Clinical Inquiries

Evidence summary

Asplenic individuals are known to be at an elevated risk for infection with encapsulated bacteria. The lifetime risk of post-splenectomy sepsis is estimated to be approximately 1% to 2%. The overwhelming majority of these cases are caused by Streptococcus pneumoniae, Haemophilus influenzae, and Neisseria meningitides.^1-4

In 2 recent RCTs, the 23-valent pneumococcal polysaccharide vaccine was tested on patients 1, 7, 14, and 28 days after splenectomy.^1,2 The studies demonstrated that the immunogenicity of the vaccine was best when given at or after day 14 after the operation. In both studies, patients immunized at day 14 had immunoglobulin G (IgG) antibody levels approaching those of control subjects with intact spleens. There were no differences in antibody levels among those patients immunized at day 14 compared with those immunized on day 28. However, those subjects immunized on days 1 and 7 had significant lower antibody levels than the control subjects or those immunized on day 14.

In another study, 130 asplenic individuals were compared with 48 age-matched controls after receiving a meningococcal vaccine.³ The majority (93%) achieved bactericidal immunoglobulin levels following immunization. This study demonstrated the need to have antibody titers drawn to ensure immunization response, as 20% of the subjects required a second dose of vaccine to achieve adequate levels. No clear evidence supports the timing of the meningococcal vaccine post-splenectomy.

Two recent studies look at the immunogenicity of the Hib for asplenic patients. The first study demonstrated increased antibody titers to Hib at 2, 6, 12, 24, and 36 months after immunization.⁴ Fifty of the 57 patients in the study (88%) maintained adequate antibody titers 3 years after immunization. No symptomatic infections were observed during the 3-year study period. In a study of 561 Danes, those vaccinated within 14 days of splenectomy (before or after) had a significantly higher need for revaccination than those who were vaccinated more than 14 days before or after surgery.⁵

Recommendations from others

The most common infections occurring among asplenic patients are due to encapsulated organisms. The incidence is 10 to 50 times higher than in the general population.

The Advisory Committee on Immunization Practices for the Centers for Disease Control and Prevention (CDC) and the Society of Surgery for the Alimentary Tract recommends all patients that undergo splenectomy have the pneumococcal polysaccharide vaccine.^6-7 In addition, this organization also recommends that all asplenic patients receive meningococcal vaccination and be considered for the Hib vaccine. Both groups recommend that these vaccinations occur at the same time as the pneumococcal vaccine.^6-8

The CDC also recommends annual influenza vaccine in addition to the pneumococcal, meningococcal, and Hib vaccines, because secondary bacterial infections can lead to severe disease in this patient population. Boosters are recommended for all the bacterial vaccines every 5 years for asplenic patients.

Evidence summary

Asplenic individuals are known to be at an elevated risk for infection with encapsulated bacteria. The lifetime risk of post-splenectomy sepsis is estimated to be approximately 1% to 2%. The overwhelming majority of these cases are caused by Streptococcus pneumoniae, Haemophilus influenzae, and Neisseria meningitides.^1-4

In 2 recent RCTs, the 23-valent pneumococcal polysaccharide vaccine was tested on patients 1, 7, 14, and 28 days after splenectomy.^1,2 The studies demonstrated that the immunogenicity of the vaccine was best when given at or after day 14 after the operation. In both studies, patients immunized at day 14 had immunoglobulin G (IgG) antibody levels approaching those of control subjects with intact spleens. There were no differences in antibody levels among those patients immunized at day 14 compared with those immunized on day 28. However, those subjects immunized on days 1 and 7 had significant lower antibody levels than the control subjects or those immunized on day 14.

In another study, 130 asplenic individuals were compared with 48 age-matched controls after receiving a meningococcal vaccine.³ The majority (93%) achieved bactericidal immunoglobulin levels following immunization. This study demonstrated the need to have antibody titers drawn to ensure immunization response, as 20% of the subjects required a second dose of vaccine to achieve adequate levels. No clear evidence supports the timing of the meningococcal vaccine post-splenectomy.

Two recent studies look at the immunogenicity of the Hib for asplenic patients. The first study demonstrated increased antibody titers to Hib at 2, 6, 12, 24, and 36 months after immunization.⁴ Fifty of the 57 patients in the study (88%) maintained adequate antibody titers 3 years after immunization. No symptomatic infections were observed during the 3-year study period. In a study of 561 Danes, those vaccinated within 14 days of splenectomy (before or after) had a significantly higher need for revaccination than those who were vaccinated more than 14 days before or after surgery.⁵

Recommendations from others

The most common infections occurring among asplenic patients are due to encapsulated organisms. The incidence is 10 to 50 times higher than in the general population.

The Advisory Committee on Immunization Practices for the Centers for Disease Control and Prevention (CDC) and the Society of Surgery for the Alimentary Tract recommends all patients that undergo splenectomy have the pneumococcal polysaccharide vaccine.^6-7 In addition, this organization also recommends that all asplenic patients receive meningococcal vaccination and be considered for the Hib vaccine. Both groups recommend that these vaccinations occur at the same time as the pneumococcal vaccine.^6-8

The CDC also recommends annual influenza vaccine in addition to the pneumococcal, meningococcal, and Hib vaccines, because secondary bacterial infections can lead to severe disease in this patient population. Boosters are recommended for all the bacterial vaccines every 5 years for asplenic patients.

Evidence summary

Asplenic individuals are known to be at an elevated risk for infection with encapsulated bacteria. The lifetime risk of post-splenectomy sepsis is estimated to be approximately 1% to 2%. The overwhelming majority of these cases are caused by Streptococcus pneumoniae, Haemophilus influenzae, and Neisseria meningitides.^1-4

In 2 recent RCTs, the 23-valent pneumococcal polysaccharide vaccine was tested on patients 1, 7, 14, and 28 days after splenectomy.^1,2 The studies demonstrated that the immunogenicity of the vaccine was best when given at or after day 14 after the operation. In both studies, patients immunized at day 14 had immunoglobulin G (IgG) antibody levels approaching those of control subjects with intact spleens. There were no differences in antibody levels among those patients immunized at day 14 compared with those immunized on day 28. However, those subjects immunized on days 1 and 7 had significant lower antibody levels than the control subjects or those immunized on day 14.

In another study, 130 asplenic individuals were compared with 48 age-matched controls after receiving a meningococcal vaccine.³ The majority (93%) achieved bactericidal immunoglobulin levels following immunization. This study demonstrated the need to have antibody titers drawn to ensure immunization response, as 20% of the subjects required a second dose of vaccine to achieve adequate levels. No clear evidence supports the timing of the meningococcal vaccine post-splenectomy.

Two recent studies look at the immunogenicity of the Hib for asplenic patients. The first study demonstrated increased antibody titers to Hib at 2, 6, 12, 24, and 36 months after immunization.⁴ Fifty of the 57 patients in the study (88%) maintained adequate antibody titers 3 years after immunization. No symptomatic infections were observed during the 3-year study period. In a study of 561 Danes, those vaccinated within 14 days of splenectomy (before or after) had a significantly higher need for revaccination than those who were vaccinated more than 14 days before or after surgery.⁵

Recommendations from others

The most common infections occurring among asplenic patients are due to encapsulated organisms. The incidence is 10 to 50 times higher than in the general population.

The Advisory Committee on Immunization Practices for the Centers for Disease Control and Prevention (CDC) and the Society of Surgery for the Alimentary Tract recommends all patients that undergo splenectomy have the pneumococcal polysaccharide vaccine.^6-7 In addition, this organization also recommends that all asplenic patients receive meningococcal vaccination and be considered for the Hib vaccine. Both groups recommend that these vaccinations occur at the same time as the pneumococcal vaccine.^6-8

The CDC also recommends annual influenza vaccine in addition to the pneumococcal, meningococcal, and Hib vaccines, because secondary bacterial infections can lead to severe disease in this patient population. Boosters are recommended for all the bacterial vaccines every 5 years for asplenic patients.

TABLE
Questions to consider before tablet-splitting

Evidence summary

Few studies have looked at the clinical effects of pill-splitting. A randomized trial (n=29) evaluated tablet-splitting by patients taking lisinopril for hypertension.¹ Patients were randomized to split tablets or whole tablets for 2 weeks, then crossed over to the other group for 2 weeks. There was no difference in blood pressures between groups.

A retrospective study of simvastatin evaluated 1098 patients taking whole tablets and 1098 patients converted to split tablets of the same dose.² There was no difference in average final low-density lipoprotein (LDL) cholesterol (111±30 mg/dL vs 112±32 mg/dL) or mean ala-nine aminotransferase (ALT) level.

Another retrospective study evaluated tablet-splitting by 512 patients taking statins (atorvastatin, lovastatin, simvastatin).³ Cholesterol values after 12 or more weeks on a stable whole-tablet dose were compared with those 6 to 52 weeks after initiating tablet-splitting; no significant change was seen in total cholesterol or triglycerides. There was a statistically significant decrease in LDL (102±28 vs 97±29 mg/dL, P<.001), an increase in high-density lipoprotein (HDL) cholesterol (46±12 vs 48±12 mg/dL, P<.001), and an increase in aspartate aminotransferase (AST) (26±8 vs 28±10 units/L, P<.001), which was attributed to higher medication dosage from accidental ingestion of whole tablets and to diet and lifestyle modifications. Another retrospective evaluation of 109 patients with split atorvastatin or simvastatin found no significant difference in total cholesterol or LDL values after initiating the tablet-splitting program.⁴

Thirty patients aged 50 years or older, half of whom received instruction and a demonstration, evaluated 2 tablet-splitters with different blade positions and types of guide.⁵ One device (Apex Pill Splitter) produced more accurate results by 10% to 20% (P value not provided) with metoprolol, warfarin, and lisinopril tablets. Instructed patients were 1% to 10% more accurate, as were those with experience splitting warfarin tablets (P=.003).

In another study, 94 healthy volunteers (mean age, 46.2 years) each split 10 hydrochlorothiazide 25 mg tablets by hand. Forty-one percent of the split products were more than 10% off ideal weight; 12% of tablets were more than 20% off.⁶ Manufacturing regulations require that medication doses vary by less than 10% of the nominal dose. Another study using 5 medications found that 0% to 44% of split tablets deviated from ideal weight by 20%, depending on tablet shape.⁷

Surveys of patient acceptance of tablet-splitting report varied rates (3%*#8211;74%),^3,6 In 1 study,¹ 89% and 97% said they would split tablets to save money for themselves or their health facility, respectively.

Experts recommend assessing patients for their physical (dexterity, strength, visual acuity) and cognitive ability to split tablets, as well as whether doing so saves money.⁸

Recommendations from others

The American Medical Society and American Pharmacists Association oppose mandatory tablet-splitting and recommend against splitting tablets that are modified-release, combination products, unscored, film-coated, friable, or dose-critical.

