Cannavaping—the inhalation of a cannabis-containing aerosol, created by a battery-driven, heated atomizer in e-cigarettes or similar devices¹—is touted as a less expensive and safer alternative to smoking marijuana. It’s also gaining in popularity.² One study of Connecticut high school students found that 5.4% had used e-cigarettes to vaporize cannabis.³ But what do we know about this new way to get high?

We know that those who wish to cannavape can easily obtain e-cigarettes from gas stations and tobacco shops. They then have to obtain a cartridge, filled with either hash oil or tetrahydrocannabinol-infused wax, to attach to the e-cigarette. These cartridges are available for purchase in states that have legalized the sale of marijuana. They also find their way into states where the sale of marijuana is not legal, and are purchased illegally for the purpose of cannavaping.

And while cannavaping does appear to reduce the cost of smoking marijuana,⁴ it has not been widely researched, nor determined to be safe.⁵

In fact, although marijuana has several important therapeutic and medicinal purposes, cannavaping the substance can result in medical concerns.⁶ The vaping aerosols of some compounds can induce lung pathology and may be carcinogenic, since they often contain a number of dangerous toxins.⁴

Chronic marijuana use can increase the likelihood of motor vehicles accidents, cognitive impairment, psychoses, and demotivation.⁴ It may predispose certain individuals to use other drugs and tobacco products and could increase the consumption of marijuana.^4,5 Increased consumption could have a detrimental effect on intellect and behavior when used chronically—especially in youngsters, whose nervous systems are not yet fully matured.^7-9

Because cannavaping has potentially deleterious effects, more regulations on the manufacture, distribution, access, and use are indicated—at least until research sheds more light on issues surrounding this practice.

Steven Lippman, MD; Devina Singh, MD
Louisville, KY

Cannavaping—the inhalation of a cannabis-containing aerosol, created by a battery-driven, heated atomizer in e-cigarettes or similar devices¹—is touted as a less expensive and safer alternative to smoking marijuana. It’s also gaining in popularity.² One study of Connecticut high school students found that 5.4% had used e-cigarettes to vaporize cannabis.³ But what do we know about this new way to get high?

We know that those who wish to cannavape can easily obtain e-cigarettes from gas stations and tobacco shops. They then have to obtain a cartridge, filled with either hash oil or tetrahydrocannabinol-infused wax, to attach to the e-cigarette. These cartridges are available for purchase in states that have legalized the sale of marijuana. They also find their way into states where the sale of marijuana is not legal, and are purchased illegally for the purpose of cannavaping.

And while cannavaping does appear to reduce the cost of smoking marijuana,⁴ it has not been widely researched, nor determined to be safe.⁵

In fact, although marijuana has several important therapeutic and medicinal purposes, cannavaping the substance can result in medical concerns.⁶ The vaping aerosols of some compounds can induce lung pathology and may be carcinogenic, since they often contain a number of dangerous toxins.⁴

Chronic marijuana use can increase the likelihood of motor vehicles accidents, cognitive impairment, psychoses, and demotivation.⁴ It may predispose certain individuals to use other drugs and tobacco products and could increase the consumption of marijuana.^4,5 Increased consumption could have a detrimental effect on intellect and behavior when used chronically—especially in youngsters, whose nervous systems are not yet fully matured.^7-9

Because cannavaping has potentially deleterious effects, more regulations on the manufacture, distribution, access, and use are indicated—at least until research sheds more light on issues surrounding this practice.

Steven Lippman, MD; Devina Singh, MD
Louisville, KY

Cannavaping—the inhalation of a cannabis-containing aerosol, created by a battery-driven, heated atomizer in e-cigarettes or similar devices¹—is touted as a less expensive and safer alternative to smoking marijuana. It’s also gaining in popularity.² One study of Connecticut high school students found that 5.4% had used e-cigarettes to vaporize cannabis.³ But what do we know about this new way to get high?

We know that those who wish to cannavape can easily obtain e-cigarettes from gas stations and tobacco shops. They then have to obtain a cartridge, filled with either hash oil or tetrahydrocannabinol-infused wax, to attach to the e-cigarette. These cartridges are available for purchase in states that have legalized the sale of marijuana. They also find their way into states where the sale of marijuana is not legal, and are purchased illegally for the purpose of cannavaping.

And while cannavaping does appear to reduce the cost of smoking marijuana,⁴ it has not been widely researched, nor determined to be safe.⁵

In fact, although marijuana has several important therapeutic and medicinal purposes, cannavaping the substance can result in medical concerns.⁶ The vaping aerosols of some compounds can induce lung pathology and may be carcinogenic, since they often contain a number of dangerous toxins.⁴

Chronic marijuana use can increase the likelihood of motor vehicles accidents, cognitive impairment, psychoses, and demotivation.⁴ It may predispose certain individuals to use other drugs and tobacco products and could increase the consumption of marijuana.^4,5 Increased consumption could have a detrimental effect on intellect and behavior when used chronically—especially in youngsters, whose nervous systems are not yet fully matured.^7-9

Because cannavaping has potentially deleterious effects, more regulations on the manufacture, distribution, access, and use are indicated—at least until research sheds more light on issues surrounding this practice.

Steven Lippman, MD; Devina Singh, MD
Louisville, KY

ILLUSTRATIVE CASE

A 60-year-old man establishes care with you. He has well-controlled osteoarthritis (as long as he takes his low-dose daily aspirin) and chronic atrial fibrillation, for which he takes warfarin. His international normalized ratio (INR) is consistently within the recommended target range of 2 to 3. He feels well and has never had gastroesophageal reflux disease (GERD) or a gastrointestinal (GI) bleed. Should you recommend a proton pump inhibitor (PPI) to decrease the likelihood of a future upper GI bleed?

Anticoagulation therapy creates a dilemma—the need to balance the benefits of preventing embolization with the risks of serious bleeding. Concurrent use of nonsteroidal anti-inflammatory drugs (NSAIDs), aspirin, and other antiplatelet agents further increases the risk of the latter.²

Physicians have long used PPIs to treat upper GI bleeds. They prevent acid secretion and are the most efficacious drugs for healing peptic ulcers.^3,4 However, while previous case-control studies show that PPIs reduce the risk of upper GI bleeds in patients taking antiplatelet agents or NSAIDs, they do not show a statistically significant benefit for patients taking warfarin.^5,6 Further reflecting the confusion and uncertainty surrounding this issue is that while one expert consensus report recommends that patients taking dual warfarin and antiplatelet agent/NSAID therapy take a PPI to decrease the risk of upper GI bleeding,² other guidelines regarding anticoagulant therapy do not address this clinical question.^2,7,8

[polldaddy:9860876]

STUDY SUMMARY

Study lends support to PPI use in a high-risk group

This retrospective cohort study sought to answer the question: “Does PPI co-therapy decrease the rate of serious upper GI bleeds in patients taking warfarin?” Researchers examined rates of hospitalization for upper GI bleeding for Medicare and Medicaid patients taking warfarin with and without PPI co-therapy (tracked via prescription fill dates). They also evaluated concomitant use of NSAIDs and antiplatelet agents.

The authors excluded patients with a recent history of a severe bleed or certain illnesses that would predispose a patient to GI bleeding (such as esophageal varices). Patients with risk factors for an upper GI bleed (such as abdominal pain, peptic ulcer disease, anemia, etc.) were more likely to be taking PPI co-therapy. Researchers analyzed the effect of PPI co-therapy in patients with and without these additional risk factors.

Results. The study followed over 75,000 person-years of active warfarin therapy (more than 52,000 person-years in the Medicaid cohort and more than 23,000 person-years in the Medicare cohort). Hospitalizations due to upper GI bleeding occurred at a rate of 127/10,000 person-years (incidence was similar in both the Medicaid and Medicare groups).

Looking at all patients taking warfarin (regardless of whether or not they were also taking an NSAID or antiplatelet agent), PPI co-therapy reduced the risk of hospitalization for upper GI bleeding by 24% (adjusted hazard ratio [HR]=0.76; 95% confidence interval [CI], 0.63 to 0.91), which translates into 29 fewer hospitalizations per 10,000 person-years. The number needed to treat (NNT) was 345 person-years, meaning 345 patients taking warfarin would have to take a PPI for one year to prevent one hospitalization for an upper GI bleed. As one might expect, PPI co-therapy did not significantly reduce the risk of lower GI, other GI, or non-GI bleeding.

Further research is warranted to determine if PPI therapy is beneficial to patients taking direct oral anticoagulants.

In patients taking both warfarin and concurrent antiplatelet agents or NSAIDs, PPI co-therapy reduced the risk of hospitalization for upper GI bleeding by about half (HR=0.55; 95% CI, 0.39-0.77). Hospitalizations decreased by 128/10,000 person-years (95% CI, -66 to -173), yielding an NNT of 78 person-years. For patients taking warfarin but not antiplatelet agents or NSAIDs, PPI co-therapy did not significantly decrease the risk of hospitalization for upper GI bleeding (HR=0.86; 95% CI, 0.70-1.06).

Additional risk factors for GI bleeds. Researchers also looked at patients who had additional risk factors for GI bleeds (other than the exclusion criteria). For patients taking both warfarin and an antiplatelet agent/NSAID, PPI co-therapy decreased the risk of upper GI bleeding whether or not the patients had other bleeding risk factors. Again, for patients who had additional bleeding risk factors, but were not taking an antiplatelet agent or NSAID, PPI therapy showed no statistically significant effect.

WHAT’S NEW

PPIs offer benefits, but not to patients taking warfarin alone

The statistically significant results in this large observational study suggest that PPI co-therapy is beneficial in reducing the risk of upper GI bleeding in patients taking warfarin plus an antiplatelet agent/NSAID, but that PPI co-therapy provides no benefit to patients taking warfarin exclusively.

CAVEATS

Study was good, but it wasn’t a randomized controlled trial

This study is observational, and not a randomized control trial (RCT). Therefore, unknown confounding variables may have skewed results. For example, patients could have taken over-the-counter medications that influenced or obscured results, but were not included in the data analysis (misclassification bias).

At best, we can infer a correlation between PPIs and decreased risk of upper GI bleeds. We need RCTs to determine whether PPIs cause a lower risk.

Don’t overlook the risk of PPIs. This study assessed the ability of PPIs to prevent bleeds, but did not address the risks of long-term PPI therapy. Adverse effects of PPIs include an increased risk of pneumonia, infection with Clostridium difficile, hip and spine fractures, anemia, and possibly chronic kidney disease and dementia.^9-11 In addition, cost-analysis studies of PPI therapy are limited and inconsistent.¹² Therefore, it’s best to make decisions regarding PPIs after discussing other risks and benefits.

What about DOACs? Another consideration is the option to prescribe a direct oral anticoagulant (DOAC), such as dabigatran, rivaroxaban, or apixaban, instead of warfarin. DOACs are at least as effective as warfarin at preventing stroke in patients with atrial fibrillation and may even be safer.¹³ Dabigatran 110 mg causes fewer “major bleeding” events than warfarin.¹³ Rivaroxaban has been shown to result in fewer fatal bleeding events than warfarin due to fatal intracranial bleeds, although it is associated with more GI bleedding.¹³ Compared with warfarin, apixaban is associated with fewer GI bleeds and lower bleeding rates overall.¹³ Further research is warranted to determine if PPI therapy is beneficial to patients taking DOACs.

CHALLENGES TO IMPLEMENTATION

It’s still a balancing act

When chronic anticoagulation is necessary, physicians and patients must attempt to prevent thrombotic events while minimizing the risk of GI bleeds. PPIs may be beneficial in preventing upper GI bleeds in patients taking dual warfarin and antiplatelet therapy, but the long-term consequences of PPI therapy should not be ignored.

ACKNOWLEDGEMENT

The PURLs Surveillance System was supported in part by Grant Number UL1RR024999 from the National Center For Research Resources, a Clinical Translational Science Award to the University of Chicago. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Center For Research Resources or the National Institutes of Health.

ILLUSTRATIVE CASE

A 60-year-old man establishes care with you. He has well-controlled osteoarthritis (as long as he takes his low-dose daily aspirin) and chronic atrial fibrillation, for which he takes warfarin. His international normalized ratio (INR) is consistently within the recommended target range of 2 to 3. He feels well and has never had gastroesophageal reflux disease (GERD) or a gastrointestinal (GI) bleed. Should you recommend a proton pump inhibitor (PPI) to decrease the likelihood of a future upper GI bleed?

Anticoagulation therapy creates a dilemma—the need to balance the benefits of preventing embolization with the risks of serious bleeding. Concurrent use of nonsteroidal anti-inflammatory drugs (NSAIDs), aspirin, and other antiplatelet agents further increases the risk of the latter.²

Physicians have long used PPIs to treat upper GI bleeds. They prevent acid secretion and are the most efficacious drugs for healing peptic ulcers.^3,4 However, while previous case-control studies show that PPIs reduce the risk of upper GI bleeds in patients taking antiplatelet agents or NSAIDs, they do not show a statistically significant benefit for patients taking warfarin.^5,6 Further reflecting the confusion and uncertainty surrounding this issue is that while one expert consensus report recommends that patients taking dual warfarin and antiplatelet agent/NSAID therapy take a PPI to decrease the risk of upper GI bleeding,² other guidelines regarding anticoagulant therapy do not address this clinical question.^2,7,8

[polldaddy:9860876]

STUDY SUMMARY

Study lends support to PPI use in a high-risk group

This retrospective cohort study sought to answer the question: “Does PPI co-therapy decrease the rate of serious upper GI bleeds in patients taking warfarin?” Researchers examined rates of hospitalization for upper GI bleeding for Medicare and Medicaid patients taking warfarin with and without PPI co-therapy (tracked via prescription fill dates). They also evaluated concomitant use of NSAIDs and antiplatelet agents.

The authors excluded patients with a recent history of a severe bleed or certain illnesses that would predispose a patient to GI bleeding (such as esophageal varices). Patients with risk factors for an upper GI bleed (such as abdominal pain, peptic ulcer disease, anemia, etc.) were more likely to be taking PPI co-therapy. Researchers analyzed the effect of PPI co-therapy in patients with and without these additional risk factors.

Results. The study followed over 75,000 person-years of active warfarin therapy (more than 52,000 person-years in the Medicaid cohort and more than 23,000 person-years in the Medicare cohort). Hospitalizations due to upper GI bleeding occurred at a rate of 127/10,000 person-years (incidence was similar in both the Medicaid and Medicare groups).

Looking at all patients taking warfarin (regardless of whether or not they were also taking an NSAID or antiplatelet agent), PPI co-therapy reduced the risk of hospitalization for upper GI bleeding by 24% (adjusted hazard ratio [HR]=0.76; 95% confidence interval [CI], 0.63 to 0.91), which translates into 29 fewer hospitalizations per 10,000 person-years. The number needed to treat (NNT) was 345 person-years, meaning 345 patients taking warfarin would have to take a PPI for one year to prevent one hospitalization for an upper GI bleed. As one might expect, PPI co-therapy did not significantly reduce the risk of lower GI, other GI, or non-GI bleeding.

Further research is warranted to determine if PPI therapy is beneficial to patients taking direct oral anticoagulants.

In patients taking both warfarin and concurrent antiplatelet agents or NSAIDs, PPI co-therapy reduced the risk of hospitalization for upper GI bleeding by about half (HR=0.55; 95% CI, 0.39-0.77). Hospitalizations decreased by 128/10,000 person-years (95% CI, -66 to -173), yielding an NNT of 78 person-years. For patients taking warfarin but not antiplatelet agents or NSAIDs, PPI co-therapy did not significantly decrease the risk of hospitalization for upper GI bleeding (HR=0.86; 95% CI, 0.70-1.06).

Additional risk factors for GI bleeds. Researchers also looked at patients who had additional risk factors for GI bleeds (other than the exclusion criteria). For patients taking both warfarin and an antiplatelet agent/NSAID, PPI co-therapy decreased the risk of upper GI bleeding whether or not the patients had other bleeding risk factors. Again, for patients who had additional bleeding risk factors, but were not taking an antiplatelet agent or NSAID, PPI therapy showed no statistically significant effect.

WHAT’S NEW

PPIs offer benefits, but not to patients taking warfarin alone

The statistically significant results in this large observational study suggest that PPI co-therapy is beneficial in reducing the risk of upper GI bleeding in patients taking warfarin plus an antiplatelet agent/NSAID, but that PPI co-therapy provides no benefit to patients taking warfarin exclusively.

CAVEATS

Study was good, but it wasn’t a randomized controlled trial

This study is observational, and not a randomized control trial (RCT). Therefore, unknown confounding variables may have skewed results. For example, patients could have taken over-the-counter medications that influenced or obscured results, but were not included in the data analysis (misclassification bias).

At best, we can infer a correlation between PPIs and decreased risk of upper GI bleeds. We need RCTs to determine whether PPIs cause a lower risk.

Don’t overlook the risk of PPIs. This study assessed the ability of PPIs to prevent bleeds, but did not address the risks of long-term PPI therapy. Adverse effects of PPIs include an increased risk of pneumonia, infection with Clostridium difficile, hip and spine fractures, anemia, and possibly chronic kidney disease and dementia.^9-11 In addition, cost-analysis studies of PPI therapy are limited and inconsistent.¹² Therefore, it’s best to make decisions regarding PPIs after discussing other risks and benefits.

What about DOACs? Another consideration is the option to prescribe a direct oral anticoagulant (DOAC), such as dabigatran, rivaroxaban, or apixaban, instead of warfarin. DOACs are at least as effective as warfarin at preventing stroke in patients with atrial fibrillation and may even be safer.¹³ Dabigatran 110 mg causes fewer “major bleeding” events than warfarin.¹³ Rivaroxaban has been shown to result in fewer fatal bleeding events than warfarin due to fatal intracranial bleeds, although it is associated with more GI bleedding.¹³ Compared with warfarin, apixaban is associated with fewer GI bleeds and lower bleeding rates overall.¹³ Further research is warranted to determine if PPI therapy is beneficial to patients taking DOACs.

CHALLENGES TO IMPLEMENTATION

It’s still a balancing act

When chronic anticoagulation is necessary, physicians and patients must attempt to prevent thrombotic events while minimizing the risk of GI bleeds. PPIs may be beneficial in preventing upper GI bleeds in patients taking dual warfarin and antiplatelet therapy, but the long-term consequences of PPI therapy should not be ignored.

ACKNOWLEDGEMENT

The PURLs Surveillance System was supported in part by Grant Number UL1RR024999 from the National Center For Research Resources, a Clinical Translational Science Award to the University of Chicago. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Center For Research Resources or the National Institutes of Health.

ILLUSTRATIVE CASE

A 60-year-old man establishes care with you. He has well-controlled osteoarthritis (as long as he takes his low-dose daily aspirin) and chronic atrial fibrillation, for which he takes warfarin. His international normalized ratio (INR) is consistently within the recommended target range of 2 to 3. He feels well and has never had gastroesophageal reflux disease (GERD) or a gastrointestinal (GI) bleed. Should you recommend a proton pump inhibitor (PPI) to decrease the likelihood of a future upper GI bleed?

Anticoagulation therapy creates a dilemma—the need to balance the benefits of preventing embolization with the risks of serious bleeding. Concurrent use of nonsteroidal anti-inflammatory drugs (NSAIDs), aspirin, and other antiplatelet agents further increases the risk of the latter.²

Physicians have long used PPIs to treat upper GI bleeds. They prevent acid secretion and are the most efficacious drugs for healing peptic ulcers.^3,4 However, while previous case-control studies show that PPIs reduce the risk of upper GI bleeds in patients taking antiplatelet agents or NSAIDs, they do not show a statistically significant benefit for patients taking warfarin.^5,6 Further reflecting the confusion and uncertainty surrounding this issue is that while one expert consensus report recommends that patients taking dual warfarin and antiplatelet agent/NSAID therapy take a PPI to decrease the risk of upper GI bleeding,² other guidelines regarding anticoagulant therapy do not address this clinical question.^2,7,8

[polldaddy:9860876]

STUDY SUMMARY

Study lends support to PPI use in a high-risk group

This retrospective cohort study sought to answer the question: “Does PPI co-therapy decrease the rate of serious upper GI bleeds in patients taking warfarin?” Researchers examined rates of hospitalization for upper GI bleeding for Medicare and Medicaid patients taking warfarin with and without PPI co-therapy (tracked via prescription fill dates). They also evaluated concomitant use of NSAIDs and antiplatelet agents.

The authors excluded patients with a recent history of a severe bleed or certain illnesses that would predispose a patient to GI bleeding (such as esophageal varices). Patients with risk factors for an upper GI bleed (such as abdominal pain, peptic ulcer disease, anemia, etc.) were more likely to be taking PPI co-therapy. Researchers analyzed the effect of PPI co-therapy in patients with and without these additional risk factors.

Results. The study followed over 75,000 person-years of active warfarin therapy (more than 52,000 person-years in the Medicaid cohort and more than 23,000 person-years in the Medicare cohort). Hospitalizations due to upper GI bleeding occurred at a rate of 127/10,000 person-years (incidence was similar in both the Medicaid and Medicare groups).

Looking at all patients taking warfarin (regardless of whether or not they were also taking an NSAID or antiplatelet agent), PPI co-therapy reduced the risk of hospitalization for upper GI bleeding by 24% (adjusted hazard ratio [HR]=0.76; 95% confidence interval [CI], 0.63 to 0.91), which translates into 29 fewer hospitalizations per 10,000 person-years. The number needed to treat (NNT) was 345 person-years, meaning 345 patients taking warfarin would have to take a PPI for one year to prevent one hospitalization for an upper GI bleed. As one might expect, PPI co-therapy did not significantly reduce the risk of lower GI, other GI, or non-GI bleeding.

Further research is warranted to determine if PPI therapy is beneficial to patients taking direct oral anticoagulants.

In patients taking both warfarin and concurrent antiplatelet agents or NSAIDs, PPI co-therapy reduced the risk of hospitalization for upper GI bleeding by about half (HR=0.55; 95% CI, 0.39-0.77). Hospitalizations decreased by 128/10,000 person-years (95% CI, -66 to -173), yielding an NNT of 78 person-years. For patients taking warfarin but not antiplatelet agents or NSAIDs, PPI co-therapy did not significantly decrease the risk of hospitalization for upper GI bleeding (HR=0.86; 95% CI, 0.70-1.06).

Additional risk factors for GI bleeds. Researchers also looked at patients who had additional risk factors for GI bleeds (other than the exclusion criteria). For patients taking both warfarin and an antiplatelet agent/NSAID, PPI co-therapy decreased the risk of upper GI bleeding whether or not the patients had other bleeding risk factors. Again, for patients who had additional bleeding risk factors, but were not taking an antiplatelet agent or NSAID, PPI therapy showed no statistically significant effect.

WHAT’S NEW

PPIs offer benefits, but not to patients taking warfarin alone

The statistically significant results in this large observational study suggest that PPI co-therapy is beneficial in reducing the risk of upper GI bleeding in patients taking warfarin plus an antiplatelet agent/NSAID, but that PPI co-therapy provides no benefit to patients taking warfarin exclusively.

CAVEATS

Study was good, but it wasn’t a randomized controlled trial

This study is observational, and not a randomized control trial (RCT). Therefore, unknown confounding variables may have skewed results. For example, patients could have taken over-the-counter medications that influenced or obscured results, but were not included in the data analysis (misclassification bias).

At best, we can infer a correlation between PPIs and decreased risk of upper GI bleeds. We need RCTs to determine whether PPIs cause a lower risk.

Don’t overlook the risk of PPIs. This study assessed the ability of PPIs to prevent bleeds, but did not address the risks of long-term PPI therapy. Adverse effects of PPIs include an increased risk of pneumonia, infection with Clostridium difficile, hip and spine fractures, anemia, and possibly chronic kidney disease and dementia.^9-11 In addition, cost-analysis studies of PPI therapy are limited and inconsistent.¹² Therefore, it’s best to make decisions regarding PPIs after discussing other risks and benefits.

