematic review showed that the physical exam is not sensitive for effusions <300 mL and can have even lower utility in obese patients.²⁸ While an upright lateral chest x-ray can accurately detect effusions as small as 50 mL, portable x-rays have sensitivities of only 53% to 71% for small- or moderate-sized effusions.^29,30 Ultrasound, however, has a sensitivity of 97% for small effusions.³¹

A 2016 meta-analysis showed that POCUS had a pooled sensitivity and specificity of 94% and 98%, respectively, for pleural effusions, while chest x-ray had a pooled sensitivity and specificity of 51% and 91%, respectively, when compared with computed tomography (CT) and expert sonography.³² POCUS evaluation for pleural effusion is technically simple, and at least one study showed that even novice users can achieve high diagnostic accuracy after only 3 hours of training.³³

Pneumonia is the eighth leading cause of death in the United States and the single leading cause of infectious disease death in children worldwide.^34-36 Pneumonia is a difficult diagnosis to make based on a history and physical examination alone, and the Infectious Diseases Society of America recommends diagnostic imaging to make the diagnosis.³⁷

The adult and pediatric literature clearly demonstrate that lung ultrasound is accurate at diagnosing pneumonia. In a 2015 meta-analysis of the pediatric literature, lung ultrasound had a sensitivity of 96% and a specificity of 93% and positive and negative likelihood ratios of 15.3 and 0.06, respectively.³⁸ In adults, a 2016 meta-analysis of lung ultrasound showed a pooled sensitivity and specificity of 90% and 88%, respectively, with positive and negative likelihood ratios of 6.6 and 0.08, respectively.³⁹

In 2015, a prospective study compared the accuracy of lung ultrasound and chest x-ray using CT as the gold standard.⁴⁰ Lung ultrasound had a significantly better sensitivity of 82% compared to a sensitivity of 64% for chest x-ray. Specificities were comparable at 94% for ultrasound and 90% for chest x-ray.⁴⁰

At least one study found novice sonographers to be accurate with lung POCUS for the diagnosis of pneumonia after only two 90-minute training sessions.⁴¹ Moreover, ultrasound has a more favorable safety profile, greater portability, and lower cost compared with chest x-ray and CT.

Pulmonary edema. Lung ultrasound can identify interstitial pulmonary edema via artifacts called B lines, which are produced by the reverberation of sound waves from the pleura due to the widening of the fluid-filled interlobular septa. These are distinctly different from the A-line pattern of repeating horizontal lines that is seen with normal lungs, making lung ultrasound more accurate than chest x-ray for identification of pulmonary edema.^42,43 When final diagnosis via blinded chart review is used as the reference standard, bilateral B lines on a lung ultrasound image have a sensitivity of 86% to 100% and a specificity of 92% to 98% for the diagnosis of pulmonary edema compared to chest x-ray’s sensitivity of 56.9% and specificity of 89.2%.⁴⁴ There is also a linear correlation between the number of B lines present and the extent of pulmonary edema.^42,45,46 The number of B lines decreases in real time as volume is removed in dialysis patients.⁴⁷

POCUS evaluation for B lines can be learned very quickly. Exams of novices who have performed only 5 prior exams correlate highly with those of experts who have performed more than 100 exams.⁴⁸

Simple, efficient screening method for abdominal aortic aneurysm

AAAs are present in up to 7% of men over the age of 50.⁴⁹ The mortality rate of a ruptured AAA is as high as 80% to 95%.⁵⁰ There is, however, a long prodromal period when interventions can make a significant difference, which is why accurate screening is so important.

AAA screening with ultrasound has been shown to decrease mortality.⁵¹ The current recommendation of the US Preventive Services Task Force (USPSTF) is a one-time AAA screening for all men ages 65 to 75 years who have ever smoked (Grade B).⁵² Despite the recommendations of the USPSTF, screening rates are low. One study found that only 9% of eligible patients in primary care practices received appropriate screening.⁵¹

Ultrasound performed by specialists is known to be an excellent screening test for AAA with a sensitivity of 98.9% and a specificity of 99.9%.⁵³ POCUS use by emergency medicine physicians for the evaluation of symptomatic AAA is well established in the literature. A meta-analysis including 7 studies and 655 patients showed a pooled sensitivity of 99% and a specificity of 98%.⁵⁴ Multiple studies also support primary care physicians performing POCUS AAA screening in the clinic setting.

For example, a 2012 prospective, observational study performed in Canada compared office-based ultrasound screening exams performed by a rural FP to scans performed in the hospital on the same patients.⁵⁵ The physician completed 50 training examinations. The average discrepancy in aorta diameters between the 2 was only 2 mm, which is clinically insignificant, and the office-based scans had a sensitivity and specificity of 100%.

Similarly, a second FP study performed in Barcelona, showed that an FP who performed POCUS AAA screening had 100% concordance with a radiologist.⁵⁶ Additionally, POCUS screening for AAA was not time consuming; it was performed in under 4 minutes per patient.^55,57

Ruling out DVT

DVT is a relatively rare occurrence in the ambulatory setting. However, patients who present with a painful, swollen lower extremity are much more common, and DVT must be considered and ruled out in these situations.

Although isolated distal DVTs that occur in the calf veins are usually self-limited and have a very low risk of embolization, they can progress to proximal DVTs of the thigh veins up to 20% of time.^58,59 Similarly, thrombophlebitis of the superficial lower extremity veins rarely embolizes, but can progress to a proximal DVT, especially if large segments are involved or if the segments are within 5 cm of the junction to the deep venous system.⁵⁹ The risk of missing a proximal leg DVT is high because embolization occurs up to 60% of the time if the DVT is left untreated.⁶⁰

The current standard for diagnosis of DVT is the lower extremity Doppler ultrasound examination, but obtaining same-day Doppler evaluations can be difficult in the ambulatory setting. In these instances, the American College of Chest Physicians (ACCP) recommends that even low-risk patients receive anticoagulation pending the evaluation if it cannot be obtained in the first 24 hours.⁵⁹ This approach not only increases the cost of care, but also exposes patients—many of whom will not be diagnosed with thrombosis in the end—to the risks of anticoagulation.

D-dimer blood tests have drawbacks, too. While a negative high-sensitivity D-dimer blood test in a patient with a low pre-test probability of DVT can effectively rule out a DVT, laboratory testing is not always immediately available in the ambulatory setting either.⁶¹ Additionally, false-positive rates are high, and positive D-dimer exams still require evaluation by Doppler ultrasound.

Given these limitations, performing an ultrasound at the bedside or in the exam room can allow for more timely and cost-effective care. In fact, research shows that a limited ultrasound, called the 2-region compression exam, which follows along the course of the common femoral vein and popliteal vein only, ignoring the femoral and calf veins, is highly accurate in assessing for proximal leg DVTs. As such, it has been adopted for POCUS use by emergency medicine physicians.⁶²

Multiple studies show that physicians with minimal training can perform the 2-region compression exam with a high degree of accuracy when full-leg Doppler ultrasound was used as the gold standard.^63,64 In these studies, hands-on training times ranged from only 10 minutes to 5 hours, and the exam could be performed in less than 4 minutes. A systematic review of 6 studies comparing emergency physician-performed ultrasound with radiology-performed ultrasound calculated an overall sensitivity of 0.95 (95% CI, 0.87-0.99) and specificity of 0.96 (95% CI, 0.87-0.99) for those performed by emergency physicians.⁶⁵

The main concern with the 2-region compression exam is that it can miss a distal leg DVT. As stated earlier, distal DVTs are relatively benign and tend to resolve without treatment; however, up to 20% can progress to become a dangerous proximal leg DVT.⁵⁸ Researchers have validated several methods by prospective trials to address this limitation.

Point-of-care ultrasound screening for abdominal aortic aneurysm can be performed in less than 4 minutes.

Specifically, researchers have demonstrated that patients with a low pre-test probability of DVT per the Wells scoring system could have DVT effectively ruled out with a single 2-region compression ultrasound without further evaluation.⁶⁶ In another study, researchers evaluated all patients (regardless of pretest probability) with a 2-point compression exam and found that those with negative exams could be followed with a second exam in 7 to 10 days without initiating anticoagulation. If the second one was negative, no further evaluation was needed.^67,68

And finally, researchers demonstrated that a negative 2-point compression ultrasound in combination with a concurrent negative D-dimer test was effective at ruling out DVT, regardless of pre-test probability.^69,70

A preferred approach

Given this data and the fact that in the ambulatory setting it is often easier and faster to perform a 2-region compression examination than to obtain a D-dimer laboratory test or a formal full-leg Doppler ultrasound, what follows is our preferred approach to a patient with suspected DVT in the outpatient setting (FIGURE 10).

We first assess pre-test probability using the Wells scoring system. We then perform the 2-region compression ultrasound. If the patient has low pre-test risk according to the Wells score, we rule out DVT. If the patient has moderate or high risk with a negative 2-region compression ultrasound, the patient gets a D-dimer test. If the D-dimer test is negative, we rule out DVT. If the D-dimer test is positive, we schedule the patient for a repeat 2-region compression ultrasound in 7 to 10 days. If at any time the 2-region compression evaluation is positive, we treat the patient for DVT.

CORRESPONDENCE
Paul Bornemann, MD, Palmetto Health Family Medicine Residency, Department of Family and Preventive Medicine, University of South Carolina School of Medicine, 3209 Colonial Drive, Columbia, SC 29203; [email protected].

