Applied Evidence

The sobering truth about major depression is that too often it goes unrecognized or undertreated.^1,2 And even when it is treated correctly, up to 34% of patients fail to respond to treatment.³ In the United States alone, the lifetime prevalence of the disease is 16.2%, and more than 6% of adults experience symptoms of major depression in any given year.² Obviously, we need to do more for these patients.

Treatment-resistant depression has been defined as the failure to achieve remission after continuous therapy for about 6 to 12 weeks with an adequate dose of a single antidepressant.⁴ Remission is typically defined as a 50% reduction of scores on depression severity scales, with the 17-item Hamilton Rating Scale of Depression (HRSD17) and the 16-item Quick Inventory of Depressive Symptomatology–Self-Report (QIDS-SR16) being the most often used. An adequate dose is the lowest effective dose that doesn’t cause intolerable side effects.

What are our best options for treatment-resistant depression?

To answer this question, we reviewed all English language studies in PubMed or Medline that were performed among adults using the search terms “augmentation, antidepressants, major depression.” We excluded studies involving patients with comorbid anxiety, bipolar disorder, or other major mental illnesses.

Based on our review of the literature, we found support for several augmentative treatments for patients with treatment-resistant depression (TABLE). Of note, though: Most of these studies were randomized trials dating back nearly 2 decades and had limitations. Most were not blinded, nor did they have consistent placebo controls. The studies were typically small (albeit frequently still showing efficacy for the agent despite lower statistical power) and of relatively short duration (typically 6-14 weeks). There were few studies that looked at treatment approaches over longer periods or that considered indications and timelines for tapering of medications once remission had been achieved. A major exception to the rule was the Sequenced Treatment Alternatives to Relieve Depression (STAR*D) study, which we’ll discuss below.⁵

The research points to several viable options: Clinicians can switch antidepressants, augment these agents by adding others—usually nonantidepressants—or completely change the therapeutic approach. Since most of the research discusses augmentation, we’ll focus on that here. Because the decision to use electroconvulsive therapy, vagus nerve stimulators, or other nonpharmaceutical approaches is made in consultation with a psychiatrist, we will not discuss these options.

TABLE
How these agents compare for treatment-resistant depression

Lithium: Good results, but with side effects
Lithium is one of the oldest and most well studied agents.^6-9 Most studies used 900 mg divided into 3 daily doses and titrated to lithium plasma levels of 0.5 to 1 mmol/L. Unfortunately, though, lithium is more likely to cause side effects than many other psychotropic agents (>25%).^5-9 And while the research has found that lithium produces significant remission rates, it’s not very effective in patients who have failed multiple antidepressant trials, with only 12.5% to 15.9% of patients in that cohort achieving remission.^10,11

Thyroid hormone has only mild side effects
Low-dose thyroid supplementation has been used for decades in euthyroid patients with treatment-resistant depression; remission rates range from 25% to 59%.^10-13 Most studies used a low fixed-dose therapy between 25 and 50 mcg daily; side effects were mild and occurred at rates similar to placebo.¹³

Atypical antipsychotics: Many pluses, but weight gain is an issue
Atypical antipsychotics are an attractive alternative to typical antipsychotics for treatment-resistant depression. They are much less likely to cause extrapyramidal symptoms, tardive dyskinesia, and other motor symptoms, but as a trade-off they often cause weight gain, abnormal glucose metabolism, dyslipidemia, and hyperprolactinemia.