What about DOACs? Another consideration is the option to prescribe a direct oral anticoagulant (DOAC), such as dabigatran, rivaroxaban, or apixaban, instead of warfarin. DOACs are at least as effective as warfarin at preventing stroke in patients with atrial fibrillation and may even be safer.¹³ Dabigatran 110 mg causes fewer “major bleeding” events than warfarin.¹³ Rivaroxaban has been shown to result in fewer fatal bleeding events than warfarin due to fatal intracranial bleeds, although it is associated with more GI bleedding.¹³ Compared with warfarin, apixaban is associated with fewer GI bleeds and lower bleeding rates overall.¹³ Further research is warranted to determine if PPI therapy is beneficial to patients taking DOACs.

CHALLENGES TO IMPLEMENTATION

It’s still a balancing act

When chronic anticoagulation is necessary, physicians and patients must attempt to prevent thrombotic events while minimizing the risk of GI bleeds. PPIs may be beneficial in preventing upper GI bleeds in patients taking dual warfarin and antiplatelet therapy, but the long-term consequences of PPI therapy should not be ignored.

ACKNOWLEDGEMENT

The PURLs Surveillance System was supported in part by Grant Number UL1RR024999 from the National Center For Research Resources, a Clinical Translational Science Award to the University of Chicago. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Center For Research Resources or the National Institutes of Health.

Last month’s mass shooting in Las Vegas, which killed 59 people and wounded 500, was committed by a single individual who legally purchased an arsenal that allowed him to fire hundreds of high-caliber bullets within minutes into a large crowd. This is just the latest in a series of high-profile mass killings that appear to be increasing in frequency.¹

As terrifying as mass murders are, they account for only a small fraction of gun-related mortality. Everyday about 80 people in the United States are killed by a gun, usually by someone they know or by themselves (almost two-thirds of gun-related mortality involves suicide).² No other developed country even comes close to our rate of gun-related violence.²

What to do? Recall anti-smoking efforts. Gun violence is a public health issue that should be addressed with tried and proven public health methods. A couple of examples from history hold valuable lessons. While tobacco-related mortality and morbidity remain public health concerns, we have made marked improvements and saved many lives through a series of public health interventions including increasing the price of tobacco products, restricting advertising and sales to minors, and prohibiting smoking in public areas, to name a few.³

These interventions occurred because the public recognized the threat of tobacco and was willing to adopt them. This was not always the case. During the first half of my life, smoking in public, including indoors at public events and even on airplanes, was accepted, and the “rights of smokers” were respected. This now seems inconceivable. Public health interventions work, and public perceptions and attitudes can change.

If we gather data and fund research, we can make changes to reduce deaths while maintaining the right to own a firearm.

Consider inroads made in driver safety, too. We have also made marked improvements in motor vehicle crash-related deaths and injuries.⁴ For decades, we have recorded hundreds of data points on every car crash resulting in a death in a comprehensive database—the Fatality Analysis Reporting System (FARS). These data have been used by researchers to identify causes of crashes and crash-related deaths and have led to improvements in car design and road safety. Additional factors leading to improved road safety include restrictions on the age at which one can drive and on drinking alcohol and driving.

We can achieve similar improvements in gun-related mortality if we establish and maintain a comprehensive database, encourage and fund research, and are willing to adopt some commonsense product improvements and ownership restrictions that, nevertheless, preserve the right for most to responsibly own a firearm.

Don’t you think it’s time?

Last month’s mass shooting in Las Vegas, which killed 59 people and wounded 500, was committed by a single individual who legally purchased an arsenal that allowed him to fire hundreds of high-caliber bullets within minutes into a large crowd. This is just the latest in a series of high-profile mass killings that appear to be increasing in frequency.¹

As terrifying as mass murders are, they account for only a small fraction of gun-related mortality. Everyday about 80 people in the United States are killed by a gun, usually by someone they know or by themselves (almost two-thirds of gun-related mortality involves suicide).² No other developed country even comes close to our rate of gun-related violence.²

What to do? Recall anti-smoking efforts. Gun violence is a public health issue that should be addressed with tried and proven public health methods. A couple of examples from history hold valuable lessons. While tobacco-related mortality and morbidity remain public health concerns, we have made marked improvements and saved many lives through a series of public health interventions including increasing the price of tobacco products, restricting advertising and sales to minors, and prohibiting smoking in public areas, to name a few.³

These interventions occurred because the public recognized the threat of tobacco and was willing to adopt them. This was not always the case. During the first half of my life, smoking in public, including indoors at public events and even on airplanes, was accepted, and the “rights of smokers” were respected. This now seems inconceivable. Public health interventions work, and public perceptions and attitudes can change.

If we gather data and fund research, we can make changes to reduce deaths while maintaining the right to own a firearm.

Consider inroads made in driver safety, too. We have also made marked improvements in motor vehicle crash-related deaths and injuries.⁴ For decades, we have recorded hundreds of data points on every car crash resulting in a death in a comprehensive database—the Fatality Analysis Reporting System (FARS). These data have been used by researchers to identify causes of crashes and crash-related deaths and have led to improvements in car design and road safety. Additional factors leading to improved road safety include restrictions on the age at which one can drive and on drinking alcohol and driving.

We can achieve similar improvements in gun-related mortality if we establish and maintain a comprehensive database, encourage and fund research, and are willing to adopt some commonsense product improvements and ownership restrictions that, nevertheless, preserve the right for most to responsibly own a firearm.

Don’t you think it’s time?

Last month’s mass shooting in Las Vegas, which killed 59 people and wounded 500, was committed by a single individual who legally purchased an arsenal that allowed him to fire hundreds of high-caliber bullets within minutes into a large crowd. This is just the latest in a series of high-profile mass killings that appear to be increasing in frequency.¹

As terrifying as mass murders are, they account for only a small fraction of gun-related mortality. Everyday about 80 people in the United States are killed by a gun, usually by someone they know or by themselves (almost two-thirds of gun-related mortality involves suicide).² No other developed country even comes close to our rate of gun-related violence.²

What to do? Recall anti-smoking efforts. Gun violence is a public health issue that should be addressed with tried and proven public health methods. A couple of examples from history hold valuable lessons. While tobacco-related mortality and morbidity remain public health concerns, we have made marked improvements and saved many lives through a series of public health interventions including increasing the price of tobacco products, restricting advertising and sales to minors, and prohibiting smoking in public areas, to name a few.³

These interventions occurred because the public recognized the threat of tobacco and was willing to adopt them. This was not always the case. During the first half of my life, smoking in public, including indoors at public events and even on airplanes, was accepted, and the “rights of smokers” were respected. This now seems inconceivable. Public health interventions work, and public perceptions and attitudes can change.

If we gather data and fund research, we can make changes to reduce deaths while maintaining the right to own a firearm.

Consider inroads made in driver safety, too. We have also made marked improvements in motor vehicle crash-related deaths and injuries.⁴ For decades, we have recorded hundreds of data points on every car crash resulting in a death in a comprehensive database—the Fatality Analysis Reporting System (FARS). These data have been used by researchers to identify causes of crashes and crash-related deaths and have led to improvements in car design and road safety. Additional factors leading to improved road safety include restrictions on the age at which one can drive and on drinking alcohol and driving.

We can achieve similar improvements in gun-related mortality if we establish and maintain a comprehensive database, encourage and fund research, and are willing to adopt some commonsense product improvements and ownership restrictions that, nevertheless, preserve the right for most to responsibly own a firearm.

Don’t you think it’s time?

EVIDENCE SUMMARY

A systematic review featured a single RCT (n=185) comparing the effect of a 12-month vs 24-month interval between dental visits on dental caries in low-risk 3- to 5-year-old children with primary teeth and young adults, ages 16 to 20 years, with permanent teeth.¹ The outcomes of caries (ie, decayed, missing, filled surfaces increment) between the 12- and 24-month visits both in younger children (mean difference [MD]= -0.90; 95% confidence interval [CI], -1.96 to 0.16) and young adults (MD= -0.86; 95% CI, -1.75 to 0.03) did not differ.

Gingivitis: Not an issue when visits were delayed in healthy adults

Another systematic review (3 RCTs; N=836) evaluated the benefits associated with scaling and polishing in the prevention of gingivitis (primary outcome measure).² One RCT (n=207) compared scaling and polishing at 6- and 12-month intervals to no treatment for 24 months in adults with healthy dental histories. There was no difference in the percentage of index teeth with bleeding in the 6-month or 12-month treatment groups compared to the group that received no treatment for 24 months (MD= -2%; 95% CI, -10% to 6% and MD= -1%; 95% CI, -9% to 7%, respectively).

2 visits/year prevents tooth loss in high-risk patients

A retrospective cohort study (N=5117) using 16 years of data evaluated the association between one or 2 preventive dental visits per year and tooth extraction events in adults at low risk and those at high risk for progressive periodontitis.³ Those at high risk had at least one of the following risk factors: smoking, diabetes, or interleukin-1 genotype. Low-risk patients had no difference in tooth loss with one visit compared to 2 visits annually (absolute risk reduction [ARR]=2.6%; 95% CI, 0.5%-5.8%; P=.092); however, high-risk patients had fewer events with 2 annual visits (number needed to treat [NNT]=19; ARR 5.2%; 95% CI, 1.8%-8.4%; P=.002).

Visits before age 3 likely benefit only those at high risk

A systematic review of 4 retrospective cohort studies (N=77,291) analyzed the impact of early preventive dental visits (EPDV) on the frequency of future preventive and non-preventive dental visits and related expenditures using data from insurance claims and a kindergarten state dental registry.⁴ One study (n=11,394) used dental disease status at kindergarten (defined as the count of decayed, missing [molar teeth only], and filled primary teeth) as an outcome measure. Children who received EPDV before age 24 months had a comparable number of caries to those who had EPDV at 24 to 36 months. The authors concluded that EPDV before age 3 years is likely to benefit only children at high risk, and that evidence for a first dental visit by age one year is weak.

RECOMMENDATIONS

The National Institute for Health and Care Excellence recommends preventive dental visit intervals based on individual risk.

The National Institute for Health and Care Excellence recommends preventive dental visit intervals based on individual risk (12 months as the longest interval under age 18 years and 24 months as the longest interval for those 18 years and older at low risk).⁵ The American Dental Association recommends preventive dental visits at intervals determined by individual risk.⁶ The American Academy of Pediatric Dentistry recommends a first exam by age one year and preventive dental visits every 6 months through adolescence or as indicated by individual risk.⁷ The US Preventive Services Task Force states there is insufficient evidence to recommend routine dental screening by primary care physicians in children up to age 5 years.⁸

EVIDENCE SUMMARY

A systematic review featured a single RCT (n=185) comparing the effect of a 12-month vs 24-month interval between dental visits on dental caries in low-risk 3- to 5-year-old children with primary teeth and young adults, ages 16 to 20 years, with permanent teeth.¹ The outcomes of caries (ie, decayed, missing, filled surfaces increment) between the 12- and 24-month visits both in younger children (mean difference [MD]= -0.90; 95% confidence interval [CI], -1.96 to 0.16) and young adults (MD= -0.86; 95% CI, -1.75 to 0.03) did not differ.

Gingivitis: Not an issue when visits were delayed in healthy adults

Another systematic review (3 RCTs; N=836) evaluated the benefits associated with scaling and polishing in the prevention of gingivitis (primary outcome measure).² One RCT (n=207) compared scaling and polishing at 6- and 12-month intervals to no treatment for 24 months in adults with healthy dental histories. There was no difference in the percentage of index teeth with bleeding in the 6-month or 12-month treatment groups compared to the group that received no treatment for 24 months (MD= -2%; 95% CI, -10% to 6% and MD= -1%; 95% CI, -9% to 7%, respectively).

2 visits/year prevents tooth loss in high-risk patients

A retrospective cohort study (N=5117) using 16 years of data evaluated the association between one or 2 preventive dental visits per year and tooth extraction events in adults at low risk and those at high risk for progressive periodontitis.³ Those at high risk had at least one of the following risk factors: smoking, diabetes, or interleukin-1 genotype. Low-risk patients had no difference in tooth loss with one visit compared to 2 visits annually (absolute risk reduction [ARR]=2.6%; 95% CI, 0.5%-5.8%; P=.092); however, high-risk patients had fewer events with 2 annual visits (number needed to treat [NNT]=19; ARR 5.2%; 95% CI, 1.8%-8.4%; P=.002).

Visits before age 3 likely benefit only those at high risk

A systematic review of 4 retrospective cohort studies (N=77,291) analyzed the impact of early preventive dental visits (EPDV) on the frequency of future preventive and non-preventive dental visits and related expenditures using data from insurance claims and a kindergarten state dental registry.⁴ One study (n=11,394) used dental disease status at kindergarten (defined as the count of decayed, missing [molar teeth only], and filled primary teeth) as an outcome measure. Children who received EPDV before age 24 months had a comparable number of caries to those who had EPDV at 24 to 36 months. The authors concluded that EPDV before age 3 years is likely to benefit only children at high risk, and that evidence for a first dental visit by age one year is weak.

RECOMMENDATIONS

The National Institute for Health and Care Excellence recommends preventive dental visit intervals based on individual risk.

The National Institute for Health and Care Excellence recommends preventive dental visit intervals based on individual risk (12 months as the longest interval under age 18 years and 24 months as the longest interval for those 18 years and older at low risk).⁵ The American Dental Association recommends preventive dental visits at intervals determined by individual risk.⁶ The American Academy of Pediatric Dentistry recommends a first exam by age one year and preventive dental visits every 6 months through adolescence or as indicated by individual risk.⁷ The US Preventive Services Task Force states there is insufficient evidence to recommend routine dental screening by primary care physicians in children up to age 5 years.⁸

EVIDENCE SUMMARY

A systematic review featured a single RCT (n=185) comparing the effect of a 12-month vs 24-month interval between dental visits on dental caries in low-risk 3- to 5-year-old children with primary teeth and young adults, ages 16 to 20 years, with permanent teeth.¹ The outcomes of caries (ie, decayed, missing, filled surfaces increment) between the 12- and 24-month visits both in younger children (mean difference [MD]= -0.90; 95% confidence interval [CI], -1.96 to 0.16) and young adults (MD= -0.86; 95% CI, -1.75 to 0.03) did not differ.

Gingivitis: Not an issue when visits were delayed in healthy adults

Another systematic review (3 RCTs; N=836) evaluated the benefits associated with scaling and polishing in the prevention of gingivitis (primary outcome measure).² One RCT (n=207) compared scaling and polishing at 6- and 12-month intervals to no treatment for 24 months in adults with healthy dental histories. There was no difference in the percentage of index teeth with bleeding in the 6-month or 12-month treatment groups compared to the group that received no treatment for 24 months (MD= -2%; 95% CI, -10% to 6% and MD= -1%; 95% CI, -9% to 7%, respectively).

2 visits/year prevents tooth loss in high-risk patients

A retrospective cohort study (N=5117) using 16 years of data evaluated the association between one or 2 preventive dental visits per year and tooth extraction events in adults at low risk and those at high risk for progressive periodontitis.³ Those at high risk had at least one of the following risk factors: smoking, diabetes, or interleukin-1 genotype. Low-risk patients had no difference in tooth loss with one visit compared to 2 visits annually (absolute risk reduction [ARR]=2.6%; 95% CI, 0.5%-5.8%; P=.092); however, high-risk patients had fewer events with 2 annual visits (number needed to treat [NNT]=19; ARR 5.2%; 95% CI, 1.8%-8.4%; P=.002).

Visits before age 3 likely benefit only those at high risk

A systematic review of 4 retrospective cohort studies (N=77,291) analyzed the impact of early preventive dental visits (EPDV) on the frequency of future preventive and non-preventive dental visits and related expenditures using data from insurance claims and a kindergarten state dental registry.⁴ One study (n=11,394) used dental disease status at kindergarten (defined as the count of decayed, missing [molar teeth only], and filled primary teeth) as an outcome measure. Children who received EPDV before age 24 months had a comparable number of caries to those who had EPDV at 24 to 36 months. The authors concluded that EPDV before age 3 years is likely to benefit only children at high risk, and that evidence for a first dental visit by age one year is weak.

RECOMMENDATIONS

The National Institute for Health and Care Excellence recommends preventive dental visit intervals based on individual risk.

The National Institute for Health and Care Excellence recommends preventive dental visit intervals based on individual risk (12 months as the longest interval under age 18 years and 24 months as the longest interval for those 18 years and older at low risk).⁵ The American Dental Association recommends preventive dental visits at intervals determined by individual risk.⁶ The American Academy of Pediatric Dentistry recommends a first exam by age one year and preventive dental visits every 6 months through adolescence or as indicated by individual risk.⁷ The US Preventive Services Task Force states there is insufficient evidence to recommend routine dental screening by primary care physicians in children up to age 5 years.⁸

A 74-year-old Caucasian man presented to the hospital with intractable back and chest pain, a diffuse skin rash, and altered mental status. He said that 2 days ago, he’d gone to a different local hospital for treatment of back pain and a headache that had begun 3 days earlier. He was treated with intravenous hydromorphone and sent home with a prescription for meperidine. He said that several hours after being treated with the hydromorphone, the rash developed on his head and then spread to his trunk and upper extremities.

On physical examination, the patient was afebrile. He had numerous erythematous papules and vesicles in various stages of development on his scalp, face, neck, chest (FIGURE), abdomen, back, upper extremities, and groin. The lesions continued to spread and eventually involved his posterior oropharynx. The patient also developed conjunctivitis.

Laboratory findings included a white blood cell count of 4000/mcL (normal: 4500-11,000/mcL) with 65.9% segmented neutrophils (normal: 40%-60%), and 16.7% lymphocytes (normal: 20%-40%). Lab tests also revealed an aspartate aminotransferase level of 263 U/L (normal: 10-40 U/L), alanine aminotransferase of 236 U/L (normal: 7-56 U/L), and lactate dehydrogenase of 628 U/L (normal: 140-280 U/L).

The patient’s medical history was significant for hypertension, osteoarthritis, and IgG-kappa multiple myeloma, which had been treated with multiple chemotherapy regimens that included lenalidomide. Five years earlier, he’d undergone an autologous bone marrow transplant (BMT). At the time of presentation, the patient was being treated with daratumumab; he received his most recent treatment approximately one month earlier. Other medications included amlodipine, esomeprazole, and escitalopram.

WHAT IS YOUR DIAGNOSIS?
HOW WOULD YOU TREAT THIS PATIENT?

Diagnosis: Disseminated varicella-zoster virus infection

Because of the patient’s immunocompromised state, his presentation with altered mental status and diffuse rash was concerning. On hospital Day 2, a sample was taken from one of his skin lesions. Polymerase chain reaction (PCR) detected varicella-zoster virus (VZV), and we diagnosed disseminated VZV infection. On hospital Day 3, we performed a lumbar puncture because of worsening confusion and discovered that the cerebrospinal fluid was also positive for VZV.

About half of all patients who receive bone marrow transplants develop varicella-zoster virus infection within 5 years.

Disseminated VZV is the most common cause of late infection in patients who have received an allogenic BMT; it is usually due to reactivation of the virus.¹ In one study of 1186 patients who underwent BMT, 52% developed VZV infection within 5 years.² Disseminated VZV may also involve visceral organs, causing pneumonitis, pancreatitis, hepatitis, or encephalitis. Mortality rates for disseminated VZV are as high as 50%.³ Because of this, physicians should be vigilant when patients who have received a BMT present with a rash and signs of systemic involvement.

Two reliable tests. Even when lesions are classic for VZV, the diagnosis must be confirmed by laboratory testing. Real-time PCR assay is a rapid and highly sensitive test for diagnosing VZV.⁴ Another rapid test that can be used to confirm the clinical diagnosis of VZV is a direct fluorescent antibody assay, which is becoming more widely available.

In contrast, the sensitivity of viral culture for VZV has been reported to be as low as 20%.⁵ Viral culture also takes much longer and has a significantly lower yield compared with newer methods.⁶ A biopsy of skin lesions will reveal multinucleated giant cells, but cannot differentiate between herpes simplex virus (HSV) and VZV.⁷

These lesions can be mimicked

When a rash develops following the use of intravenous hydromorphone, as occurred with our patient, a drug reaction must be ruled out. A drug reaction can cause almost any skin manifestation and may present as vesicles, a macular rash, a papular rash, or diffuse erythema. In this case, drug rash was ruled out by the positive VZV PCR.

Viral exanthems can also present in a variety of ways. They may cause a macular, papular, or vesicular rash.

Prompt management is crucial

Prompt treatment of VZV with acyclovir improves outcomes, but death may still occur, even with early diagnosis.³ Immunocompromised patients with VZV should be closely monitored for secondary infections, which may rapidly progress and become fatal.⁸ The Centers for Disease Control and Prevention recommends both airborne and contact precautions for patients with disseminated VZV until all lesions are dry and crusted.⁹

While the live zoster vaccine is approved for prevention of shingles in patients <60 years of age, it is contraindicated in patients with a history of primary or acquired immunodeficiency states including leukemia, lymphoma, or other malignant neoplasms affecting bone marrow.

Our patient. On admission, he was treated with intravenous (IV) acyclovir 10 mg/kg TID; IV vancomycin 15 mg/kg every 12 hours; and IV ceftriaxone 2 g/d. Slowly, his mental status returned to baseline, and his rash and conjunctivitis resolved. We discharged him on hospital Day 12. He was transitioned to oral valacyclovir 1000 mg TID. Including both inpatient and outpatient treatment, the patient received 3 weeks (total) of acyclovir/valacyclovir therapy.

CORRESPONDENCE
Caitlyn T. Reed, MD, School of Medicine, University of Mississippi Medical Center, 2500 North State Street, Jackson, MS 39216; [email protected].

A 74-year-old Caucasian man presented to the hospital with intractable back and chest pain, a diffuse skin rash, and altered mental status. He said that 2 days ago, he’d gone to a different local hospital for treatment of back pain and a headache that had begun 3 days earlier. He was treated with intravenous hydromorphone and sent home with a prescription for meperidine. He said that several hours after being treated with the hydromorphone, the rash developed on his head and then spread to his trunk and upper extremities.

On physical examination, the patient was afebrile. He had numerous erythematous papules and vesicles in various stages of development on his scalp, face, neck, chest (FIGURE), abdomen, back, upper extremities, and groin. The lesions continued to spread and eventually involved his posterior oropharynx. The patient also developed conjunctivitis.

Laboratory findings included a white blood cell count of 4000/mcL (normal: 4500-11,000/mcL) with 65.9% segmented neutrophils (normal: 40%-60%), and 16.7% lymphocytes (normal: 20%-40%). Lab tests also revealed an aspartate aminotransferase level of 263 U/L (normal: 10-40 U/L), alanine aminotransferase of 236 U/L (normal: 7-56 U/L), and lactate dehydrogenase of 628 U/L (normal: 140-280 U/L).

The patient’s medical history was significant for hypertension, osteoarthritis, and IgG-kappa multiple myeloma, which had been treated with multiple chemotherapy regimens that included lenalidomide. Five years earlier, he’d undergone an autologous bone marrow transplant (BMT). At the time of presentation, the patient was being treated with daratumumab; he received his most recent treatment approximately one month earlier. Other medications included amlodipine, esomeprazole, and escitalopram.

WHAT IS YOUR DIAGNOSIS?
HOW WOULD YOU TREAT THIS PATIENT?

Diagnosis: Disseminated varicella-zoster virus infection

Because of the patient’s immunocompromised state, his presentation with altered mental status and diffuse rash was concerning. On hospital Day 2, a sample was taken from one of his skin lesions. Polymerase chain reaction (PCR) detected varicella-zoster virus (VZV), and we diagnosed disseminated VZV infection. On hospital Day 3, we performed a lumbar puncture because of worsening confusion and discovered that the cerebrospinal fluid was also positive for VZV.

About half of all patients who receive bone marrow transplants develop varicella-zoster virus infection within 5 years.

Disseminated VZV is the most common cause of late infection in patients who have received an allogenic BMT; it is usually due to reactivation of the virus.¹ In one study of 1186 patients who underwent BMT, 52% developed VZV infection within 5 years.² Disseminated VZV may also involve visceral organs, causing pneumonitis, pancreatitis, hepatitis, or encephalitis. Mortality rates for disseminated VZV are as high as 50%.³ Because of this, physicians should be vigilant when patients who have received a BMT present with a rash and signs of systemic involvement.