Point-of-care ultrasound (POCUS) has been gaining greater traction in recent years as a way to quickly (and cost-effectively) assess for conditions including systolic dysfunction, pleural effusion, abdominal aortic aneurysms (AAAs), and deep vein thrombosis (DVT). It involves limited and specific ultrasound protocols performed at the bedside by the health care provider who is trying to answer a specific question and, thus, help guide treatment of the patient.

POCUS was first widely used by emergency physicians starting in the early 1990s with the widespread adoption of the Focused Assessment with Sonography in Trauma (FAST) scan.^1,2 Since that time, POCUS has expanded beyond trauma applications and into family medicine.

One study assessed physicians’ perceptions of POCUS after its integration into a military family medicine clinic. The study showed that physicians perceived POCUS to be relatively easy to use, not overly time consuming, and of high value to the practice.³ In fact, the literature tells us that POCUS can help decrease the cost of health care and improve outcomes,^4-7 while requiring a relatively brief training period.

If residencies are any indication, POCUS may be headed your way

Ultrasound units are becoming smaller and more affordable, and medical schools are increasingly incorporating ultrasound curricula into medical student training.⁸ As of 2016, only 6% of practicing FPs reported using non-obstetric POCUS in their practices.⁹ Similarly, a survey from 2015 reported that only 2% of family medicine residency programs had established POCUS curricula.¹⁰ However, 50% of respondents in the 2015 survey reported early-stage development or interest in developing a POCUS curriculum.

Since then a validated family medicine residency curriculum has been published,¹¹ and the American Academy of Family Physicians (AAFP) recently released a POCUS Curriculum Guideline for residencies (https://www.aafp.org/dam/AAFP/documents/medical_education_residency/program_directors/Reprint290D_POCUS.pdf).

[polldaddy:9928416]

The potential applications of POCUS in family medicine are numerous and have been reviewed in several recent publications.^12,13 In this article, we will review the evidence for the use of POCUS in 4 areas: the cardiovascular exam (FIGURES 1 and 2), the lung exam (FIGURES 3-6), the screening exam for AAAs (FIGURE 7), and the evaluation for DVT (FIGURES 8 and 9). (Obstetric and musculoskeletal applications have been sufficiently covered elsewhere.^14-17) For all of these applications, POCUS is safe, accurate, and beneficial and can be performed with a relatively small amount of training by non-radiology specialists, including FPs (TABLEs 1 and 2).

Just 2 hours of cardio POCUS training enhanced Dx accuracy

The American Society of Echocardiography (ASE) issued an expert consensus statement for focused cardiac ultrasound in 2013.¹⁸ The guideline supports non-cardiologists utilizing POCUS to assess for pericardial effusion and right and left ventricular enlargement, as well as to review global cardiac systolic function and intravascular volume status. Cardiovascular POCUS protocols are relatively easy to learn; even small amounts of training and practice can yield competency.

Point-of-care ultrasound is safe, accurate, and beneficial and can be performed with a relatively small amount of training by family physicians.

For example, a 2013 study showed that after 2 hours of training with a pocket ultrasound device, medical students and junior physicians inexperienced with POCUS were able to improve their diagnostic accuracy for heart failure from 50% to 75%.¹⁹ In another study, internal medicine residents with limited cardiac ultrasound training (ie, 20 practice exams) were able to detect decreased left ventricular ejection fraction using a handheld ultrasound device with 94% sensitivity and specificity in patients admitted to the hospital with acute decompensated heart failure.²⁰ Similarly, after only 8 hours of training, a group of Norwegian general practitioners were able to obtain measurements of systolic function with a pocket ultrasound device that were not statistically different from a cardiologist’s measurements.²¹

In another study, rural FPs attended a 4-day course and then performed focused cardiac ultrasounds on primary care patients with a clinical indication for an echocardiogram.²² The scans were uploaded to a Web-based program for remote interpretation by a cardiologist. There was high concordance between the FPs’ interpretations of the focused cardiac ultrasounds and the cardiologist’s interpretations. Only 32% of the patients in the study group required a formal follow-up echocardiogram.

Kimura et al published a POCUS protocol for the rapid assessment of patients with heart failure, called the Cardiopulmonary Limited Ultrasound Exam (CLUE).²³ The CLUE protocol utilizes 4 views to assess left ventricular systolic and diastolic function along with signs of pulmonary edema or systemic volume overload (TABLE 3²³). The presence of pulmonary edema or a plethoric inferior vena cava (IVC) was highly prognostic of in-hospital mortality. The CLUE protocol has been successfully used by novices including internal medicine residents after brief training (ie, up to 60 supervised scans) and can be performed in less than 5 minutes.^24,25

More sensitive, specific than x-rays for pulmonary diagnoses

The chest x-ray has traditionally been the imaging modality of choice to evaluate primary care pulmonary complaints. However, POCUS can be more sensitive and specific than a chest x-ray for evaluating several pulmonary diagnoses including pleural effusion, pneumonia, and pulmonary edema.

Pleural effusion can be difficult to detect with a physical exam alone. A systematic review showed that the physical exam is not sensitive for effusions <300 mL and can have even lower utility in obese patients.²⁸ While an upright lateral chest x-ray can accurately detect effusions as small as 50 mL, portable x-rays have sensitivities of only 53% to 71% for small- or moderate-sized effusions.^29,30 Ultrasound, however, has a sensitivity of 97% for small effusions.³¹

A 2016 meta-analysis showed that POCUS had a pooled sensitivity and specificity of 94% and 98%, respectively, for pleural effusions, while chest x-ray had a pooled sensitivity and specificity of 51% and 91%, respectively, when compared with computed tomography (CT) and expert sonography.³² POCUS evaluation for pleural effusion is technically simple, and at least one study showed that even novice users can achieve high diagnostic accuracy after only 3 hours of training.³³

Pneumonia is the eighth leading cause of death in the United States and the single leading cause of infectious disease death in children worldwide.^34-36 Pneumonia is a difficult diagnosis to make based on a history and physical examination alone, and the Infectious Diseases Society of America recommends diagnostic imaging to make the diagnosis.³⁷

The adult and pediatric literature clearly demonstrate that lung ultrasound is accurate at diagnosing pneumonia. In a 2015 meta-analysis of the pediatric literature, lung ultrasound had a sensitivity of 96% and a specificity of 93% and positive and negative likelihood ratios of 15.3 and 0.06, respectively.³⁸ In adults, a 2016 meta-analysis of lung ultrasound showed a pooled sensitivity and specificity of 90% and 88%, respectively, with positive and negative likelihood ratios of 6.6 and 0.08, respectively.³⁹

In 2015, a prospective study compared the accuracy of lung ultrasound and chest x-ray using CT as the gold standard.⁴⁰ Lung ultrasound had a significantly better sensitivity of 82% compared to a sensitivity of 64% for chest x-ray. Specificities were comparable at 94% for ultrasound and 90% for chest x-ray.⁴⁰

At least one study found novice sonographers to be accurate with lung POCUS for the diagnosis of pneumonia after only two 90-minute training sessions.⁴¹ Moreover, ultrasound has a more favorable safety profile, greater portability, and lower cost compared with chest x-ray and CT.

Pulmonary edema. Lung ultrasound can identify interstitial pulmonary edema via artifacts called B lines, which are produced by the reverberation of sound waves from the pleura due to the widening of the fluid-filled interlobular septa. These are distinctly different from the A-line pattern of repeating horizontal lines that is seen with normal lungs, making lung ultrasound more accurate than chest x-ray for identification of pulmonary edema.^42,43 When final diagnosis via blinded chart review is used as the reference standard, bilateral B lines on a lung ultrasound image have a sensitivity of 86% to 100% and a specificity of 92% to 98% for the diagnosis of pulmonary edema compared to chest x-ray’s sensitivity of 56.9% and specificity of 89.2%.⁴⁴ There is also a linear correlation between the number of B lines present and the extent of pulmonary edema.^42,45,46 The number of B lines decreases in real time as volume is removed in dialysis patients.⁴⁷

POCUS evaluation for B lines can be learned very quickly. Exams of novices who have performed only 5 prior exams correlate highly with those of experts who have performed more than 100 exams.⁴⁸

Simple, efficient screening method for abdominal aortic aneurysm

AAAs are present in up to 7% of men over the age of 50.⁴⁹ The mortality rate of a ruptured AAA is as high as 80% to 95%.⁵⁰ There is, however, a long prodromal period when interventions can make a significant difference, which is why accurate screening is so important.

AAA screening with ultrasound has been shown to decrease mortality.⁵¹ The current recommendation of the US Preventive Services Task Force (USPSTF) is a one-time AAA screening for all men ages 65 to 75 years who have ever smoked (Grade B).⁵² Despite the recommendations of the USPSTF, screening rates are low. One study found that only 9% of eligible patients in primary care practices received appropriate screening.⁵¹

Ultrasound performed by specialists is known to be an excellent screening test for AAA with a sensitivity of 98.9% and a specificity of 99.9%.⁵³ POCUS use by emergency medicine physicians for the evaluation of symptomatic AAA is well established in the literature. A meta-analysis including 7 studies and 655 patients showed a pooled sensitivity of 99% and a specificity of 98%.⁵⁴ Multiple studies also support primary care physicians performing POCUS AAA screening in the clinic setting.