Aripiprazole (Abilify) is 1 of 2 medications approved by the Food and Drug Administration (FDA) for treatment-resistant depression. Studies have shown remission rates of 25% to 37%, with side effects in 16% to 25% of patients.^14-16 Akathisia, the major side effect, can be reduced by lowering the starting dose to 2.5 mg.¹⁷

Olanzapine, combined with fluoxetine in a fixed-dose pill (Symbyax), is the other FDA-approved agent for treatment-resistant depression. Trivedi and colleagues showed a remission rate of 25.5% and side effects ranging from 10% to 40%.¹⁸

Risperidone (Risperdal) rivals the efficacy of the medications previously discussed, but starting doses and titration schedules vary widely, making it difficult to determine which treatment course would be most efficacious.^19-21

Quetiapine (Seroquel) has produced mixed results in treatment-resistant patients. That may have been because some studies used lower daily doses—25 to 100 mg—vs 150 to 600 mg in other trials.²² Bauer found higher remission rates compared to placebo (36% vs 24%)²³ while Garakani²² did not (49% for quetiapine vs. 63% for placebo when using an intention-to-treat analysis; a similar lack of efficacy was found if those who dropped out of the study were excluded; P<0.29). Garakani also found that dry mouth, sedation, and other side effects occurred in up to 76% of patients. (Of note: Most studies of atypical antipsychotics are industry funded.)

Mirtazapine has not been well studied
Unfortunately, there are very little data on mirtazapine (Remeron). When the drug was added to ongoing antidepressant therapy, a single double-blind, randomized controlled trial found significantly better response rates compared with placebo.²⁴ One advantage of the drug was that it helped relieve the sexual side effects of ongoing selective serotonin reuptake inhibitor (SSRI) therapy.²⁵

Folic acid and SAMe also haven’t been well studied

Up to 50% of Americans have low levels of central nervous system L-methylfolate, which is a key co-factor in monoamine neurotransmitter production. Although lower plasma folate has been linked to depression, folate supplementation as a primary treatment for major depression has not been well studied and its use in treatment-resistant depression is limited to 1 study by Alpert and colleagues.²⁶ Using an open-label, nonplacebo-controlled design in which folinic acid—an activated form of folic acid—was compared with placebo, researchers found remission rates of 18%, which is not significantly higher than the placebo response seen in other studies.

Similarly, there’s limited research on S-adenosyl-L-methionine (SAMe). Using a similar open-label, nonplacebo-controlled design, Alpert found a 43% remission rate in patients with treatment-resistant depression. Side effects occurred in up to half of patients, prompting 6.6% of patients to leave the study.²⁷

Omega-3 fatty acids: The news is mixed
Data are contradictory on the value of omega-3 fatty acid for major depression^28,29 and the evidence to support its use in treatment-resistant depression is likewise limited and contradictory.³⁰ Until the data are more consistent and robust, it’s unclear whether omega-3 fatty acid supplementation can benefit patients.

Exercise helps
Some studies suggest that exercise can have a dose-responsive effect on clinical depression.³¹ As a result, the Treatment with Exercise Augmentation for Depression (TREAD) trial is underway to examine whether it can augment drug therapy. Preliminary evidence suggests that 30 minutes of aerobic exercise most days of the week can be effective.^32,33 (For more on exercise, see “Does exercise alleviate symptoms of depression?”)

Cognitive therapy helps patients, and there are no side effects

The STAR*D study showed equivalent efficacy in achieving remission (23% vs. 33%) when comparing augmentation with cognitive therapy and augmentation with medications, with a low rate of side effects seen for cognitive therapy.³⁴ Notably, it took longer to achieve remission (55 vs 40 days) when cognitive therapy was added to antidepressants. (For more on the STAR*D trial, see “What we learned from the STAR*D trial”) It’s also worth mentioning that other researchers did not find similar results for augmentation when they compared cognitive behavioral and brief supportive psychotherapy with medication therapy only.³⁵

What we learned from the STAR*D trial

The Sequenced Treatment Alternatives to Relieve Depression (STAR*D) study is noteworthy because of its size, duration, design, and impact on the treatment of depression.⁵ This was a multicenter study of 3671 patients with major unipolar depression. Remission was determined by 2 rating scales—HRSD17 and QIDS-SR16*—and side effects were measured at each visit. There was no placebo control, no consideration for atypical antipsychotic medications or other therapies, and randomization was limited.