TABLE
Questions to consider before tablet-splitting

Evidence summary

Few studies have looked at the clinical effects of pill-splitting. A randomized trial (n=29) evaluated tablet-splitting by patients taking lisinopril for hypertension.¹ Patients were randomized to split tablets or whole tablets for 2 weeks, then crossed over to the other group for 2 weeks. There was no difference in blood pressures between groups.

A retrospective study of simvastatin evaluated 1098 patients taking whole tablets and 1098 patients converted to split tablets of the same dose.² There was no difference in average final low-density lipoprotein (LDL) cholesterol (111±30 mg/dL vs 112±32 mg/dL) or mean ala-nine aminotransferase (ALT) level.

Another retrospective study evaluated tablet-splitting by 512 patients taking statins (atorvastatin, lovastatin, simvastatin).³ Cholesterol values after 12 or more weeks on a stable whole-tablet dose were compared with those 6 to 52 weeks after initiating tablet-splitting; no significant change was seen in total cholesterol or triglycerides. There was a statistically significant decrease in LDL (102±28 vs 97±29 mg/dL, P<.001), an increase in high-density lipoprotein (HDL) cholesterol (46±12 vs 48±12 mg/dL, P<.001), and an increase in aspartate aminotransferase (AST) (26±8 vs 28±10 units/L, P<.001), which was attributed to higher medication dosage from accidental ingestion of whole tablets and to diet and lifestyle modifications. Another retrospective evaluation of 109 patients with split atorvastatin or simvastatin found no significant difference in total cholesterol or LDL values after initiating the tablet-splitting program.⁴

Thirty patients aged 50 years or older, half of whom received instruction and a demonstration, evaluated 2 tablet-splitters with different blade positions and types of guide.⁵ One device (Apex Pill Splitter) produced more accurate results by 10% to 20% (P value not provided) with metoprolol, warfarin, and lisinopril tablets. Instructed patients were 1% to 10% more accurate, as were those with experience splitting warfarin tablets (P=.003).

In another study, 94 healthy volunteers (mean age, 46.2 years) each split 10 hydrochlorothiazide 25 mg tablets by hand. Forty-one percent of the split products were more than 10% off ideal weight; 12% of tablets were more than 20% off.⁶ Manufacturing regulations require that medication doses vary by less than 10% of the nominal dose. Another study using 5 medications found that 0% to 44% of split tablets deviated from ideal weight by 20%, depending on tablet shape.⁷

Surveys of patient acceptance of tablet-splitting report varied rates (3%*#8211;74%),^3,6 In 1 study,¹ 89% and 97% said they would split tablets to save money for themselves or their health facility, respectively.

Experts recommend assessing patients for their physical (dexterity, strength, visual acuity) and cognitive ability to split tablets, as well as whether doing so saves money.⁸

Recommendations from others

The American Medical Society and American Pharmacists Association oppose mandatory tablet-splitting and recommend against splitting tablets that are modified-release, combination products, unscored, film-coated, friable, or dose-critical.

TABLE
Questions to consider before tablet-splitting

Evidence summary

Few studies have looked at the clinical effects of pill-splitting. A randomized trial (n=29) evaluated tablet-splitting by patients taking lisinopril for hypertension.¹ Patients were randomized to split tablets or whole tablets for 2 weeks, then crossed over to the other group for 2 weeks. There was no difference in blood pressures between groups.

A retrospective study of simvastatin evaluated 1098 patients taking whole tablets and 1098 patients converted to split tablets of the same dose.² There was no difference in average final low-density lipoprotein (LDL) cholesterol (111±30 mg/dL vs 112±32 mg/dL) or mean ala-nine aminotransferase (ALT) level.

Another retrospective study evaluated tablet-splitting by 512 patients taking statins (atorvastatin, lovastatin, simvastatin).³ Cholesterol values after 12 or more weeks on a stable whole-tablet dose were compared with those 6 to 52 weeks after initiating tablet-splitting; no significant change was seen in total cholesterol or triglycerides. There was a statistically significant decrease in LDL (102±28 vs 97±29 mg/dL, P<.001), an increase in high-density lipoprotein (HDL) cholesterol (46±12 vs 48±12 mg/dL, P<.001), and an increase in aspartate aminotransferase (AST) (26±8 vs 28±10 units/L, P<.001), which was attributed to higher medication dosage from accidental ingestion of whole tablets and to diet and lifestyle modifications. Another retrospective evaluation of 109 patients with split atorvastatin or simvastatin found no significant difference in total cholesterol or LDL values after initiating the tablet-splitting program.⁴

Thirty patients aged 50 years or older, half of whom received instruction and a demonstration, evaluated 2 tablet-splitters with different blade positions and types of guide.⁵ One device (Apex Pill Splitter) produced more accurate results by 10% to 20% (P value not provided) with metoprolol, warfarin, and lisinopril tablets. Instructed patients were 1% to 10% more accurate, as were those with experience splitting warfarin tablets (P=.003).

In another study, 94 healthy volunteers (mean age, 46.2 years) each split 10 hydrochlorothiazide 25 mg tablets by hand. Forty-one percent of the split products were more than 10% off ideal weight; 12% of tablets were more than 20% off.⁶ Manufacturing regulations require that medication doses vary by less than 10% of the nominal dose. Another study using 5 medications found that 0% to 44% of split tablets deviated from ideal weight by 20%, depending on tablet shape.⁷

Surveys of patient acceptance of tablet-splitting report varied rates (3%*#8211;74%),^3,6 In 1 study,¹ 89% and 97% said they would split tablets to save money for themselves or their health facility, respectively.

Experts recommend assessing patients for their physical (dexterity, strength, visual acuity) and cognitive ability to split tablets, as well as whether doing so saves money.⁸

Recommendations from others

The American Medical Society and American Pharmacists Association oppose mandatory tablet-splitting and recommend against splitting tablets that are modified-release, combination products, unscored, film-coated, friable, or dose-critical.

Evidence summary

No randomized controlled trials have been published regarding percutaneous vertebroplasty. A 2005 Technology Assessment by the Centers for Medicare and Medicaid Services (CMS) provides the best evidence about indications and efficacy of percutaneous vertebroplasty for vertebral fractures.¹ The CMS report is based on a search of Medline and Current Contents through April 2005 for relevant studies, along with hand searches of retrieved articles’ references and of recent pertinent journals. Study inclusion criteria included English language, vertebral fractures due to osteoporosis or malignancy, consecutive patient enrollment, outcomes reported for pain, functional status, and quality of life, and study size ≥20 patients for studies of osteoporosis or ≥10 patients for studies of malignancy. There was no description of a formal study validity assessment or attempts to control bias by use of multiple reviewers.

Fifteen studies were included, representing 1056 patients. Fourteen of the studies were observational and 1 was a nonrandomized controlled trial. The common inclusion requirement was severe pain attributable to vertebral fracture. Nine of the studies further specified failure of analgesics or conservative treatments. The studies showed statistically significant decreases in comparative visual analog pain scale scores in the short term. Four studies showed pain reduction lasting up to 1 year. These results favor the conclusion that percutaneous vertebroplasty provides short- and long-term pain reduction for patients meeting the inclusion criteria. However, the lack of randomized trials cannot control for the placebo effect, the natural history of vertebral fractures, and regression to the mean as possible reasons for the apparent efficacy of percutaneous vertebroplasty.

Two structured, but not systematic, reviews of percutaneous vertebroplasty in vertebral fractures^2,3 included 15 small observational studies, of which only 1 was included in the CMS report. These reviews examined outcomes of vertebroplasty performed from less than 1 month to a mean of 7 months after fracture, using similar inclusion criteria to those used in the CMS report. The studies’ common patient exclusion criteria were uncorrectable coagulation disorder, infection in the area, spinal cord compression, destruction of the posterior vertebral wall, and severe degrees of vertebral body collapse (defined as >67% collapse). The 2 reviews found between 67% and 100% of patients reported pain reduction after vertebroplasty in follow-up periods ranging from 24 hours to up to 10 years. Based on this limited evidence, 1 review suggested that the likelihood of alleviation of pain decreases over time and is low for fractures occurring more than 6 months in the past.³ In contradistinction, 3 subsequent observational studies, involving a total of 233 patients with 365 vertebral compression fractures failed to find an association between postprocedural pain and age of fracture (ranging from less than 2 weeks to more than 24 months from injury).^4-6

Recommendations from others

In their guideline on rehabilitation of the patient with osteoporosis, the National Osteoporosis Foundation states an experienced practitioner may perform percutaneous vertebroplasty on a patient with unremitting pain for whom conservative medical therapy has not helped.⁷ They qualify this recommendation by further stating long-term clinical studies are required before vertebroplasty becomes standard of care. The Medicare Coverage Advisory Committee, in its review of the 2005 CMS report, suggested that percutaneous vertebroplasty produces a clinically important net health benefit for patients with vertebral compression fracture compared to conservative care for both acute and chronic fractures.⁸

Evidence summary

No randomized controlled trials have been published regarding percutaneous vertebroplasty. A 2005 Technology Assessment by the Centers for Medicare and Medicaid Services (CMS) provides the best evidence about indications and efficacy of percutaneous vertebroplasty for vertebral fractures.¹ The CMS report is based on a search of Medline and Current Contents through April 2005 for relevant studies, along with hand searches of retrieved articles’ references and of recent pertinent journals. Study inclusion criteria included English language, vertebral fractures due to osteoporosis or malignancy, consecutive patient enrollment, outcomes reported for pain, functional status, and quality of life, and study size ≥20 patients for studies of osteoporosis or ≥10 patients for studies of malignancy. There was no description of a formal study validity assessment or attempts to control bias by use of multiple reviewers.

Fifteen studies were included, representing 1056 patients. Fourteen of the studies were observational and 1 was a nonrandomized controlled trial. The common inclusion requirement was severe pain attributable to vertebral fracture. Nine of the studies further specified failure of analgesics or conservative treatments. The studies showed statistically significant decreases in comparative visual analog pain scale scores in the short term. Four studies showed pain reduction lasting up to 1 year. These results favor the conclusion that percutaneous vertebroplasty provides short- and long-term pain reduction for patients meeting the inclusion criteria. However, the lack of randomized trials cannot control for the placebo effect, the natural history of vertebral fractures, and regression to the mean as possible reasons for the apparent efficacy of percutaneous vertebroplasty.