For example, a 2012 prospective, observational study performed in Canada compared office-based ultrasound screening exams performed by a rural FP to scans performed in the hospital on the same patients.⁵⁵ The physician completed 50 training examinations. The average discrepancy in aorta diameters between the 2 was only 2 mm, which is clinically insignificant, and the office-based scans had a sensitivity and specificity of 100%.

Similarly, a second FP study performed in Barcelona, showed that an FP who performed POCUS AAA screening had 100% concordance with a radiologist.⁵⁶ Additionally, POCUS screening for AAA was not time consuming; it was performed in under 4 minutes per patient.^55,57

Ruling out DVT

DVT is a relatively rare occurrence in the ambulatory setting. However, patients who present with a painful, swollen lower extremity are much more common, and DVT must be considered and ruled out in these situations.

Although isolated distal DVTs that occur in the calf veins are usually self-limited and have a very low risk of embolization, they can progress to proximal DVTs of the thigh veins up to 20% of time.^58,59 Similarly, thrombophlebitis of the superficial lower extremity veins rarely embolizes, but can progress to a proximal DVT, especially if large segments are involved or if the segments are within 5 cm of the junction to the deep venous system.⁵⁹ The risk of missing a proximal leg DVT is high because embolization occurs up to 60% of the time if the DVT is left untreated.⁶⁰

The current standard for diagnosis of DVT is the lower extremity Doppler ultrasound examination, but obtaining same-day Doppler evaluations can be difficult in the ambulatory setting. In these instances, the American College of Chest Physicians (ACCP) recommends that even low-risk patients receive anticoagulation pending the evaluation if it cannot be obtained in the first 24 hours.⁵⁹ This approach not only increases the cost of care, but also exposes patients—many of whom will not be diagnosed with thrombosis in the end—to the risks of anticoagulation.

D-dimer blood tests have drawbacks, too. While a negative high-sensitivity D-dimer blood test in a patient with a low pre-test probability of DVT can effectively rule out a DVT, laboratory testing is not always immediately available in the ambulatory setting either.⁶¹ Additionally, false-positive rates are high, and positive D-dimer exams still require evaluation by Doppler ultrasound.

Given these limitations, performing an ultrasound at the bedside or in the exam room can allow for more timely and cost-effective care. In fact, research shows that a limited ultrasound, called the 2-region compression exam, which follows along the course of the common femoral vein and popliteal vein only, ignoring the femoral and calf veins, is highly accurate in assessing for proximal leg DVTs. As such, it has been adopted for POCUS use by emergency medicine physicians.⁶²

Multiple studies show that physicians with minimal training can perform the 2-region compression exam with a high degree of accuracy when full-leg Doppler ultrasound was used as the gold standard.^63,64 In these studies, hands-on training times ranged from only 10 minutes to 5 hours, and the exam could be performed in less than 4 minutes. A systematic review of 6 studies comparing emergency physician-performed ultrasound with radiology-performed ultrasound calculated an overall sensitivity of 0.95 (95% CI, 0.87-0.99) and specificity of 0.96 (95% CI, 0.87-0.99) for those performed by emergency physicians.⁶⁵

The main concern with the 2-region compression exam is that it can miss a distal leg DVT. As stated earlier, distal DVTs are relatively benign and tend to resolve without treatment; however, up to 20% can progress to become a dangerous proximal leg DVT.⁵⁸ Researchers have validated several methods by prospective trials to address this limitation.

Point-of-care ultrasound screening for abdominal aortic aneurysm can be performed in less than 4 minutes.

Specifically, researchers have demonstrated that patients with a low pre-test probability of DVT per the Wells scoring system could have DVT effectively ruled out with a single 2-region compression ultrasound without further evaluation.⁶⁶ In another study, researchers evaluated all patients (regardless of pretest probability) with a 2-point compression exam and found that those with negative exams could be followed with a second exam in 7 to 10 days without initiating anticoagulation. If the second one was negative, no further evaluation was needed.^67,68

And finally, researchers demonstrated that a negative 2-point compression ultrasound in combination with a concurrent negative D-dimer test was effective at ruling out DVT, regardless of pre-test probability.^69,70

A preferred approach

Given this data and the fact that in the ambulatory setting it is often easier and faster to perform a 2-region compression examination than to obtain a D-dimer laboratory test or a formal full-leg Doppler ultrasound, what follows is our preferred approach to a patient with suspected DVT in the outpatient setting (FIGURE 10).

We first assess pre-test probability using the Wells scoring system. We then perform the 2-region compression ultrasound. If the patient has low pre-test risk according to the Wells score, we rule out DVT. If the patient has moderate or high risk with a negative 2-region compression ultrasound, the patient gets a D-dimer test. If the D-dimer test is negative, we rule out DVT. If the D-dimer test is positive, we schedule the patient for a repeat 2-region compression ultrasound in 7 to 10 days. If at any time the 2-region compression evaluation is positive, we treat the patient for DVT.

CORRESPONDENCE
Paul Bornemann, MD, Palmetto Health Family Medicine Residency, Department of Family and Preventive Medicine, University of South Carolina School of Medicine, 3209 Colonial Drive, Columbia, SC 29203; [email protected].

Point-of-care ultrasound (POCUS) has been gaining greater traction in recent years as a way to quickly (and cost-effectively) assess for conditions including systolic dysfunction, pleural effusion, abdominal aortic aneurysms (AAAs), and deep vein thrombosis (DVT). It involves limited and specific ultrasound protocols performed at the bedside by the health care provider who is trying to answer a specific question and, thus, help guide treatment of the patient.

POCUS was first widely used by emergency physicians starting in the early 1990s with the widespread adoption of the Focused Assessment with Sonography in Trauma (FAST) scan.^1,2 Since that time, POCUS has expanded beyond trauma applications and into family medicine.

One study assessed physicians’ perceptions of POCUS after its integration into a military family medicine clinic. The study showed that physicians perceived POCUS to be relatively easy to use, not overly time consuming, and of high value to the practice.³ In fact, the literature tells us that POCUS can help decrease the cost of health care and improve outcomes,^4-7 while requiring a relatively brief training period.

If residencies are any indication, POCUS may be headed your way

Ultrasound units are becoming smaller and more affordable, and medical schools are increasingly incorporating ultrasound curricula into medical student training.⁸ As of 2016, only 6% of practicing FPs reported using non-obstetric POCUS in their practices.⁹ Similarly, a survey from 2015 reported that only 2% of family medicine residency programs had established POCUS curricula.¹⁰ However, 50% of respondents in the 2015 survey reported early-stage development or interest in developing a POCUS curriculum.

Since then a validated family medicine residency curriculum has been published,¹¹ and the American Academy of Family Physicians (AAFP) recently released a POCUS Curriculum Guideline for residencies (https://www.aafp.org/dam/AAFP/documents/medical_education_residency/program_directors/Reprint290D_POCUS.pdf).

[polldaddy:9928416]

The potential applications of POCUS in family medicine are numerous and have been reviewed in several recent publications.^12,13 In this article, we will review the evidence for the use of POCUS in 4 areas: the cardiovascular exam (FIGURES 1 and 2), the lung exam (FIGURES 3-6), the screening exam for AAAs (FIGURE 7), and the evaluation for DVT (FIGURES 8 and 9). (Obstetric and musculoskeletal applications have been sufficiently covered elsewhere.^14-17) For all of these applications, POCUS is safe, accurate, and beneficial and can be performed with a relatively small amount of training by non-radiology specialists, including FPs (TABLEs 1 and 2).

Just 2 hours of cardio POCUS training enhanced Dx accuracy

The American Society of Echocardiography (ASE) issued an expert consensus statement for focused cardiac ultrasound in 2013.¹⁸ The guideline supports non-cardiologists utilizing POCUS to assess for pericardial effusion and right and left ventricular enlargement, as well as to review global cardiac systolic function and intravascular volume status. Cardiovascular POCUS protocols are relatively easy to learn; even small amounts of training and practice can yield competency.

Point-of-care ultrasound is safe, accurate, and beneficial and can be performed with a relatively small amount of training by family physicians.

For example, a 2013 study showed that after 2 hours of training with a pocket ultrasound device, medical students and junior physicians inexperienced with POCUS were able to improve their diagnostic accuracy for heart failure from 50% to 75%.¹⁹ In another study, internal medicine residents with limited cardiac ultrasound training (ie, 20 practice exams) were able to detect decreased left ventricular ejection fraction using a handheld ultrasound device with 94% sensitivity and specificity in patients admitted to the hospital with acute decompensated heart failure.²⁰ Similarly, after only 8 hours of training, a group of Norwegian general practitioners were able to obtain measurements of systolic function with a pocket ultrasound device that were not statistically different from a cardiologist’s measurements.²¹

In another study, rural FPs attended a 4-day course and then performed focused cardiac ultrasounds on primary care patients with a clinical indication for an echocardiogram.²² The scans were uploaded to a Web-based program for remote interpretation by a cardiologist. There was high concordance between the FPs’ interpretations of the focused cardiac ultrasounds and the cardiologist’s interpretations. Only 32% of the patients in the study group required a formal follow-up echocardiogram.

Kimura et al published a POCUS protocol for the rapid assessment of patients with heart failure, called the Cardiopulmonary Limited Ultrasound Exam (CLUE).²³ The CLUE protocol utilizes 4 views to assess left ventricular systolic and diastolic function along with signs of pulmonary edema or systemic volume overload (TABLE 3²³). The presence of pulmonary edema or a plethoric inferior vena cava (IVC) was highly prognostic of in-hospital mortality. The CLUE protocol has been successfully used by novices including internal medicine residents after brief training (ie, up to 60 supervised scans) and can be performed in less than 5 minutes.^24,25

More sensitive, specific than x-rays for pulmonary diagnoses

The chest x-ray has traditionally been the imaging modality of choice to evaluate primary care pulmonary complaints. However, POCUS can be more sensitive and specific than a chest x-ray for evaluating several pulmonary diagnoses including pleural effusion, pneumonia, and pulmonary edema.