Two reliable tests. Even when lesions are classic for VZV, the diagnosis must be confirmed by laboratory testing. Real-time PCR assay is a rapid and highly sensitive test for diagnosing VZV.⁴ Another rapid test that can be used to confirm the clinical diagnosis of VZV is a direct fluorescent antibody assay, which is becoming more widely available.

In contrast, the sensitivity of viral culture for VZV has been reported to be as low as 20%.⁵ Viral culture also takes much longer and has a significantly lower yield compared with newer methods.⁶ A biopsy of skin lesions will reveal multinucleated giant cells, but cannot differentiate between herpes simplex virus (HSV) and VZV.⁷

These lesions can be mimicked

When a rash develops following the use of intravenous hydromorphone, as occurred with our patient, a drug reaction must be ruled out. A drug reaction can cause almost any skin manifestation and may present as vesicles, a macular rash, a papular rash, or diffuse erythema. In this case, drug rash was ruled out by the positive VZV PCR.

Viral exanthems can also present in a variety of ways. They may cause a macular, papular, or vesicular rash.

Prompt management is crucial

Prompt treatment of VZV with acyclovir improves outcomes, but death may still occur, even with early diagnosis.³ Immunocompromised patients with VZV should be closely monitored for secondary infections, which may rapidly progress and become fatal.⁸ The Centers for Disease Control and Prevention recommends both airborne and contact precautions for patients with disseminated VZV until all lesions are dry and crusted.⁹

While the live zoster vaccine is approved for prevention of shingles in patients <60 years of age, it is contraindicated in patients with a history of primary or acquired immunodeficiency states including leukemia, lymphoma, or other malignant neoplasms affecting bone marrow.

Our patient. On admission, he was treated with intravenous (IV) acyclovir 10 mg/kg TID; IV vancomycin 15 mg/kg every 12 hours; and IV ceftriaxone 2 g/d. Slowly, his mental status returned to baseline, and his rash and conjunctivitis resolved. We discharged him on hospital Day 12. He was transitioned to oral valacyclovir 1000 mg TID. Including both inpatient and outpatient treatment, the patient received 3 weeks (total) of acyclovir/valacyclovir therapy.

CORRESPONDENCE
Caitlyn T. Reed, MD, School of Medicine, University of Mississippi Medical Center, 2500 North State Street, Jackson, MS 39216; [email protected].

A 74-year-old Caucasian man presented to the hospital with intractable back and chest pain, a diffuse skin rash, and altered mental status. He said that 2 days ago, he’d gone to a different local hospital for treatment of back pain and a headache that had begun 3 days earlier. He was treated with intravenous hydromorphone and sent home with a prescription for meperidine. He said that several hours after being treated with the hydromorphone, the rash developed on his head and then spread to his trunk and upper extremities.

On physical examination, the patient was afebrile. He had numerous erythematous papules and vesicles in various stages of development on his scalp, face, neck, chest (FIGURE), abdomen, back, upper extremities, and groin. The lesions continued to spread and eventually involved his posterior oropharynx. The patient also developed conjunctivitis.

Laboratory findings included a white blood cell count of 4000/mcL (normal: 4500-11,000/mcL) with 65.9% segmented neutrophils (normal: 40%-60%), and 16.7% lymphocytes (normal: 20%-40%). Lab tests also revealed an aspartate aminotransferase level of 263 U/L (normal: 10-40 U/L), alanine aminotransferase of 236 U/L (normal: 7-56 U/L), and lactate dehydrogenase of 628 U/L (normal: 140-280 U/L).

The patient’s medical history was significant for hypertension, osteoarthritis, and IgG-kappa multiple myeloma, which had been treated with multiple chemotherapy regimens that included lenalidomide. Five years earlier, he’d undergone an autologous bone marrow transplant (BMT). At the time of presentation, the patient was being treated with daratumumab; he received his most recent treatment approximately one month earlier. Other medications included amlodipine, esomeprazole, and escitalopram.

WHAT IS YOUR DIAGNOSIS?
HOW WOULD YOU TREAT THIS PATIENT?

Diagnosis: Disseminated varicella-zoster virus infection

Because of the patient’s immunocompromised state, his presentation with altered mental status and diffuse rash was concerning. On hospital Day 2, a sample was taken from one of his skin lesions. Polymerase chain reaction (PCR) detected varicella-zoster virus (VZV), and we diagnosed disseminated VZV infection. On hospital Day 3, we performed a lumbar puncture because of worsening confusion and discovered that the cerebrospinal fluid was also positive for VZV.

About half of all patients who receive bone marrow transplants develop varicella-zoster virus infection within 5 years.

Disseminated VZV is the most common cause of late infection in patients who have received an allogenic BMT; it is usually due to reactivation of the virus.¹ In one study of 1186 patients who underwent BMT, 52% developed VZV infection within 5 years.² Disseminated VZV may also involve visceral organs, causing pneumonitis, pancreatitis, hepatitis, or encephalitis. Mortality rates for disseminated VZV are as high as 50%.³ Because of this, physicians should be vigilant when patients who have received a BMT present with a rash and signs of systemic involvement.

Two reliable tests. Even when lesions are classic for VZV, the diagnosis must be confirmed by laboratory testing. Real-time PCR assay is a rapid and highly sensitive test for diagnosing VZV.⁴ Another rapid test that can be used to confirm the clinical diagnosis of VZV is a direct fluorescent antibody assay, which is becoming more widely available.

In contrast, the sensitivity of viral culture for VZV has been reported to be as low as 20%.⁵ Viral culture also takes much longer and has a significantly lower yield compared with newer methods.⁶ A biopsy of skin lesions will reveal multinucleated giant cells, but cannot differentiate between herpes simplex virus (HSV) and VZV.⁷

These lesions can be mimicked

When a rash develops following the use of intravenous hydromorphone, as occurred with our patient, a drug reaction must be ruled out. A drug reaction can cause almost any skin manifestation and may present as vesicles, a macular rash, a papular rash, or diffuse erythema. In this case, drug rash was ruled out by the positive VZV PCR.

Viral exanthems can also present in a variety of ways. They may cause a macular, papular, or vesicular rash.

Prompt management is crucial

Prompt treatment of VZV with acyclovir improves outcomes, but death may still occur, even with early diagnosis.³ Immunocompromised patients with VZV should be closely monitored for secondary infections, which may rapidly progress and become fatal.⁸ The Centers for Disease Control and Prevention recommends both airborne and contact precautions for patients with disseminated VZV until all lesions are dry and crusted.⁹

While the live zoster vaccine is approved for prevention of shingles in patients <60 years of age, it is contraindicated in patients with a history of primary or acquired immunodeficiency states including leukemia, lymphoma, or other malignant neoplasms affecting bone marrow.

Our patient. On admission, he was treated with intravenous (IV) acyclovir 10 mg/kg TID; IV vancomycin 15 mg/kg every 12 hours; and IV ceftriaxone 2 g/d. Slowly, his mental status returned to baseline, and his rash and conjunctivitis resolved. We discharged him on hospital Day 12. He was transitioned to oral valacyclovir 1000 mg TID. Including both inpatient and outpatient treatment, the patient received 3 weeks (total) of acyclovir/valacyclovir therapy.

CORRESPONDENCE
Caitlyn T. Reed, MD, School of Medicine, University of Mississippi Medical Center, 2500 North State Street, Jackson, MS 39216; [email protected].

A 15-month-old black male was brought to the pediatric emergency department by his grandmother because she was concerned about his 2 swollen big toes. The patient’s grandmother said that the swelling began 36 hours prior and that her grandson’s big toes had continued to increase in size. She denied trauma, bites, or unusual exposures and said that although her grandson had been fussier than usual that day, he was eating and drinking normally and had normal urine output.

The patient had a history of developmental delay, but was otherwise healthy. He had no rashes, and there was no recent history of vomiting, diarrhea, difficulty breathing, or fever.

Examination of the patient’s skin revealed diffuse edema and erythema of the bilateral great toes (FIGURE 1A), with large overlying bullae extending from the dorsal surface of the base of the great toes around to the plantar (volar) surface of the foot (FIGURE 1B). The bullae on the plantar surface were approximately 4 cm long, extending from the tip of the toes proximally to the region of the head of the first metatarsal.

The patient’s vital signs were notable for a rectal temperature of 100.2° F and a heart rate of 180 beats per minute.

Initial lab tests included a complete blood count (CBC), blood cultures, and urinalysis with urine culture. The CBC revealed a white blood count of 27,000/mcL (normal: 6000-17,500/mcL). Both wound culture and herpes simplex viral culture were negative. An intranasal surveillance culture for methicillin-resistant Staphylococcus aureus (MRSA) was also negative.

Given the patient’s fever and leukocytosis, a 100-mg dose of intravenous clindamycin was administered.

WHAT IS YOUR DIAGNOSIS?
HOW WOULD YOU TREAT THIS PATIENT?

Diagnosis: Blistering distal dactylitis

We made a clinical diagnosis of blistering distal dactylitis (BDD), a condition typically caused by infection with Gram-positive bacteria. BDD is generally described as a localized infection of the volar fat pads of one or more fingers. The infection may also occur more proximally on the hand or involve the thumbs or toes.¹

Who’s at risk? BDD occurs among children ages 2 to 16 years, although it has been reported in infants as young as 6 months and in adults. No cases have occurred among the elderly.^2-7

The most common etiologic agents are group A beta-hemolytic Streptococci. Less commonly reported agents include Staphylococcus aureus, S. epidermidis, group B Streptococci, and MRSA.^1,6,8 The presence of multiple bullae may be predictive of infection with S. aureus.⁹

A clinical diagnosis

While blistering distal dactylitis typically affects the volar fat pads of the fingers, it may also occur more proximally on the hand or involve the thumbs or toes.

Diagnosis is usually made on clinical grounds based on the presence of large, tense, superficial, and typically painful bullae, the base of which may be erythematous. Culture of the blister fluid and the base of an unroofed blister may confirm the presence of a Streptococcus or Staphylococcus species.

Lab tests are typically not required to confirm a diagnosis of BDD. However, wound cultures of blister fluid, rapid antigen testing for group A beta-hemolytic Streptococci, and viral culture or polymerase chain reaction testing for herpes simplex virus may be considered.

Rule these conditions out

Lesions similar to those seen with BDD can be caused by the following infections and irritants:^4,5,8

Herpetic whitlow is caused by a herpes simplex virus infection. It presents as a cluster of painful vesicles or ulcers with an erythematous base on the distal part of a finger or toe.

Bullous impetigo is the result of a staphylococcal infection, which produces an epidermolytic toxin leading to bulla formation. Lesions may occur anywhere on the body but are most common on the face.

Irritant or allergic contact dermatitis results from an external topical exposure and is typically localized to the area of contact. The reaction is an eczematous eruption that may include bullae.

Treatment is typically empiric

Treatment of BDD includes wound care with wet-to-dry saline dressings, incision and drainage of the bulla(e), and a systemic beta-lactamase-resistant antibiotic. Topical antibiotics alone are not recommended.⁷

Our patient was transitioned from intravenous to oral clindamycin, 100 mg every 8 hours, and the bullae were incised and drained. His leukocytosis resolved within 24 hours, and he continued to do well. At follow-up one week later, the patient’s blisters were healing well, and he was playful and eating and drinking normally.

CORRESPONDENCE
C. Randall Clinch, DO, MS, Wake Forest University School of Medicine, 1 Medical Center Blvd, Winston-Salem, NC 27157; [email protected].

A 15-month-old black male was brought to the pediatric emergency department by his grandmother because she was concerned about his 2 swollen big toes. The patient’s grandmother said that the swelling began 36 hours prior and that her grandson’s big toes had continued to increase in size. She denied trauma, bites, or unusual exposures and said that although her grandson had been fussier than usual that day, he was eating and drinking normally and had normal urine output.

The patient had a history of developmental delay, but was otherwise healthy. He had no rashes, and there was no recent history of vomiting, diarrhea, difficulty breathing, or fever.

Examination of the patient’s skin revealed diffuse edema and erythema of the bilateral great toes (FIGURE 1A), with large overlying bullae extending from the dorsal surface of the base of the great toes around to the plantar (volar) surface of the foot (FIGURE 1B). The bullae on the plantar surface were approximately 4 cm long, extending from the tip of the toes proximally to the region of the head of the first metatarsal.

The patient’s vital signs were notable for a rectal temperature of 100.2° F and a heart rate of 180 beats per minute.

Initial lab tests included a complete blood count (CBC), blood cultures, and urinalysis with urine culture. The CBC revealed a white blood count of 27,000/mcL (normal: 6000-17,500/mcL). Both wound culture and herpes simplex viral culture were negative. An intranasal surveillance culture for methicillin-resistant Staphylococcus aureus (MRSA) was also negative.

Given the patient’s fever and leukocytosis, a 100-mg dose of intravenous clindamycin was administered.

WHAT IS YOUR DIAGNOSIS?
HOW WOULD YOU TREAT THIS PATIENT?

Diagnosis: Blistering distal dactylitis

We made a clinical diagnosis of blistering distal dactylitis (BDD), a condition typically caused by infection with Gram-positive bacteria. BDD is generally described as a localized infection of the volar fat pads of one or more fingers. The infection may also occur more proximally on the hand or involve the thumbs or toes.¹

Who’s at risk? BDD occurs among children ages 2 to 16 years, although it has been reported in infants as young as 6 months and in adults. No cases have occurred among the elderly.^2-7

The most common etiologic agents are group A beta-hemolytic Streptococci. Less commonly reported agents include Staphylococcus aureus, S. epidermidis, group B Streptococci, and MRSA.^1,6,8 The presence of multiple bullae may be predictive of infection with S. aureus.⁹

A clinical diagnosis

While blistering distal dactylitis typically affects the volar fat pads of the fingers, it may also occur more proximally on the hand or involve the thumbs or toes.

Diagnosis is usually made on clinical grounds based on the presence of large, tense, superficial, and typically painful bullae, the base of which may be erythematous. Culture of the blister fluid and the base of an unroofed blister may confirm the presence of a Streptococcus or Staphylococcus species.

Lab tests are typically not required to confirm a diagnosis of BDD. However, wound cultures of blister fluid, rapid antigen testing for group A beta-hemolytic Streptococci, and viral culture or polymerase chain reaction testing for herpes simplex virus may be considered.

Rule these conditions out

Lesions similar to those seen with BDD can be caused by the following infections and irritants:^4,5,8

Herpetic whitlow is caused by a herpes simplex virus infection. It presents as a cluster of painful vesicles or ulcers with an erythematous base on the distal part of a finger or toe.

Bullous impetigo is the result of a staphylococcal infection, which produces an epidermolytic toxin leading to bulla formation. Lesions may occur anywhere on the body but are most common on the face.

Irritant or allergic contact dermatitis results from an external topical exposure and is typically localized to the area of contact. The reaction is an eczematous eruption that may include bullae.

Treatment is typically empiric

Treatment of BDD includes wound care with wet-to-dry saline dressings, incision and drainage of the bulla(e), and a systemic beta-lactamase-resistant antibiotic. Topical antibiotics alone are not recommended.⁷

Our patient was transitioned from intravenous to oral clindamycin, 100 mg every 8 hours, and the bullae were incised and drained. His leukocytosis resolved within 24 hours, and he continued to do well. At follow-up one week later, the patient’s blisters were healing well, and he was playful and eating and drinking normally.

CORRESPONDENCE
C. Randall Clinch, DO, MS, Wake Forest University School of Medicine, 1 Medical Center Blvd, Winston-Salem, NC 27157; [email protected].

A 15-month-old black male was brought to the pediatric emergency department by his grandmother because she was concerned about his 2 swollen big toes. The patient’s grandmother said that the swelling began 36 hours prior and that her grandson’s big toes had continued to increase in size. She denied trauma, bites, or unusual exposures and said that although her grandson had been fussier than usual that day, he was eating and drinking normally and had normal urine output.

The patient had a history of developmental delay, but was otherwise healthy. He had no rashes, and there was no recent history of vomiting, diarrhea, difficulty breathing, or fever.

Examination of the patient’s skin revealed diffuse edema and erythema of the bilateral great toes (FIGURE 1A), with large overlying bullae extending from the dorsal surface of the base of the great toes around to the plantar (volar) surface of the foot (FIGURE 1B). The bullae on the plantar surface were approximately 4 cm long, extending from the tip of the toes proximally to the region of the head of the first metatarsal.

The patient’s vital signs were notable for a rectal temperature of 100.2° F and a heart rate of 180 beats per minute.

Initial lab tests included a complete blood count (CBC), blood cultures, and urinalysis with urine culture. The CBC revealed a white blood count of 27,000/mcL (normal: 6000-17,500/mcL). Both wound culture and herpes simplex viral culture were negative. An intranasal surveillance culture for methicillin-resistant Staphylococcus aureus (MRSA) was also negative.

Given the patient’s fever and leukocytosis, a 100-mg dose of intravenous clindamycin was administered.

WHAT IS YOUR DIAGNOSIS?
HOW WOULD YOU TREAT THIS PATIENT?

Diagnosis: Blistering distal dactylitis

We made a clinical diagnosis of blistering distal dactylitis (BDD), a condition typically caused by infection with Gram-positive bacteria. BDD is generally described as a localized infection of the volar fat pads of one or more fingers. The infection may also occur more proximally on the hand or involve the thumbs or toes.¹

Who’s at risk? BDD occurs among children ages 2 to 16 years, although it has been reported in infants as young as 6 months and in adults. No cases have occurred among the elderly.^2-7

The most common etiologic agents are group A beta-hemolytic Streptococci. Less commonly reported agents include Staphylococcus aureus, S. epidermidis, group B Streptococci, and MRSA.^1,6,8 The presence of multiple bullae may be predictive of infection with S. aureus.⁹

A clinical diagnosis

While blistering distal dactylitis typically affects the volar fat pads of the fingers, it may also occur more proximally on the hand or involve the thumbs or toes.

Diagnosis is usually made on clinical grounds based on the presence of large, tense, superficial, and typically painful bullae, the base of which may be erythematous. Culture of the blister fluid and the base of an unroofed blister may confirm the presence of a Streptococcus or Staphylococcus species.

Lab tests are typically not required to confirm a diagnosis of BDD. However, wound cultures of blister fluid, rapid antigen testing for group A beta-hemolytic Streptococci, and viral culture or polymerase chain reaction testing for herpes simplex virus may be considered.

Rule these conditions out

Lesions similar to those seen with BDD can be caused by the following infections and irritants:^4,5,8

Herpetic whitlow is caused by a herpes simplex virus infection. It presents as a cluster of painful vesicles or ulcers with an erythematous base on the distal part of a finger or toe.

Bullous impetigo is the result of a staphylococcal infection, which produces an epidermolytic toxin leading to bulla formation. Lesions may occur anywhere on the body but are most common on the face.

Irritant or allergic contact dermatitis results from an external topical exposure and is typically localized to the area of contact. The reaction is an eczematous eruption that may include bullae.

Treatment is typically empiric

Treatment of BDD includes wound care with wet-to-dry saline dressings, incision and drainage of the bulla(e), and a systemic beta-lactamase-resistant antibiotic. Topical antibiotics alone are not recommended.⁷

Our patient was transitioned from intravenous to oral clindamycin, 100 mg every 8 hours, and the bullae were incised and drained. His leukocytosis resolved within 24 hours, and he continued to do well. At follow-up one week later, the patient’s blisters were healing well, and he was playful and eating and drinking normally.

CORRESPONDENCE
C. Randall Clinch, DO, MS, Wake Forest University School of Medicine, 1 Medical Center Blvd, Winston-Salem, NC 27157; [email protected].

THE CASE

A 14-year-old Caucasian boy presented to our clinic with a complaint of left anterior hip pain. The patient had been running during a flag football match when he suddenly developed a sharp, stabbing pain in his left hip. He said he felt a “pop” in his left groin while his left foot was planted and he was cutting to the right. The patient said this was followed by worsening pain with ambulation and hip flexion.

The patient had considerable difficulty walking into the exam room. On physical examination, he had significant tenderness to palpation along the anteromedial thigh and groin. The patient’s strength was 1/5 with left hip flexion. There was apparent muscle firing, but no significant leg movement. He had full passive range of motion and there was no soft-tissue swelling, erythema, or other integumentary changes.

THE DIAGNOSIS

Plain radiographs revealed a lesser trochanter avulsion fracture with a 2-cm displacement (FIGURE 1).

DISCUSSION

Pelvic and proximal femur avulsion fractures tend to occur during the second decade of life.^1,2 They’re more frequently seen in boys and adolescent athletes, especially those involved in soccer and gymnastics.^3,4

Anterior superior iliac spine (ASIS), ischial tuberosity (IT), and anterior inferior iliac spine (AIIS) avulsion fractures are more prevalent,⁴ while lesser trochanter avulsion fractures are more rare. In one review of 1126 children with femoral neck and proximal 1/3 femoral shaft fractures, only 3 of them had lesser trochanter avulsion fractures.⁵

Clinical presentation. Presenting symptoms of lesser trochanter avulsion fractures can be vague, but are usually localized to the groin and medial hip region. Patients will demonstrate pain and weakness with hip flexion.^3,6 There may be signs of inflammation, tenderness, and ecchymosis near the site of injury.

On physical exam, a positive Ludloff sign helps localize the injury to the iliopsoas muscle, which inserts at the lesser trochanter and is involved in hip flexion.^3,6,7 The Ludloff test is performed by flexing the patient’s hip while he/she is in a seated position.

BIOMECHANICS OF AVULSION FRACTURES

Perhaps surprisingly, the majority of avulsion injuries in children and adolescents are the result of non-contact athletic movement and indirect trauma.⁴ In children, muscles and tendons are often stronger than their bones,⁷ and physes—structurally weak regions—are particularly predisposed to fractures.^2,4,6

The mechanism of injury in children and adolescents is commonly a sudden, forceful contraction of the iliopsoas muscle.^6,7 While similar movement in adults will produce tendon sprains and muscle strains, children often experience a complete avulsion fracture.⁷ So uncommon are these fractures among adults that an adult patient presenting with one should receive further work-up for underlying pathology such as malignancy.^8,9

While other hip and femur avulsion fractures in children and adolescents involve different muscle groups, the etiologic mechanism—forceful muscle contraction—is usually the same.^2,4,7 IT injuries are often seen with sudden, aggressive lengthening of the hamstring muscles, whereas injuries to the ASIS and AIIS are the result of abrupt eccentric contraction of hip extensor muscles while the knee is flexed.⁴

DIFFERENTIAL DIAGNOSIS

There are several entities that can mimic a lesser trochanter avulsion fracture including Legg-Calve-Perthes disease (LCPD), slipped capital femoral epiphysis (SCFE), snapping hip with the iliofemoral ligament, iliopsoas tendonitis, referred pain from the gastrointestinal region, and a genito-urologic etiology.^1,7,10 The work-up and treatment for these alternative diagnoses are quite different and can lead to unnecessary testing.

Diagnostic studies. Physical exam findings of severe pain and reduced strength are clear indications for obtaining baseline imaging. Baseline radiographs are key to the diagnosis of avulsion fractures. They help differentiate between more benign fractures, such as a nondisplaced avulsion fracture, and more substantial conditions, such as LCPD and SCFE, which require significantly different approaches to treatment and follow-up.^1,7

Anteroposterior, oblique, and axial views of the pelvis all assist in assessing avulsion fractures radiographically.^3,4,7 In the event that an avulsion fracture is not radiographically visible, but is still suspected, additional imaging should be obtained.¹⁰ A computerized tomography (CT) scan is an appropriate follow-up, given its meticulous detail of bony anatomy.^3,10 Alternatively, if physes have yet to ossify or there are concerns about soft tissue injury, magnetic resonance imaging can be useful.^3,7,10

MANAGEMENT

The majority of lesser trochanter avulsion fractures are managed conservatively with rest, nonsteroidal anti-inflammatory drugs (NSAIDs), and physical therapy. Patients are often placed on non-weight bearing activity for up to 6 weeks while the fracture repairs and forms a new union.⁷ Current management strategies have moved away from immobilization with splints and braces.

In rare instances when the fragment is displaced >2 cm, or there is inadequate healing or pain relief after 3 months of supportive care, surgery may be required.¹ With appropriate diagnosis and medical care, the injured athlete should fully recover with no impairment or chronic pain.²

Our patient was placed on non-weight-bearing activity and treated with NSAIDs and acetaminophen. We advanced him to weight-bearing activities 4 weeks after injury. After 8 weeks of conservative management, he returned to competitive play with no further complications (FIGURE 2).