Two structured, but not systematic, reviews of percutaneous vertebroplasty in vertebral fractures^2,3 included 15 small observational studies, of which only 1 was included in the CMS report. These reviews examined outcomes of vertebroplasty performed from less than 1 month to a mean of 7 months after fracture, using similar inclusion criteria to those used in the CMS report. The studies’ common patient exclusion criteria were uncorrectable coagulation disorder, infection in the area, spinal cord compression, destruction of the posterior vertebral wall, and severe degrees of vertebral body collapse (defined as >67% collapse). The 2 reviews found between 67% and 100% of patients reported pain reduction after vertebroplasty in follow-up periods ranging from 24 hours to up to 10 years. Based on this limited evidence, 1 review suggested that the likelihood of alleviation of pain decreases over time and is low for fractures occurring more than 6 months in the past.³ In contradistinction, 3 subsequent observational studies, involving a total of 233 patients with 365 vertebral compression fractures failed to find an association between postprocedural pain and age of fracture (ranging from less than 2 weeks to more than 24 months from injury).^4-6

Recommendations from others

In their guideline on rehabilitation of the patient with osteoporosis, the National Osteoporosis Foundation states an experienced practitioner may perform percutaneous vertebroplasty on a patient with unremitting pain for whom conservative medical therapy has not helped.⁷ They qualify this recommendation by further stating long-term clinical studies are required before vertebroplasty becomes standard of care. The Medicare Coverage Advisory Committee, in its review of the 2005 CMS report, suggested that percutaneous vertebroplasty produces a clinically important net health benefit for patients with vertebral compression fracture compared to conservative care for both acute and chronic fractures.⁸

Evidence summary

No randomized controlled trials have been published regarding percutaneous vertebroplasty. A 2005 Technology Assessment by the Centers for Medicare and Medicaid Services (CMS) provides the best evidence about indications and efficacy of percutaneous vertebroplasty for vertebral fractures.¹ The CMS report is based on a search of Medline and Current Contents through April 2005 for relevant studies, along with hand searches of retrieved articles’ references and of recent pertinent journals. Study inclusion criteria included English language, vertebral fractures due to osteoporosis or malignancy, consecutive patient enrollment, outcomes reported for pain, functional status, and quality of life, and study size ≥20 patients for studies of osteoporosis or ≥10 patients for studies of malignancy. There was no description of a formal study validity assessment or attempts to control bias by use of multiple reviewers.

Fifteen studies were included, representing 1056 patients. Fourteen of the studies were observational and 1 was a nonrandomized controlled trial. The common inclusion requirement was severe pain attributable to vertebral fracture. Nine of the studies further specified failure of analgesics or conservative treatments. The studies showed statistically significant decreases in comparative visual analog pain scale scores in the short term. Four studies showed pain reduction lasting up to 1 year. These results favor the conclusion that percutaneous vertebroplasty provides short- and long-term pain reduction for patients meeting the inclusion criteria. However, the lack of randomized trials cannot control for the placebo effect, the natural history of vertebral fractures, and regression to the mean as possible reasons for the apparent efficacy of percutaneous vertebroplasty.

Two structured, but not systematic, reviews of percutaneous vertebroplasty in vertebral fractures^2,3 included 15 small observational studies, of which only 1 was included in the CMS report. These reviews examined outcomes of vertebroplasty performed from less than 1 month to a mean of 7 months after fracture, using similar inclusion criteria to those used in the CMS report. The studies’ common patient exclusion criteria were uncorrectable coagulation disorder, infection in the area, spinal cord compression, destruction of the posterior vertebral wall, and severe degrees of vertebral body collapse (defined as >67% collapse). The 2 reviews found between 67% and 100% of patients reported pain reduction after vertebroplasty in follow-up periods ranging from 24 hours to up to 10 years. Based on this limited evidence, 1 review suggested that the likelihood of alleviation of pain decreases over time and is low for fractures occurring more than 6 months in the past.³ In contradistinction, 3 subsequent observational studies, involving a total of 233 patients with 365 vertebral compression fractures failed to find an association between postprocedural pain and age of fracture (ranging from less than 2 weeks to more than 24 months from injury).^4-6

Recommendations from others

In their guideline on rehabilitation of the patient with osteoporosis, the National Osteoporosis Foundation states an experienced practitioner may perform percutaneous vertebroplasty on a patient with unremitting pain for whom conservative medical therapy has not helped.⁷ They qualify this recommendation by further stating long-term clinical studies are required before vertebroplasty becomes standard of care. The Medicare Coverage Advisory Committee, in its review of the 2005 CMS report, suggested that percutaneous vertebroplasty produces a clinically important net health benefit for patients with vertebral compression fracture compared to conservative care for both acute and chronic fractures.⁸

Evidence summary

Colic has been described using the “rule of 3”: crying for at least 3 hours per day on at least 3 days per week for at least 3 weeks.¹ The distinction can be subtle; a normal infant can cry more than 2 hours per day. This syndrome has its onset typically in the first few weeks of life. It spontaneously resolves by age 4 to 6 months. Prevalence depends on the definition used for colic; approximately 5% to 25% of infants meet some reasonable definition of colic.² The cause of infantile colic is poorly understood. Although clinicians tend to focus on a likely gastrointestinal cause, neuropsychological issues, food allergy, and parenting misadventures are also potential contributing factors.

There are myriad strategies—ranging from craniosacral osteopathic manipulation to car ride simulation—offered for dealing with infantile colic. Although none of these treatments has been validated in rigorous studies, the available evidence offers tentative support for 3 strategies: (1) a trial of a hypoallergenic (protein hydrolysate) formula (for formula fed infants), (2) a low-allergen maternal diet (for breastfeeding mothers), and (3) reduced stimulation of the infant.

A systematic review analyzed controlled clinical trials lasting at least 3 days involving infants less than 6 months of age who cried excessively.³ Twenty-seven studies were included; the outcome measure was colic symptoms, typically reported as duration of crying. Two reports studying hypoallergenic (protein hydrolysate) formula in nearly 130 infants found an effect size of 0.22 (95% confidence interval [CI], 0.10–0.34) for the hypoallergenic formula. Additionally, 3 behavioral trials (involving nearly 200 infants) revealed the benefits of reduced stimulation of the colicky infant (effect size of 0.48; 95% CI, 0.23–0.74).

A more recent systematic review⁴ followed a similar high-quality search strategy and identified 22 articles, and reported a number needed to treat (NNT) of 6 for the 2 hypoallergenic formula studies identified in the previous review.⁴ Because of concern regarding the quality of the behavioral studies involving infants with colic, the authors of this second review only included 1 small (42 patients) trial of decreased stimulation, which resulted in a relative risk (RR) of 1.87 (95% CI, 1.04–3.34) and a NNT of 2. There was some inconclusive evidence to suggest benefit to dietary adjustment for breastfeeding mothers (specifically, the avoidance of cow’s milk and other potential allergens like nuts, eggs, and wheat).

A recent randomized controlled trial confirmed the value of this approach by showing significant improvement in distress scores of infants whose mothers followed a low-allergen diet (excluding dairy, soy, wheat, eggs, peanuts, tree nuts, and fish) for 7 days.⁵ This well-designed study included 107 patients (a relatively large sample in the published research about colic), and showed an absolute risk reduction of 37% (NNT=3) for those mothers following the challenge.

A small RCT (43 patients) suggested efficacy in the substitution of a whey hydrolysate formula in place of cow’s milkbased formula for infants with colic (casein hydrolysate formula has been more widely studied), but there continues to be controversy regarding the preferred protein hydrolysate formula (whey vs casein) in the treatment of colic.³

Several medications have been tested in RCTs; only dicyclomine has shown an effect in a few small RCTs.^3,4 However, there have been reports of apnea and other serious, although infrequent, adverse effects. For that reason, the manufacturer has contraindicated the use of this medication in infants aged <6 months.

A small (n=68) study of an herbal tea showed reduced symptoms (RR=0.57 favoring the active tea), although the mean volume of tea consumption (32 mL/kg/d) is a nutritional concern in this age group.⁶ No adverse events were noted, but the small sample size limits the ability to detect any but the most common events.

Recommendations from others

The American Gastroenterological Association recommends a hypoallergenic, protein hydrolysate formula for formula fed infants or a maternal low-allergen diet as an initial strategy for infant struggling with colic symptoms if the clinician is considering a diagnosis of (cow’s milk) allergy.⁷

The American Academy of Family Physicians on their familydoctor.org web site makes no specific formula or diet adjustment recommendations.⁸ The web site does list some techniques (eg, massage or warm compress of abdomen, swing or car rides) not supported by the available evidence. The National Library of Medicine and the National Institutes of Health web site Medline Plus presents similar information.⁹ The American Academy of Pediatrics does not address the topic on its public web site.

Evidence summary

Colic has been described using the “rule of 3”: crying for at least 3 hours per day on at least 3 days per week for at least 3 weeks.¹ The distinction can be subtle; a normal infant can cry more than 2 hours per day. This syndrome has its onset typically in the first few weeks of life. It spontaneously resolves by age 4 to 6 months. Prevalence depends on the definition used for colic; approximately 5% to 25% of infants meet some reasonable definition of colic.² The cause of infantile colic is poorly understood. Although clinicians tend to focus on a likely gastrointestinal cause, neuropsychological issues, food allergy, and parenting misadventures are also potential contributing factors.

There are myriad strategies—ranging from craniosacral osteopathic manipulation to car ride simulation—offered for dealing with infantile colic. Although none of these treatments has been validated in rigorous studies, the available evidence offers tentative support for 3 strategies: (1) a trial of a hypoallergenic (protein hydrolysate) formula (for formula fed infants), (2) a low-allergen maternal diet (for breastfeeding mothers), and (3) reduced stimulation of the infant.

A systematic review analyzed controlled clinical trials lasting at least 3 days involving infants less than 6 months of age who cried excessively.³ Twenty-seven studies were included; the outcome measure was colic symptoms, typically reported as duration of crying. Two reports studying hypoallergenic (protein hydrolysate) formula in nearly 130 infants found an effect size of 0.22 (95% confidence interval [CI], 0.10–0.34) for the hypoallergenic formula. Additionally, 3 behavioral trials (involving nearly 200 infants) revealed the benefits of reduced stimulation of the colicky infant (effect size of 0.48; 95% CI, 0.23–0.74).

A more recent systematic review⁴ followed a similar high-quality search strategy and identified 22 articles, and reported a number needed to treat (NNT) of 6 for the 2 hypoallergenic formula studies identified in the previous review.⁴ Because of concern regarding the quality of the behavioral studies involving infants with colic, the authors of this second review only included 1 small (42 patients) trial of decreased stimulation, which resulted in a relative risk (RR) of 1.87 (95% CI, 1.04–3.34) and a NNT of 2. There was some inconclusive evidence to suggest benefit to dietary adjustment for breastfeeding mothers (specifically, the avoidance of cow’s milk and other potential allergens like nuts, eggs, and wheat).

A recent randomized controlled trial confirmed the value of this approach by showing significant improvement in distress scores of infants whose mothers followed a low-allergen diet (excluding dairy, soy, wheat, eggs, peanuts, tree nuts, and fish) for 7 days.⁵ This well-designed study included 107 patients (a relatively large sample in the published research about colic), and showed an absolute risk reduction of 37% (NNT=3) for those mothers following the challenge.

A small RCT (43 patients) suggested efficacy in the substitution of a whey hydrolysate formula in place of cow’s milkbased formula for infants with colic (casein hydrolysate formula has been more widely studied), but there continues to be controversy regarding the preferred protein hydrolysate formula (whey vs casein) in the treatment of colic.³

Several medications have been tested in RCTs; only dicyclomine has shown an effect in a few small RCTs.^3,4 However, there have been reports of apnea and other serious, although infrequent, adverse effects. For that reason, the manufacturer has contraindicated the use of this medication in infants aged <6 months.