Pleural effusion can be difficult to detect with a physical exam alone. A systematic review showed that the physical exam is not sensitive for effusions <300 mL and can have even lower utility in obese patients.²⁸ While an upright lateral chest x-ray can accurately detect effusions as small as 50 mL, portable x-rays have sensitivities of only 53% to 71% for small- or moderate-sized effusions.^29,30 Ultrasound, however, has a sensitivity of 97% for small effusions.³¹

A 2016 meta-analysis showed that POCUS had a pooled sensitivity and specificity of 94% and 98%, respectively, for pleural effusions, while chest x-ray had a pooled sensitivity and specificity of 51% and 91%, respectively, when compared with computed tomography (CT) and expert sonography.³² POCUS evaluation for pleural effusion is technically simple, and at least one study showed that even novice users can achieve high diagnostic accuracy after only 3 hours of training.³³

Pneumonia is the eighth leading cause of death in the United States and the single leading cause of infectious disease death in children worldwide.^34-36 Pneumonia is a difficult diagnosis to make based on a history and physical examination alone, and the Infectious Diseases Society of America recommends diagnostic imaging to make the diagnosis.³⁷

The adult and pediatric literature clearly demonstrate that lung ultrasound is accurate at diagnosing pneumonia. In a 2015 meta-analysis of the pediatric literature, lung ultrasound had a sensitivity of 96% and a specificity of 93% and positive and negative likelihood ratios of 15.3 and 0.06, respectively.³⁸ In adults, a 2016 meta-analysis of lung ultrasound showed a pooled sensitivity and specificity of 90% and 88%, respectively, with positive and negative likelihood ratios of 6.6 and 0.08, respectively.³⁹

In 2015, a prospective study compared the accuracy of lung ultrasound and chest x-ray using CT as the gold standard.⁴⁰ Lung ultrasound had a significantly better sensitivity of 82% compared to a sensitivity of 64% for chest x-ray. Specificities were comparable at 94% for ultrasound and 90% for chest x-ray.⁴⁰

At least one study found novice sonographers to be accurate with lung POCUS for the diagnosis of pneumonia after only two 90-minute training sessions.⁴¹ Moreover, ultrasound has a more favorable safety profile, greater portability, and lower cost compared with chest x-ray and CT.

Pulmonary edema. Lung ultrasound can identify interstitial pulmonary edema via artifacts called B lines, which are produced by the reverberation of sound waves from the pleura due to the widening of the fluid-filled interlobular septa. These are distinctly different from the A-line pattern of repeating horizontal lines that is seen with normal lungs, making lung ultrasound more accurate than chest x-ray for identification of pulmonary edema.^42,43 When final diagnosis via blinded chart review is used as the reference standard, bilateral B lines on a lung ultrasound image have a sensitivity of 86% to 100% and a specificity of 92% to 98% for the diagnosis of pulmonary edema compared to chest x-ray’s sensitivity of 56.9% and specificity of 89.2%.⁴⁴ There is also a linear correlation between the number of B lines present and the extent of pulmonary edema.^42,45,46 The number of B lines decreases in real time as volume is removed in dialysis patients.⁴⁷

POCUS evaluation for B lines can be learned very quickly. Exams of novices who have performed only 5 prior exams correlate highly with those of experts who have performed more than 100 exams.⁴⁸

Simple, efficient screening method for abdominal aortic aneurysm

AAAs are present in up to 7% of men over the age of 50.⁴⁹ The mortality rate of a ruptured AAA is as high as 80% to 95%.⁵⁰ There is, however, a long prodromal period when interventions can make a significant difference, which is why accurate screening is so important.

AAA screening with ultrasound has been shown to decrease mortality.⁵¹ The current recommendation of the US Preventive Services Task Force (USPSTF) is a one-time AAA screening for all men ages 65 to 75 years who have ever smoked (Grade B).⁵² Despite the recommendations of the USPSTF, screening rates are low. One study found that only 9% of eligible patients in primary care practices received appropriate screening.⁵¹

Ultrasound performed by specialists is known to be an excellent screening test for AAA with a sensitivity of 98.9% and a specificity of 99.9%.⁵³ POCUS use by emergency medicine physicians for the evaluation of symptomatic AAA is well established in the literature. A meta-analysis including 7 studies and 655 patients showed a pooled sensitivity of 99% and a specificity of 98%.⁵⁴ Multiple studies also support primary care physicians performing POCUS AAA screening in the clinic setting.

For example, a 2012 prospective, observational study performed in Canada compared office-based ultrasound screening exams performed by a rural FP to scans performed in the hospital on the same patients.⁵⁵ The physician completed 50 training examinations. The average discrepancy in aorta diameters between the 2 was only 2 mm, which is clinically insignificant, and the office-based scans had a sensitivity and specificity of 100%.

Similarly, a second FP study performed in Barcelona, showed that an FP who performed POCUS AAA screening had 100% concordance with a radiologist.⁵⁶ Additionally, POCUS screening for AAA was not time consuming; it was performed in under 4 minutes per patient.^55,57

Ruling out DVT

DVT is a relatively rare occurrence in the ambulatory setting. However, patients who present with a painful, swollen lower extremity are much more common, and DVT must be considered and ruled out in these situations.

Although isolated distal DVTs that occur in the calf veins are usually self-limited and have a very low risk of embolization, they can progress to proximal DVTs of the thigh veins up to 20% of time.^58,59 Similarly, thrombophlebitis of the superficial lower extremity veins rarely embolizes, but can progress to a proximal DVT, especially if large segments are involved or if the segments are within 5 cm of the junction to the deep venous system.⁵⁹ The risk of missing a proximal leg DVT is high because embolization occurs up to 60% of the time if the DVT is left untreated.⁶⁰

The current standard for diagnosis of DVT is the lower extremity Doppler ultrasound examination, but obtaining same-day Doppler evaluations can be difficult in the ambulatory setting. In these instances, the American College of Chest Physicians (ACCP) recommends that even low-risk patients receive anticoagulation pending the evaluation if it cannot be obtained in the first 24 hours.⁵⁹ This approach not only increases the cost of care, but also exposes patients—many of whom will not be diagnosed with thrombosis in the end—to the risks of anticoagulation.

D-dimer blood tests have drawbacks, too. While a negative high-sensitivity D-dimer blood test in a patient with a low pre-test probability of DVT can effectively rule out a DVT, laboratory testing is not always immediately available in the ambulatory setting either.⁶¹ Additionally, false-positive rates are high, and positive D-dimer exams still require evaluation by Doppler ultrasound.

Given these limitations, performing an ultrasound at the bedside or in the exam room can allow for more timely and cost-effective care. In fact, research shows that a limited ultrasound, called the 2-region compression exam, which follows along the course of the common femoral vein and popliteal vein only, ignoring the femoral and calf veins, is highly accurate in assessing for proximal leg DVTs. As such, it has been adopted for POCUS use by emergency medicine physicians.⁶²

Multiple studies show that physicians with minimal training can perform the 2-region compression exam with a high degree of accuracy when full-leg Doppler ultrasound was used as the gold standard.^63,64 In these studies, hands-on training times ranged from only 10 minutes to 5 hours, and the exam could be performed in less than 4 minutes. A systematic review of 6 studies comparing emergency physician-performed ultrasound with radiology-performed ultrasound calculated an overall sensitivity of 0.95 (95% CI, 0.87-0.99) and specificity of 0.96 (95% CI, 0.87-0.99) for those performed by emergency physicians.⁶⁵

The main concern with the 2-region compression exam is that it can miss a distal leg DVT. As stated earlier, distal DVTs are relatively benign and tend to resolve without treatment; however, up to 20% can progress to become a dangerous proximal leg DVT.⁵⁸ Researchers have validated several methods by prospective trials to address this limitation.

Point-of-care ultrasound screening for abdominal aortic aneurysm can be performed in less than 4 minutes.

Specifically, researchers have demonstrated that patients with a low pre-test probability of DVT per the Wells scoring system could have DVT effectively ruled out with a single 2-region compression ultrasound without further evaluation.⁶⁶ In another study, researchers evaluated all patients (regardless of pretest probability) with a 2-point compression exam and found that those with negative exams could be followed with a second exam in 7 to 10 days without initiating anticoagulation. If the second one was negative, no further evaluation was needed.^67,68

And finally, researchers demonstrated that a negative 2-point compression ultrasound in combination with a concurrent negative D-dimer test was effective at ruling out DVT, regardless of pre-test probability.^69,70

A preferred approach

Given this data and the fact that in the ambulatory setting it is often easier and faster to perform a 2-region compression examination than to obtain a D-dimer laboratory test or a formal full-leg Doppler ultrasound, what follows is our preferred approach to a patient with suspected DVT in the outpatient setting (FIGURE 10).

We first assess pre-test probability using the Wells scoring system. We then perform the 2-region compression ultrasound. If the patient has low pre-test risk according to the Wells score, we rule out DVT. If the patient has moderate or high risk with a negative 2-region compression ultrasound, the patient gets a D-dimer test. If the D-dimer test is negative, we rule out DVT. If the D-dimer test is positive, we schedule the patient for a repeat 2-region compression ultrasound in 7 to 10 days. If at any time the 2-region compression evaluation is positive, we treat the patient for DVT.