THE TAKEAWAY

Pelvic and proximal femur avulsion fractures occur more often in child and adolescent athletes. As this population becomes increasingly competitive in athletics, the risk of injury increases. Infrequent fractures such as lesser trochanter avulsion fractures may become more common, as well. The majority of avulsion fractures don’t require surgical intervention, but it’s important to obtain baseline radiographs to rule out other injuries or pathologies that may lead to poor prognoses if they are left untreated.

THE CASE

A 14-year-old Caucasian boy presented to our clinic with a complaint of left anterior hip pain. The patient had been running during a flag football match when he suddenly developed a sharp, stabbing pain in his left hip. He said he felt a “pop” in his left groin while his left foot was planted and he was cutting to the right. The patient said this was followed by worsening pain with ambulation and hip flexion.

The patient had considerable difficulty walking into the exam room. On physical examination, he had significant tenderness to palpation along the anteromedial thigh and groin. The patient’s strength was 1/5 with left hip flexion. There was apparent muscle firing, but no significant leg movement. He had full passive range of motion and there was no soft-tissue swelling, erythema, or other integumentary changes.

THE DIAGNOSIS

Plain radiographs revealed a lesser trochanter avulsion fracture with a 2-cm displacement (FIGURE 1).

DISCUSSION

Pelvic and proximal femur avulsion fractures tend to occur during the second decade of life.^1,2 They’re more frequently seen in boys and adolescent athletes, especially those involved in soccer and gymnastics.^3,4

Anterior superior iliac spine (ASIS), ischial tuberosity (IT), and anterior inferior iliac spine (AIIS) avulsion fractures are more prevalent,⁴ while lesser trochanter avulsion fractures are more rare. In one review of 1126 children with femoral neck and proximal 1/3 femoral shaft fractures, only 3 of them had lesser trochanter avulsion fractures.⁵

Clinical presentation. Presenting symptoms of lesser trochanter avulsion fractures can be vague, but are usually localized to the groin and medial hip region. Patients will demonstrate pain and weakness with hip flexion.^3,6 There may be signs of inflammation, tenderness, and ecchymosis near the site of injury.

On physical exam, a positive Ludloff sign helps localize the injury to the iliopsoas muscle, which inserts at the lesser trochanter and is involved in hip flexion.^3,6,7 The Ludloff test is performed by flexing the patient’s hip while he/she is in a seated position.

BIOMECHANICS OF AVULSION FRACTURES

Perhaps surprisingly, the majority of avulsion injuries in children and adolescents are the result of non-contact athletic movement and indirect trauma.⁴ In children, muscles and tendons are often stronger than their bones,⁷ and physes—structurally weak regions—are particularly predisposed to fractures.^2,4,6

The mechanism of injury in children and adolescents is commonly a sudden, forceful contraction of the iliopsoas muscle.^6,7 While similar movement in adults will produce tendon sprains and muscle strains, children often experience a complete avulsion fracture.⁷ So uncommon are these fractures among adults that an adult patient presenting with one should receive further work-up for underlying pathology such as malignancy.^8,9

While other hip and femur avulsion fractures in children and adolescents involve different muscle groups, the etiologic mechanism—forceful muscle contraction—is usually the same.^2,4,7 IT injuries are often seen with sudden, aggressive lengthening of the hamstring muscles, whereas injuries to the ASIS and AIIS are the result of abrupt eccentric contraction of hip extensor muscles while the knee is flexed.⁴

DIFFERENTIAL DIAGNOSIS

There are several entities that can mimic a lesser trochanter avulsion fracture including Legg-Calve-Perthes disease (LCPD), slipped capital femoral epiphysis (SCFE), snapping hip with the iliofemoral ligament, iliopsoas tendonitis, referred pain from the gastrointestinal region, and a genito-urologic etiology.^1,7,10 The work-up and treatment for these alternative diagnoses are quite different and can lead to unnecessary testing.

Diagnostic studies. Physical exam findings of severe pain and reduced strength are clear indications for obtaining baseline imaging. Baseline radiographs are key to the diagnosis of avulsion fractures. They help differentiate between more benign fractures, such as a nondisplaced avulsion fracture, and more substantial conditions, such as LCPD and SCFE, which require significantly different approaches to treatment and follow-up.^1,7

Anteroposterior, oblique, and axial views of the pelvis all assist in assessing avulsion fractures radiographically.^3,4,7 In the event that an avulsion fracture is not radiographically visible, but is still suspected, additional imaging should be obtained.¹⁰ A computerized tomography (CT) scan is an appropriate follow-up, given its meticulous detail of bony anatomy.^3,10 Alternatively, if physes have yet to ossify or there are concerns about soft tissue injury, magnetic resonance imaging can be useful.^3,7,10

MANAGEMENT

The majority of lesser trochanter avulsion fractures are managed conservatively with rest, nonsteroidal anti-inflammatory drugs (NSAIDs), and physical therapy. Patients are often placed on non-weight bearing activity for up to 6 weeks while the fracture repairs and forms a new union.⁷ Current management strategies have moved away from immobilization with splints and braces.

In rare instances when the fragment is displaced >2 cm, or there is inadequate healing or pain relief after 3 months of supportive care, surgery may be required.¹ With appropriate diagnosis and medical care, the injured athlete should fully recover with no impairment or chronic pain.²

Our patient was placed on non-weight-bearing activity and treated with NSAIDs and acetaminophen. We advanced him to weight-bearing activities 4 weeks after injury. After 8 weeks of conservative management, he returned to competitive play with no further complications (FIGURE 2).

THE TAKEAWAY

Pelvic and proximal femur avulsion fractures occur more often in child and adolescent athletes. As this population becomes increasingly competitive in athletics, the risk of injury increases. Infrequent fractures such as lesser trochanter avulsion fractures may become more common, as well. The majority of avulsion fractures don’t require surgical intervention, but it’s important to obtain baseline radiographs to rule out other injuries or pathologies that may lead to poor prognoses if they are left untreated.

THE CASE

A 14-year-old Caucasian boy presented to our clinic with a complaint of left anterior hip pain. The patient had been running during a flag football match when he suddenly developed a sharp, stabbing pain in his left hip. He said he felt a “pop” in his left groin while his left foot was planted and he was cutting to the right. The patient said this was followed by worsening pain with ambulation and hip flexion.

The patient had considerable difficulty walking into the exam room. On physical examination, he had significant tenderness to palpation along the anteromedial thigh and groin. The patient’s strength was 1/5 with left hip flexion. There was apparent muscle firing, but no significant leg movement. He had full passive range of motion and there was no soft-tissue swelling, erythema, or other integumentary changes.

THE DIAGNOSIS

Plain radiographs revealed a lesser trochanter avulsion fracture with a 2-cm displacement (FIGURE 1).

DISCUSSION

Pelvic and proximal femur avulsion fractures tend to occur during the second decade of life.^1,2 They’re more frequently seen in boys and adolescent athletes, especially those involved in soccer and gymnastics.^3,4

Anterior superior iliac spine (ASIS), ischial tuberosity (IT), and anterior inferior iliac spine (AIIS) avulsion fractures are more prevalent,⁴ while lesser trochanter avulsion fractures are more rare. In one review of 1126 children with femoral neck and proximal 1/3 femoral shaft fractures, only 3 of them had lesser trochanter avulsion fractures.⁵

Clinical presentation. Presenting symptoms of lesser trochanter avulsion fractures can be vague, but are usually localized to the groin and medial hip region. Patients will demonstrate pain and weakness with hip flexion.^3,6 There may be signs of inflammation, tenderness, and ecchymosis near the site of injury.

On physical exam, a positive Ludloff sign helps localize the injury to the iliopsoas muscle, which inserts at the lesser trochanter and is involved in hip flexion.^3,6,7 The Ludloff test is performed by flexing the patient’s hip while he/she is in a seated position.

BIOMECHANICS OF AVULSION FRACTURES

Perhaps surprisingly, the majority of avulsion injuries in children and adolescents are the result of non-contact athletic movement and indirect trauma.⁴ In children, muscles and tendons are often stronger than their bones,⁷ and physes—structurally weak regions—are particularly predisposed to fractures.^2,4,6

The mechanism of injury in children and adolescents is commonly a sudden, forceful contraction of the iliopsoas muscle.^6,7 While similar movement in adults will produce tendon sprains and muscle strains, children often experience a complete avulsion fracture.⁷ So uncommon are these fractures among adults that an adult patient presenting with one should receive further work-up for underlying pathology such as malignancy.^8,9

While other hip and femur avulsion fractures in children and adolescents involve different muscle groups, the etiologic mechanism—forceful muscle contraction—is usually the same.^2,4,7 IT injuries are often seen with sudden, aggressive lengthening of the hamstring muscles, whereas injuries to the ASIS and AIIS are the result of abrupt eccentric contraction of hip extensor muscles while the knee is flexed.⁴

DIFFERENTIAL DIAGNOSIS

There are several entities that can mimic a lesser trochanter avulsion fracture including Legg-Calve-Perthes disease (LCPD), slipped capital femoral epiphysis (SCFE), snapping hip with the iliofemoral ligament, iliopsoas tendonitis, referred pain from the gastrointestinal region, and a genito-urologic etiology.^1,7,10 The work-up and treatment for these alternative diagnoses are quite different and can lead to unnecessary testing.

Diagnostic studies. Physical exam findings of severe pain and reduced strength are clear indications for obtaining baseline imaging. Baseline radiographs are key to the diagnosis of avulsion fractures. They help differentiate between more benign fractures, such as a nondisplaced avulsion fracture, and more substantial conditions, such as LCPD and SCFE, which require significantly different approaches to treatment and follow-up.^1,7

Anteroposterior, oblique, and axial views of the pelvis all assist in assessing avulsion fractures radiographically.^3,4,7 In the event that an avulsion fracture is not radiographically visible, but is still suspected, additional imaging should be obtained.¹⁰ A computerized tomography (CT) scan is an appropriate follow-up, given its meticulous detail of bony anatomy.^3,10 Alternatively, if physes have yet to ossify or there are concerns about soft tissue injury, magnetic resonance imaging can be useful.^3,7,10

MANAGEMENT

The majority of lesser trochanter avulsion fractures are managed conservatively with rest, nonsteroidal anti-inflammatory drugs (NSAIDs), and physical therapy. Patients are often placed on non-weight bearing activity for up to 6 weeks while the fracture repairs and forms a new union.⁷ Current management strategies have moved away from immobilization with splints and braces.

In rare instances when the fragment is displaced >2 cm, or there is inadequate healing or pain relief after 3 months of supportive care, surgery may be required.¹ With appropriate diagnosis and medical care, the injured athlete should fully recover with no impairment or chronic pain.²

Our patient was placed on non-weight-bearing activity and treated with NSAIDs and acetaminophen. We advanced him to weight-bearing activities 4 weeks after injury. After 8 weeks of conservative management, he returned to competitive play with no further complications (FIGURE 2).

THE TAKEAWAY

Pelvic and proximal femur avulsion fractures occur more often in child and adolescent athletes. As this population becomes increasingly competitive in athletics, the risk of injury increases. Infrequent fractures such as lesser trochanter avulsion fractures may become more common, as well. The majority of avulsion fractures don’t require surgical intervention, but it’s important to obtain baseline radiographs to rule out other injuries or pathologies that may lead to poor prognoses if they are left untreated.

There is a strong relationship between the number of medications taken and the likelihood of a potentially serious drug-drug interaction.^1,2 Drug interaction software programs can help alert prescribers to potential problems, but these programs sometimes fail to detect important interactions or generate so many clinically insignificant alerts that they become a nuisance.³ This review provides guidance about 5 clinically relevant drug interactions, including those that are common (TABLE 1^4-6)—and those that are less common, but no less important (TABLE 2^6-10).

1. Antiepileptics & contraceptives

Many antiepileptic medications decrease the efficacy of certain contraceptives

Contraception management in women with epilepsy is critical due to potential maternal and fetal complications. Many antiepileptic drugs (AEDs), including carbamazepine, ethosuximide, fosphenytoin, phenobarbital, phenytoin, primidone, topiramate, and valproate, are potentially teratogenic.¹¹ A retrospective, observational study of 115 women of childbearing age who had epilepsy and were seen at a neurology clinic found that 74% were not using documented contraception.¹¹ Of the minority of study participants using contraception, most were using oral contraceptives (OCs) that could potentially interact with AEDs.

CYP inducers. Estrogen and progesterone are metabolized by the cytochrome P450 3A4 enzyme. Some AEDs induce this enzyme, which can enhance the metabolism of OCs, thus reducing their efficacy.¹² It is not known, however, if this interaction results in increased pregnancy rates.¹³ Most newer AEDs (TABLE 3⁶) do not induce cytochrome P450 3A4 and, thus, do not appear to affect OC efficacy, and may be safer for women with seizure disorders.¹² While enzyme-inducing AEDs may decrease the efficacy of progesterone-only OCs and the morning-after pill,^12,14,15 progesterone-containing intrauterine devices (IUDs), long-acting progesterone injections, and non-hormonal contraceptive methods appear to be unaffected.^14-17

OCs and seizure frequency. There is no strong evidence that OCs affect seizure frequency in epileptic women, although changes in hormone levels during the menstrual cycle do affect seizure susceptibility.¹² Combination OCs decrease lamotrigine levels and, therefore, may increase the risk of seizures, but progesterone-only pills do not produce this effect.^12,16

Do guidelines exist? There are no specific evidence-based guidelines that pertain to the use of AEDs and contraception together, but some organizations have issued recommendations.

The American College of Obstetricians and Gynecologists recommends using a 30- to 35-mcg estrogen-containing OC rather than a lower dose in women taking an enzyme-inducing AED. The group also recommends using condoms with OCs or using IUDs.¹⁸

The American Academy of Neurology suggests that women taking OCs and enzyme-inducing AEDs use an OC containing at least 50 mcg estrogen.¹⁹

The National Institute for Health and Care Excellence recommends that women taking enzyme-inducing AEDs avoid progestin-only pills.²⁰

The Faculty of Sexual and Reproductive Healthcare agrees that enzyme-inducing drugs may decrease efficacy and recommend considering IUDs and injectable contraceptive methods.²¹

2. SSRIs & NSAIDs.

SSRIs increase the GI bleeding risk associated with NSAIDs alone

Nonsteroidal anti-inflammatory drugs (NSAIDs) and selective serotonin reuptake inhibitors (SSRIs) are commonly prescribed worldwide.^22,23 A well-established adverse effect of NSAIDs is gastrointestinal (GI) bleeding, and there is increasing evidence that concomitant use of an SSRI can further increase that risk through a variety of mechanisms.²³

SSRIs decrease platelet serotonin levels resulting in defective platelet aggregation and impaired hemostasis. Studies have also shown that SSRIs increase gastric acidity, which leads to increased risk of peptic ulcer disease and GI bleeding.²³ These mechanisms, combined with the inhibition of gastroprotective prostaglandin cyclooxygenase-1 and platelets by NSAIDs, further potentiate GI bleeding risk.²⁴

Patients at high risk for bleeding with concomitant SSRIs and NSAIDs include older patients, patients with other risk factors for GI bleeding (eg, chronic steroid use), and patients with a history of GI bleeding.²³

The evidence. A 2014 meta-analysis found that when SSRIs were used in combination with NSAIDs, the risk of GI bleeding was significantly increased, compared with SSRI monotherapy.²³

Case control studies found the risk of upper GI bleeding with SSRIs had a number needed to harm (NNH) of 3177 for a low-risk population and 881 for a high-risk population with an odds ratio (OR) of 1.66 (95% confidence interval [CI], 1.44-1.92; P<.00001).²³ When SSRIs were used in combination with NSAIDs, the NNH decreased to 645 for a low-risk population and 179 for a high-risk population (OR=4.25; 95% CI, 2.82-6.42; P<.0001).²³

Another meta-analysis found that the OR for bleeding risk increased to 6.33 (95% CI, 3.40-11.8; P<.00001; NNH=106) with concomitant use of NSAIDs and SSRIs, compared with 2.36 (95% CI, 1.44-3.85; P=.0006; NNH=411) for SSRI use alone.²⁵

The studies did not evaluate results based on the indication, dose, or duration of SSRI or NSAID treatment. If both an SSRI and an NSAID must be used, select a cyclooxygenase-2 selective NSAID at the lowest effective dose and consider the addition of a proton pump inhibitor to decrease the risk of a GI bleed.^23,26

3. Direct oral anticoagulants and antiepileptics

Don’t use DOACs in patients taking certain antiepileptic medications

Drug interactions with anticoagulants, such as warfarin, are well documented and have been publicized for years, but physicians must also be aware of the potential for interaction between the direct oral anticoagulants (DOACs) and AEDs.

Apixaban, rivaroxaban, and dabigatran appear to interact withthe AEDs carbamazepine, phenytoin, and phenobarbital.^27,28 These interactions occur due to AED induction of the CYP3A4 enzyme and effects on the P-glycoprotein (P-gp) efflux pump.^27,29 When taken together, the AED induces metabolism and elimination of the DOAC medication to occur more quickly than it would normally, resulting in subtherapeutic concentrations of the DOAC. This could theoretically result in a venous thromboembolic event or stroke.

A caveat. One thing to consider is that studies demonstrating interaction between the DOAC and AED drug classes have been performed in healthy volunteers, making it difficult to extrapolate how this interaction may increase the risk for thrombotic events in other patients.

Some studies demonstrated reductions in drug levels of up to 50% with strong CYP3A4 and P-glycoprotein inducers.³⁰ Common inducers include carbamazepine, rifampin, and St. John’s Wort.⁶ Patients taking such agents could theoretically have decreased exposure to the DOAC, resulting in an increase in thromboembolic risk.³¹

4. Statins & certain CYP inhibitors

Combining simvastatin with fibrates warrants extra attention

The efficacy of statin medications in the prevention of atherosclerotic cardiovascular disease (ASCVD) is clear. However, the clinical significance of many identified drug interactions involving statins is difficult to interpret. Interactions that cause increased serum concentrations of statins can increase the risk for liver enzyme elevations and skeletal muscle abnormalities (myalgias to rhabdomyolysis).³² Strong inhibitors of CYP3A4 (amiodarone, cyclosporine, ketoconazole, etc.) significantly increase concentrations of lovastatin, simvastatin, and atorvastatin. Pitavastatin, pravastatin, and rosuvastatin are not susceptible to any CYP-mediated drug interactions;³³ therefore, rosuvastatin (a high-intensity statin) is usually recommended over other statins for patients taking strong inhibitors of CYP3A4.

When to limit simvastatin. Doses of simvastatin should not exceed 10 mg/d when combined with diltiazem, dronedarone, or verapamil, and doses should not exceed 20 mg/d when used with amiodarone, amlodipine, or ranolazine.⁶ These recommendations are in response to results from the SEARCH (Study of the Effectiveness of Additional Reductions in cholesterol and homocysteine) trial, which found a higher incidence of myopathies and rhabdomyolysis in patients taking 80 mg of simvastatin compared with those taking 20-mg doses.³⁴ CYP3A4-inducing medications, especially diltiazem, were thought to also contribute to an increased risk.³⁴

Avoid gemfibrozil with statins. Using fibrates with statins is beneficial for some patients; however, gemfibrozil significantly interacts with statins by inhibiting CYP2C8 and organic anion transporting polypeptide 1B1 (OATP1B1).³³ The safer choice is fenofibrate because it does not interfere with statin metabolism and can be safely used in combination with statins.⁶

Enzyme-inducing antiepileptic drugs can enhance the metabolism of oral contraceptives, thus reducing their efficacy.

A retrospective review of the FDA Adverse Event Reporting System (AERS) database found that 88% of fibrate and statin combinations that resulted in rhabdomyolysis were associated with gemfibrozil/cerivastatin (cerivastatin is no longer available in the United States).³⁵

5. One serotonergic drug & another

Serotonin syndrome is associated with more than just SSRIs

Serotonin syndrome is a constellation of symptoms (hyperthermia, hyperreflexia, muscle clonus, tremor and altered mental status) caused by increases in serotonin levels in the central and peripheral nervous systems that can lead to mild or life-threatening complications such as seizures, muscle breakdown, or hyperthermia. Serotonin syndrome is most likely to occur within 24 hours after a dose increase, after starting a new medication that increases serotonin levels, or after a drug overdose.³⁶

SSRIs are the most commonly reported drug associated with serotonin syndrome; however, other medications (TABLE 4³⁷) may be responsible, especially when used in combination with agents that act on serotonin receptors or in patients with impaired metabolism of the drugs being used.³⁷

Other culprits. Serotonergic effects can also be associated with illicit drugs, some nonprescription medications, and supplements. And in March 2016, the FDA issued a warning about the risks of taking opioids with serotonergic medications.³⁸ Although labeling changes have been recommended for all opioids, the cases of serotonin syndrome were reported more often with normal doses of fentanyl and methadone.

There are 2 mechanisms by which drugs may increase a patient’s risk for serotonin syndrome. The first is a pharmacodynamic interaction, which can occur when 2 or more medications act at the same receptor site (serotonin receptors in this example), which may result in an additive or synergistic effect.³⁹

There is increasing evidence that concomitant use of an SSRI and an NSAID increases the risk of gastrointestinal bleeding.

The second mechanism is a pharmacokinetic alteration (an agent alters absorption, distribution, metabolism, or excretion) of CYP enzymes.⁴⁰ Of the more commonly used antidepressants, citalopram, escitalopram, venlafaxine, and mirtazapine seem to have the least potential for clinically significant pharmacokinetic interactions.⁴¹

Guidelines? Currently there are no guidelines for preventing serotonin syndrome. Clinicians should exercise caution in patients at high risk for drug adverse events, such as the elderly, patients taking multiple medications, and patients with comorbidities. Healthy low-risk patients can generally take 2 or 3 serotonergic medications at therapeutic doses without a major risk of harm.

CORRESPONDENCE
Mary Onysko, PharmD, BCPS, 191 East Orchard Road, Suite 200, Littleton, CO 80121; [email protected].

There is a strong relationship between the number of medications taken and the likelihood of a potentially serious drug-drug interaction.^1,2 Drug interaction software programs can help alert prescribers to potential problems, but these programs sometimes fail to detect important interactions or generate so many clinically insignificant alerts that they become a nuisance.³ This review provides guidance about 5 clinically relevant drug interactions, including those that are common (TABLE 1^4-6)—and those that are less common, but no less important (TABLE 2^6-10).

1. Antiepileptics & contraceptives

Many antiepileptic medications decrease the efficacy of certain contraceptives

Contraception management in women with epilepsy is critical due to potential maternal and fetal complications. Many antiepileptic drugs (AEDs), including carbamazepine, ethosuximide, fosphenytoin, phenobarbital, phenytoin, primidone, topiramate, and valproate, are potentially teratogenic.¹¹ A retrospective, observational study of 115 women of childbearing age who had epilepsy and were seen at a neurology clinic found that 74% were not using documented contraception.¹¹ Of the minority of study participants using contraception, most were using oral contraceptives (OCs) that could potentially interact with AEDs.

CYP inducers. Estrogen and progesterone are metabolized by the cytochrome P450 3A4 enzyme. Some AEDs induce this enzyme, which can enhance the metabolism of OCs, thus reducing their efficacy.¹² It is not known, however, if this interaction results in increased pregnancy rates.¹³ Most newer AEDs (TABLE 3⁶) do not induce cytochrome P450 3A4 and, thus, do not appear to affect OC efficacy, and may be safer for women with seizure disorders.¹² While enzyme-inducing AEDs may decrease the efficacy of progesterone-only OCs and the morning-after pill,^12,14,15 progesterone-containing intrauterine devices (IUDs), long-acting progesterone injections, and non-hormonal contraceptive methods appear to be unaffected.^14-17

OCs and seizure frequency. There is no strong evidence that OCs affect seizure frequency in epileptic women, although changes in hormone levels during the menstrual cycle do affect seizure susceptibility.¹² Combination OCs decrease lamotrigine levels and, therefore, may increase the risk of seizures, but progesterone-only pills do not produce this effect.^12,16

Do guidelines exist? There are no specific evidence-based guidelines that pertain to the use of AEDs and contraception together, but some organizations have issued recommendations.