A small (n=68) study of an herbal tea showed reduced symptoms (RR=0.57 favoring the active tea), although the mean volume of tea consumption (32 mL/kg/d) is a nutritional concern in this age group.⁶ No adverse events were noted, but the small sample size limits the ability to detect any but the most common events.

Recommendations from others

The American Gastroenterological Association recommends a hypoallergenic, protein hydrolysate formula for formula fed infants or a maternal low-allergen diet as an initial strategy for infant struggling with colic symptoms if the clinician is considering a diagnosis of (cow’s milk) allergy.⁷

The American Academy of Family Physicians on their familydoctor.org web site makes no specific formula or diet adjustment recommendations.⁸ The web site does list some techniques (eg, massage or warm compress of abdomen, swing or car rides) not supported by the available evidence. The National Library of Medicine and the National Institutes of Health web site Medline Plus presents similar information.⁹ The American Academy of Pediatrics does not address the topic on its public web site.

Evidence summary

Colic has been described using the “rule of 3”: crying for at least 3 hours per day on at least 3 days per week for at least 3 weeks.¹ The distinction can be subtle; a normal infant can cry more than 2 hours per day. This syndrome has its onset typically in the first few weeks of life. It spontaneously resolves by age 4 to 6 months. Prevalence depends on the definition used for colic; approximately 5% to 25% of infants meet some reasonable definition of colic.² The cause of infantile colic is poorly understood. Although clinicians tend to focus on a likely gastrointestinal cause, neuropsychological issues, food allergy, and parenting misadventures are also potential contributing factors.

There are myriad strategies—ranging from craniosacral osteopathic manipulation to car ride simulation—offered for dealing with infantile colic. Although none of these treatments has been validated in rigorous studies, the available evidence offers tentative support for 3 strategies: (1) a trial of a hypoallergenic (protein hydrolysate) formula (for formula fed infants), (2) a low-allergen maternal diet (for breastfeeding mothers), and (3) reduced stimulation of the infant.

A systematic review analyzed controlled clinical trials lasting at least 3 days involving infants less than 6 months of age who cried excessively.³ Twenty-seven studies were included; the outcome measure was colic symptoms, typically reported as duration of crying. Two reports studying hypoallergenic (protein hydrolysate) formula in nearly 130 infants found an effect size of 0.22 (95% confidence interval [CI], 0.10–0.34) for the hypoallergenic formula. Additionally, 3 behavioral trials (involving nearly 200 infants) revealed the benefits of reduced stimulation of the colicky infant (effect size of 0.48; 95% CI, 0.23–0.74).

A more recent systematic review⁴ followed a similar high-quality search strategy and identified 22 articles, and reported a number needed to treat (NNT) of 6 for the 2 hypoallergenic formula studies identified in the previous review.⁴ Because of concern regarding the quality of the behavioral studies involving infants with colic, the authors of this second review only included 1 small (42 patients) trial of decreased stimulation, which resulted in a relative risk (RR) of 1.87 (95% CI, 1.04–3.34) and a NNT of 2. There was some inconclusive evidence to suggest benefit to dietary adjustment for breastfeeding mothers (specifically, the avoidance of cow’s milk and other potential allergens like nuts, eggs, and wheat).

A recent randomized controlled trial confirmed the value of this approach by showing significant improvement in distress scores of infants whose mothers followed a low-allergen diet (excluding dairy, soy, wheat, eggs, peanuts, tree nuts, and fish) for 7 days.⁵ This well-designed study included 107 patients (a relatively large sample in the published research about colic), and showed an absolute risk reduction of 37% (NNT=3) for those mothers following the challenge.

A small RCT (43 patients) suggested efficacy in the substitution of a whey hydrolysate formula in place of cow’s milkbased formula for infants with colic (casein hydrolysate formula has been more widely studied), but there continues to be controversy regarding the preferred protein hydrolysate formula (whey vs casein) in the treatment of colic.³

Several medications have been tested in RCTs; only dicyclomine has shown an effect in a few small RCTs.^3,4 However, there have been reports of apnea and other serious, although infrequent, adverse effects. For that reason, the manufacturer has contraindicated the use of this medication in infants aged <6 months.

A small (n=68) study of an herbal tea showed reduced symptoms (RR=0.57 favoring the active tea), although the mean volume of tea consumption (32 mL/kg/d) is a nutritional concern in this age group.⁶ No adverse events were noted, but the small sample size limits the ability to detect any but the most common events.

Recommendations from others

The American Gastroenterological Association recommends a hypoallergenic, protein hydrolysate formula for formula fed infants or a maternal low-allergen diet as an initial strategy for infant struggling with colic symptoms if the clinician is considering a diagnosis of (cow’s milk) allergy.⁷

The American Academy of Family Physicians on their familydoctor.org web site makes no specific formula or diet adjustment recommendations.⁸ The web site does list some techniques (eg, massage or warm compress of abdomen, swing or car rides) not supported by the available evidence. The National Library of Medicine and the National Institutes of Health web site Medline Plus presents similar information.⁹ The American Academy of Pediatrics does not address the topic on its public web site.

Evidence summary

Depletion of iron stores leads to IDA, which, among children, is associated with motor and cognitive deficits that may be irreversible. Little is known about whether iron deficiency, in the absence of anemia, results in physiologic sequelae. A Cochrane review of iron therapy for children with IDA aged >3 years found no short-term (5–11 days) improvement in Bayley scores of mental and motor development following iron therapy.¹ A 10-year longitudinal cohort study in Costa Rica found that adolescents treated for severe chronic IDA in infancy (n=48) scored 0.4 to 0.7 standard deviations lower on cognitive and motor testing relative to controls (n=114).² In an Indonesian randomized controlled trial (RCT), baseline Bayley scores were 10% to 15% lower (P<.01) for infants (12–18 months) with IDA compared with both nonanemic iron-deficient and iron-sufficient infants.³ Following treatment with ferrous sulfate (3 mg/kg/d of elemental iron) for 4 months, the IDA infants’ Bayley scores improved compared with those of nonanemic children.

Consensus recommendations suggest that iron deficiency should be the presumptive diagnosis in a child with anemia, and that a trial of ferrous sulfate at a dose of 3 mg/kg/d of elemental iron be instituted because of low cost, tolerability, and relative simplicity.^4,5 In a prospective study of 75 1-year-olds with anemia (hemoglobin <11.0 g/dL), 45% achieved an increase in hemoglobin ≥1 g/dL after 3 months of iron therapy (3 mg/kg/d).⁶ An RCT of 278 nonanemic 1-year-olds found no difference in adverse effects from this dose compared with placebo.⁷ However, an analysis of data from NHANES III showed that a Hgb <11.0 had a positive predictive value of just 29% and sensitivity of 30% for diagnosing iron-deficiency in children aged <3 years.⁸

The recommended dose of 3 mg/kg/d was derived from models of bioavailability and iron needs⁹; no studies compare alternative doses. An RCT of 557 anemic children under 24 months of age in Ghana demonstrated that ferrous sulfate (5 mg/kg/d) given once daily was equivalent to 3-times-daily dosing in terms of effectiveness (61% vs 56%) and tolerance.¹⁰ Less frequent dosing has been studied in developing countries with mixed results.

Because anemia may lead to developmental impairment, primary prevention is critical. In a cohort study, infants given iron-fortified formula (n=98) were less likely to become iron-deficient by their 12-month visit than infants fed whole cow milk (n=69) (11.2% vs 24.6%, number needed to treat [NNT]=8).¹¹ In a RCT of innercity children who had been switched to cows’ milk by 6 months, half (n=50) were randomized to receive iron-fortified formula for another year, resulting in a decreased risk of anemia at 24 months (0% vs 26%, NNT=4), and smaller declines in developmental functioning compared with those on cows’ milk.¹²

Recommendations from others

The CDC and the Institute of Medicine recommend parental dietary counseling, treatment with oral ferrous sulfate at 3 mg/kg/d for 3 months to restore iron stores, and monitoring of hemoglobin or hematocrit to assess response.^4,5 To prevent IDA, the American Academy of Pediatrics (AAP) recommends that all infants who are not breastfed or are partially breastfed should receive an iron-fortified formula (containing between 4.0–12 mg/L of iron) from birth to 12 months. The AAP also recommends that parents should refrain from feeding cow’s milk to infants until after age 12 months and introduce iron-enriched foods between ages 4 and 6 months.¹³

Evidence summary

Depletion of iron stores leads to IDA, which, among children, is associated with motor and cognitive deficits that may be irreversible. Little is known about whether iron deficiency, in the absence of anemia, results in physiologic sequelae. A Cochrane review of iron therapy for children with IDA aged >3 years found no short-term (5–11 days) improvement in Bayley scores of mental and motor development following iron therapy.¹ A 10-year longitudinal cohort study in Costa Rica found that adolescents treated for severe chronic IDA in infancy (n=48) scored 0.4 to 0.7 standard deviations lower on cognitive and motor testing relative to controls (n=114).² In an Indonesian randomized controlled trial (RCT), baseline Bayley scores were 10% to 15% lower (P<.01) for infants (12–18 months) with IDA compared with both nonanemic iron-deficient and iron-sufficient infants.³ Following treatment with ferrous sulfate (3 mg/kg/d of elemental iron) for 4 months, the IDA infants’ Bayley scores improved compared with those of nonanemic children.

Consensus recommendations suggest that iron deficiency should be the presumptive diagnosis in a child with anemia, and that a trial of ferrous sulfate at a dose of 3 mg/kg/d of elemental iron be instituted because of low cost, tolerability, and relative simplicity.^4,5 In a prospective study of 75 1-year-olds with anemia (hemoglobin <11.0 g/dL), 45% achieved an increase in hemoglobin ≥1 g/dL after 3 months of iron therapy (3 mg/kg/d).⁶ An RCT of 278 nonanemic 1-year-olds found no difference in adverse effects from this dose compared with placebo.⁷ However, an analysis of data from NHANES III showed that a Hgb <11.0 had a positive predictive value of just 29% and sensitivity of 30% for diagnosing iron-deficiency in children aged <3 years.⁸

The recommended dose of 3 mg/kg/d was derived from models of bioavailability and iron needs⁹; no studies compare alternative doses. An RCT of 557 anemic children under 24 months of age in Ghana demonstrated that ferrous sulfate (5 mg/kg/d) given once daily was equivalent to 3-times-daily dosing in terms of effectiveness (61% vs 56%) and tolerance.¹⁰ Less frequent dosing has been studied in developing countries with mixed results.