CORRESPONDENCE
Paul Bornemann, MD, Palmetto Health Family Medicine Residency, Department of Family and Preventive Medicine, University of South Carolina School of Medicine, 3209 Colonial Drive, Columbia, SC 29203; [email protected].

CASE 

A 68-year-old woman is admitted to the hospital from home with acute onset, unrelenting, upper abdominal pain radiating to the back and nausea/vomiting. Her medical history includes bile duct obstruction secondary to gall stones, which was managed in another facility 6 days earlier with endoscopic retrograde cholangiopancreatography and stenting. The patient has type 2 diabetes (managed with metformin and glargine insulin), hypertension (managed with lisinopril and hydrochlorothiazide), and cholesterolemia (managed with atorvastatin).

On admission, the patient's white blood cell count is 14.7 x 10³ cells/mm³, heart rate is 100 bpm, blood pressure is 90/68 mm Hg, and temperature is 101.5° F. Serum amylase and lipase are 3 and 2 times the upper limit of normal, respectively. A working diagnosis of acute pancreatitis with sepsis is made. Blood cultures are drawn. A computed tomography scan confirms acute pancreatitis. She receives one dose of meropenem, is started on intravenous fluids and morphine, and is transferred to the intensive care unit (ICU) for further management.

Her ICU course is complicated by worsening sepsis despite aggressive fluid resuscitation, nutrition, and broad-spectrum antibiotics. On post-admission Day 2, blood culture results reveal Escherichia coli that is resistant to gentamicin, amoxicillin/clavulanate, ceftriaxone, piperacillin/tazobactam, imipenem, trimethoprim/sulfamethoxazole, ciprofloxacin, and tetracycline. Additional susceptibility testing is ordered.

The Centers for Disease Control and Prevention (CDC) conservatively estimates that antibiotic-resistant bacteria are responsible for 2 billion infections annually, resulting in approximately 23,000 deaths and $20 billion in excess health care expenditures annually.¹ Infections caused by antibiotic-resistant bacteria typically require longer hospitalizations, more expensive drug therapies, and additional follow-up visits.¹ They also result in greater morbidity and mortality compared with similar infections involving non-resistant bacteria.¹ To compound the problem, antibiotic development has steadily declined over the last 3 decades, with few novel antimicrobials developed in recent years.² The most recently approved antibiotics with new mechanisms of action were linezolid in 2000 and daptomycin in 2003, preceded by the carbapenems 15 years earlier. (See “New antimicrobials in the pipeline.”)

Greater efforts aimed at using antimicrobials sparingly and appropriately, as well as developing new antimicrobials with activity against multidrug-resistant pathogens, are ultimately needed to address the threat of antimicrobial resistance. This article describes the evidence-based management of inpatient infections caused by resistant bacteria and the role family physicians (FPs) can play in reducing further development of resistance through antimicrobial stewardship practices.

S. aureus is a common culprit of hospital-acquired infections, including central line-associated bloodstream infections, catheter-associated urinary tract infections, ventilator-associated pneumonia, and nosocomial skin and soft tissue infections. In fact, nearly half of all isolates from these infections are reported to be methicillin-resistant S. aureus (MRSA).³

Nearly half of all Staphylococcus aureus isolates from hospital-acquired infections are reported to be methicillin-resistant.

Patients at greatest risk for MRSA infections include those who have been recently hospitalized, those receiving recent antibiotic therapy or surgery, long-term care residents, intravenous drug abusers, immunocompromised patients, hemodialysis patients, military personnel, and athletes who play contact sports.^4,5 Patients with these infections often require the use of an anti-MRSA agent (eg, vancomycin, linezolid) in empiric antibiotic regimens.^6,7 The focus of this discussion is on MRSA in hospital and long-term care settings; a discussion of community-acquired MRSA is addressed elsewhere. (See “Antibiotic stewardship: The FP’s role,” J Fam Pract. 2016;65:876-885.⁸)

Efforts are working, but problems remain. MRSA accounts for almost 60% of S. aureus isolates in ICUs.⁹ Thankfully, rates of health care-associated MRSA are now either static or declining nationwide, as a result of major initiatives targeted toward preventing health care-associated infection in recent years.¹⁰

Methicillin resistance in S. aureus results from expression of PBP2a, an altered penicillin-binding protein with reduced binding affinity for beta-lactam antibiotics. As a result, MRSA isolates are resistant to most beta-lactams.⁹ Resistance to macrolides, azithromycin, aminoglycosides, fluoroquinolones, and clindamycin is also common in health care-associated MRSA.⁹

The first case of true vancomycin-resistant S. aureus (VRSA) in the United States was reported in 2002.¹¹ Fortunately, both VRSA and vancomycin-intermediate S. aureus (VISA) have remained rare throughout the United States and abroad.^9,11 Heterogeneous VISA (hVISA), which is characterized by a few resistant subpopulations within a fully susceptible population of S. aureus, is more common than VRSA or VISA. Unfortunately, hVISA is difficult to detect using commercially available susceptibility tests. This can result in treatment failure with vancomycin, even though the MRSA isolate may appear fully susceptible and the patient has received clinically appropriate doses of the drug.¹²

Treatment. Vancomycin is the mainstay of therapy for many systemic health care-associated MRSA infections. Alternative therapies (daptomycin or linezolid) should be considered for isolates with a vancomycin minimum inhibitory concentration (MIC) >2 mcg/mL or in the setting of a poor clinical response.⁴ Combination therapy may be warranted in the setting of treatment failure. Because comparative efficacy data for alternative therapies is lacking, agent selection should be tailored to the site of infection and patient-specific factors such as allergies, drug interactions, and the risk for adverse events (TABLE 1^13-17).

Oritavancin and dalbavancin both have dosing regimens that may allow for earlier hospital discharge or treatment in an outpatient setting.^13,14 Telavancin, quinupristin/dalfopristin, and tigecycline are typically reserved for salvage therapy due to adverse event profiles and/or limited efficacy data.¹⁵

Vancomycin-resistant enterococci (VRE)

Enterococci are typically considered normal gastrointestinal tract flora. However, antibiotic exposure can alter gut flora allowing for VRE colonization, which in some instances, can progress to the development of a health care-associated infection.¹⁵ Therefore, it is important to distinguish whether a patient is colonized or infected with VRE because treatment of colonization is unnecessary and may lead to resistance and other adverse effects.¹⁵

It's important to distinguish whether a patient is colonized or infected with vancomycin-resistant enterococci to avoid unnecessary treatment.

Enterococci may be the culprit in nosocomially-acquired intra-abdominal infections, bacteremia, endocarditis, urinary tract infections (UTIs), and skin and skin structure infections, and can exhibit resistance to ampicillin, aminoglycosides, and vancomycin.¹⁵ VRE is predominantly a health care-associated pathogen and may account for up to 77% of all health care-associated Enterococcus faecium infections and 9% of Enterococcus faecalis infections.¹

Treatment. Antibiotic selection for VRE infections depends upon the site of infection, patient comorbidities, the potential for drug interactions, and treatment duration. Current treatment options include linezolid, daptomycin, quinupristin/dalfopristin (for E. faecium only), tigecycline, and ampicillin if the organism is susceptible (TABLE 1^13-17).¹⁵ For cystitis caused by VRE (not urinary colonization), fosfomycin and nitrofurantoin are additional options.¹⁶

Resistant Enterobacteriaceae

Resistant Enterobacteriaceae such as Escherichia coli and Klebsiella pneumoniae have emerged as a result of increased broad-spectrum antibiotic utilization and have been implicated in health care-associated UTIs, intra-abdominal infections, bacteremia, and even pneumonia.¹ Patients with prolonged hospital stays and invasive medical devices, such as urinary and vascular catheters, endotracheal tubes, and endoscopy scopes, have the highest risk for infection with these organisms.²⁰

The genotypic profiles of resistance among the Enterobacteriaceae are diverse and complex, resulting in different levels of activity for the various beta-lactam agents (TABLE 2^21-24).²⁵ Furthermore, extended-spectrum beta-lactamase (ESBL)-producers and carbapenem-resistant Enterobacteriaceae (CRE) are often resistant to other classes of antibiotics, too, including aminoglycosides and fluoroquinolones.^20,25 The increasing diversity among beta-lactamase enzymes has made the selection of appropriate antibiotic therapy challenging, since the ability to identify specific beta-lactamase genes is not yet available in the clinical setting.

Cefepime may retain activity against some ESBL-producing isolates, but it is also susceptible to the inoculum effect and should only be used for non–life-threatening infections and at higher doses.²³ Fosfomycin has activity against ESBL-producing bacteria, but is only approved for oral use in UTIs in the United States.^20,27 Ceftolozane/tazobactam (Zerbaxa) and ceftazidime/avibactam (Avycaz) were approved in 2014 and 2015, respectively, by the FDA for the management of complicated urinary tract and intra-abdominal infections caused by susceptible ESBL-producing Enterobacteriaceae. In order to preserve the antimicrobial efficacy of these 2 newer agents, however, they are typically reserved for definitive therapy when in vitro susceptibility is demonstrated and there are no other viable options.