The FIGURE below shows how STAR*D was structured. In Level 1, patients were treated with citalopram (Celexa). If they didn’t go into remission, they were encouraged to proceed to Level 2, which involved 3 arms: switching agents, augmenting agents, or cognitive therapy. Patients chose which arm they wanted and were randomized to the specific medications used in that level. Switch agents included bupropion SR (Wellbutrin SR), sertraline (Zoloft), and venlafaxine XR (Effexor XR). Augmentation was with buspirone (BuSpar) and bupropion. Cognitive therapy was with or without continued citalopram with a second augmentation level of bupropion SR or venlafaxine XR. Level 3 also involved switch and augmentation arms using nortriptyline (Aventyl, Pamelor) or mirtazapine (Remeron) for switch agents and lithium or thyroxine for augmentation. Level 4 involved switching to tranylcypromine (Parnate) or venlafaxine XR plus mirtazapine.

Remission rates were 36.8% for those in Level 1, 30.6% in Level 2, 13.7% in Level 3, and 13% in Level 4. The cumulative remission rate was 67%. While remission rates for switch strategies in Levels 2 and 3 appeared to be lower for pharmacologic agents compared with augmentation strategies (27% vs 35% in Level 2 and 10.7% vs 20.5% in Level 3), the sample sizes were too small for the differences to be statistically significant, and the study was underpowered to establish the superiority of switch vs augmentation strategies.

Finally, it’s important to mention that patients of primary care physicians had similar remission rates, when compared with psychiatrists’ patients.⁴² Keep in mind, however, that the care provided in the STAR*D trial was structured and protocol-driven. After randomization into 1 medication group or another, dose adjustments were standardized at set intervals and based on inadequate response to validated depression severity assessment tools. Outcomes were patient-based and uniformly applied. This study structure may explain why remission rates were identical regardless of the specialty of the treating physician.

*The 17-item Hamilton Rating Scale of Depression and the 16-item Quick Inventory of Depressive Symptomatology–Self Report.

FIGURE
Structure of the Sequenced Treatment Alternatives to Relieve Depression (STAR*D)

These agents don’t appear helpful in augmentation efforts

Several other agents have been studied with no effect on remission rates. These include pindolol,³⁶ modafinil,³⁷ buspirone, lamotrigine,³⁸ stimulants, and estrogen replacement.

Consider side effects, cost, and patient preference

Because most studies showed that the efficacy of the tested drugs was similar, how do you decide which augmenting agent to prescribe? The usual standard is lithium, which offers rates of remission that are high, but not statistically significantly better when compared with thyroid supplementation and cognitive therapy. Quetiapine, while not more efficacious, may lower scores on depression rating scales more quickly than lithium.³⁹ Additionally, the STAR*D trial suggests that many agents may be used in augmentation with similar results.

In the final analysis, the family physician has to consider factors other than efficacy. You also have to factor in the costs of medicines and lab testing, patient preference, and side effects. Lithium is most likely to cause side effects. Atypical antipsychotics seem to have lower short-term side effect profiles with efficacy similar to cognitive therapy. However, the potential drawbacks of antipsychotics, including aberrations in glucose metabolism, weight, and lipid profiles, are not typically seen in short-term studies. Obviously, more long-term studies are necessary before 1 agent can be deemed superior.

When should you stop therapy?

That question still doesn’t have a definitive answer. The STAR*D trial found that patients who achieved full remission were less likely to relapse/worsen than those who had only a partial response. The time to relapse ranged from 2.5 to 4.5 months and was shorter for those patients requiring 2 or more levels of treatment. As there was no control of the therapeutic interventions used during this time, we can’t be certain about what caused the relapse, but burden of disease, income levels, and ethnicity all played roles in symptom severity, decreasing remission rates, and increasing relapse rates.^40,41

While we cannot identify the ideal or even preferred duration of augmentation in patients with treatment-resistant depression, it seems clear that relapse is extremely common and tends to occur relatively early after achieving remission.