Because anemia may lead to developmental impairment, primary prevention is critical. In a cohort study, infants given iron-fortified formula (n=98) were less likely to become iron-deficient by their 12-month visit than infants fed whole cow milk (n=69) (11.2% vs 24.6%, number needed to treat [NNT]=8).¹¹ In a RCT of innercity children who had been switched to cows’ milk by 6 months, half (n=50) were randomized to receive iron-fortified formula for another year, resulting in a decreased risk of anemia at 24 months (0% vs 26%, NNT=4), and smaller declines in developmental functioning compared with those on cows’ milk.¹²

Recommendations from others

The CDC and the Institute of Medicine recommend parental dietary counseling, treatment with oral ferrous sulfate at 3 mg/kg/d for 3 months to restore iron stores, and monitoring of hemoglobin or hematocrit to assess response.^4,5 To prevent IDA, the American Academy of Pediatrics (AAP) recommends that all infants who are not breastfed or are partially breastfed should receive an iron-fortified formula (containing between 4.0–12 mg/L of iron) from birth to 12 months. The AAP also recommends that parents should refrain from feeding cow’s milk to infants until after age 12 months and introduce iron-enriched foods between ages 4 and 6 months.¹³

Evidence summary

Depletion of iron stores leads to IDA, which, among children, is associated with motor and cognitive deficits that may be irreversible. Little is known about whether iron deficiency, in the absence of anemia, results in physiologic sequelae. A Cochrane review of iron therapy for children with IDA aged >3 years found no short-term (5–11 days) improvement in Bayley scores of mental and motor development following iron therapy.¹ A 10-year longitudinal cohort study in Costa Rica found that adolescents treated for severe chronic IDA in infancy (n=48) scored 0.4 to 0.7 standard deviations lower on cognitive and motor testing relative to controls (n=114).² In an Indonesian randomized controlled trial (RCT), baseline Bayley scores were 10% to 15% lower (P<.01) for infants (12–18 months) with IDA compared with both nonanemic iron-deficient and iron-sufficient infants.³ Following treatment with ferrous sulfate (3 mg/kg/d of elemental iron) for 4 months, the IDA infants’ Bayley scores improved compared with those of nonanemic children.

Consensus recommendations suggest that iron deficiency should be the presumptive diagnosis in a child with anemia, and that a trial of ferrous sulfate at a dose of 3 mg/kg/d of elemental iron be instituted because of low cost, tolerability, and relative simplicity.^4,5 In a prospective study of 75 1-year-olds with anemia (hemoglobin <11.0 g/dL), 45% achieved an increase in hemoglobin ≥1 g/dL after 3 months of iron therapy (3 mg/kg/d).⁶ An RCT of 278 nonanemic 1-year-olds found no difference in adverse effects from this dose compared with placebo.⁷ However, an analysis of data from NHANES III showed that a Hgb <11.0 had a positive predictive value of just 29% and sensitivity of 30% for diagnosing iron-deficiency in children aged <3 years.⁸

The recommended dose of 3 mg/kg/d was derived from models of bioavailability and iron needs⁹; no studies compare alternative doses. An RCT of 557 anemic children under 24 months of age in Ghana demonstrated that ferrous sulfate (5 mg/kg/d) given once daily was equivalent to 3-times-daily dosing in terms of effectiveness (61% vs 56%) and tolerance.¹⁰ Less frequent dosing has been studied in developing countries with mixed results.

Because anemia may lead to developmental impairment, primary prevention is critical. In a cohort study, infants given iron-fortified formula (n=98) were less likely to become iron-deficient by their 12-month visit than infants fed whole cow milk (n=69) (11.2% vs 24.6%, number needed to treat [NNT]=8).¹¹ In a RCT of innercity children who had been switched to cows’ milk by 6 months, half (n=50) were randomized to receive iron-fortified formula for another year, resulting in a decreased risk of anemia at 24 months (0% vs 26%, NNT=4), and smaller declines in developmental functioning compared with those on cows’ milk.¹²

Recommendations from others

The CDC and the Institute of Medicine recommend parental dietary counseling, treatment with oral ferrous sulfate at 3 mg/kg/d for 3 months to restore iron stores, and monitoring of hemoglobin or hematocrit to assess response.^4,5 To prevent IDA, the American Academy of Pediatrics (AAP) recommends that all infants who are not breastfed or are partially breastfed should receive an iron-fortified formula (containing between 4.0–12 mg/L of iron) from birth to 12 months. The AAP also recommends that parents should refrain from feeding cow’s milk to infants until after age 12 months and introduce iron-enriched foods between ages 4 and 6 months.¹³

Evidence summary

Patients who experience TIAs are at high risk for stroke and need adequate preventative therapies. A meta-analysis evaluated 158 randomized trials involving primary and secondary prevention of stroke, concluding that antiplatelet therapy results in a 30% reduction in occurrence of ischemic stroke (95% confidence interval [CI], 24–35; P<.0001).¹ Data that evaluate the antiplatelet efficacy following a TIA while on aspirin therapy are limited.

Combination aspirin and dipyridamole (Aggrenox) therapy reduces the risk of secondary stroke. Several randomized controlled trials (RCTs) that evaluated this combination for prevention of stroke included patients who had TIAs. Although the combination reduced the occurrence of a subsequent stroke, the difference was not significant compared with aspirin alone, possibly due to the use of high-dose aspirin in the comparison group.^2,3 A large-scale RCT including patients with previous stroke or TIA concluded that the combination of aspirin and extended-release dipyridamole reduced the occurrence of stroke by 23% (P<.001) compared with aspirin (number needed to treat [NNT]=35; 95% CI, 20–130).⁴ In this trial, patients with a prior TIA comprised only one quarter of the patients studied.⁴ Subgroup analyses of patients on aspirin prior to experiencing a TIA have not been reported.

Thienopyridines may be considered for secondary stroke prevention for patients at high risk for a cardioembolic event. An RCT studying secondary prevention of stroke, in which 50% of the study population experienced a TIA as their qualifying event, concluded that ticlopidine (Ticlid) reduced the risk of stroke by 21% (95% CI, 0.04–0.38; P=.024).⁵ Subgroup analysis indicated that ticlopidine provided superior benefit for patients on aspirin or anticoagulant therapy at the time of their qualifying event.⁶ Unlike ticlopidine, clopidogrel (Plavix) does not have significant hematologic side effects. An RCT comparing clopidogrel with aspirin found a nonsignificant risk reduction of 0.3% (95% CI, -0.03 to 0.9; P=.26) in occurrence of stroke when compared with aspirin for patients with previous stroke, myocardial infarction, or peripheral arterial disease. A 0.9% absolute reduction in combined risk of cardioembolic events was reported for patients randomized to clopidogrel (NNT=111; 95% CI, 57–2454; P=.043), but patients with TIA were excluded from the study population.⁷

A large-scale RCT concluded that the combination of clopidogrel and aspirin did not provide additional benefit in reducing a combined endpoint of cardioembolic events in comparison with aspirin alone for patients with prior stroke or TIA. The combination resulted in a significantly greater number of life-threatening and major bleeding complications.⁸

Recommendations from others

The American Heart Association addresses the lack of evidence in treating patients who experience a TIA while on aspirin therapy. They recommend therapy be individualized for each patient to receive either the combination of extended-release dipyridamole and aspirin or clopidogrel daily for secondary prevention.⁹ Clopidogrel 75 mg daily is recommended over ticlopidine 250 mg twice daily due to its favorable safety profile.¹⁰ Similarly, the American College of Chest Physicians recommends use of dipyridamole and aspirin 200/25 mg twice daily or clopidogrel 75 mg daily.¹¹

Evidence summary

Patients who experience TIAs are at high risk for stroke and need adequate preventative therapies. A meta-analysis evaluated 158 randomized trials involving primary and secondary prevention of stroke, concluding that antiplatelet therapy results in a 30% reduction in occurrence of ischemic stroke (95% confidence interval [CI], 24–35; P<.0001).¹ Data that evaluate the antiplatelet efficacy following a TIA while on aspirin therapy are limited.

Combination aspirin and dipyridamole (Aggrenox) therapy reduces the risk of secondary stroke. Several randomized controlled trials (RCTs) that evaluated this combination for prevention of stroke included patients who had TIAs. Although the combination reduced the occurrence of a subsequent stroke, the difference was not significant compared with aspirin alone, possibly due to the use of high-dose aspirin in the comparison group.^2,3 A large-scale RCT including patients with previous stroke or TIA concluded that the combination of aspirin and extended-release dipyridamole reduced the occurrence of stroke by 23% (P<.001) compared with aspirin (number needed to treat [NNT]=35; 95% CI, 20–130).⁴ In this trial, patients with a prior TIA comprised only one quarter of the patients studied.⁴ Subgroup analyses of patients on aspirin prior to experiencing a TIA have not been reported.

Thienopyridines may be considered for secondary stroke prevention for patients at high risk for a cardioembolic event. An RCT studying secondary prevention of stroke, in which 50% of the study population experienced a TIA as their qualifying event, concluded that ticlopidine (Ticlid) reduced the risk of stroke by 21% (95% CI, 0.04–0.38; P=.024).⁵ Subgroup analysis indicated that ticlopidine provided superior benefit for patients on aspirin or anticoagulant therapy at the time of their qualifying event.⁶ Unlike ticlopidine, clopidogrel (Plavix) does not have significant hematologic side effects. An RCT comparing clopidogrel with aspirin found a nonsignificant risk reduction of 0.3% (95% CI, -0.03 to 0.9; P=.26) in occurrence of stroke when compared with aspirin for patients with previous stroke, myocardial infarction, or peripheral arterial disease. A 0.9% absolute reduction in combined risk of cardioembolic events was reported for patients randomized to clopidogrel (NNT=111; 95% CI, 57–2454; P=.043), but patients with TIA were excluded from the study population.⁷

A large-scale RCT concluded that the combination of clopidogrel and aspirin did not provide additional benefit in reducing a combined endpoint of cardioembolic events in comparison with aspirin alone for patients with prior stroke or TIA. The combination resulted in a significantly greater number of life-threatening and major bleeding complications.⁸

Recommendations from others

The American Heart Association addresses the lack of evidence in treating patients who experience a TIA while on aspirin therapy. They recommend therapy be individualized for each patient to receive either the combination of extended-release dipyridamole and aspirin or clopidogrel daily for secondary prevention.⁹ Clopidogrel 75 mg daily is recommended over ticlopidine 250 mg twice daily due to its favorable safety profile.¹⁰ Similarly, the American College of Chest Physicians recommends use of dipyridamole and aspirin 200/25 mg twice daily or clopidogrel 75 mg daily.¹¹

Evidence summary

Patients who experience TIAs are at high risk for stroke and need adequate preventative therapies. A meta-analysis evaluated 158 randomized trials involving primary and secondary prevention of stroke, concluding that antiplatelet therapy results in a 30% reduction in occurrence of ischemic stroke (95% confidence interval [CI], 24–35; P<.0001).¹ Data that evaluate the antiplatelet efficacy following a TIA while on aspirin therapy are limited.