AmpC beta-lactamases are resistant to similar agents as the ESBLs, in addition to cefoxitin and the beta-lactam/beta-lactamase inhibitor combinations containing clavulanic acid, sulbactam, and in some cases, tazobactam. Resistance can be induced and emerges in certain pathogens while patients are on therapy.²⁸ Fluoroquinolones and aminoglycosides have a low risk of developing resistance while patients are on therapy, but are more likely to cause adverse effects and toxicity compared with the beta-lactams.²⁸ Carbapenems have the lowest risk of emerging resistance and are the empiric treatment of choice for known AmpC-producing Enterobacteriaceae in serious infections.^20,28 Cefepime may also be an option in less severe infections, such as UTIs or those in which adequate source control has been achieved.^28,29

Carbapenem-resistant Enterobacteriaceae (CRE) have become a serious threat as a result of increased carbapenem use. While carbapenem resistance is less common in the United States than worldwide, rates have increased nearly 4-fold (1.2% to 4.2%) in the last decade, with some regions of the country experiencing substantially higher rates.²⁴ The most commonly reported CRE genotypes identified in the United States include the serine carbapenemase (K. pneumoniae carbapenemase, or KPC), and the metallo-beta-lactamases (Verona integrin-encoded metallo-beta-lactamase, or VIM, and the New Dehli metallo-beta-lactamase, or NDM), with each class conferring slightly different resistance patterns (TABLE 2^21-24).^20,30

Few treatment options exist for Enterobacteriaceae producing a serine carbapenemase, and, unfortunately, evidence to support these therapies is extremely limited. Some CRE isolates retain susceptibility to the polymyxins, the aminoglycosides, and tigecycline.³⁰ Even fewer options exist for treating Enterobacteriaceae producing metallo-beta-lactamases, which are typically only susceptible to the polymyxins and tigecycline.^43-45

The increasing diversity among beta-lactamase enzymes has made the selection of appropriate antibiotics more challenging in recent years.

Several studies have demonstrated lower mortality rates when combination therapy is utilized for CRE bloodstream infections.^31,32 Furthermore, the combination of colistin, tigecycline, and meropenem was found to have a significant mortality advantage.³² Double carbapenem therapy has been effective in several cases of invasive KPC-producing K. pneumoniae infections.^33,34 However, it is important to note that current clinical evidence comes from small, single-center, retrospective studies, and additional research is needed to determine optimal combinations and dosing strategies for these infections.

Lastly, ceftazidime/avibactam (Avycaz) was recently approved for the treatment of complicated urinary tract and intra-abdominal infections, and has activity against KPC-producing Enterobacteriaceae, but not those producing metallo-beta-lactamases, like VIM or NDM. In the absence of strong evidence to support one therapy over another, it may be reasonable to select at least 2 active agents when treating serious CRE infections. Agent selection should be based on the site of the infection, susceptibility data, and patient-specific factors (TABLE 3^{20,22,,23,25,27-42}). The CDC also recommends contact precautions for patients who are colonized or infected with CRE.³⁵

Multi-drug resistant Pseudomonas aeruginosa

Pseudomonas aeruginosa is a gram-negative rod that can be isolated from nosocomial infections such as UTIs, bacteremias, pneumonias, skin and skin structure infections, and burn infections.²⁰ Pseudomonal infections are associated with high morbidity and mortality and can cause recurrent infections in patients with cystic fibrosis.²⁰ Multidrug-resistant P. aeruginosa (MDR-P) infections account for approximately 13% of all health care-associated pseudomonal infections nationally.¹ Both fluoroquinolone and aminoglycoside resistance has emerged, and multiple types of beta-lactamases (ESBL, AmpC, carbapenemases) have resulted in organisms that are resistant to nearly all anti-pseudomonal beta-lactams.²⁰

Treatment. For patients at risk for MDR-P, some clinical practice guidelines have recommended using an empiric therapy regimen that contains antimicrobial agents from 2 different classes with activity against P. aeruginosa to increase the likelihood of susceptibility to at least one agent.⁶ De-escalation can occur once culture and susceptibility results are available.⁶ Dose optimization based on pharmacodynamic principles is critical for ensuring clinical efficacy and minimizing resistance.³⁶ The use of high-dose, prolonged-infusion beta-lactams (piperacillin/tazobactam, cefepime, ceftazidime, and carbapenems) is becoming common practice at institutions with higher rates of resistance.^36-38

A resurgence of polymyxin (colistin) use for MDR-P isolates has occurred, and may be warranted empirically in select patients, based on local resistance patterns and patient history. Newer pharmacokinetic data are available, resulting in improved dosing strategies that may enhance efficacy while alleviating some of the nephrotoxicity concerns associated with colistin therapy.³⁹

Ceftolozane/tazobactam (Zerbaxa) and ceftazidime/avibactam (Avycaz) are options for complicated urinary tract and intra-abdominal infections caused by susceptible P. aeruginosa isolates. Given the lack of comparative efficacy data available for the management of MDR-P infections, agent selection should be based on site of infection, susceptibility data, and patient-specific factors.

Multi-drug resistant Acinetobacter baumannii

A. baumannii is a lactose-fermenting, gram-negative rod sometimes implicated in nosocomial pneumonias, line-related bloodstream infections, UTIs, and surgical site infections.²⁰ Resistance has been documented for nearly all classes of antibiotics, including carbapenems.^1,20 Over half of all health care-associated A. baumannii isolates in the United States are multidrug resistant.¹

Treatment. Therapy options for A. baumannii infections are often limited to polymyxins, tigecycline, carbapenems (except ertapenem), aminoglycosides, and high-dose ampicillin/sulbactam, depending on in vitro susceptibilities.^40,41 When using ampicillin/sulbactam for A. baumannii infections, sulbactam is the active ingredient. Doses of 2 to 4 g/d of sulbactam have demonstrated efficacy in non-critically ill patients, while critically ill patients may require higher doses (up to 12 g/d).⁴⁰ Colistin is considered the mainstay of therapy for carbapenem-resistant A. baumannii. It should be used in combination with either a carbapenem, rifampin, an aminoglycoside, or tigecycline.⁴²

Drug therapies for nosocomial-resistant gram-negative infections, along with clinical pearls for use, are summarized in TABLE 3.^{20,22,23,25,27-42} Because efficacy data are limited for treating infections caused by these pathogens, appropriate antimicrobial selection is frequently guided by location of infection, susceptibility patterns, and patient-specific factors such as allergies and the risk for adverse effects.

Antimicrobial stewardship

Antibiotic misuse has been a significant driver of antibiotic resistance.⁴⁶ Efforts to improve and measure the appropriate use of antibiotics have historically focused on acute care settings. Broad interventions to reduce antibiotic use include prospective audit with intervention and feedback, formulary restriction and preauthorization, and antibiotic time-outs.^47,48

Multidrug-resistant Pseudomonas aeruginosa infections account for approximately 13% of all health care-associated pseudomonal infections nationally.

Pharmacy-driven interventions include intravenous-to-oral conversions, dose adjustments for organ dysfunction, pharmacokinetic or pharmacodynamic interventions to optimize treatment for organisms with reduced susceptibility, therapeutic duplication alerts, and automatic-stop orders.^47,48

Diagnosis-specific interventions include order sets for common infections and the use of rapid diagnostic assays (TABLE 4^49,50). Rapid diagnostic testing is increasingly being considered an essential component of stewardship programs because it permits significantly shortened time to organism identification and susceptibility testing and allows for improved antibiotic utilization and patient outcomes when coupled with other effective stewardship strategies.⁴⁹

The core elements. The CDC has defined the core elements of successful inpatient ASPs.⁴⁶ These include:

• commitment from hospital leadership
• a physician leader who is responsible for overall program outcomes
• a pharmacist leader who co-leads the program and is accountable for enterprise-wide improvements in antibiotic use
• implementation of at least one systemic intervention (broad, pharmacy-driven, or infection/syndrome-specific)
• monitoring of prescribing and resistance patterns
• reporting antibiotic use and resistance patterns to all involved in the medication use process
• Education directed at the health care team about optimal antibiotic use.

Above all, success with antibiotic stewardship is dependent on identified leadership and an enterprise-wide multidisciplinary approach.

The FP’s role in hospital ASPs can take a number of forms. FPs who practice inpatient medicine should work with all members of their department and be supportive of efforts to improve antibiotic use. Prescribers should help develop and implement hospital-specific treatment recommendations, as well as be responsive to measurements and audits aimed at determining the quantity and quality of antibiotic use. Hospital-specific updates on antibiotic prescribing and antibiotic resistance should be shared widely through formal and informal settings. FPs should know if patients with resistant organisms are hospitalized at institutions where they practice, and should remain abreast of infection rates and resistance patterns.

Over half of all health care-associated Acinetobacter baumannii isolates in the United States are multidrug resistant.

When admitting a patient, the FP should ask if the patient has received medical care elsewhere, including in another country. When caring for patients known to be currently or previously colonized or infected with resistant organisms, the FP should follow the appropriate precautions and insist that all members of the health care team follow suit.

CASE

A diagnosis of carbapenem-resistant E.coli sepsis is eventually made. Additional susceptibility test results reported later the same day revealed sensitivity to tigecycline and colistin, with intermediate sensitivity to doripenem. An infectious disease expert recommended contact precautions and combination treatment with tigecycline and doripenem for at least 7 days. The addition of a polymyxin was also considered; however, the patient’s renal function was not favorable enough to support a course of that agent. Longer duration of therapy may be required if adequate source control is not achieved.