The American College of Obstetricians and Gynecologists recommends using a 30- to 35-mcg estrogen-containing OC rather than a lower dose in women taking an enzyme-inducing AED. The group also recommends using condoms with OCs or using IUDs.¹⁸

The American Academy of Neurology suggests that women taking OCs and enzyme-inducing AEDs use an OC containing at least 50 mcg estrogen.¹⁹

The National Institute for Health and Care Excellence recommends that women taking enzyme-inducing AEDs avoid progestin-only pills.²⁰

The Faculty of Sexual and Reproductive Healthcare agrees that enzyme-inducing drugs may decrease efficacy and recommend considering IUDs and injectable contraceptive methods.²¹

2. SSRIs & NSAIDs.

SSRIs increase the GI bleeding risk associated with NSAIDs alone

Nonsteroidal anti-inflammatory drugs (NSAIDs) and selective serotonin reuptake inhibitors (SSRIs) are commonly prescribed worldwide.^22,23 A well-established adverse effect of NSAIDs is gastrointestinal (GI) bleeding, and there is increasing evidence that concomitant use of an SSRI can further increase that risk through a variety of mechanisms.²³

SSRIs decrease platelet serotonin levels resulting in defective platelet aggregation and impaired hemostasis. Studies have also shown that SSRIs increase gastric acidity, which leads to increased risk of peptic ulcer disease and GI bleeding.²³ These mechanisms, combined with the inhibition of gastroprotective prostaglandin cyclooxygenase-1 and platelets by NSAIDs, further potentiate GI bleeding risk.²⁴

Patients at high risk for bleeding with concomitant SSRIs and NSAIDs include older patients, patients with other risk factors for GI bleeding (eg, chronic steroid use), and patients with a history of GI bleeding.²³

The evidence. A 2014 meta-analysis found that when SSRIs were used in combination with NSAIDs, the risk of GI bleeding was significantly increased, compared with SSRI monotherapy.²³

Case control studies found the risk of upper GI bleeding with SSRIs had a number needed to harm (NNH) of 3177 for a low-risk population and 881 for a high-risk population with an odds ratio (OR) of 1.66 (95% confidence interval [CI], 1.44-1.92; P<.00001).²³ When SSRIs were used in combination with NSAIDs, the NNH decreased to 645 for a low-risk population and 179 for a high-risk population (OR=4.25; 95% CI, 2.82-6.42; P<.0001).²³

Another meta-analysis found that the OR for bleeding risk increased to 6.33 (95% CI, 3.40-11.8; P<.00001; NNH=106) with concomitant use of NSAIDs and SSRIs, compared with 2.36 (95% CI, 1.44-3.85; P=.0006; NNH=411) for SSRI use alone.²⁵

The studies did not evaluate results based on the indication, dose, or duration of SSRI or NSAID treatment. If both an SSRI and an NSAID must be used, select a cyclooxygenase-2 selective NSAID at the lowest effective dose and consider the addition of a proton pump inhibitor to decrease the risk of a GI bleed.^23,26

3. Direct oral anticoagulants and antiepileptics

Don’t use DOACs in patients taking certain antiepileptic medications

Drug interactions with anticoagulants, such as warfarin, are well documented and have been publicized for years, but physicians must also be aware of the potential for interaction between the direct oral anticoagulants (DOACs) and AEDs.

Apixaban, rivaroxaban, and dabigatran appear to interact withthe AEDs carbamazepine, phenytoin, and phenobarbital.^27,28 These interactions occur due to AED induction of the CYP3A4 enzyme and effects on the P-glycoprotein (P-gp) efflux pump.^27,29 When taken together, the AED induces metabolism and elimination of the DOAC medication to occur more quickly than it would normally, resulting in subtherapeutic concentrations of the DOAC. This could theoretically result in a venous thromboembolic event or stroke.

A caveat. One thing to consider is that studies demonstrating interaction between the DOAC and AED drug classes have been performed in healthy volunteers, making it difficult to extrapolate how this interaction may increase the risk for thrombotic events in other patients.

Some studies demonstrated reductions in drug levels of up to 50% with strong CYP3A4 and P-glycoprotein inducers.³⁰ Common inducers include carbamazepine, rifampin, and St. John’s Wort.⁶ Patients taking such agents could theoretically have decreased exposure to the DOAC, resulting in an increase in thromboembolic risk.³¹

4. Statins & certain CYP inhibitors

Combining simvastatin with fibrates warrants extra attention

The efficacy of statin medications in the prevention of atherosclerotic cardiovascular disease (ASCVD) is clear. However, the clinical significance of many identified drug interactions involving statins is difficult to interpret. Interactions that cause increased serum concentrations of statins can increase the risk for liver enzyme elevations and skeletal muscle abnormalities (myalgias to rhabdomyolysis).³² Strong inhibitors of CYP3A4 (amiodarone, cyclosporine, ketoconazole, etc.) significantly increase concentrations of lovastatin, simvastatin, and atorvastatin. Pitavastatin, pravastatin, and rosuvastatin are not susceptible to any CYP-mediated drug interactions;³³ therefore, rosuvastatin (a high-intensity statin) is usually recommended over other statins for patients taking strong inhibitors of CYP3A4.

When to limit simvastatin. Doses of simvastatin should not exceed 10 mg/d when combined with diltiazem, dronedarone, or verapamil, and doses should not exceed 20 mg/d when used with amiodarone, amlodipine, or ranolazine.⁶ These recommendations are in response to results from the SEARCH (Study of the Effectiveness of Additional Reductions in cholesterol and homocysteine) trial, which found a higher incidence of myopathies and rhabdomyolysis in patients taking 80 mg of simvastatin compared with those taking 20-mg doses.³⁴ CYP3A4-inducing medications, especially diltiazem, were thought to also contribute to an increased risk.³⁴

Avoid gemfibrozil with statins. Using fibrates with statins is beneficial for some patients; however, gemfibrozil significantly interacts with statins by inhibiting CYP2C8 and organic anion transporting polypeptide 1B1 (OATP1B1).³³ The safer choice is fenofibrate because it does not interfere with statin metabolism and can be safely used in combination with statins.⁶

Enzyme-inducing antiepileptic drugs can enhance the metabolism of oral contraceptives, thus reducing their efficacy.

A retrospective review of the FDA Adverse Event Reporting System (AERS) database found that 88% of fibrate and statin combinations that resulted in rhabdomyolysis were associated with gemfibrozil/cerivastatin (cerivastatin is no longer available in the United States).³⁵

5. One serotonergic drug & another

Serotonin syndrome is associated with more than just SSRIs

Serotonin syndrome is a constellation of symptoms (hyperthermia, hyperreflexia, muscle clonus, tremor and altered mental status) caused by increases in serotonin levels in the central and peripheral nervous systems that can lead to mild or life-threatening complications such as seizures, muscle breakdown, or hyperthermia. Serotonin syndrome is most likely to occur within 24 hours after a dose increase, after starting a new medication that increases serotonin levels, or after a drug overdose.³⁶

SSRIs are the most commonly reported drug associated with serotonin syndrome; however, other medications (TABLE 4³⁷) may be responsible, especially when used in combination with agents that act on serotonin receptors or in patients with impaired metabolism of the drugs being used.³⁷

Other culprits. Serotonergic effects can also be associated with illicit drugs, some nonprescription medications, and supplements. And in March 2016, the FDA issued a warning about the risks of taking opioids with serotonergic medications.³⁸ Although labeling changes have been recommended for all opioids, the cases of serotonin syndrome were reported more often with normal doses of fentanyl and methadone.

There are 2 mechanisms by which drugs may increase a patient’s risk for serotonin syndrome. The first is a pharmacodynamic interaction, which can occur when 2 or more medications act at the same receptor site (serotonin receptors in this example), which may result in an additive or synergistic effect.³⁹

There is increasing evidence that concomitant use of an SSRI and an NSAID increases the risk of gastrointestinal bleeding.

The second mechanism is a pharmacokinetic alteration (an agent alters absorption, distribution, metabolism, or excretion) of CYP enzymes.⁴⁰ Of the more commonly used antidepressants, citalopram, escitalopram, venlafaxine, and mirtazapine seem to have the least potential for clinically significant pharmacokinetic interactions.⁴¹

Guidelines? Currently there are no guidelines for preventing serotonin syndrome. Clinicians should exercise caution in patients at high risk for drug adverse events, such as the elderly, patients taking multiple medications, and patients with comorbidities. Healthy low-risk patients can generally take 2 or 3 serotonergic medications at therapeutic doses without a major risk of harm.

CORRESPONDENCE
Mary Onysko, PharmD, BCPS, 191 East Orchard Road, Suite 200, Littleton, CO 80121; [email protected].

There is a strong relationship between the number of medications taken and the likelihood of a potentially serious drug-drug interaction.^1,2 Drug interaction software programs can help alert prescribers to potential problems, but these programs sometimes fail to detect important interactions or generate so many clinically insignificant alerts that they become a nuisance.³ This review provides guidance about 5 clinically relevant drug interactions, including those that are common (TABLE 1^4-6)—and those that are less common, but no less important (TABLE 2^6-10).

1. Antiepileptics & contraceptives

Many antiepileptic medications decrease the efficacy of certain contraceptives

Contraception management in women with epilepsy is critical due to potential maternal and fetal complications. Many antiepileptic drugs (AEDs), including carbamazepine, ethosuximide, fosphenytoin, phenobarbital, phenytoin, primidone, topiramate, and valproate, are potentially teratogenic.¹¹ A retrospective, observational study of 115 women of childbearing age who had epilepsy and were seen at a neurology clinic found that 74% were not using documented contraception.¹¹ Of the minority of study participants using contraception, most were using oral contraceptives (OCs) that could potentially interact with AEDs.

CYP inducers. Estrogen and progesterone are metabolized by the cytochrome P450 3A4 enzyme. Some AEDs induce this enzyme, which can enhance the metabolism of OCs, thus reducing their efficacy.¹² It is not known, however, if this interaction results in increased pregnancy rates.¹³ Most newer AEDs (TABLE 3⁶) do not induce cytochrome P450 3A4 and, thus, do not appear to affect OC efficacy, and may be safer for women with seizure disorders.¹² While enzyme-inducing AEDs may decrease the efficacy of progesterone-only OCs and the morning-after pill,^12,14,15 progesterone-containing intrauterine devices (IUDs), long-acting progesterone injections, and non-hormonal contraceptive methods appear to be unaffected.^14-17

OCs and seizure frequency. There is no strong evidence that OCs affect seizure frequency in epileptic women, although changes in hormone levels during the menstrual cycle do affect seizure susceptibility.¹² Combination OCs decrease lamotrigine levels and, therefore, may increase the risk of seizures, but progesterone-only pills do not produce this effect.^12,16

Do guidelines exist? There are no specific evidence-based guidelines that pertain to the use of AEDs and contraception together, but some organizations have issued recommendations.

The American College of Obstetricians and Gynecologists recommends using a 30- to 35-mcg estrogen-containing OC rather than a lower dose in women taking an enzyme-inducing AED. The group also recommends using condoms with OCs or using IUDs.¹⁸

The American Academy of Neurology suggests that women taking OCs and enzyme-inducing AEDs use an OC containing at least 50 mcg estrogen.¹⁹

The National Institute for Health and Care Excellence recommends that women taking enzyme-inducing AEDs avoid progestin-only pills.²⁰

The Faculty of Sexual and Reproductive Healthcare agrees that enzyme-inducing drugs may decrease efficacy and recommend considering IUDs and injectable contraceptive methods.²¹

2. SSRIs & NSAIDs.

SSRIs increase the GI bleeding risk associated with NSAIDs alone

Nonsteroidal anti-inflammatory drugs (NSAIDs) and selective serotonin reuptake inhibitors (SSRIs) are commonly prescribed worldwide.^22,23 A well-established adverse effect of NSAIDs is gastrointestinal (GI) bleeding, and there is increasing evidence that concomitant use of an SSRI can further increase that risk through a variety of mechanisms.²³

SSRIs decrease platelet serotonin levels resulting in defective platelet aggregation and impaired hemostasis. Studies have also shown that SSRIs increase gastric acidity, which leads to increased risk of peptic ulcer disease and GI bleeding.²³ These mechanisms, combined with the inhibition of gastroprotective prostaglandin cyclooxygenase-1 and platelets by NSAIDs, further potentiate GI bleeding risk.²⁴

Patients at high risk for bleeding with concomitant SSRIs and NSAIDs include older patients, patients with other risk factors for GI bleeding (eg, chronic steroid use), and patients with a history of GI bleeding.²³

The evidence. A 2014 meta-analysis found that when SSRIs were used in combination with NSAIDs, the risk of GI bleeding was significantly increased, compared with SSRI monotherapy.²³

Case control studies found the risk of upper GI bleeding with SSRIs had a number needed to harm (NNH) of 3177 for a low-risk population and 881 for a high-risk population with an odds ratio (OR) of 1.66 (95% confidence interval [CI], 1.44-1.92; P<.00001).²³ When SSRIs were used in combination with NSAIDs, the NNH decreased to 645 for a low-risk population and 179 for a high-risk population (OR=4.25; 95% CI, 2.82-6.42; P<.0001).²³

Another meta-analysis found that the OR for bleeding risk increased to 6.33 (95% CI, 3.40-11.8; P<.00001; NNH=106) with concomitant use of NSAIDs and SSRIs, compared with 2.36 (95% CI, 1.44-3.85; P=.0006; NNH=411) for SSRI use alone.²⁵

The studies did not evaluate results based on the indication, dose, or duration of SSRI or NSAID treatment. If both an SSRI and an NSAID must be used, select a cyclooxygenase-2 selective NSAID at the lowest effective dose and consider the addition of a proton pump inhibitor to decrease the risk of a GI bleed.^23,26

3. Direct oral anticoagulants and antiepileptics

Don’t use DOACs in patients taking certain antiepileptic medications

Drug interactions with anticoagulants, such as warfarin, are well documented and have been publicized for years, but physicians must also be aware of the potential for interaction between the direct oral anticoagulants (DOACs) and AEDs.

Apixaban, rivaroxaban, and dabigatran appear to interact withthe AEDs carbamazepine, phenytoin, and phenobarbital.^27,28 These interactions occur due to AED induction of the CYP3A4 enzyme and effects on the P-glycoprotein (P-gp) efflux pump.^27,29 When taken together, the AED induces metabolism and elimination of the DOAC medication to occur more quickly than it would normally, resulting in subtherapeutic concentrations of the DOAC. This could theoretically result in a venous thromboembolic event or stroke.

A caveat. One thing to consider is that studies demonstrating interaction between the DOAC and AED drug classes have been performed in healthy volunteers, making it difficult to extrapolate how this interaction may increase the risk for thrombotic events in other patients.

Some studies demonstrated reductions in drug levels of up to 50% with strong CYP3A4 and P-glycoprotein inducers.³⁰ Common inducers include carbamazepine, rifampin, and St. John’s Wort.⁶ Patients taking such agents could theoretically have decreased exposure to the DOAC, resulting in an increase in thromboembolic risk.³¹

4. Statins & certain CYP inhibitors

Combining simvastatin with fibrates warrants extra attention

The efficacy of statin medications in the prevention of atherosclerotic cardiovascular disease (ASCVD) is clear. However, the clinical significance of many identified drug interactions involving statins is difficult to interpret. Interactions that cause increased serum concentrations of statins can increase the risk for liver enzyme elevations and skeletal muscle abnormalities (myalgias to rhabdomyolysis).³² Strong inhibitors of CYP3A4 (amiodarone, cyclosporine, ketoconazole, etc.) significantly increase concentrations of lovastatin, simvastatin, and atorvastatin. Pitavastatin, pravastatin, and rosuvastatin are not susceptible to any CYP-mediated drug interactions;³³ therefore, rosuvastatin (a high-intensity statin) is usually recommended over other statins for patients taking strong inhibitors of CYP3A4.

When to limit simvastatin. Doses of simvastatin should not exceed 10 mg/d when combined with diltiazem, dronedarone, or verapamil, and doses should not exceed 20 mg/d when used with amiodarone, amlodipine, or ranolazine.⁶ These recommendations are in response to results from the SEARCH (Study of the Effectiveness of Additional Reductions in cholesterol and homocysteine) trial, which found a higher incidence of myopathies and rhabdomyolysis in patients taking 80 mg of simvastatin compared with those taking 20-mg doses.³⁴ CYP3A4-inducing medications, especially diltiazem, were thought to also contribute to an increased risk.³⁴

Avoid gemfibrozil with statins. Using fibrates with statins is beneficial for some patients; however, gemfibrozil significantly interacts with statins by inhibiting CYP2C8 and organic anion transporting polypeptide 1B1 (OATP1B1).³³ The safer choice is fenofibrate because it does not interfere with statin metabolism and can be safely used in combination with statins.⁶

Enzyme-inducing antiepileptic drugs can enhance the metabolism of oral contraceptives, thus reducing their efficacy.

A retrospective review of the FDA Adverse Event Reporting System (AERS) database found that 88% of fibrate and statin combinations that resulted in rhabdomyolysis were associated with gemfibrozil/cerivastatin (cerivastatin is no longer available in the United States).³⁵

5. One serotonergic drug & another

Serotonin syndrome is associated with more than just SSRIs

Serotonin syndrome is a constellation of symptoms (hyperthermia, hyperreflexia, muscle clonus, tremor and altered mental status) caused by increases in serotonin levels in the central and peripheral nervous systems that can lead to mild or life-threatening complications such as seizures, muscle breakdown, or hyperthermia. Serotonin syndrome is most likely to occur within 24 hours after a dose increase, after starting a new medication that increases serotonin levels, or after a drug overdose.³⁶

SSRIs are the most commonly reported drug associated with serotonin syndrome; however, other medications (TABLE 4³⁷) may be responsible, especially when used in combination with agents that act on serotonin receptors or in patients with impaired metabolism of the drugs being used.³⁷

Other culprits. Serotonergic effects can also be associated with illicit drugs, some nonprescription medications, and supplements. And in March 2016, the FDA issued a warning about the risks of taking opioids with serotonergic medications.³⁸ Although labeling changes have been recommended for all opioids, the cases of serotonin syndrome were reported more often with normal doses of fentanyl and methadone.

There are 2 mechanisms by which drugs may increase a patient’s risk for serotonin syndrome. The first is a pharmacodynamic interaction, which can occur when 2 or more medications act at the same receptor site (serotonin receptors in this example), which may result in an additive or synergistic effect.³⁹

There is increasing evidence that concomitant use of an SSRI and an NSAID increases the risk of gastrointestinal bleeding.

The second mechanism is a pharmacokinetic alteration (an agent alters absorption, distribution, metabolism, or excretion) of CYP enzymes.⁴⁰ Of the more commonly used antidepressants, citalopram, escitalopram, venlafaxine, and mirtazapine seem to have the least potential for clinically significant pharmacokinetic interactions.⁴¹

Guidelines? Currently there are no guidelines for preventing serotonin syndrome. Clinicians should exercise caution in patients at high risk for drug adverse events, such as the elderly, patients taking multiple medications, and patients with comorbidities. Healthy low-risk patients can generally take 2 or 3 serotonergic medications at therapeutic doses without a major risk of harm.

CORRESPONDENCE
Mary Onysko, PharmD, BCPS, 191 East Orchard Road, Suite 200, Littleton, CO 80121; [email protected].

While the unintended pregnancy rate for women ages 15 to 44 years decreased by 18% between 2008 and 2011, almost half of pregnancies in the United States remain unintended.¹ On a more positive note, however, women who use birth control consistently and correctly account for only 5% of unintended pregnancies.² As family physicians (FPs), we can support and facilitate our female patients’ efforts to consistently use highly effective forms of contraception. The 5 initiatives detailed here can help toward that end.

1. Routinely screen patients for their reproductive intentions

All women of reproductive age should be screened routinely for their pregnancy intentions. The American College of Obstetricians and Gynecologists (ACOG) encourages clinicians to ask women about pregnancy intendedness and encourages patients to develop a reproductive life plan, or a set of personal goals about whether or when to have children.³ The Centers for Disease Control and Prevention (CDC) has also developed a reproductive life plan tool for health professionals to encourage women and men to reflect upon their plans.⁴ So just as we regularly screen and document cigarette use and blood pressure (BP), so too, should we routinely screen women for their reproductive goals.

Ask women this one question. The Oregon Foundation for Reproductive Health launched the One Key Question Initiative, which proposes that the care team ask women ages 18 to 50: “Would you like to become pregnant in the next year?”⁵ A common workflow includes the medical assistant asking women about pregnancy intentions and providing a preconception and/or contraceptive handout, if appropriate. The physician provides additional counseling as needed. Pilot studies of One Key Question indicate that 30% to 40% of women screened needed follow-up counseling, suggesting the need for clinicians to be proactive in asking about reproductive plans. (Additional information on the Initiative is available on the Foundation’s Web site at http://www.orfrh.org/.)

This approach assumes women feel in control of their reproduction; however, this may not be the reality for many, especially low-income women.⁶ Additionally, women commonly cite planning a pregnancy as appropriate only when they are in an ideal relationship and when they are living in a financially stable environment—conditions that some women may never achieve.

ACOG encourages clinicians to ask women about pregnancy intendedness and encourages patients to develop a reproductive plan.

Another caveat is that women may not have explicit pregnancy intentions, in which case, this particular approach may not be effective. A study of low-income women found only 60% intended to use the method prescribed after contraception counseling, with 37% of those stopping because of adverse effects, 23% saying they wanted another method, and 17% citing method complexity.⁷

Reproductive coercion from male partners, ranging from pressure to become pregnant to method sabotage, is also common in low-income women.⁸ Regular conversations that prioritize a woman’s values and experience are needed to promote reproductive autonomy.

2. Decouple provision of contraception from unnecessary exams

Pelvic exams and pap smears should not be required prior to offering patients hormonal contraception, according to the Choosing Wisely campaign of the American Board of Internal Medicine and ACOG.^9,10 Hormonal contraception may instead be provided safely based on a medical history and BP assessment. Adolescents, minority groups, obese women, and victims of sexual trauma, in particular may avoid asking about birth control because of anxiety and fear of pain from these exams.¹¹ The American College of Physicians recommends against speculum and bimanual exams in asymptomatic, non-pregnant, adult women.¹² Pap smears and sexually transmitted infection (STI) testing should be performed at their normally scheduled intervals as recommended by the US Preventive Services Task Force (USPSTF) and not be tied to contraceptive provision.¹³

Assess pregnancy status using criteria,rather than a pregnancy text

Use the CDC’s criteria to assess pregnancy status rather than relying on a urine pregnancy test prior to providing contraception. Once you are reasonably sure that a woman is not pregnant (TABLE 1¹⁴), contraception may be started. Some physicians have traditionally requested that a woman delay starting contraception until the next menses to ensure that she is not already pregnant. However, given the evidence that hormonal contraception does not cause birth defects, such a delay is not warranted and puts the woman at risk of an unintended pregnancy during the gap.¹⁵

Furthermore, there is an approximate 2-week window in which a woman could have a negative urine pregnancy test despite being pregnant, so the test alone is not completely reliable. In addition, obese women may experience irregular cycles, further complicating the traditional approach.¹⁶

Another largely unnecessary step … The US Selected Practice Recommendations (US SPR) from the CDC notes that additional STI screening prior to an intrauterine device (IUD) insertion is unnecessary for most women if appropriate screening guidelines have been previously followed.¹⁴ For those who have not been screened according to guidelines, the CDC recommends same-day screening and IUD insertion. You can then treat an STI without removing the IUD. Women with purulent cervicitis or a current chlamydial or gonorrheal infection should delay IUD insertion until after treatment.