Combination aspirin and dipyridamole (Aggrenox) therapy reduces the risk of secondary stroke. Several randomized controlled trials (RCTs) that evaluated this combination for prevention of stroke included patients who had TIAs. Although the combination reduced the occurrence of a subsequent stroke, the difference was not significant compared with aspirin alone, possibly due to the use of high-dose aspirin in the comparison group.^2,3 A large-scale RCT including patients with previous stroke or TIA concluded that the combination of aspirin and extended-release dipyridamole reduced the occurrence of stroke by 23% (P<.001) compared with aspirin (number needed to treat [NNT]=35; 95% CI, 20–130).⁴ In this trial, patients with a prior TIA comprised only one quarter of the patients studied.⁴ Subgroup analyses of patients on aspirin prior to experiencing a TIA have not been reported.

Thienopyridines may be considered for secondary stroke prevention for patients at high risk for a cardioembolic event. An RCT studying secondary prevention of stroke, in which 50% of the study population experienced a TIA as their qualifying event, concluded that ticlopidine (Ticlid) reduced the risk of stroke by 21% (95% CI, 0.04–0.38; P=.024).⁵ Subgroup analysis indicated that ticlopidine provided superior benefit for patients on aspirin or anticoagulant therapy at the time of their qualifying event.⁶ Unlike ticlopidine, clopidogrel (Plavix) does not have significant hematologic side effects. An RCT comparing clopidogrel with aspirin found a nonsignificant risk reduction of 0.3% (95% CI, -0.03 to 0.9; P=.26) in occurrence of stroke when compared with aspirin for patients with previous stroke, myocardial infarction, or peripheral arterial disease. A 0.9% absolute reduction in combined risk of cardioembolic events was reported for patients randomized to clopidogrel (NNT=111; 95% CI, 57–2454; P=.043), but patients with TIA were excluded from the study population.⁷

A large-scale RCT concluded that the combination of clopidogrel and aspirin did not provide additional benefit in reducing a combined endpoint of cardioembolic events in comparison with aspirin alone for patients with prior stroke or TIA. The combination resulted in a significantly greater number of life-threatening and major bleeding complications.⁸

Recommendations from others

The American Heart Association addresses the lack of evidence in treating patients who experience a TIA while on aspirin therapy. They recommend therapy be individualized for each patient to receive either the combination of extended-release dipyridamole and aspirin or clopidogrel daily for secondary prevention.⁹ Clopidogrel 75 mg daily is recommended over ticlopidine 250 mg twice daily due to its favorable safety profile.¹⁰ Similarly, the American College of Chest Physicians recommends use of dipyridamole and aspirin 200/25 mg twice daily or clopidogrel 75 mg daily.¹¹

Evidence summary

Rotator cuff tears can cause shoulder pain, decreased strength, and decreased range of motion. Clinical findings associated with a rotator cuff injury can vary. Full-thickness and partial-thickness tears may present differently, and it is important to test clinically for both of these conditions.

A meta-analysis of 10 cohort studies found the overall sensitivity and specificity of a clinical exam to rule out a full-thickness rotator cuff tear to be 0.9 (95% confidence interval [CI], 0.87–0.93) and 0.54 (95% CI, 0.47–0.61).¹ However, no single physical exam finding provided comparable accuracy. Another prospective study of 400 patients comparing 23 different clinical exams found that 3 simple clinical tests—supraspinatus weakness, weakness in external rotation, and the presence of impingement—were highly predictive of rotator cuff tear. When all 3 tests were positive, or when 2 tests were positive for a patient aged >60 years, there was a 98% chance of the patient having a rotator cuff tear.²

Ultrasound can be used to evaluate both suspected full- and partial-thickness rotator cuff tears. In a systematic review of 38 cohort studies, the overall sensitivity and specificity of ultrasound for full-thickness rotator cuff tears was 0.87 (95% CI, 0.84–0.89) and 0.96 (95% CI, 0.94–0.97).¹ For partial-thickness tears, ultrasound sensitivity was 0.67 (95% CI, 0.61–0.73).¹ The incidence of rotator cuff tears increases with age and with athletic activity.⁵

Positive and negative predictive values of a test depend on the prevalence of the condition in the study population. In the case of rotator cuff tears, such differences in prevalence of rotator cuff tears in the 38 cohort studies left it unclear whether a negative ultrasound could reliably rule out a tear.

A meta-analysis of 29 cohort studies of MRI for the diagnosis of full-thickness tears found a pooled sensitivity of 0.89 (95% CI, 0.86–0.92) and a pooled specificity of 0.93 (95% CI, 0.91–0.95), respectively.¹ For partial-thickness tears, the pooled MRI sensitivity was lower at 0.44 (95% CI, 0.36–0.51), but with a high specificity of 0.90 (95% CI, 0.87–0.92).¹ This implies that MRI is the most valuable test to rule out a partial-thickness tear. However, we found no studies that directly compared the test characteristics of ultrasound and MRI.

Conventional arthrography can be used as an invasive alternative to MRI imaging for full-thickness tears, particularly when an implanted device precludes the use of MRI. One prospective trial (in which patients were randomized to the order in which MRI or arthrography were performed) of 38 patients showed arthrography to have a sensitivity of 0.50 and a specificity of 0.96 when used to diagnose full-thickness tears.^3,6

Magnetic resonance arthrography (MRA), based on 6 cohort studies, may be accurate in the diagnosis of a full-thickness tear, with a sensitivity of 0.95 (95% CI, 0.82–0.98) and specificity of 0.93 (95% CI, 0.84–0.97).¹ In these studies, diagnosis of partial-thickness tears with MRA was inconsistent.¹ The invasiveness of MRA limits its utility as compared with MRI and ultrasound. The TABLE summarizes these findings.

TABLE
Summary of test characteristics of diagnostic studies for rotator cuff injuries

Recommendations from others

The American Academy of Orthopaedic Surgeons has a clinical guideline on shoulder pain,⁴ and the Brigham and Women’s Hospital has a guide to the prevention, diagnosis and treatment of upper extremity musculoskeletal disorders.⁵ These guidelines emphasize the importance and utility of physical examination of the shoulder. A patient with a full-thickness tear will likely demonstrate compromised strength in shoulder active mid-arc abduction and resisted external rotation with elbow flexed at patient’s side. However, a partial tear might not compromise strength. Atrophy of the infraspinatus or supra-spinatus muscles is sometimes seen with a full-thickness tear that is several weeks old.⁵

Following a clinical assessment, the guidelines give no preference to any of the diagnostic tests mentioned above, with the exception of arthrography in the presence of implantable devices. Plain X-rays are typically unrevealing, but could be used to rule out other reasons for pain, such as calcific tendonitis.

Evidence summary

Rotator cuff tears can cause shoulder pain, decreased strength, and decreased range of motion. Clinical findings associated with a rotator cuff injury can vary. Full-thickness and partial-thickness tears may present differently, and it is important to test clinically for both of these conditions.

A meta-analysis of 10 cohort studies found the overall sensitivity and specificity of a clinical exam to rule out a full-thickness rotator cuff tear to be 0.9 (95% confidence interval [CI], 0.87–0.93) and 0.54 (95% CI, 0.47–0.61).¹ However, no single physical exam finding provided comparable accuracy. Another prospective study of 400 patients comparing 23 different clinical exams found that 3 simple clinical tests—supraspinatus weakness, weakness in external rotation, and the presence of impingement—were highly predictive of rotator cuff tear. When all 3 tests were positive, or when 2 tests were positive for a patient aged >60 years, there was a 98% chance of the patient having a rotator cuff tear.²

Ultrasound can be used to evaluate both suspected full- and partial-thickness rotator cuff tears. In a systematic review of 38 cohort studies, the overall sensitivity and specificity of ultrasound for full-thickness rotator cuff tears was 0.87 (95% CI, 0.84–0.89) and 0.96 (95% CI, 0.94–0.97).¹ For partial-thickness tears, ultrasound sensitivity was 0.67 (95% CI, 0.61–0.73).¹ The incidence of rotator cuff tears increases with age and with athletic activity.⁵

Positive and negative predictive values of a test depend on the prevalence of the condition in the study population. In the case of rotator cuff tears, such differences in prevalence of rotator cuff tears in the 38 cohort studies left it unclear whether a negative ultrasound could reliably rule out a tear.

A meta-analysis of 29 cohort studies of MRI for the diagnosis of full-thickness tears found a pooled sensitivity of 0.89 (95% CI, 0.86–0.92) and a pooled specificity of 0.93 (95% CI, 0.91–0.95), respectively.¹ For partial-thickness tears, the pooled MRI sensitivity was lower at 0.44 (95% CI, 0.36–0.51), but with a high specificity of 0.90 (95% CI, 0.87–0.92).¹ This implies that MRI is the most valuable test to rule out a partial-thickness tear. However, we found no studies that directly compared the test characteristics of ultrasound and MRI.

Conventional arthrography can be used as an invasive alternative to MRI imaging for full-thickness tears, particularly when an implanted device precludes the use of MRI. One prospective trial (in which patients were randomized to the order in which MRI or arthrography were performed) of 38 patients showed arthrography to have a sensitivity of 0.50 and a specificity of 0.96 when used to diagnose full-thickness tears.^3,6

Magnetic resonance arthrography (MRA), based on 6 cohort studies, may be accurate in the diagnosis of a full-thickness tear, with a sensitivity of 0.95 (95% CI, 0.82–0.98) and specificity of 0.93 (95% CI, 0.84–0.97).¹ In these studies, diagnosis of partial-thickness tears with MRA was inconsistent.¹ The invasiveness of MRA limits its utility as compared with MRI and ultrasound. The TABLE summarizes these findings.

TABLE
Summary of test characteristics of diagnostic studies for rotator cuff injuries

Recommendations from others

The American Academy of Orthopaedic Surgeons has a clinical guideline on shoulder pain,⁴ and the Brigham and Women’s Hospital has a guide to the prevention, diagnosis and treatment of upper extremity musculoskeletal disorders.⁵ These guidelines emphasize the importance and utility of physical examination of the shoulder. A patient with a full-thickness tear will likely demonstrate compromised strength in shoulder active mid-arc abduction and resisted external rotation with elbow flexed at patient’s side. However, a partial tear might not compromise strength. Atrophy of the infraspinatus or supra-spinatus muscles is sometimes seen with a full-thickness tear that is several weeks old.⁵

Following a clinical assessment, the guidelines give no preference to any of the diagnostic tests mentioned above, with the exception of arthrography in the presence of implantable devices. Plain X-rays are typically unrevealing, but could be used to rule out other reasons for pain, such as calcific tendonitis.

Evidence summary

Rotator cuff tears can cause shoulder pain, decreased strength, and decreased range of motion. Clinical findings associated with a rotator cuff injury can vary. Full-thickness and partial-thickness tears may present differently, and it is important to test clinically for both of these conditions.