After a complicated ICU stay, including the need for surgical wound drainage, the patient responded satisfactorily and was transferred to a medical step-down unit for continued recovery and eventual discharge.

CORRESPONDENCE
Dora E. Wiskirchen, PharmD, BCPS, Department of Pharmacy, St. Francis Hospital and Medical Center, 114 Woodland St., Hartford, CT 06105; Email: [email protected].

CASE 

A 68-year-old woman is admitted to the hospital from home with acute onset, unrelenting, upper abdominal pain radiating to the back and nausea/vomiting. Her medical history includes bile duct obstruction secondary to gall stones, which was managed in another facility 6 days earlier with endoscopic retrograde cholangiopancreatography and stenting. The patient has type 2 diabetes (managed with metformin and glargine insulin), hypertension (managed with lisinopril and hydrochlorothiazide), and cholesterolemia (managed with atorvastatin).

On admission, the patient's white blood cell count is 14.7 x 10³ cells/mm³, heart rate is 100 bpm, blood pressure is 90/68 mm Hg, and temperature is 101.5° F. Serum amylase and lipase are 3 and 2 times the upper limit of normal, respectively. A working diagnosis of acute pancreatitis with sepsis is made. Blood cultures are drawn. A computed tomography scan confirms acute pancreatitis. She receives one dose of meropenem, is started on intravenous fluids and morphine, and is transferred to the intensive care unit (ICU) for further management.

Her ICU course is complicated by worsening sepsis despite aggressive fluid resuscitation, nutrition, and broad-spectrum antibiotics. On post-admission Day 2, blood culture results reveal Escherichia coli that is resistant to gentamicin, amoxicillin/clavulanate, ceftriaxone, piperacillin/tazobactam, imipenem, trimethoprim/sulfamethoxazole, ciprofloxacin, and tetracycline. Additional susceptibility testing is ordered.

The Centers for Disease Control and Prevention (CDC) conservatively estimates that antibiotic-resistant bacteria are responsible for 2 billion infections annually, resulting in approximately 23,000 deaths and $20 billion in excess health care expenditures annually.¹ Infections caused by antibiotic-resistant bacteria typically require longer hospitalizations, more expensive drug therapies, and additional follow-up visits.¹ They also result in greater morbidity and mortality compared with similar infections involving non-resistant bacteria.¹ To compound the problem, antibiotic development has steadily declined over the last 3 decades, with few novel antimicrobials developed in recent years.² The most recently approved antibiotics with new mechanisms of action were linezolid in 2000 and daptomycin in 2003, preceded by the carbapenems 15 years earlier. (See “New antimicrobials in the pipeline.”)

Greater efforts aimed at using antimicrobials sparingly and appropriately, as well as developing new antimicrobials with activity against multidrug-resistant pathogens, are ultimately needed to address the threat of antimicrobial resistance. This article describes the evidence-based management of inpatient infections caused by resistant bacteria and the role family physicians (FPs) can play in reducing further development of resistance through antimicrobial stewardship practices.

S. aureus is a common culprit of hospital-acquired infections, including central line-associated bloodstream infections, catheter-associated urinary tract infections, ventilator-associated pneumonia, and nosocomial skin and soft tissue infections. In fact, nearly half of all isolates from these infections are reported to be methicillin-resistant S. aureus (MRSA).³

Nearly half of all Staphylococcus aureus isolates from hospital-acquired infections are reported to be methicillin-resistant.

Patients at greatest risk for MRSA infections include those who have been recently hospitalized, those receiving recent antibiotic therapy or surgery, long-term care residents, intravenous drug abusers, immunocompromised patients, hemodialysis patients, military personnel, and athletes who play contact sports.^4,5 Patients with these infections often require the use of an anti-MRSA agent (eg, vancomycin, linezolid) in empiric antibiotic regimens.^6,7 The focus of this discussion is on MRSA in hospital and long-term care settings; a discussion of community-acquired MRSA is addressed elsewhere. (See “Antibiotic stewardship: The FP’s role,” J Fam Pract. 2016;65:876-885.⁸)

Efforts are working, but problems remain. MRSA accounts for almost 60% of S. aureus isolates in ICUs.⁹ Thankfully, rates of health care-associated MRSA are now either static or declining nationwide, as a result of major initiatives targeted toward preventing health care-associated infection in recent years.¹⁰

Methicillin resistance in S. aureus results from expression of PBP2a, an altered penicillin-binding protein with reduced binding affinity for beta-lactam antibiotics. As a result, MRSA isolates are resistant to most beta-lactams.⁹ Resistance to macrolides, azithromycin, aminoglycosides, fluoroquinolones, and clindamycin is also common in health care-associated MRSA.⁹

The first case of true vancomycin-resistant S. aureus (VRSA) in the United States was reported in 2002.¹¹ Fortunately, both VRSA and vancomycin-intermediate S. aureus (VISA) have remained rare throughout the United States and abroad.^9,11 Heterogeneous VISA (hVISA), which is characterized by a few resistant subpopulations within a fully susceptible population of S. aureus, is more common than VRSA or VISA. Unfortunately, hVISA is difficult to detect using commercially available susceptibility tests. This can result in treatment failure with vancomycin, even though the MRSA isolate may appear fully susceptible and the patient has received clinically appropriate doses of the drug.¹²

Treatment. Vancomycin is the mainstay of therapy for many systemic health care-associated MRSA infections. Alternative therapies (daptomycin or linezolid) should be considered for isolates with a vancomycin minimum inhibitory concentration (MIC) >2 mcg/mL or in the setting of a poor clinical response.⁴ Combination therapy may be warranted in the setting of treatment failure. Because comparative efficacy data for alternative therapies is lacking, agent selection should be tailored to the site of infection and patient-specific factors such as allergies, drug interactions, and the risk for adverse events (TABLE 1^13-17).

Oritavancin and dalbavancin both have dosing regimens that may allow for earlier hospital discharge or treatment in an outpatient setting.^13,14 Telavancin, quinupristin/dalfopristin, and tigecycline are typically reserved for salvage therapy due to adverse event profiles and/or limited efficacy data.¹⁵

Vancomycin-resistant enterococci (VRE)

Enterococci are typically considered normal gastrointestinal tract flora. However, antibiotic exposure can alter gut flora allowing for VRE colonization, which in some instances, can progress to the development of a health care-associated infection.¹⁵ Therefore, it is important to distinguish whether a patient is colonized or infected with VRE because treatment of colonization is unnecessary and may lead to resistance and other adverse effects.¹⁵

It's important to distinguish whether a patient is colonized or infected with vancomycin-resistant enterococci to avoid unnecessary treatment.

Enterococci may be the culprit in nosocomially-acquired intra-abdominal infections, bacteremia, endocarditis, urinary tract infections (UTIs), and skin and skin structure infections, and can exhibit resistance to ampicillin, aminoglycosides, and vancomycin.¹⁵ VRE is predominantly a health care-associated pathogen and may account for up to 77% of all health care-associated Enterococcus faecium infections and 9% of Enterococcus faecalis infections.¹

Treatment. Antibiotic selection for VRE infections depends upon the site of infection, patient comorbidities, the potential for drug interactions, and treatment duration. Current treatment options include linezolid, daptomycin, quinupristin/dalfopristin (for E. faecium only), tigecycline, and ampicillin if the organism is susceptible (TABLE 1^13-17).¹⁵ For cystitis caused by VRE (not urinary colonization), fosfomycin and nitrofurantoin are additional options.¹⁶

Resistant Enterobacteriaceae

Resistant Enterobacteriaceae such as Escherichia coli and Klebsiella pneumoniae have emerged as a result of increased broad-spectrum antibiotic utilization and have been implicated in health care-associated UTIs, intra-abdominal infections, bacteremia, and even pneumonia.¹ Patients with prolonged hospital stays and invasive medical devices, such as urinary and vascular catheters, endotracheal tubes, and endoscopy scopes, have the highest risk for infection with these organisms.²⁰

The genotypic profiles of resistance among the Enterobacteriaceae are diverse and complex, resulting in different levels of activity for the various beta-lactam agents (TABLE 2^21-24).²⁵ Furthermore, extended-spectrum beta-lactamase (ESBL)-producers and carbapenem-resistant Enterobacteriaceae (CRE) are often resistant to other classes of antibiotics, too, including aminoglycosides and fluoroquinolones.^20,25 The increasing diversity among beta-lactamase enzymes has made the selection of appropriate antibiotic therapy challenging, since the ability to identify specific beta-lactamase genes is not yet available in the clinical setting.

Cefepime may retain activity against some ESBL-producing isolates, but it is also susceptible to the inoculum effect and should only be used for non–life-threatening infections and at higher doses.²³ Fosfomycin has activity against ESBL-producing bacteria, but is only approved for oral use in UTIs in the United States.^20,27 Ceftolozane/tazobactam (Zerbaxa) and ceftazidime/avibactam (Avycaz) were approved in 2014 and 2015, respectively, by the FDA for the management of complicated urinary tract and intra-abdominal infections caused by susceptible ESBL-producing Enterobacteriaceae. In order to preserve the antimicrobial efficacy of these 2 newer agents, however, they are typically reserved for definitive therapy when in vitro susceptibility is demonstrated and there are no other viable options.