3. Expand long-acting reversible contraception counseling and access

Offer long-acting reversible contraception (LARC), such as IUDs and implants, as first-line options for most women. ACOG endorses LARC as the most effective reversible method for most women, including those who have not given birth and adolescents.¹⁷ Unfortunately, a 2012 study found that family physicians were less likely than OB-GYNs to have enough time for contraceptive counseling and fewer than half felt competent inserting IUDs.¹⁸ While 79% of OB-GYNs routinely discussed IUDs with their patients, only 47% of family physicians did. In 2014, the American Academy of Pediatrics (AAP) endorsed a LARC-first tiered counseling approach for adolescents.¹⁹

A test of LARC-first counseling

The Contraceptive CHOICE project, a St. Louis, Missouri-based initiative, was launched to reduce unintended pregnancies in women ages 14 to 45 years by offering LARC-first counseling and free contraception of their choice.²⁰ This project involved more than 9000 women at high risk for unintended pregnancy. Same-day LARC insertion was available. Seventy-five percent of women chose a LARC method and they reported greater continuation at 12 and 24 months, when compared to women who did not choose a LARC method. LARC users also reported higher satisfaction at one year. Provision of contraception through the project contributed to a reduction in repeat abortions as well as decreased rates of teenage pregnancy, birth, and abortion. Three years after the start of the project, IUDs had continuation rates of nearly 70%, implants of 56%, and non-LARC methods of 31%.²¹

When counseling women, it’s important to remember that effectiveness may not be the only criterium a woman uses when choosing a method. A 2010 study found that for 91% of women at high risk for unintended pregnancy, no single method possessed all the features they deemed “extremely important.”²² Clinicians should take a patient-centered approach to find birth control that fits each patient’s priorities.

Clinicians need proper training in LARC methods

Only 20% of FPs regularly insert IUDs, and 11% offer contraceptive implants, according to estimates from physicians recertifying with the American Board of Family Medicine in 2014.²³ Access to training during residency is a key component to increasing these rates. FPs who practice obstetrics should be trained in postpartum LARC insertion and offer this option prior to hospital discharge as well as during the postpartum office visit.

Performing LARC insertions on the same day as counseling is ideal, and clinics should strive to reduce barriers to same-day procedures. Time constraints may be addressed by shifting tasks among the medical team. In the CHOICE project, contraceptive counselors—half of whom had no clinical experience—were trained to provide tiered counseling to participants. By working with a cross-trained health care team and offering prepared resources, clinicians can save time and improve access.

Physicians may want to incorporate the free online resources Bedsider.org or Stayteen.org to help women learn about contraceptive methods.²⁴ The user-friendly Web sites, operated by the National Campaign to Prevent Teen and Unplanned Pregnancy, describe various forms of contraception and offer text and email reminders. Incorporating Bedsider into the counseling workflow and discussing the various reminder tools available may improve patients’ knowledge and enhance their compliance.

Pelvic exams and pap smears should not be required before offering patients hormonal contraception.

Additional barriers for practices may include high upfront costs associated with stocking devices. Practices that may be unable to sustain the costs surrounding enhanced contraception counseling and provision can collaborate with family planning clinics that are able to offer same-day services. A study of clinics in California found that Title X clinics were more likely to provide on-site LARC services than non-Title X public and private providers.²⁵

4. Follow CDC guidelines for initiating and continuing contraception

Follow the US SPR for guidance on initiating and continuing contraceptive methods.¹⁴ The CDC’s Medical Eligibility Criteria for Contraceptive Use is another vital resource, providing recommendations for contraceptive methods to patients who have specific medical conditions or characteristics.²⁶

Utilize the “quick start” method for hormonal contraception, where birth control is started on the same day as its prescription regardless of timing of the menstrual cycle. If you can’t be reasonably certain that a woman is not pregnant based on the criteria listed in TABLE 1,¹⁴ conduct a pregnancy test (while recognizing the aforementioned 2-week window of limitations) and counsel the patient to use back-up protection for the first 7 days along with repeating a pregnancy test in 2 weeks’ time.

The quick start method may lead to higher adherence than delayed initiation.²⁷ Differences in continuation rates between women who use the quick start method and those who follow the delayed approach may disappear over time.²⁸

Prescribe and provide a year’s supply of oral contraceptive pills (OCPs) as recommended by the CDC US SPR.¹⁴ It is important to note that pharmacists are usually restricted by insurance companies to only fill a one or 3 month’s supply.

One study found that FPs were less likely than OB-GYNs to have enough time for contraceptive counseling and fewer than half felt competent inserting IUDs.

In January 2016, Oregon began requiring private and state health insurance providers to reimburse for a year’s supply of prescription contraception; in January 2017, insurers in Washington, DC, were also required to offer women a year’s supply of prescription contraception.^29,30 Several other states have followed suit. The California Health Benefits Review Program estimates a savings of $42.8 million a year from fewer office visits and 15,000 fewer unintended pregnancies if their state enacts a similar policy.³¹

Pharmacist initiatives are worth watching. In January 2016, Oregon pharmacists with additional training were allowed to prescribe OCs and hormonal patches to women 18 years and older.³² In April 2016, a similar law went into effect in California, but without a minimum age requirement and with the additional coverage of vaginal rings and Depo-Provera (depo) injections.³³ Pharmacists in both states must review a health questionnaire completed by the woman and can refer to a physician as necessary.

The CDC recommends that clinicians extend the allowed window for repeat depo injections to 15 weeks.¹⁴ Common institutional protocol is to give repeat injections every 11 to 13 weeks. If past that window, protocol often dictates the woman abstain from unprotected sex for 2 weeks and then return for a negative pregnancy test (or await menses) before the next injection. However, the CDC notes that depo is effective for longer than the 13-week period.¹⁴ No additional birth control or pregnancy testing is needed and the woman can receive the next depo shot if she is up to 15 weeks from the previous shot.

One study found no additional pregnancy risks for those who were up to 4 weeks “late” for their next shot, suggesting there is potential for an even larger grace period.³⁴ The World Health Organization advises allowing a repeat injection up to 4 weeks late.³⁵ We encourage institutions to change their policies to comply with the CDC’s 15-week window.

Another initiative is over-the-counter (OTC) access to OCs, which the American Academy of Family Physicians (AAFP) and ACOG support.^36,37 ACOG notes that “no drug or intervention is completely without risk of harm” and that the risk of venous thromboembolism for OC users is lower than the risk of pregnancy.³⁷ Women can successfully self-screen for contraindications using a checklist. Concerns about women potentially being less adherent or less likely to choose LARCs are not reasons to preclude access to other methods. The AAFP supports insurance coverage of OCs, regardless of prescription status.³⁶

5. Routinely counsel about, and advance-prescribe, emergency contraception pills

Physicians should counsel and advance-prescribe emergency contraception pills (ECPs) to women, including adolescents, using less reliable contraception, as recommended by ACOG, AAP, and the CDC.^14,37,38 It’s also important to provide information on the copper IUD as the most effective method of emergency contraception, with nearly 100% efficacy if placed within 5 days.³⁹ An easy-to-read patient hand-out in English and Spanish on EC options can be found at http://beyondthepill.ucsf.edu/tools-materials.

Only 3% of respondents participating in the 2006-2010 National Survey of Family Growth received counseling about emergency contraception in the past year.⁴⁰ ECPs are most effective when used within 24 hours but have some efficacy up to 5 days.³⁷ Due to the Affordable Care Act, most insurance plans will cover ECPs if purchased with a prescription, but coverage varies by state.⁴¹ Ulipristal acetate (UPA) ECP is only available with a prescription. Advance prescriptions can alleviate financial burdens on women when they need to access ECPs quickly.

Women should wait at least 5 days before resuming or starting hormonal contraception after taking UPA-based ECP, as it may reduce the ovulation-delaying effect of the ECP.¹⁴ For IUDs, implants, and depo, which require a visit to a health care provider, physicians evaluating earlier provision should consider the risks of reduced efficacy against the many barriers to access.

UPA-based ECPs (such as ella) may be more effective for overweight and obese women than levonorgestrel-based ECPs (such as Plan B and Next Choice).¹⁴ Consider advance-prescribing UPA ECPs to women with a body mass index (BMI) >25 kg/m².⁴² Such considerations are important as the prevalence of obesity in women between 2013 and 2014 was 40.4%.⁴³

A copper IUD may be the only effective form of emergency contraception for patients with a BMI ≥35 kg/m².

In May 2016, the FDA noted that while current data are insufficient regarding whether the effectiveness of levonorgestrel ECPs is reduced in overweight or obese women, there are no safety concerns regarding their use in this population.⁴⁴ Therefore, a woman with a BMI >25 kg/m² should use UPA ECPs if available; but if not, she can still use levonorgestrel ECPs. One study, however, has found that UPA ECPs are only as effective as a placebo when BMI is ≥35 kg/m², at which point a copper IUD may be the only effective form of emergency contraception.⁴⁵

Transitioning from customary practices to best practices

Following these practical steps, FPs can improve contraceptive care for women. However, to make a significant impact, clinicians must be willing to change customary practices that are based on tradition, routines, or outdated protocols in favor of those based on current evidence.

One good place to start the transition to best practices is to familiarize yourself with the 2016 US Medical Eligibility Criteria for Contraceptive Use²⁶ and Selected Practice Recommendations for Contraceptive Use.¹⁴ TABLES 2^14,26,46,47 and 3 offer additional resources that can enhance contraceptive counseling and further promote access to contraceptive care.

The contraceptive coverage guarantee under the Affordable Care Act has allowed many women to make contraceptive choices based on personal needs and preferences rather than cost. The new contraceptive coverage exemptions issued under the Trump administration will bring cost back as the driving decision factor for women whose employers choose not to provide contraceptive coverage. Providers should be aware of the typical costs associated with the various contraceptive options offered in their practice and community.

CORRESPONDENCE
Jessica Dalby, MD, Department of Family Medicine and Community Health, University of Wisconsin School of Medicine and Public Health, 1102 South Park St, Suite 100, Madison, WI 53715; [email protected].

While the unintended pregnancy rate for women ages 15 to 44 years decreased by 18% between 2008 and 2011, almost half of pregnancies in the United States remain unintended.¹ On a more positive note, however, women who use birth control consistently and correctly account for only 5% of unintended pregnancies.² As family physicians (FPs), we can support and facilitate our female patients’ efforts to consistently use highly effective forms of contraception. The 5 initiatives detailed here can help toward that end.

1. Routinely screen patients for their reproductive intentions

All women of reproductive age should be screened routinely for their pregnancy intentions. The American College of Obstetricians and Gynecologists (ACOG) encourages clinicians to ask women about pregnancy intendedness and encourages patients to develop a reproductive life plan, or a set of personal goals about whether or when to have children.³ The Centers for Disease Control and Prevention (CDC) has also developed a reproductive life plan tool for health professionals to encourage women and men to reflect upon their plans.⁴ So just as we regularly screen and document cigarette use and blood pressure (BP), so too, should we routinely screen women for their reproductive goals.

Ask women this one question. The Oregon Foundation for Reproductive Health launched the One Key Question Initiative, which proposes that the care team ask women ages 18 to 50: “Would you like to become pregnant in the next year?”⁵ A common workflow includes the medical assistant asking women about pregnancy intentions and providing a preconception and/or contraceptive handout, if appropriate. The physician provides additional counseling as needed. Pilot studies of One Key Question indicate that 30% to 40% of women screened needed follow-up counseling, suggesting the need for clinicians to be proactive in asking about reproductive plans. (Additional information on the Initiative is available on the Foundation’s Web site at http://www.orfrh.org/.)

This approach assumes women feel in control of their reproduction; however, this may not be the reality for many, especially low-income women.⁶ Additionally, women commonly cite planning a pregnancy as appropriate only when they are in an ideal relationship and when they are living in a financially stable environment—conditions that some women may never achieve.

ACOG encourages clinicians to ask women about pregnancy intendedness and encourages patients to develop a reproductive plan.

Another caveat is that women may not have explicit pregnancy intentions, in which case, this particular approach may not be effective. A study of low-income women found only 60% intended to use the method prescribed after contraception counseling, with 37% of those stopping because of adverse effects, 23% saying they wanted another method, and 17% citing method complexity.⁷

Reproductive coercion from male partners, ranging from pressure to become pregnant to method sabotage, is also common in low-income women.⁸ Regular conversations that prioritize a woman’s values and experience are needed to promote reproductive autonomy.

2. Decouple provision of contraception from unnecessary exams

Pelvic exams and pap smears should not be required prior to offering patients hormonal contraception, according to the Choosing Wisely campaign of the American Board of Internal Medicine and ACOG.^9,10 Hormonal contraception may instead be provided safely based on a medical history and BP assessment. Adolescents, minority groups, obese women, and victims of sexual trauma, in particular may avoid asking about birth control because of anxiety and fear of pain from these exams.¹¹ The American College of Physicians recommends against speculum and bimanual exams in asymptomatic, non-pregnant, adult women.¹² Pap smears and sexually transmitted infection (STI) testing should be performed at their normally scheduled intervals as recommended by the US Preventive Services Task Force (USPSTF) and not be tied to contraceptive provision.¹³

Assess pregnancy status using criteria,rather than a pregnancy text

Use the CDC’s criteria to assess pregnancy status rather than relying on a urine pregnancy test prior to providing contraception. Once you are reasonably sure that a woman is not pregnant (TABLE 1¹⁴), contraception may be started. Some physicians have traditionally requested that a woman delay starting contraception until the next menses to ensure that she is not already pregnant. However, given the evidence that hormonal contraception does not cause birth defects, such a delay is not warranted and puts the woman at risk of an unintended pregnancy during the gap.¹⁵

Furthermore, there is an approximate 2-week window in which a woman could have a negative urine pregnancy test despite being pregnant, so the test alone is not completely reliable. In addition, obese women may experience irregular cycles, further complicating the traditional approach.¹⁶

Another largely unnecessary step … The US Selected Practice Recommendations (US SPR) from the CDC notes that additional STI screening prior to an intrauterine device (IUD) insertion is unnecessary for most women if appropriate screening guidelines have been previously followed.¹⁴ For those who have not been screened according to guidelines, the CDC recommends same-day screening and IUD insertion. You can then treat an STI without removing the IUD. Women with purulent cervicitis or a current chlamydial or gonorrheal infection should delay IUD insertion until after treatment.

3. Expand long-acting reversible contraception counseling and access

Offer long-acting reversible contraception (LARC), such as IUDs and implants, as first-line options for most women. ACOG endorses LARC as the most effective reversible method for most women, including those who have not given birth and adolescents.¹⁷ Unfortunately, a 2012 study found that family physicians were less likely than OB-GYNs to have enough time for contraceptive counseling and fewer than half felt competent inserting IUDs.¹⁸ While 79% of OB-GYNs routinely discussed IUDs with their patients, only 47% of family physicians did. In 2014, the American Academy of Pediatrics (AAP) endorsed a LARC-first tiered counseling approach for adolescents.¹⁹

A test of LARC-first counseling

The Contraceptive CHOICE project, a St. Louis, Missouri-based initiative, was launched to reduce unintended pregnancies in women ages 14 to 45 years by offering LARC-first counseling and free contraception of their choice.²⁰ This project involved more than 9000 women at high risk for unintended pregnancy. Same-day LARC insertion was available. Seventy-five percent of women chose a LARC method and they reported greater continuation at 12 and 24 months, when compared to women who did not choose a LARC method. LARC users also reported higher satisfaction at one year. Provision of contraception through the project contributed to a reduction in repeat abortions as well as decreased rates of teenage pregnancy, birth, and abortion. Three years after the start of the project, IUDs had continuation rates of nearly 70%, implants of 56%, and non-LARC methods of 31%.²¹

When counseling women, it’s important to remember that effectiveness may not be the only criterium a woman uses when choosing a method. A 2010 study found that for 91% of women at high risk for unintended pregnancy, no single method possessed all the features they deemed “extremely important.”²² Clinicians should take a patient-centered approach to find birth control that fits each patient’s priorities.

Clinicians need proper training in LARC methods

Only 20% of FPs regularly insert IUDs, and 11% offer contraceptive implants, according to estimates from physicians recertifying with the American Board of Family Medicine in 2014.²³ Access to training during residency is a key component to increasing these rates. FPs who practice obstetrics should be trained in postpartum LARC insertion and offer this option prior to hospital discharge as well as during the postpartum office visit.

Performing LARC insertions on the same day as counseling is ideal, and clinics should strive to reduce barriers to same-day procedures. Time constraints may be addressed by shifting tasks among the medical team. In the CHOICE project, contraceptive counselors—half of whom had no clinical experience—were trained to provide tiered counseling to participants. By working with a cross-trained health care team and offering prepared resources, clinicians can save time and improve access.

Physicians may want to incorporate the free online resources Bedsider.org or Stayteen.org to help women learn about contraceptive methods.²⁴ The user-friendly Web sites, operated by the National Campaign to Prevent Teen and Unplanned Pregnancy, describe various forms of contraception and offer text and email reminders. Incorporating Bedsider into the counseling workflow and discussing the various reminder tools available may improve patients’ knowledge and enhance their compliance.

Pelvic exams and pap smears should not be required before offering patients hormonal contraception.

Additional barriers for practices may include high upfront costs associated with stocking devices. Practices that may be unable to sustain the costs surrounding enhanced contraception counseling and provision can collaborate with family planning clinics that are able to offer same-day services. A study of clinics in California found that Title X clinics were more likely to provide on-site LARC services than non-Title X public and private providers.²⁵

4. Follow CDC guidelines for initiating and continuing contraception

Follow the US SPR for guidance on initiating and continuing contraceptive methods.¹⁴ The CDC’s Medical Eligibility Criteria for Contraceptive Use is another vital resource, providing recommendations for contraceptive methods to patients who have specific medical conditions or characteristics.²⁶

Utilize the “quick start” method for hormonal contraception, where birth control is started on the same day as its prescription regardless of timing of the menstrual cycle. If you can’t be reasonably certain that a woman is not pregnant based on the criteria listed in TABLE 1,¹⁴ conduct a pregnancy test (while recognizing the aforementioned 2-week window of limitations) and counsel the patient to use back-up protection for the first 7 days along with repeating a pregnancy test in 2 weeks’ time.

The quick start method may lead to higher adherence than delayed initiation.²⁷ Differences in continuation rates between women who use the quick start method and those who follow the delayed approach may disappear over time.²⁸

Prescribe and provide a year’s supply of oral contraceptive pills (OCPs) as recommended by the CDC US SPR.¹⁴ It is important to note that pharmacists are usually restricted by insurance companies to only fill a one or 3 month’s supply.

One study found that FPs were less likely than OB-GYNs to have enough time for contraceptive counseling and fewer than half felt competent inserting IUDs.

In January 2016, Oregon began requiring private and state health insurance providers to reimburse for a year’s supply of prescription contraception; in January 2017, insurers in Washington, DC, were also required to offer women a year’s supply of prescription contraception.^29,30 Several other states have followed suit. The California Health Benefits Review Program estimates a savings of $42.8 million a year from fewer office visits and 15,000 fewer unintended pregnancies if their state enacts a similar policy.³¹

Pharmacist initiatives are worth watching. In January 2016, Oregon pharmacists with additional training were allowed to prescribe OCs and hormonal patches to women 18 years and older.³² In April 2016, a similar law went into effect in California, but without a minimum age requirement and with the additional coverage of vaginal rings and Depo-Provera (depo) injections.³³ Pharmacists in both states must review a health questionnaire completed by the woman and can refer to a physician as necessary.

The CDC recommends that clinicians extend the allowed window for repeat depo injections to 15 weeks.¹⁴ Common institutional protocol is to give repeat injections every 11 to 13 weeks. If past that window, protocol often dictates the woman abstain from unprotected sex for 2 weeks and then return for a negative pregnancy test (or await menses) before the next injection. However, the CDC notes that depo is effective for longer than the 13-week period.¹⁴ No additional birth control or pregnancy testing is needed and the woman can receive the next depo shot if she is up to 15 weeks from the previous shot.

One study found no additional pregnancy risks for those who were up to 4 weeks “late” for their next shot, suggesting there is potential for an even larger grace period.³⁴ The World Health Organization advises allowing a repeat injection up to 4 weeks late.³⁵ We encourage institutions to change their policies to comply with the CDC’s 15-week window.

Another initiative is over-the-counter (OTC) access to OCs, which the American Academy of Family Physicians (AAFP) and ACOG support.^36,37 ACOG notes that “no drug or intervention is completely without risk of harm” and that the risk of venous thromboembolism for OC users is lower than the risk of pregnancy.³⁷ Women can successfully self-screen for contraindications using a checklist. Concerns about women potentially being less adherent or less likely to choose LARCs are not reasons to preclude access to other methods. The AAFP supports insurance coverage of OCs, regardless of prescription status.³⁶

5. Routinely counsel about, and advance-prescribe, emergency contraception pills

Physicians should counsel and advance-prescribe emergency contraception pills (ECPs) to women, including adolescents, using less reliable contraception, as recommended by ACOG, AAP, and the CDC.^14,37,38 It’s also important to provide information on the copper IUD as the most effective method of emergency contraception, with nearly 100% efficacy if placed within 5 days.³⁹ An easy-to-read patient hand-out in English and Spanish on EC options can be found at http://beyondthepill.ucsf.edu/tools-materials.

Only 3% of respondents participating in the 2006-2010 National Survey of Family Growth received counseling about emergency contraception in the past year.⁴⁰ ECPs are most effective when used within 24 hours but have some efficacy up to 5 days.³⁷ Due to the Affordable Care Act, most insurance plans will cover ECPs if purchased with a prescription, but coverage varies by state.⁴¹ Ulipristal acetate (UPA) ECP is only available with a prescription. Advance prescriptions can alleviate financial burdens on women when they need to access ECPs quickly.

Women should wait at least 5 days before resuming or starting hormonal contraception after taking UPA-based ECP, as it may reduce the ovulation-delaying effect of the ECP.¹⁴ For IUDs, implants, and depo, which require a visit to a health care provider, physicians evaluating earlier provision should consider the risks of reduced efficacy against the many barriers to access.

UPA-based ECPs (such as ella) may be more effective for overweight and obese women than levonorgestrel-based ECPs (such as Plan B and Next Choice).¹⁴ Consider advance-prescribing UPA ECPs to women with a body mass index (BMI) >25 kg/m².⁴² Such considerations are important as the prevalence of obesity in women between 2013 and 2014 was 40.4%.⁴³

A copper IUD may be the only effective form of emergency contraception for patients with a BMI ≥35 kg/m².

In May 2016, the FDA noted that while current data are insufficient regarding whether the effectiveness of levonorgestrel ECPs is reduced in overweight or obese women, there are no safety concerns regarding their use in this population.⁴⁴ Therefore, a woman with a BMI >25 kg/m² should use UPA ECPs if available; but if not, she can still use levonorgestrel ECPs. One study, however, has found that UPA ECPs are only as effective as a placebo when BMI is ≥35 kg/m², at which point a copper IUD may be the only effective form of emergency contraception.⁴⁵

Transitioning from customary practices to best practices

Following these practical steps, FPs can improve contraceptive care for women. However, to make a significant impact, clinicians must be willing to change customary practices that are based on tradition, routines, or outdated protocols in favor of those based on current evidence.

One good place to start the transition to best practices is to familiarize yourself with the 2016 US Medical Eligibility Criteria for Contraceptive Use²⁶ and Selected Practice Recommendations for Contraceptive Use.¹⁴ TABLES 2^14,26,46,47 and 3 offer additional resources that can enhance contraceptive counseling and further promote access to contraceptive care.

The contraceptive coverage guarantee under the Affordable Care Act has allowed many women to make contraceptive choices based on personal needs and preferences rather than cost. The new contraceptive coverage exemptions issued under the Trump administration will bring cost back as the driving decision factor for women whose employers choose not to provide contraceptive coverage. Providers should be aware of the typical costs associated with the various contraceptive options offered in their practice and community.

CORRESPONDENCE
Jessica Dalby, MD, Department of Family Medicine and Community Health, University of Wisconsin School of Medicine and Public Health, 1102 South Park St, Suite 100, Madison, WI 53715; [email protected].

While the unintended pregnancy rate for women ages 15 to 44 years decreased by 18% between 2008 and 2011, almost half of pregnancies in the United States remain unintended.¹ On a more positive note, however, women who use birth control consistently and correctly account for only 5% of unintended pregnancies.² As family physicians (FPs), we can support and facilitate our female patients’ efforts to consistently use highly effective forms of contraception. The 5 initiatives detailed here can help toward that end.