A meta-analysis of 10 cohort studies found the overall sensitivity and specificity of a clinical exam to rule out a full-thickness rotator cuff tear to be 0.9 (95% confidence interval [CI], 0.87–0.93) and 0.54 (95% CI, 0.47–0.61).¹ However, no single physical exam finding provided comparable accuracy. Another prospective study of 400 patients comparing 23 different clinical exams found that 3 simple clinical tests—supraspinatus weakness, weakness in external rotation, and the presence of impingement—were highly predictive of rotator cuff tear. When all 3 tests were positive, or when 2 tests were positive for a patient aged >60 years, there was a 98% chance of the patient having a rotator cuff tear.²

Ultrasound can be used to evaluate both suspected full- and partial-thickness rotator cuff tears. In a systematic review of 38 cohort studies, the overall sensitivity and specificity of ultrasound for full-thickness rotator cuff tears was 0.87 (95% CI, 0.84–0.89) and 0.96 (95% CI, 0.94–0.97).¹ For partial-thickness tears, ultrasound sensitivity was 0.67 (95% CI, 0.61–0.73).¹ The incidence of rotator cuff tears increases with age and with athletic activity.⁵

Positive and negative predictive values of a test depend on the prevalence of the condition in the study population. In the case of rotator cuff tears, such differences in prevalence of rotator cuff tears in the 38 cohort studies left it unclear whether a negative ultrasound could reliably rule out a tear.

A meta-analysis of 29 cohort studies of MRI for the diagnosis of full-thickness tears found a pooled sensitivity of 0.89 (95% CI, 0.86–0.92) and a pooled specificity of 0.93 (95% CI, 0.91–0.95), respectively.¹ For partial-thickness tears, the pooled MRI sensitivity was lower at 0.44 (95% CI, 0.36–0.51), but with a high specificity of 0.90 (95% CI, 0.87–0.92).¹ This implies that MRI is the most valuable test to rule out a partial-thickness tear. However, we found no studies that directly compared the test characteristics of ultrasound and MRI.

Conventional arthrography can be used as an invasive alternative to MRI imaging for full-thickness tears, particularly when an implanted device precludes the use of MRI. One prospective trial (in which patients were randomized to the order in which MRI or arthrography were performed) of 38 patients showed arthrography to have a sensitivity of 0.50 and a specificity of 0.96 when used to diagnose full-thickness tears.^3,6

Magnetic resonance arthrography (MRA), based on 6 cohort studies, may be accurate in the diagnosis of a full-thickness tear, with a sensitivity of 0.95 (95% CI, 0.82–0.98) and specificity of 0.93 (95% CI, 0.84–0.97).¹ In these studies, diagnosis of partial-thickness tears with MRA was inconsistent.¹ The invasiveness of MRA limits its utility as compared with MRI and ultrasound. The TABLE summarizes these findings.

TABLE
Summary of test characteristics of diagnostic studies for rotator cuff injuries

Recommendations from others

The American Academy of Orthopaedic Surgeons has a clinical guideline on shoulder pain,⁴ and the Brigham and Women’s Hospital has a guide to the prevention, diagnosis and treatment of upper extremity musculoskeletal disorders.⁵ These guidelines emphasize the importance and utility of physical examination of the shoulder. A patient with a full-thickness tear will likely demonstrate compromised strength in shoulder active mid-arc abduction and resisted external rotation with elbow flexed at patient’s side. However, a partial tear might not compromise strength. Atrophy of the infraspinatus or supra-spinatus muscles is sometimes seen with a full-thickness tear that is several weeks old.⁵

Following a clinical assessment, the guidelines give no preference to any of the diagnostic tests mentioned above, with the exception of arthrography in the presence of implantable devices. Plain X-rays are typically unrevealing, but could be used to rule out other reasons for pain, such as calcific tendonitis.

Evidence summary

The decreased TSH level seen in SCH results from the pituitary’s response to minor elevations in serum or tissue T4 and T3 levels.¹ Although these level remain within the normal range, the increases are sufficient to decrease the serum TSH level. The prevalence of SCH depends on the level of TSH used as a threshold. When the lower limit of TSH is set at 0.4 mIU/L, the prevalence was 3.2%.² When followed up at 1 year, 40% to 60% of subjects with suppressed TSH levels will have normal TSH values.³ Progression to overt hyperthyroidism is uncommon, occurring in 4.3% of subjects at 4 years.⁴ It is worth noting that individuals treated with levothyroxine have a prevalence of iatrogenic SCH from 14% to 21%.⁵

In patients with SCH aged >60 years, the cumulative incidence of atrial fibrillation after 10 years varied with the serum TSH level: it was 28% in those with serum TSH <0.1 mIU/L; 16% in those with values between 0.1 and 0.4 mIU/L, and 11 % in those with normal values.⁶ Patients with SCH have been reported to have increased heart rate, contractility, left ventricular mass, and increased risk of diastolic dysfunction and atrial arrhythmias.⁷ Patients aged >60 years with at least 1 suppressed TSH value have an increase in mortality over 5 years (standardized mortality ratio [SMR]=1.8; 95% confidence interval [CI], 1.2–2.7). At 10 years, the SMR was 1.2 (95% CI, 0.9–1.7). It appears that this is primarily related to cardiovascular mortality.⁸

There are little data on the effects of treating SCH. One study of postmenopausal women with endogenous SCH (defined as TSH <0.1 mIU/L) randomly assigned women to take methimazole (Tapazole) or placebo. Both groups were followed for 2 years and none received any medication with known effects on bone metabolism in the past or during the study period. The untreated patients with SCH had significantly higher bone mineral density loss (>5%) at both 18 and 24 months.⁹

Recommendations from others

A systematic review suggests the following regarding the evaluation and treatment of SCH.¹⁰

1. Exclude other causes of subnormal serum TSH concentration (TABLE)
2. Retest patients. Patients with atrial fibrillation, and cardiac disease, or a TSH <0.1 mIU/L should be retested in 2 to 4 weeks. Other patents can be retested in 3 months.
3. Patients whose TSH remains <0.1 mIU/L should undergo a radioactive iodine uptake scan. If the uptake is high (consistent with Graves’s disease or a focal nodule), treat as appropriate for that disease.

Younger patients (<60 years), with mild TSH suppression (0.1–0.45 mIU/L) or low radioactive iodine uptake can be followed with serial TSH testing at 3- to 12-month intervals. However, for these patients who also have cardiac disease, decreased bone mineral density, or symptoms suggestive of hyperthyroidism, thyroid suppression is recommended.

In patients aged >60 years with TSH <0.1 mIU/L, antithyroid treatment should be considered to decrease cardiac and bone loss complications.

Patients receiving thyroid replacement therapy should have their dose adjusted to maintain a normal serum TSH concentration. However, when thyroid hormone therapy is used for TSH suppression to prevent or reduce goiter growth or prevent recurrence of thyroid cancer, then a lower TSH may be unavoidable. The adverse effects can be minimized by treatment with the least level of suppression necessary to meet the desired goal.

Evidence summary

The decreased TSH level seen in SCH results from the pituitary’s response to minor elevations in serum or tissue T4 and T3 levels.¹ Although these level remain within the normal range, the increases are sufficient to decrease the serum TSH level. The prevalence of SCH depends on the level of TSH used as a threshold. When the lower limit of TSH is set at 0.4 mIU/L, the prevalence was 3.2%.² When followed up at 1 year, 40% to 60% of subjects with suppressed TSH levels will have normal TSH values.³ Progression to overt hyperthyroidism is uncommon, occurring in 4.3% of subjects at 4 years.⁴ It is worth noting that individuals treated with levothyroxine have a prevalence of iatrogenic SCH from 14% to 21%.⁵

In patients with SCH aged >60 years, the cumulative incidence of atrial fibrillation after 10 years varied with the serum TSH level: it was 28% in those with serum TSH <0.1 mIU/L; 16% in those with values between 0.1 and 0.4 mIU/L, and 11 % in those with normal values.⁶ Patients with SCH have been reported to have increased heart rate, contractility, left ventricular mass, and increased risk of diastolic dysfunction and atrial arrhythmias.⁷ Patients aged >60 years with at least 1 suppressed TSH value have an increase in mortality over 5 years (standardized mortality ratio [SMR]=1.8; 95% confidence interval [CI], 1.2–2.7). At 10 years, the SMR was 1.2 (95% CI, 0.9–1.7). It appears that this is primarily related to cardiovascular mortality.⁸

There are little data on the effects of treating SCH. One study of postmenopausal women with endogenous SCH (defined as TSH <0.1 mIU/L) randomly assigned women to take methimazole (Tapazole) or placebo. Both groups were followed for 2 years and none received any medication with known effects on bone metabolism in the past or during the study period. The untreated patients with SCH had significantly higher bone mineral density loss (>5%) at both 18 and 24 months.⁹

Recommendations from others

A systematic review suggests the following regarding the evaluation and treatment of SCH.¹⁰

1. Exclude other causes of subnormal serum TSH concentration (TABLE)
2. Retest patients. Patients with atrial fibrillation, and cardiac disease, or a TSH <0.1 mIU/L should be retested in 2 to 4 weeks. Other patents can be retested in 3 months.
3. Patients whose TSH remains <0.1 mIU/L should undergo a radioactive iodine uptake scan. If the uptake is high (consistent with Graves’s disease or a focal nodule), treat as appropriate for that disease.

Younger patients (<60 years), with mild TSH suppression (0.1–0.45 mIU/L) or low radioactive iodine uptake can be followed with serial TSH testing at 3- to 12-month intervals. However, for these patients who also have cardiac disease, decreased bone mineral density, or symptoms suggestive of hyperthyroidism, thyroid suppression is recommended.

In patients aged >60 years with TSH <0.1 mIU/L, antithyroid treatment should be considered to decrease cardiac and bone loss complications.

Patients receiving thyroid replacement therapy should have their dose adjusted to maintain a normal serum TSH concentration. However, when thyroid hormone therapy is used for TSH suppression to prevent or reduce goiter growth or prevent recurrence of thyroid cancer, then a lower TSH may be unavoidable. The adverse effects can be minimized by treatment with the least level of suppression necessary to meet the desired goal.

Evidence summary

The decreased TSH level seen in SCH results from the pituitary’s response to minor elevations in serum or tissue T4 and T3 levels.¹ Although these level remain within the normal range, the increases are sufficient to decrease the serum TSH level. The prevalence of SCH depends on the level of TSH used as a threshold. When the lower limit of TSH is set at 0.4 mIU/L, the prevalence was 3.2%.² When followed up at 1 year, 40% to 60% of subjects with suppressed TSH levels will have normal TSH values.³ Progression to overt hyperthyroidism is uncommon, occurring in 4.3% of subjects at 4 years.⁴ It is worth noting that individuals treated with levothyroxine have a prevalence of iatrogenic SCH from 14% to 21%.⁵

In patients with SCH aged >60 years, the cumulative incidence of atrial fibrillation after 10 years varied with the serum TSH level: it was 28% in those with serum TSH <0.1 mIU/L; 16% in those with values between 0.1 and 0.4 mIU/L, and 11 % in those with normal values.⁶ Patients with SCH have been reported to have increased heart rate, contractility, left ventricular mass, and increased risk of diastolic dysfunction and atrial arrhythmias.⁷ Patients aged >60 years with at least 1 suppressed TSH value have an increase in mortality over 5 years (standardized mortality ratio [SMR]=1.8; 95% confidence interval [CI], 1.2–2.7). At 10 years, the SMR was 1.2 (95% CI, 0.9–1.7). It appears that this is primarily related to cardiovascular mortality.⁸

There are little data on the effects of treating SCH. One study of postmenopausal women with endogenous SCH (defined as TSH <0.1 mIU/L) randomly assigned women to take methimazole (Tapazole) or placebo. Both groups were followed for 2 years and none received any medication with known effects on bone metabolism in the past or during the study period. The untreated patients with SCH had significantly higher bone mineral density loss (>5%) at both 18 and 24 months.⁹

Recommendations from others

A systematic review suggests the following regarding the evaluation and treatment of SCH.¹⁰

1. Exclude other causes of subnormal serum TSH concentration (TABLE)
2. Retest patients. Patients with atrial fibrillation, and cardiac disease, or a TSH <0.1 mIU/L should be retested in 2 to 4 weeks. Other patents can be retested in 3 months.
3. Patients whose TSH remains <0.1 mIU/L should undergo a radioactive iodine uptake scan. If the uptake is high (consistent with Graves’s disease or a focal nodule), treat as appropriate for that disease.