AmpC beta-lactamases are resistant to similar agents as the ESBLs, in addition to cefoxitin and the beta-lactam/beta-lactamase inhibitor combinations containing clavulanic acid, sulbactam, and in some cases, tazobactam. Resistance can be induced and emerges in certain pathogens while patients are on therapy.²⁸ Fluoroquinolones and aminoglycosides have a low risk of developing resistance while patients are on therapy, but are more likely to cause adverse effects and toxicity compared with the beta-lactams.²⁸ Carbapenems have the lowest risk of emerging resistance and are the empiric treatment of choice for known AmpC-producing Enterobacteriaceae in serious infections.^20,28 Cefepime may also be an option in less severe infections, such as UTIs or those in which adequate source control has been achieved.^28,29

Carbapenem-resistant Enterobacteriaceae (CRE) have become a serious threat as a result of increased carbapenem use. While carbapenem resistance is less common in the United States than worldwide, rates have increased nearly 4-fold (1.2% to 4.2%) in the last decade, with some regions of the country experiencing substantially higher rates.²⁴ The most commonly reported CRE genotypes identified in the United States include the serine carbapenemase (K. pneumoniae carbapenemase, or KPC), and the metallo-beta-lactamases (Verona integrin-encoded metallo-beta-lactamase, or VIM, and the New Dehli metallo-beta-lactamase, or NDM), with each class conferring slightly different resistance patterns (TABLE 2^21-24).^20,30

Few treatment options exist for Enterobacteriaceae producing a serine carbapenemase, and, unfortunately, evidence to support these therapies is extremely limited. Some CRE isolates retain susceptibility to the polymyxins, the aminoglycosides, and tigecycline.³⁰ Even fewer options exist for treating Enterobacteriaceae producing metallo-beta-lactamases, which are typically only susceptible to the polymyxins and tigecycline.^43-45

The increasing diversity among beta-lactamase enzymes has made the selection of appropriate antibiotics more challenging in recent years.

Several studies have demonstrated lower mortality rates when combination therapy is utilized for CRE bloodstream infections.^31,32 Furthermore, the combination of colistin, tigecycline, and meropenem was found to have a significant mortality advantage.³² Double carbapenem therapy has been effective in several cases of invasive KPC-producing K. pneumoniae infections.^33,34 However, it is important to note that current clinical evidence comes from small, single-center, retrospective studies, and additional research is needed to determine optimal combinations and dosing strategies for these infections.

Lastly, ceftazidime/avibactam (Avycaz) was recently approved for the treatment of complicated urinary tract and intra-abdominal infections, and has activity against KPC-producing Enterobacteriaceae, but not those producing metallo-beta-lactamases, like VIM or NDM. In the absence of strong evidence to support one therapy over another, it may be reasonable to select at least 2 active agents when treating serious CRE infections. Agent selection should be based on the site of the infection, susceptibility data, and patient-specific factors (TABLE 3^{20,22,,23,25,27-42}). The CDC also recommends contact precautions for patients who are colonized or infected with CRE.³⁵

Multi-drug resistant Pseudomonas aeruginosa

Pseudomonas aeruginosa is a gram-negative rod that can be isolated from nosocomial infections such as UTIs, bacteremias, pneumonias, skin and skin structure infections, and burn infections.²⁰ Pseudomonal infections are associated with high morbidity and mortality and can cause recurrent infections in patients with cystic fibrosis.²⁰ Multidrug-resistant P. aeruginosa (MDR-P) infections account for approximately 13% of all health care-associated pseudomonal infections nationally.¹ Both fluoroquinolone and aminoglycoside resistance has emerged, and multiple types of beta-lactamases (ESBL, AmpC, carbapenemases) have resulted in organisms that are resistant to nearly all anti-pseudomonal beta-lactams.²⁰

Treatment. For patients at risk for MDR-P, some clinical practice guidelines have recommended using an empiric therapy regimen that contains antimicrobial agents from 2 different classes with activity against P. aeruginosa to increase the likelihood of susceptibility to at least one agent.⁶ De-escalation can occur once culture and susceptibility results are available.⁶ Dose optimization based on pharmacodynamic principles is critical for ensuring clinical efficacy and minimizing resistance.³⁶ The use of high-dose, prolonged-infusion beta-lactams (piperacillin/tazobactam, cefepime, ceftazidime, and carbapenems) is becoming common practice at institutions with higher rates of resistance.^36-38

A resurgence of polymyxin (colistin) use for MDR-P isolates has occurred, and may be warranted empirically in select patients, based on local resistance patterns and patient history. Newer pharmacokinetic data are available, resulting in improved dosing strategies that may enhance efficacy while alleviating some of the nephrotoxicity concerns associated with colistin therapy.³⁹

Ceftolozane/tazobactam (Zerbaxa) and ceftazidime/avibactam (Avycaz) are options for complicated urinary tract and intra-abdominal infections caused by susceptible P. aeruginosa isolates. Given the lack of comparative efficacy data available for the management of MDR-P infections, agent selection should be based on site of infection, susceptibility data, and patient-specific factors.

Multi-drug resistant Acinetobacter baumannii

A. baumannii is a lactose-fermenting, gram-negative rod sometimes implicated in nosocomial pneumonias, line-related bloodstream infections, UTIs, and surgical site infections.²⁰ Resistance has been documented for nearly all classes of antibiotics, including carbapenems.^1,20 Over half of all health care-associated A. baumannii isolates in the United States are multidrug resistant.¹

Treatment. Therapy options for A. baumannii infections are often limited to polymyxins, tigecycline, carbapenems (except ertapenem), aminoglycosides, and high-dose ampicillin/sulbactam, depending on in vitro susceptibilities.^40,41 When using ampicillin/sulbactam for A. baumannii infections, sulbactam is the active ingredient. Doses of 2 to 4 g/d of sulbactam have demonstrated efficacy in non-critically ill patients, while critically ill patients may require higher doses (up to 12 g/d).⁴⁰ Colistin is considered the mainstay of therapy for carbapenem-resistant A. baumannii. It should be used in combination with either a carbapenem, rifampin, an aminoglycoside, or tigecycline.⁴²

Drug therapies for nosocomial-resistant gram-negative infections, along with clinical pearls for use, are summarized in TABLE 3.^{20,22,23,25,27-42} Because efficacy data are limited for treating infections caused by these pathogens, appropriate antimicrobial selection is frequently guided by location of infection, susceptibility patterns, and patient-specific factors such as allergies and the risk for adverse effects.

Antimicrobial stewardship

Antibiotic misuse has been a significant driver of antibiotic resistance.⁴⁶ Efforts to improve and measure the appropriate use of antibiotics have historically focused on acute care settings. Broad interventions to reduce antibiotic use include prospective audit with intervention and feedback, formulary restriction and preauthorization, and antibiotic time-outs.^47,48

Multidrug-resistant Pseudomonas aeruginosa infections account for approximately 13% of all health care-associated pseudomonal infections nationally.

Pharmacy-driven interventions include intravenous-to-oral conversions, dose adjustments for organ dysfunction, pharmacokinetic or pharmacodynamic interventions to optimize treatment for organisms with reduced susceptibility, therapeutic duplication alerts, and automatic-stop orders.^47,48

Diagnosis-specific interventions include order sets for common infections and the use of rapid diagnostic assays (TABLE 4^49,50). Rapid diagnostic testing is increasingly being considered an essential component of stewardship programs because it permits significantly shortened time to organism identification and susceptibility testing and allows for improved antibiotic utilization and patient outcomes when coupled with other effective stewardship strategies.⁴⁹

The core elements. The CDC has defined the core elements of successful inpatient ASPs.⁴⁶ These include:

• commitment from hospital leadership
• a physician leader who is responsible for overall program outcomes
• a pharmacist leader who co-leads the program and is accountable for enterprise-wide improvements in antibiotic use
• implementation of at least one systemic intervention (broad, pharmacy-driven, or infection/syndrome-specific)
• monitoring of prescribing and resistance patterns
• reporting antibiotic use and resistance patterns to all involved in the medication use process
• Education directed at the health care team about optimal antibiotic use.

Above all, success with antibiotic stewardship is dependent on identified leadership and an enterprise-wide multidisciplinary approach.

The FP’s role in hospital ASPs can take a number of forms. FPs who practice inpatient medicine should work with all members of their department and be supportive of efforts to improve antibiotic use. Prescribers should help develop and implement hospital-specific treatment recommendations, as well as be responsive to measurements and audits aimed at determining the quantity and quality of antibiotic use. Hospital-specific updates on antibiotic prescribing and antibiotic resistance should be shared widely through formal and informal settings. FPs should know if patients with resistant organisms are hospitalized at institutions where they practice, and should remain abreast of infection rates and resistance patterns.

Over half of all health care-associated Acinetobacter baumannii isolates in the United States are multidrug resistant.

When admitting a patient, the FP should ask if the patient has received medical care elsewhere, including in another country. When caring for patients known to be currently or previously colonized or infected with resistant organisms, the FP should follow the appropriate precautions and insist that all members of the health care team follow suit.

CASE

A diagnosis of carbapenem-resistant E.coli sepsis is eventually made. Additional susceptibility test results reported later the same day revealed sensitivity to tigecycline and colistin, with intermediate sensitivity to doripenem. An infectious disease expert recommended contact precautions and combination treatment with tigecycline and doripenem for at least 7 days. The addition of a polymyxin was also considered; however, the patient’s renal function was not favorable enough to support a course of that agent. Longer duration of therapy may be required if adequate source control is not achieved.

After a complicated ICU stay, including the need for surgical wound drainage, the patient responded satisfactorily and was transferred to a medical step-down unit for continued recovery and eventual discharge.