1. Routinely screen patients for their reproductive intentions

All women of reproductive age should be screened routinely for their pregnancy intentions. The American College of Obstetricians and Gynecologists (ACOG) encourages clinicians to ask women about pregnancy intendedness and encourages patients to develop a reproductive life plan, or a set of personal goals about whether or when to have children.³ The Centers for Disease Control and Prevention (CDC) has also developed a reproductive life plan tool for health professionals to encourage women and men to reflect upon their plans.⁴ So just as we regularly screen and document cigarette use and blood pressure (BP), so too, should we routinely screen women for their reproductive goals.

Ask women this one question. The Oregon Foundation for Reproductive Health launched the One Key Question Initiative, which proposes that the care team ask women ages 18 to 50: “Would you like to become pregnant in the next year?”⁵ A common workflow includes the medical assistant asking women about pregnancy intentions and providing a preconception and/or contraceptive handout, if appropriate. The physician provides additional counseling as needed. Pilot studies of One Key Question indicate that 30% to 40% of women screened needed follow-up counseling, suggesting the need for clinicians to be proactive in asking about reproductive plans. (Additional information on the Initiative is available on the Foundation’s Web site at http://www.orfrh.org/.)

This approach assumes women feel in control of their reproduction; however, this may not be the reality for many, especially low-income women.⁶ Additionally, women commonly cite planning a pregnancy as appropriate only when they are in an ideal relationship and when they are living in a financially stable environment—conditions that some women may never achieve.

ACOG encourages clinicians to ask women about pregnancy intendedness and encourages patients to develop a reproductive plan.

Another caveat is that women may not have explicit pregnancy intentions, in which case, this particular approach may not be effective. A study of low-income women found only 60% intended to use the method prescribed after contraception counseling, with 37% of those stopping because of adverse effects, 23% saying they wanted another method, and 17% citing method complexity.⁷

Reproductive coercion from male partners, ranging from pressure to become pregnant to method sabotage, is also common in low-income women.⁸ Regular conversations that prioritize a woman’s values and experience are needed to promote reproductive autonomy.

2. Decouple provision of contraception from unnecessary exams

Pelvic exams and pap smears should not be required prior to offering patients hormonal contraception, according to the Choosing Wisely campaign of the American Board of Internal Medicine and ACOG.^9,10 Hormonal contraception may instead be provided safely based on a medical history and BP assessment. Adolescents, minority groups, obese women, and victims of sexual trauma, in particular may avoid asking about birth control because of anxiety and fear of pain from these exams.¹¹ The American College of Physicians recommends against speculum and bimanual exams in asymptomatic, non-pregnant, adult women.¹² Pap smears and sexually transmitted infection (STI) testing should be performed at their normally scheduled intervals as recommended by the US Preventive Services Task Force (USPSTF) and not be tied to contraceptive provision.¹³

Assess pregnancy status using criteria,rather than a pregnancy text

Use the CDC’s criteria to assess pregnancy status rather than relying on a urine pregnancy test prior to providing contraception. Once you are reasonably sure that a woman is not pregnant (TABLE 1¹⁴), contraception may be started. Some physicians have traditionally requested that a woman delay starting contraception until the next menses to ensure that she is not already pregnant. However, given the evidence that hormonal contraception does not cause birth defects, such a delay is not warranted and puts the woman at risk of an unintended pregnancy during the gap.¹⁵

Furthermore, there is an approximate 2-week window in which a woman could have a negative urine pregnancy test despite being pregnant, so the test alone is not completely reliable. In addition, obese women may experience irregular cycles, further complicating the traditional approach.¹⁶

Another largely unnecessary step … The US Selected Practice Recommendations (US SPR) from the CDC notes that additional STI screening prior to an intrauterine device (IUD) insertion is unnecessary for most women if appropriate screening guidelines have been previously followed.¹⁴ For those who have not been screened according to guidelines, the CDC recommends same-day screening and IUD insertion. You can then treat an STI without removing the IUD. Women with purulent cervicitis or a current chlamydial or gonorrheal infection should delay IUD insertion until after treatment.

3. Expand long-acting reversible contraception counseling and access

Offer long-acting reversible contraception (LARC), such as IUDs and implants, as first-line options for most women. ACOG endorses LARC as the most effective reversible method for most women, including those who have not given birth and adolescents.¹⁷ Unfortunately, a 2012 study found that family physicians were less likely than OB-GYNs to have enough time for contraceptive counseling and fewer than half felt competent inserting IUDs.¹⁸ While 79% of OB-GYNs routinely discussed IUDs with their patients, only 47% of family physicians did. In 2014, the American Academy of Pediatrics (AAP) endorsed a LARC-first tiered counseling approach for adolescents.¹⁹

A test of LARC-first counseling

The Contraceptive CHOICE project, a St. Louis, Missouri-based initiative, was launched to reduce unintended pregnancies in women ages 14 to 45 years by offering LARC-first counseling and free contraception of their choice.²⁰ This project involved more than 9000 women at high risk for unintended pregnancy. Same-day LARC insertion was available. Seventy-five percent of women chose a LARC method and they reported greater continuation at 12 and 24 months, when compared to women who did not choose a LARC method. LARC users also reported higher satisfaction at one year. Provision of contraception through the project contributed to a reduction in repeat abortions as well as decreased rates of teenage pregnancy, birth, and abortion. Three years after the start of the project, IUDs had continuation rates of nearly 70%, implants of 56%, and non-LARC methods of 31%.²¹

When counseling women, it’s important to remember that effectiveness may not be the only criterium a woman uses when choosing a method. A 2010 study found that for 91% of women at high risk for unintended pregnancy, no single method possessed all the features they deemed “extremely important.”²² Clinicians should take a patient-centered approach to find birth control that fits each patient’s priorities.

Clinicians need proper training in LARC methods

Only 20% of FPs regularly insert IUDs, and 11% offer contraceptive implants, according to estimates from physicians recertifying with the American Board of Family Medicine in 2014.²³ Access to training during residency is a key component to increasing these rates. FPs who practice obstetrics should be trained in postpartum LARC insertion and offer this option prior to hospital discharge as well as during the postpartum office visit.

Performing LARC insertions on the same day as counseling is ideal, and clinics should strive to reduce barriers to same-day procedures. Time constraints may be addressed by shifting tasks among the medical team. In the CHOICE project, contraceptive counselors—half of whom had no clinical experience—were trained to provide tiered counseling to participants. By working with a cross-trained health care team and offering prepared resources, clinicians can save time and improve access.

Physicians may want to incorporate the free online resources Bedsider.org or Stayteen.org to help women learn about contraceptive methods.²⁴ The user-friendly Web sites, operated by the National Campaign to Prevent Teen and Unplanned Pregnancy, describe various forms of contraception and offer text and email reminders. Incorporating Bedsider into the counseling workflow and discussing the various reminder tools available may improve patients’ knowledge and enhance their compliance.

Pelvic exams and pap smears should not be required before offering patients hormonal contraception.

Additional barriers for practices may include high upfront costs associated with stocking devices. Practices that may be unable to sustain the costs surrounding enhanced contraception counseling and provision can collaborate with family planning clinics that are able to offer same-day services. A study of clinics in California found that Title X clinics were more likely to provide on-site LARC services than non-Title X public and private providers.²⁵

4. Follow CDC guidelines for initiating and continuing contraception

Follow the US SPR for guidance on initiating and continuing contraceptive methods.¹⁴ The CDC’s Medical Eligibility Criteria for Contraceptive Use is another vital resource, providing recommendations for contraceptive methods to patients who have specific medical conditions or characteristics.²⁶

Utilize the “quick start” method for hormonal contraception, where birth control is started on the same day as its prescription regardless of timing of the menstrual cycle. If you can’t be reasonably certain that a woman is not pregnant based on the criteria listed in TABLE 1,¹⁴ conduct a pregnancy test (while recognizing the aforementioned 2-week window of limitations) and counsel the patient to use back-up protection for the first 7 days along with repeating a pregnancy test in 2 weeks’ time.

The quick start method may lead to higher adherence than delayed initiation.²⁷ Differences in continuation rates between women who use the quick start method and those who follow the delayed approach may disappear over time.²⁸

Prescribe and provide a year’s supply of oral contraceptive pills (OCPs) as recommended by the CDC US SPR.¹⁴ It is important to note that pharmacists are usually restricted by insurance companies to only fill a one or 3 month’s supply.

One study found that FPs were less likely than OB-GYNs to have enough time for contraceptive counseling and fewer than half felt competent inserting IUDs.

In January 2016, Oregon began requiring private and state health insurance providers to reimburse for a year’s supply of prescription contraception; in January 2017, insurers in Washington, DC, were also required to offer women a year’s supply of prescription contraception.^29,30 Several other states have followed suit. The California Health Benefits Review Program estimates a savings of $42.8 million a year from fewer office visits and 15,000 fewer unintended pregnancies if their state enacts a similar policy.³¹

Pharmacist initiatives are worth watching. In January 2016, Oregon pharmacists with additional training were allowed to prescribe OCs and hormonal patches to women 18 years and older.³² In April 2016, a similar law went into effect in California, but without a minimum age requirement and with the additional coverage of vaginal rings and Depo-Provera (depo) injections.³³ Pharmacists in both states must review a health questionnaire completed by the woman and can refer to a physician as necessary.

The CDC recommends that clinicians extend the allowed window for repeat depo injections to 15 weeks.¹⁴ Common institutional protocol is to give repeat injections every 11 to 13 weeks. If past that window, protocol often dictates the woman abstain from unprotected sex for 2 weeks and then return for a negative pregnancy test (or await menses) before the next injection. However, the CDC notes that depo is effective for longer than the 13-week period.¹⁴ No additional birth control or pregnancy testing is needed and the woman can receive the next depo shot if she is up to 15 weeks from the previous shot.

One study found no additional pregnancy risks for those who were up to 4 weeks “late” for their next shot, suggesting there is potential for an even larger grace period.³⁴ The World Health Organization advises allowing a repeat injection up to 4 weeks late.³⁵ We encourage institutions to change their policies to comply with the CDC’s 15-week window.

Another initiative is over-the-counter (OTC) access to OCs, which the American Academy of Family Physicians (AAFP) and ACOG support.^36,37 ACOG notes that “no drug or intervention is completely without risk of harm” and that the risk of venous thromboembolism for OC users is lower than the risk of pregnancy.³⁷ Women can successfully self-screen for contraindications using a checklist. Concerns about women potentially being less adherent or less likely to choose LARCs are not reasons to preclude access to other methods. The AAFP supports insurance coverage of OCs, regardless of prescription status.³⁶

5. Routinely counsel about, and advance-prescribe, emergency contraception pills

Physicians should counsel and advance-prescribe emergency contraception pills (ECPs) to women, including adolescents, using less reliable contraception, as recommended by ACOG, AAP, and the CDC.^14,37,38 It’s also important to provide information on the copper IUD as the most effective method of emergency contraception, with nearly 100% efficacy if placed within 5 days.³⁹ An easy-to-read patient hand-out in English and Spanish on EC options can be found at http://beyondthepill.ucsf.edu/tools-materials.

Only 3% of respondents participating in the 2006-2010 National Survey of Family Growth received counseling about emergency contraception in the past year.⁴⁰ ECPs are most effective when used within 24 hours but have some efficacy up to 5 days.³⁷ Due to the Affordable Care Act, most insurance plans will cover ECPs if purchased with a prescription, but coverage varies by state.⁴¹ Ulipristal acetate (UPA) ECP is only available with a prescription. Advance prescriptions can alleviate financial burdens on women when they need to access ECPs quickly.

Women should wait at least 5 days before resuming or starting hormonal contraception after taking UPA-based ECP, as it may reduce the ovulation-delaying effect of the ECP.¹⁴ For IUDs, implants, and depo, which require a visit to a health care provider, physicians evaluating earlier provision should consider the risks of reduced efficacy against the many barriers to access.

UPA-based ECPs (such as ella) may be more effective for overweight and obese women than levonorgestrel-based ECPs (such as Plan B and Next Choice).¹⁴ Consider advance-prescribing UPA ECPs to women with a body mass index (BMI) >25 kg/m².⁴² Such considerations are important as the prevalence of obesity in women between 2013 and 2014 was 40.4%.⁴³

A copper IUD may be the only effective form of emergency contraception for patients with a BMI ≥35 kg/m².

In May 2016, the FDA noted that while current data are insufficient regarding whether the effectiveness of levonorgestrel ECPs is reduced in overweight or obese women, there are no safety concerns regarding their use in this population.⁴⁴ Therefore, a woman with a BMI >25 kg/m² should use UPA ECPs if available; but if not, she can still use levonorgestrel ECPs. One study, however, has found that UPA ECPs are only as effective as a placebo when BMI is ≥35 kg/m², at which point a copper IUD may be the only effective form of emergency contraception.⁴⁵

Transitioning from customary practices to best practices

Following these practical steps, FPs can improve contraceptive care for women. However, to make a significant impact, clinicians must be willing to change customary practices that are based on tradition, routines, or outdated protocols in favor of those based on current evidence.

One good place to start the transition to best practices is to familiarize yourself with the 2016 US Medical Eligibility Criteria for Contraceptive Use²⁶ and Selected Practice Recommendations for Contraceptive Use.¹⁴ TABLES 2^14,26,46,47 and 3 offer additional resources that can enhance contraceptive counseling and further promote access to contraceptive care.

The contraceptive coverage guarantee under the Affordable Care Act has allowed many women to make contraceptive choices based on personal needs and preferences rather than cost. The new contraceptive coverage exemptions issued under the Trump administration will bring cost back as the driving decision factor for women whose employers choose not to provide contraceptive coverage. Providers should be aware of the typical costs associated with the various contraceptive options offered in their practice and community.

CORRESPONDENCE
Jessica Dalby, MD, Department of Family Medicine and Community Health, University of Wisconsin School of Medicine and Public Health, 1102 South Park St, Suite 100, Madison, WI 53715; [email protected].

Meniscectomy is the most common orthopedic procedure performed in the United States with 700,000 meniscectomies performed every year.¹ More than half of these procedures are performed in patients ≥45 years of age,² giving rise to the question: Does arthroscopic surgery have a role in the treatment of patients who may have osteoarthritis (OA) and another knee condition, such as a symptomatic meniscal tear? Determining the answer is especially important when you consider that the number of true and incidental tears diagnosed on magnetic resonance imaging (MRI) has been on the rise³—a result of the routine use of MRIs to identify the cause of patients’ chronic knee pain.

At a cost of roughly $5,000 per procedure, some experts have suggested that at least a portion of the approximately $4 billion annual direct medical costs associated with meniscectomy could be put to better use.^4,5 This prompted us to wonder what the literature tells us about the management of degenerative meniscus tears in middle-aged patients with OA and whether these patients would benefit from nonoperative management with optimized physical and medical therapy as a first-line approach. Our findings follow.

But first, a word about the connection between OA and meniscal tears.

What we’ve learned about meniscal damage

Research has shown that over one-third of individuals >50 years of age and three-quarters of people with knee OA have degenerative meniscal tears.⁶ In the past, the relative paucity of epidemiologic data on the prevalence of meniscal tears in the general population made it difficult to interpret the diagnostic information provided by MRI.

More recently, experts found that meniscal damage is especially prevalent among individuals with OA, and they began treating with arthroscopic partial meniscectomy (APM), as the meniscal damage was thought to be the anatomical foundation for the complaint of knee pain.⁷

However, researchers then began realizing that many patients with findings of a meniscal tear visualized on MRI reported no knee symptoms. In one study, adults in a large community-based sample found to have a meniscal tear on MRI were no more likely to have knee pain than subjects without a meniscal tear.⁶ Similarly, subjects with a meniscal tear and OA had no more severe pain than subjects with OA and no meniscal tear.⁶

Over one-third of people >50 years of age and three-quarters of people with knee osteoarthritis have degenerative meniscal tears.

In addition, the landmark Fairbank study from 1948⁸ and others since have shown that meniscectomy can lead to other problems. Removal of meniscal tissue decreases the contact stress area, which increases stress on the articular cartilage, and inevitably leads to degeneration of the involved joint.⁹ Researchers have shown that even partial meniscectomy produces late articular cartilage changes.¹⁰

Which interventions and when?

We conducted an in-depth literature review to determine which approaches were best for the treatment of OA and meniscal tears, and summarize our findings below, according to OA severity. (For details of how the literature review was conducted and an at-a-glance summary of the key findings, see the TABLE.^4,5,11-17) Of note: All of the studies reviewed here included patients with chronic knee pain and excluded patients with sudden onset pain from a single physical event.

The findings: Early OA and meniscal tears

The first 2 studies we identified in the literature, both published by Herrlin et al,^11,12 examined the efficacy of APM in middle-aged patients with early OA (≤grade 1), according to the Ahlbäck classification.

In the first Herrlin study, a 6-month prospective randomized trial, 90 middle-aged patients with a medial meniscal tear (without traumatic history) were assigned to either APM followed by supervised exercise or 8 weeks of supervised exercise alone.¹¹ Exercise consisted of activities for improving muscle strength, endurance, and flexibility, as well as for balance and proprioception. The authors concluded that a combination of APM and supervised exercise did not lead to greater improvements in knee function compared with supervised exercise alone.¹¹

In the second Herrlin study, a prospective randomized study involving 96 middle-aged patients with an MRI-verified medial meniscal tear and radiographic OA, the authors concluded that arthroscopic surgery followed by exercise therapy was not superior to the same exercise therapy alone.¹² The results were gleaned from both patient-reported outcomes and radiographic assessment at 2 and 5 years. Both groups reported significant improvements at 5 years, but participants did not reach the level of fitness and quality of life of similarly-aged healthy controls.

Perhaps one of the most interesting aspects of this study was that approximately one-third of patients from the exercise-only group still had disabling knee symptoms after exercise therapy, but improved to the same level as the rest of the patients after crossing over and undergoing APM.¹² Part of the observed benefit of arthroscopy in these patients has the potential to be explained by the placebo effect, especially given that invasive procedures have a stronger placebo effect than do noninvasive ones, and due to the lack of blinding.¹⁸ Additionally, limitations of the above studies include small sample sizes, lack of a control group, and short-term follow-up.

Next, a 2013 study by Yim et al looked specifically at APM vs nonoperative treatment with strengthening exercises.¹³ A total of 102 patients with an average age of 53.8 years, a Kellgren-Lawrence Classification of Osteoarthritis of <2, and an MRI-confirmed degenerative horizontal tear of the posterior horn of the medial meniscus were randomized to the 2 intervention groups. The 2 groups were highly comparable, giving the study high internal validity. These patients were then assessed at 3 months, one year, and at 2 years after treatment.

Although most patients at the outset of the study had intense knee pain with mechanical symptoms, both groups reported a decrease in experienced knee pain, improved function, and a high level of satisfaction with their prescribed treatment, with no significant difference in any of these values after 2 years of follow-up.¹³ A limitation of the study was that it used subjective questionnaires to assess pain, swelling, and activities of daily living (ADLs).

A fourth study, a 2013 multicenter, randomized, sham-controlled trial, looked at 146 patients ages 35 to 65 years who had knee symptoms consistent with a degenerative medial meniscus tear (confirmed by MRI and arthroscopic evaluation) and no knee arthritis.⁴ The subjects were assigned to either APM or sham surgery (skin incisions only). The results showed that APM was not superior to sham surgery with regard to outcomes assessed during a 12-month follow-up period.⁴

Most recently (2016), Kise et al¹⁴ published the results of a randomized controlled superiority trial conducted in Norway comparing 12 weeks of supervised exercise therapy with APM for patients with degenerative meniscus tears. Their study included 140 patients ages 36 to 60 years. Notably, most (96%), but not all, of their patients had no radiographic evidence of OA.

Many patients with MRI-verified meniscal tears have no symptoms, and many with OA and tears have no worse symptoms than those with OA alone.

At the 2-year follow-up, there were no differences in patient-reported pain or functional outcomes, with all patients improving significantly from baseline. Muscle strength was also measured and found to be significantly greater at 3 and 12 months in the exercise group. Limitations of this study were a lack of patient blinding and a 19% crossover from the exercise to the APM group.

The findings: Mild to moderate OA and meniscal tears

The authors of a 2002 double-blind, placebo-controlled trial randomly assigned 180 patients with degenerative meniscus tears and knee OA to either arthroscopic debridement, arthroscopic lavage, or placebo surgery consisting of superficial skin incisions without insertion of an arthroscope.⁵ Patients were eligible if they were ≤75 years of age and had not undergone arthroscopy of the knee within the previous 2 years. Arthritis was graded utilizing the Kellgren-Lawrence scale (0-4) and by calculating a severity score (0-12) by adding the individual scores for each of the 3 compartments of the knee; patients were excluded if they had a severity grade ≥9. One-quarter of the participants had severe arthritis, with scores of 7 or 8. Outcomes were measured over a 24-month period. At no point did either of the intervention groups report less pain or better function than the sham-surgery group.⁵

Additionally, in a single-center RCT involving 178 patients with mild to severe OA, subjects were randomly assigned to surgical lavage and arthroscopic debridement together with optimized physical and medical therapy or to treatment with physical and medical therapy alone.¹⁵ Patients were excluded if they had “bucket handle” meniscal tears detected by physical exam or by MRI, previous arthroscopic surgery, previous major knee trauma, or steroid injections in the last 3 months. All participants were required to have Kellgren-Lawrence grade 2, 3, or 4 arthritis.

The researchers used 2 validated outcome measures to evaluate pain, symptoms, and functional ability and followed the patients for 2 years after the initiation of treatment. This study failed to show that arthroscopic surgery provided any additional benefit to optimized physical and medical therapy.¹⁵

A 2013 multicenter, RCT randomly assigned 351 patients ≥45 years with a meniscal tear and radiographic evidence of mild to moderate OA to either surgery and postoperative physical therapy or to a standardized physical therapy regimen.¹⁶ Patients were required to have at least one symptom that was consistent with a meniscal tear for approximately one month. About 30% of the patients who were assigned to physical therapy alone underwent surgery within 6 months. There were no significant differences between the 2 groups in the magnitude of improvement in functional status or pain at 6 or 12 months.

Finally, a prospective Scottish study consisting of 270 patients with everything from no signs of OA to advanced OA who underwent APM were sent preoperative and 6-month postop questionnaires evaluating ADLs, pain, symptoms, quality of life, and body mass index (BMI).¹⁷ Their OA was graded via preop MRI or radiographs and confirmed by the operating surgeon. The investigators were unable to demonstrate any significant benefit associated with arthroscopic meniscectomy, and, therefore, could not recommend the procedure for patients with moderate to advanced OA.

However, in their analysis, 3 subgroup populations were found to benefit from APM: those with greater body habitus (BMI >30 kg/m²), those without signs of OA, and those with early OA. Limitations of this study included the lack of randomization, blinding, a control, and long-term follow up, and that the authors didn’t use established OA grading criteria.¹⁷

The bottom line: Nonoperative treatment benefits most patients

Physical therapy is an appropriate first-line treatment for degenerative meniscus tears in middle-aged patients. In fact, a trial of nonoperative treatment is likely to benefit the majority of patients. In addition, avoiding surgery eliminates surgical complications and decreases overall health care costs.

Reserve APM for those patients without significant OA who fail to improve after physical therapy, who have mechanical symptoms, or who have intra-articular loose bodies.

Consider a trial of nonoperative treatment, as it is likely to benefit the majority of patients with degenerative meniscal tears.

In addition, exercise therapy is an effective treatment for patients with knee OA. It improves function and limits joint pain in both acute arthritic flares and more long-term, chronic situations. There is strong evidence that strengthening plays a critical role in reducing symptoms and improving muscle strength, physical ability, and quality of life.¹⁹ It has been suggested that physical exercise 3 times a week for 4 months could lead to >35% improvement of knee function.²⁰ In contrast, other studies reported that while 91% of patients 11.5 years after APM considered their knees “normal or almost normal,” patients actually experienced a reduction in postoperative physical activity and quality of life.²¹

The most recent American Academy of Orthopaedic Surgeons (AAOS) guidelines²² do not recommend for or against arthroscopic partial meniscectomy in patients with knee OA and a torn meniscus. In middle-aged patients, MRI abnormalities of the meniscus do not consistently correlate with symptoms. Many meniscus lesions are asymptomatic or not the primary source of pain in the setting of OA.³

Potential harms and considerations. Deep venous thrombosis is the most frequently reported adverse event of arthroscopic surgery, with an approximate incidence of 4.13 per 1000 procedures, followed by less frequent complications such as infection, pulmonary embolism, and death.¹⁸

It is important to note that the degenerative meniscus tears that occur in middle age and that are associated with OA are not the same as acute, traumatic meniscus tears. All of the studies discussed here included patients with chronic knee pain and excluded patients with sudden onset pain from a single physical event. Many of the studies excluded patients with bucket-handle tears or severe mechanical symptoms (ie, locking). APM may be indicated for these meniscus tears regardless of age or OA status.