Younger patients (<60 years), with mild TSH suppression (0.1–0.45 mIU/L) or low radioactive iodine uptake can be followed with serial TSH testing at 3- to 12-month intervals. However, for these patients who also have cardiac disease, decreased bone mineral density, or symptoms suggestive of hyperthyroidism, thyroid suppression is recommended.

In patients aged >60 years with TSH <0.1 mIU/L, antithyroid treatment should be considered to decrease cardiac and bone loss complications.

Patients receiving thyroid replacement therapy should have their dose adjusted to maintain a normal serum TSH concentration. However, when thyroid hormone therapy is used for TSH suppression to prevent or reduce goiter growth or prevent recurrence of thyroid cancer, then a lower TSH may be unavoidable. The adverse effects can be minimized by treatment with the least level of suppression necessary to meet the desired goal.

Evidence summary

Electrocardiograms have been a mainstay in the evaluation for MI for many years. A systematic review of the workup of acute chest pain found that the ECG was the most useful bedside test for MI.¹ In this review, ST segment elevation and Q waves were found to be equally reliable predictors of MI (positive likelihood ratio [LR+]=22). A normal ECG was also found to be the most important bedside finding for ruling out the diagnosis of MI (LR–=0.2).

New ST segment elevation is the most important ECG feature in increasing the probability of diagnosing an MI, with LRs ranging from 5.7 to 53.9.² Another systematic review revealed similar findings where ST segment elevation (most commonly defined as at least 1 mm in 2 or more contiguous limb leads or at least 2 mm in 2 contiguous precordial leads) had a LR+=13.1 (95% confidence interval [CI], 8.28–20.6).³ This review also found that a “completely normal” ECG is reasonably useful in ruling out MI with a LR–=0.14 (95% CI, 0.11–0.20).³

In 2001, a working group of the National Heart Attack Alert Program (NHAAP) performed a systematic review to define the accuracy of “out of hospital” ECG in the diagnosis of acute cardiac ischemia (ACI) and MI. Based on the 8 studies for which data were available, the random effects pooled sensitivity for acute MI was 68% (95% CI, 59%–76%), the specificity was 97% (95% CI, 89%–92%), and the diagnostic odds ratio (DOR) was 104 (95% CI, 48–224).⁴ (The DOR is the change in post-test odds from a negative test to a positive test. It is used as a summary measure in meta-analyses of diagnostic studies. A DOR of 1 represents a useless test, with higher values representing more useful tests.)

There were sparse data available in our search results that specifically addressed the effect of serial ECGs on accuracy of diagnosis of MI. Another systematic review performed by a NHAAP working group evaluating different technologies in the emergency department diagnosis of ACI found only 1 study on the accuracy of serial ECGs in acute MI (sensitivity 39%, specificity 88%).⁵

As part of the Myocardial Infarction Triage and Intervention Project, the investigators found that when compared with a single ECG, serial exams increased the diagnostic sensitivity for acute coronary syndrome from ~34% to 46% with a reduction in specificity from 96% to 93% and positive predictive value from 88% to 84%.⁶ This particular study was unusual in that it used the hospital discharge diagnosis to define the outcome. In most other studies, cardiac enzymes were used as the gold standard for defining outcome.

Recommendations from others

In 2000, the European Society of Cardiology and the American College of Cardiology (ACC) issued a joint consensus statement redefining MI in which ECG findings such as ST segment elevation or new Q waves were insufficient for the diagnosis of MI without concomitant detection of elevated blood levels of cardiac biomarkers such as troponins.⁷

The ACC also published guidelines for management of ST elevation MIs in 2004 that recommended obtaining a 12-lead ECG on all patients presenting with symptoms suggestive of MI. If the initial ECG was not diagnostic, the guideline suggested obtaining either serial ECGs at 5- to 10-minute intervals or continuous 12-lead ST segment monitoring in order to detect the development of ST elevation.⁸

Evidence summary

Electrocardiograms have been a mainstay in the evaluation for MI for many years. A systematic review of the workup of acute chest pain found that the ECG was the most useful bedside test for MI.¹ In this review, ST segment elevation and Q waves were found to be equally reliable predictors of MI (positive likelihood ratio [LR+]=22). A normal ECG was also found to be the most important bedside finding for ruling out the diagnosis of MI (LR–=0.2).

New ST segment elevation is the most important ECG feature in increasing the probability of diagnosing an MI, with LRs ranging from 5.7 to 53.9.² Another systematic review revealed similar findings where ST segment elevation (most commonly defined as at least 1 mm in 2 or more contiguous limb leads or at least 2 mm in 2 contiguous precordial leads) had a LR+=13.1 (95% confidence interval [CI], 8.28–20.6).³ This review also found that a “completely normal” ECG is reasonably useful in ruling out MI with a LR–=0.14 (95% CI, 0.11–0.20).³

In 2001, a working group of the National Heart Attack Alert Program (NHAAP) performed a systematic review to define the accuracy of “out of hospital” ECG in the diagnosis of acute cardiac ischemia (ACI) and MI. Based on the 8 studies for which data were available, the random effects pooled sensitivity for acute MI was 68% (95% CI, 59%–76%), the specificity was 97% (95% CI, 89%–92%), and the diagnostic odds ratio (DOR) was 104 (95% CI, 48–224).⁴ (The DOR is the change in post-test odds from a negative test to a positive test. It is used as a summary measure in meta-analyses of diagnostic studies. A DOR of 1 represents a useless test, with higher values representing more useful tests.)

There were sparse data available in our search results that specifically addressed the effect of serial ECGs on accuracy of diagnosis of MI. Another systematic review performed by a NHAAP working group evaluating different technologies in the emergency department diagnosis of ACI found only 1 study on the accuracy of serial ECGs in acute MI (sensitivity 39%, specificity 88%).⁵

As part of the Myocardial Infarction Triage and Intervention Project, the investigators found that when compared with a single ECG, serial exams increased the diagnostic sensitivity for acute coronary syndrome from ~34% to 46% with a reduction in specificity from 96% to 93% and positive predictive value from 88% to 84%.⁶ This particular study was unusual in that it used the hospital discharge diagnosis to define the outcome. In most other studies, cardiac enzymes were used as the gold standard for defining outcome.

Recommendations from others

In 2000, the European Society of Cardiology and the American College of Cardiology (ACC) issued a joint consensus statement redefining MI in which ECG findings such as ST segment elevation or new Q waves were insufficient for the diagnosis of MI without concomitant detection of elevated blood levels of cardiac biomarkers such as troponins.⁷

The ACC also published guidelines for management of ST elevation MIs in 2004 that recommended obtaining a 12-lead ECG on all patients presenting with symptoms suggestive of MI. If the initial ECG was not diagnostic, the guideline suggested obtaining either serial ECGs at 5- to 10-minute intervals or continuous 12-lead ST segment monitoring in order to detect the development of ST elevation.⁸

Evidence summary

Electrocardiograms have been a mainstay in the evaluation for MI for many years. A systematic review of the workup of acute chest pain found that the ECG was the most useful bedside test for MI.¹ In this review, ST segment elevation and Q waves were found to be equally reliable predictors of MI (positive likelihood ratio [LR+]=22). A normal ECG was also found to be the most important bedside finding for ruling out the diagnosis of MI (LR–=0.2).

New ST segment elevation is the most important ECG feature in increasing the probability of diagnosing an MI, with LRs ranging from 5.7 to 53.9.² Another systematic review revealed similar findings where ST segment elevation (most commonly defined as at least 1 mm in 2 or more contiguous limb leads or at least 2 mm in 2 contiguous precordial leads) had a LR+=13.1 (95% confidence interval [CI], 8.28–20.6).³ This review also found that a “completely normal” ECG is reasonably useful in ruling out MI with a LR–=0.14 (95% CI, 0.11–0.20).³

In 2001, a working group of the National Heart Attack Alert Program (NHAAP) performed a systematic review to define the accuracy of “out of hospital” ECG in the diagnosis of acute cardiac ischemia (ACI) and MI. Based on the 8 studies for which data were available, the random effects pooled sensitivity for acute MI was 68% (95% CI, 59%–76%), the specificity was 97% (95% CI, 89%–92%), and the diagnostic odds ratio (DOR) was 104 (95% CI, 48–224).⁴ (The DOR is the change in post-test odds from a negative test to a positive test. It is used as a summary measure in meta-analyses of diagnostic studies. A DOR of 1 represents a useless test, with higher values representing more useful tests.)

There were sparse data available in our search results that specifically addressed the effect of serial ECGs on accuracy of diagnosis of MI. Another systematic review performed by a NHAAP working group evaluating different technologies in the emergency department diagnosis of ACI found only 1 study on the accuracy of serial ECGs in acute MI (sensitivity 39%, specificity 88%).⁵

As part of the Myocardial Infarction Triage and Intervention Project, the investigators found that when compared with a single ECG, serial exams increased the diagnostic sensitivity for acute coronary syndrome from ~34% to 46% with a reduction in specificity from 96% to 93% and positive predictive value from 88% to 84%.⁶ This particular study was unusual in that it used the hospital discharge diagnosis to define the outcome. In most other studies, cardiac enzymes were used as the gold standard for defining outcome.

Recommendations from others

In 2000, the European Society of Cardiology and the American College of Cardiology (ACC) issued a joint consensus statement redefining MI in which ECG findings such as ST segment elevation or new Q waves were insufficient for the diagnosis of MI without concomitant detection of elevated blood levels of cardiac biomarkers such as troponins.⁷

The ACC also published guidelines for management of ST elevation MIs in 2004 that recommended obtaining a 12-lead ECG on all patients presenting with symptoms suggestive of MI. If the initial ECG was not diagnostic, the guideline suggested obtaining either serial ECGs at 5- to 10-minute intervals or continuous 12-lead ST segment monitoring in order to detect the development of ST elevation.⁸