Making a sensible choice. Ultimately, physicians and their patients must use the best evidence available to make sensible clinical decisions. The ability to retain native meniscal tissue is of utmost importance to maintaining the longevity of their knee. According to previous studies, OA progression is more likely to occur after meniscectomy than after nonoperative treatment.²³

CORRESPONDENCE
William Bassett, MD, Department of Orthopedics, Rutgers-Robert Wood Johnson Medical School, 51 French St., PO Box 19, New Brunswick, NJ 08903; [email protected]. 

Meniscectomy is the most common orthopedic procedure performed in the United States with 700,000 meniscectomies performed every year.¹ More than half of these procedures are performed in patients ≥45 years of age,² giving rise to the question: Does arthroscopic surgery have a role in the treatment of patients who may have osteoarthritis (OA) and another knee condition, such as a symptomatic meniscal tear? Determining the answer is especially important when you consider that the number of true and incidental tears diagnosed on magnetic resonance imaging (MRI) has been on the rise³—a result of the routine use of MRIs to identify the cause of patients’ chronic knee pain.

At a cost of roughly $5,000 per procedure, some experts have suggested that at least a portion of the approximately $4 billion annual direct medical costs associated with meniscectomy could be put to better use.^4,5 This prompted us to wonder what the literature tells us about the management of degenerative meniscus tears in middle-aged patients with OA and whether these patients would benefit from nonoperative management with optimized physical and medical therapy as a first-line approach. Our findings follow.

But first, a word about the connection between OA and meniscal tears.

What we’ve learned about meniscal damage

Research has shown that over one-third of individuals >50 years of age and three-quarters of people with knee OA have degenerative meniscal tears.⁶ In the past, the relative paucity of epidemiologic data on the prevalence of meniscal tears in the general population made it difficult to interpret the diagnostic information provided by MRI.

More recently, experts found that meniscal damage is especially prevalent among individuals with OA, and they began treating with arthroscopic partial meniscectomy (APM), as the meniscal damage was thought to be the anatomical foundation for the complaint of knee pain.⁷

However, researchers then began realizing that many patients with findings of a meniscal tear visualized on MRI reported no knee symptoms. In one study, adults in a large community-based sample found to have a meniscal tear on MRI were no more likely to have knee pain than subjects without a meniscal tear.⁶ Similarly, subjects with a meniscal tear and OA had no more severe pain than subjects with OA and no meniscal tear.⁶

Over one-third of people >50 years of age and three-quarters of people with knee osteoarthritis have degenerative meniscal tears.

In addition, the landmark Fairbank study from 1948⁸ and others since have shown that meniscectomy can lead to other problems. Removal of meniscal tissue decreases the contact stress area, which increases stress on the articular cartilage, and inevitably leads to degeneration of the involved joint.⁹ Researchers have shown that even partial meniscectomy produces late articular cartilage changes.¹⁰

Which interventions and when?

We conducted an in-depth literature review to determine which approaches were best for the treatment of OA and meniscal tears, and summarize our findings below, according to OA severity. (For details of how the literature review was conducted and an at-a-glance summary of the key findings, see the TABLE.^4,5,11-17) Of note: All of the studies reviewed here included patients with chronic knee pain and excluded patients with sudden onset pain from a single physical event.

The findings: Early OA and meniscal tears

The first 2 studies we identified in the literature, both published by Herrlin et al,^11,12 examined the efficacy of APM in middle-aged patients with early OA (≤grade 1), according to the Ahlbäck classification.

In the first Herrlin study, a 6-month prospective randomized trial, 90 middle-aged patients with a medial meniscal tear (without traumatic history) were assigned to either APM followed by supervised exercise or 8 weeks of supervised exercise alone.¹¹ Exercise consisted of activities for improving muscle strength, endurance, and flexibility, as well as for balance and proprioception. The authors concluded that a combination of APM and supervised exercise did not lead to greater improvements in knee function compared with supervised exercise alone.¹¹

In the second Herrlin study, a prospective randomized study involving 96 middle-aged patients with an MRI-verified medial meniscal tear and radiographic OA, the authors concluded that arthroscopic surgery followed by exercise therapy was not superior to the same exercise therapy alone.¹² The results were gleaned from both patient-reported outcomes and radiographic assessment at 2 and 5 years. Both groups reported significant improvements at 5 years, but participants did not reach the level of fitness and quality of life of similarly-aged healthy controls.

Perhaps one of the most interesting aspects of this study was that approximately one-third of patients from the exercise-only group still had disabling knee symptoms after exercise therapy, but improved to the same level as the rest of the patients after crossing over and undergoing APM.¹² Part of the observed benefit of arthroscopy in these patients has the potential to be explained by the placebo effect, especially given that invasive procedures have a stronger placebo effect than do noninvasive ones, and due to the lack of blinding.¹⁸ Additionally, limitations of the above studies include small sample sizes, lack of a control group, and short-term follow-up.

Next, a 2013 study by Yim et al looked specifically at APM vs nonoperative treatment with strengthening exercises.¹³ A total of 102 patients with an average age of 53.8 years, a Kellgren-Lawrence Classification of Osteoarthritis of <2, and an MRI-confirmed degenerative horizontal tear of the posterior horn of the medial meniscus were randomized to the 2 intervention groups. The 2 groups were highly comparable, giving the study high internal validity. These patients were then assessed at 3 months, one year, and at 2 years after treatment.

Although most patients at the outset of the study had intense knee pain with mechanical symptoms, both groups reported a decrease in experienced knee pain, improved function, and a high level of satisfaction with their prescribed treatment, with no significant difference in any of these values after 2 years of follow-up.¹³ A limitation of the study was that it used subjective questionnaires to assess pain, swelling, and activities of daily living (ADLs).

A fourth study, a 2013 multicenter, randomized, sham-controlled trial, looked at 146 patients ages 35 to 65 years who had knee symptoms consistent with a degenerative medial meniscus tear (confirmed by MRI and arthroscopic evaluation) and no knee arthritis.⁴ The subjects were assigned to either APM or sham surgery (skin incisions only). The results showed that APM was not superior to sham surgery with regard to outcomes assessed during a 12-month follow-up period.⁴

Most recently (2016), Kise et al¹⁴ published the results of a randomized controlled superiority trial conducted in Norway comparing 12 weeks of supervised exercise therapy with APM for patients with degenerative meniscus tears. Their study included 140 patients ages 36 to 60 years. Notably, most (96%), but not all, of their patients had no radiographic evidence of OA.

Many patients with MRI-verified meniscal tears have no symptoms, and many with OA and tears have no worse symptoms than those with OA alone.

At the 2-year follow-up, there were no differences in patient-reported pain or functional outcomes, with all patients improving significantly from baseline. Muscle strength was also measured and found to be significantly greater at 3 and 12 months in the exercise group. Limitations of this study were a lack of patient blinding and a 19% crossover from the exercise to the APM group.

The findings: Mild to moderate OA and meniscal tears

The authors of a 2002 double-blind, placebo-controlled trial randomly assigned 180 patients with degenerative meniscus tears and knee OA to either arthroscopic debridement, arthroscopic lavage, or placebo surgery consisting of superficial skin incisions without insertion of an arthroscope.⁵ Patients were eligible if they were ≤75 years of age and had not undergone arthroscopy of the knee within the previous 2 years. Arthritis was graded utilizing the Kellgren-Lawrence scale (0-4) and by calculating a severity score (0-12) by adding the individual scores for each of the 3 compartments of the knee; patients were excluded if they had a severity grade ≥9. One-quarter of the participants had severe arthritis, with scores of 7 or 8. Outcomes were measured over a 24-month period. At no point did either of the intervention groups report less pain or better function than the sham-surgery group.⁵

Additionally, in a single-center RCT involving 178 patients with mild to severe OA, subjects were randomly assigned to surgical lavage and arthroscopic debridement together with optimized physical and medical therapy or to treatment with physical and medical therapy alone.¹⁵ Patients were excluded if they had “bucket handle” meniscal tears detected by physical exam or by MRI, previous arthroscopic surgery, previous major knee trauma, or steroid injections in the last 3 months. All participants were required to have Kellgren-Lawrence grade 2, 3, or 4 arthritis.

The researchers used 2 validated outcome measures to evaluate pain, symptoms, and functional ability and followed the patients for 2 years after the initiation of treatment. This study failed to show that arthroscopic surgery provided any additional benefit to optimized physical and medical therapy.¹⁵

A 2013 multicenter, RCT randomly assigned 351 patients ≥45 years with a meniscal tear and radiographic evidence of mild to moderate OA to either surgery and postoperative physical therapy or to a standardized physical therapy regimen.¹⁶ Patients were required to have at least one symptom that was consistent with a meniscal tear for approximately one month. About 30% of the patients who were assigned to physical therapy alone underwent surgery within 6 months. There were no significant differences between the 2 groups in the magnitude of improvement in functional status or pain at 6 or 12 months.

Finally, a prospective Scottish study consisting of 270 patients with everything from no signs of OA to advanced OA who underwent APM were sent preoperative and 6-month postop questionnaires evaluating ADLs, pain, symptoms, quality of life, and body mass index (BMI).¹⁷ Their OA was graded via preop MRI or radiographs and confirmed by the operating surgeon. The investigators were unable to demonstrate any significant benefit associated with arthroscopic meniscectomy, and, therefore, could not recommend the procedure for patients with moderate to advanced OA.

However, in their analysis, 3 subgroup populations were found to benefit from APM: those with greater body habitus (BMI >30 kg/m²), those without signs of OA, and those with early OA. Limitations of this study included the lack of randomization, blinding, a control, and long-term follow up, and that the authors didn’t use established OA grading criteria.¹⁷

The bottom line: Nonoperative treatment benefits most patients

Physical therapy is an appropriate first-line treatment for degenerative meniscus tears in middle-aged patients. In fact, a trial of nonoperative treatment is likely to benefit the majority of patients. In addition, avoiding surgery eliminates surgical complications and decreases overall health care costs.

Reserve APM for those patients without significant OA who fail to improve after physical therapy, who have mechanical symptoms, or who have intra-articular loose bodies.

Consider a trial of nonoperative treatment, as it is likely to benefit the majority of patients with degenerative meniscal tears.

In addition, exercise therapy is an effective treatment for patients with knee OA. It improves function and limits joint pain in both acute arthritic flares and more long-term, chronic situations. There is strong evidence that strengthening plays a critical role in reducing symptoms and improving muscle strength, physical ability, and quality of life.¹⁹ It has been suggested that physical exercise 3 times a week for 4 months could lead to >35% improvement of knee function.²⁰ In contrast, other studies reported that while 91% of patients 11.5 years after APM considered their knees “normal or almost normal,” patients actually experienced a reduction in postoperative physical activity and quality of life.²¹

The most recent American Academy of Orthopaedic Surgeons (AAOS) guidelines²² do not recommend for or against arthroscopic partial meniscectomy in patients with knee OA and a torn meniscus. In middle-aged patients, MRI abnormalities of the meniscus do not consistently correlate with symptoms. Many meniscus lesions are asymptomatic or not the primary source of pain in the setting of OA.³

Potential harms and considerations. Deep venous thrombosis is the most frequently reported adverse event of arthroscopic surgery, with an approximate incidence of 4.13 per 1000 procedures, followed by less frequent complications such as infection, pulmonary embolism, and death.¹⁸

It is important to note that the degenerative meniscus tears that occur in middle age and that are associated with OA are not the same as acute, traumatic meniscus tears. All of the studies discussed here included patients with chronic knee pain and excluded patients with sudden onset pain from a single physical event. Many of the studies excluded patients with bucket-handle tears or severe mechanical symptoms (ie, locking). APM may be indicated for these meniscus tears regardless of age or OA status.

Making a sensible choice. Ultimately, physicians and their patients must use the best evidence available to make sensible clinical decisions. The ability to retain native meniscal tissue is of utmost importance to maintaining the longevity of their knee. According to previous studies, OA progression is more likely to occur after meniscectomy than after nonoperative treatment.²³

CORRESPONDENCE
William Bassett, MD, Department of Orthopedics, Rutgers-Robert Wood Johnson Medical School, 51 French St., PO Box 19, New Brunswick, NJ 08903; [email protected]. 

Meniscectomy is the most common orthopedic procedure performed in the United States with 700,000 meniscectomies performed every year.¹ More than half of these procedures are performed in patients ≥45 years of age,² giving rise to the question: Does arthroscopic surgery have a role in the treatment of patients who may have osteoarthritis (OA) and another knee condition, such as a symptomatic meniscal tear? Determining the answer is especially important when you consider that the number of true and incidental tears diagnosed on magnetic resonance imaging (MRI) has been on the rise³—a result of the routine use of MRIs to identify the cause of patients’ chronic knee pain.

At a cost of roughly $5,000 per procedure, some experts have suggested that at least a portion of the approximately $4 billion annual direct medical costs associated with meniscectomy could be put to better use.^4,5 This prompted us to wonder what the literature tells us about the management of degenerative meniscus tears in middle-aged patients with OA and whether these patients would benefit from nonoperative management with optimized physical and medical therapy as a first-line approach. Our findings follow.

But first, a word about the connection between OA and meniscal tears.

What we’ve learned about meniscal damage

Research has shown that over one-third of individuals >50 years of age and three-quarters of people with knee OA have degenerative meniscal tears.⁶ In the past, the relative paucity of epidemiologic data on the prevalence of meniscal tears in the general population made it difficult to interpret the diagnostic information provided by MRI.

More recently, experts found that meniscal damage is especially prevalent among individuals with OA, and they began treating with arthroscopic partial meniscectomy (APM), as the meniscal damage was thought to be the anatomical foundation for the complaint of knee pain.⁷

However, researchers then began realizing that many patients with findings of a meniscal tear visualized on MRI reported no knee symptoms. In one study, adults in a large community-based sample found to have a meniscal tear on MRI were no more likely to have knee pain than subjects without a meniscal tear.⁶ Similarly, subjects with a meniscal tear and OA had no more severe pain than subjects with OA and no meniscal tear.⁶

Over one-third of people >50 years of age and three-quarters of people with knee osteoarthritis have degenerative meniscal tears.

In addition, the landmark Fairbank study from 1948⁸ and others since have shown that meniscectomy can lead to other problems. Removal of meniscal tissue decreases the contact stress area, which increases stress on the articular cartilage, and inevitably leads to degeneration of the involved joint.⁹ Researchers have shown that even partial meniscectomy produces late articular cartilage changes.¹⁰

Which interventions and when?

We conducted an in-depth literature review to determine which approaches were best for the treatment of OA and meniscal tears, and summarize our findings below, according to OA severity. (For details of how the literature review was conducted and an at-a-glance summary of the key findings, see the TABLE.^4,5,11-17) Of note: All of the studies reviewed here included patients with chronic knee pain and excluded patients with sudden onset pain from a single physical event.

The findings: Early OA and meniscal tears

The first 2 studies we identified in the literature, both published by Herrlin et al,^11,12 examined the efficacy of APM in middle-aged patients with early OA (≤grade 1), according to the Ahlbäck classification.

In the first Herrlin study, a 6-month prospective randomized trial, 90 middle-aged patients with a medial meniscal tear (without traumatic history) were assigned to either APM followed by supervised exercise or 8 weeks of supervised exercise alone.¹¹ Exercise consisted of activities for improving muscle strength, endurance, and flexibility, as well as for balance and proprioception. The authors concluded that a combination of APM and supervised exercise did not lead to greater improvements in knee function compared with supervised exercise alone.¹¹

In the second Herrlin study, a prospective randomized study involving 96 middle-aged patients with an MRI-verified medial meniscal tear and radiographic OA, the authors concluded that arthroscopic surgery followed by exercise therapy was not superior to the same exercise therapy alone.¹² The results were gleaned from both patient-reported outcomes and radiographic assessment at 2 and 5 years. Both groups reported significant improvements at 5 years, but participants did not reach the level of fitness and quality of life of similarly-aged healthy controls.

Perhaps one of the most interesting aspects of this study was that approximately one-third of patients from the exercise-only group still had disabling knee symptoms after exercise therapy, but improved to the same level as the rest of the patients after crossing over and undergoing APM.¹² Part of the observed benefit of arthroscopy in these patients has the potential to be explained by the placebo effect, especially given that invasive procedures have a stronger placebo effect than do noninvasive ones, and due to the lack of blinding.¹⁸ Additionally, limitations of the above studies include small sample sizes, lack of a control group, and short-term follow-up.

Next, a 2013 study by Yim et al looked specifically at APM vs nonoperative treatment with strengthening exercises.¹³ A total of 102 patients with an average age of 53.8 years, a Kellgren-Lawrence Classification of Osteoarthritis of <2, and an MRI-confirmed degenerative horizontal tear of the posterior horn of the medial meniscus were randomized to the 2 intervention groups. The 2 groups were highly comparable, giving the study high internal validity. These patients were then assessed at 3 months, one year, and at 2 years after treatment.

Although most patients at the outset of the study had intense knee pain with mechanical symptoms, both groups reported a decrease in experienced knee pain, improved function, and a high level of satisfaction with their prescribed treatment, with no significant difference in any of these values after 2 years of follow-up.¹³ A limitation of the study was that it used subjective questionnaires to assess pain, swelling, and activities of daily living (ADLs).

A fourth study, a 2013 multicenter, randomized, sham-controlled trial, looked at 146 patients ages 35 to 65 years who had knee symptoms consistent with a degenerative medial meniscus tear (confirmed by MRI and arthroscopic evaluation) and no knee arthritis.⁴ The subjects were assigned to either APM or sham surgery (skin incisions only). The results showed that APM was not superior to sham surgery with regard to outcomes assessed during a 12-month follow-up period.⁴

Most recently (2016), Kise et al¹⁴ published the results of a randomized controlled superiority trial conducted in Norway comparing 12 weeks of supervised exercise therapy with APM for patients with degenerative meniscus tears. Their study included 140 patients ages 36 to 60 years. Notably, most (96%), but not all, of their patients had no radiographic evidence of OA.

Many patients with MRI-verified meniscal tears have no symptoms, and many with OA and tears have no worse symptoms than those with OA alone.

At the 2-year follow-up, there were no differences in patient-reported pain or functional outcomes, with all patients improving significantly from baseline. Muscle strength was also measured and found to be significantly greater at 3 and 12 months in the exercise group. Limitations of this study were a lack of patient blinding and a 19% crossover from the exercise to the APM group.

The findings: Mild to moderate OA and meniscal tears

The authors of a 2002 double-blind, placebo-controlled trial randomly assigned 180 patients with degenerative meniscus tears and knee OA to either arthroscopic debridement, arthroscopic lavage, or placebo surgery consisting of superficial skin incisions without insertion of an arthroscope.⁵ Patients were eligible if they were ≤75 years of age and had not undergone arthroscopy of the knee within the previous 2 years. Arthritis was graded utilizing the Kellgren-Lawrence scale (0-4) and by calculating a severity score (0-12) by adding the individual scores for each of the 3 compartments of the knee; patients were excluded if they had a severity grade ≥9. One-quarter of the participants had severe arthritis, with scores of 7 or 8. Outcomes were measured over a 24-month period. At no point did either of the intervention groups report less pain or better function than the sham-surgery group.⁵

Additionally, in a single-center RCT involving 178 patients with mild to severe OA, subjects were randomly assigned to surgical lavage and arthroscopic debridement together with optimized physical and medical therapy or to treatment with physical and medical therapy alone.¹⁵ Patients were excluded if they had “bucket handle” meniscal tears detected by physical exam or by MRI, previous arthroscopic surgery, previous major knee trauma, or steroid injections in the last 3 months. All participants were required to have Kellgren-Lawrence grade 2, 3, or 4 arthritis.

The researchers used 2 validated outcome measures to evaluate pain, symptoms, and functional ability and followed the patients for 2 years after the initiation of treatment. This study failed to show that arthroscopic surgery provided any additional benefit to optimized physical and medical therapy.¹⁵

A 2013 multicenter, RCT randomly assigned 351 patients ≥45 years with a meniscal tear and radiographic evidence of mild to moderate OA to either surgery and postoperative physical therapy or to a standardized physical therapy regimen.¹⁶ Patients were required to have at least one symptom that was consistent with a meniscal tear for approximately one month. About 30% of the patients who were assigned to physical therapy alone underwent surgery within 6 months. There were no significant differences between the 2 groups in the magnitude of improvement in functional status or pain at 6 or 12 months.

Finally, a prospective Scottish study consisting of 270 patients with everything from no signs of OA to advanced OA who underwent APM were sent preoperative and 6-month postop questionnaires evaluating ADLs, pain, symptoms, quality of life, and body mass index (BMI).¹⁷ Their OA was graded via preop MRI or radiographs and confirmed by the operating surgeon. The investigators were unable to demonstrate any significant benefit associated with arthroscopic meniscectomy, and, therefore, could not recommend the procedure for patients with moderate to advanced OA.

However, in their analysis, 3 subgroup populations were found to benefit from APM: those with greater body habitus (BMI >30 kg/m²), those without signs of OA, and those with early OA. Limitations of this study included the lack of randomization, blinding, a control, and long-term follow up, and that the authors didn’t use established OA grading criteria.¹⁷

The bottom line: Nonoperative treatment benefits most patients

Physical therapy is an appropriate first-line treatment for degenerative meniscus tears in middle-aged patients. In fact, a trial of nonoperative treatment is likely to benefit the majority of patients. In addition, avoiding surgery eliminates surgical complications and decreases overall health care costs.

Reserve APM for those patients without significant OA who fail to improve after physical therapy, who have mechanical symptoms, or who have intra-articular loose bodies.

Consider a trial of nonoperative treatment, as it is likely to benefit the majority of patients with degenerative meniscal tears.

In addition, exercise therapy is an effective treatment for patients with knee OA. It improves function and limits joint pain in both acute arthritic flares and more long-term, chronic situations. There is strong evidence that strengthening plays a critical role in reducing symptoms and improving muscle strength, physical ability, and quality of life.¹⁹ It has been suggested that physical exercise 3 times a week for 4 months could lead to >35% improvement of knee function.²⁰ In contrast, other studies reported that while 91% of patients 11.5 years after APM considered their knees “normal or almost normal,” patients actually experienced a reduction in postoperative physical activity and quality of life.²¹

The most recent American Academy of Orthopaedic Surgeons (AAOS) guidelines²² do not recommend for or against arthroscopic partial meniscectomy in patients with knee OA and a torn meniscus. In middle-aged patients, MRI abnormalities of the meniscus do not consistently correlate with symptoms. Many meniscus lesions are asymptomatic or not the primary source of pain in the setting of OA.³

Potential harms and considerations. Deep venous thrombosis is the most frequently reported adverse event of arthroscopic surgery, with an approximate incidence of 4.13 per 1000 procedures, followed by less frequent complications such as infection, pulmonary embolism, and death.¹⁸

It is important to note that the degenerative meniscus tears that occur in middle age and that are associated with OA are not the same as acute, traumatic meniscus tears. All of the studies discussed here included patients with chronic knee pain and excluded patients with sudden onset pain from a single physical event. Many of the studies excluded patients with bucket-handle tears or severe mechanical symptoms (ie, locking). APM may be indicated for these meniscus tears regardless of age or OA status.

Making a sensible choice. Ultimately, physicians and their patients must use the best evidence available to make sensible clinical decisions. The ability to retain native meniscal tissue is of utmost importance to maintaining the longevity of their knee. According to previous studies, OA progression is more likely to occur after meniscectomy than after nonoperative treatment.²³

CORRESPONDENCE
William Bassett, MD, Department of Orthopedics, Rutgers-Robert Wood Johnson Medical School, 51 French St., PO Box 19, New Brunswick, NJ 08903; [email protected].