CORRESPONDENCE Paul Hicks, MD, Department of Family and Community Medicine, 1450 North Cherry Avenue, Tucson, AZ 85719; [email protected]

The sobering truth about major depression is that too often it goes unrecognized or undertreated.^1,2 And even when it is treated correctly, up to 34% of patients fail to respond to treatment.³ In the United States alone, the lifetime prevalence of the disease is 16.2%, and more than 6% of adults experience symptoms of major depression in any given year.² Obviously, we need to do more for these patients.

Treatment-resistant depression has been defined as the failure to achieve remission after continuous therapy for about 6 to 12 weeks with an adequate dose of a single antidepressant.⁴ Remission is typically defined as a 50% reduction of scores on depression severity scales, with the 17-item Hamilton Rating Scale of Depression (HRSD17) and the 16-item Quick Inventory of Depressive Symptomatology–Self-Report (QIDS-SR16) being the most often used. An adequate dose is the lowest effective dose that doesn’t cause intolerable side effects.

What are our best options for treatment-resistant depression?

To answer this question, we reviewed all English language studies in PubMed or Medline that were performed among adults using the search terms “augmentation, antidepressants, major depression.” We excluded studies involving patients with comorbid anxiety, bipolar disorder, or other major mental illnesses.

Based on our review of the literature, we found support for several augmentative treatments for patients with treatment-resistant depression (TABLE). Of note, though: Most of these studies were randomized trials dating back nearly 2 decades and had limitations. Most were not blinded, nor did they have consistent placebo controls. The studies were typically small (albeit frequently still showing efficacy for the agent despite lower statistical power) and of relatively short duration (typically 6-14 weeks). There were few studies that looked at treatment approaches over longer periods or that considered indications and timelines for tapering of medications once remission had been achieved. A major exception to the rule was the Sequenced Treatment Alternatives to Relieve Depression (STAR*D) study, which we’ll discuss below.⁵

The research points to several viable options: Clinicians can switch antidepressants, augment these agents by adding others—usually nonantidepressants—or completely change the therapeutic approach. Since most of the research discusses augmentation, we’ll focus on that here. Because the decision to use electroconvulsive therapy, vagus nerve stimulators, or other nonpharmaceutical approaches is made in consultation with a psychiatrist, we will not discuss these options.

TABLE
How these agents compare for treatment-resistant depression

Lithium: Good results, but with side effects
Lithium is one of the oldest and most well studied agents.^6-9 Most studies used 900 mg divided into 3 daily doses and titrated to lithium plasma levels of 0.5 to 1 mmol/L. Unfortunately, though, lithium is more likely to cause side effects than many other psychotropic agents (>25%).^5-9 And while the research has found that lithium produces significant remission rates, it’s not very effective in patients who have failed multiple antidepressant trials, with only 12.5% to 15.9% of patients in that cohort achieving remission.^10,11

Thyroid hormone has only mild side effects
Low-dose thyroid supplementation has been used for decades in euthyroid patients with treatment-resistant depression; remission rates range from 25% to 59%.^10-13 Most studies used a low fixed-dose therapy between 25 and 50 mcg daily; side effects were mild and occurred at rates similar to placebo.¹³

Atypical antipsychotics: Many pluses, but weight gain is an issue
Atypical antipsychotics are an attractive alternative to typical antipsychotics for treatment-resistant depression. They are much less likely to cause extrapyramidal symptoms, tardive dyskinesia, and other motor symptoms, but as a trade-off they often cause weight gain, abnormal glucose metabolism, dyslipidemia, and hyperprolactinemia.

Aripiprazole (Abilify) is 1 of 2 medications approved by the Food and Drug Administration (FDA) for treatment-resistant depression. Studies have shown remission rates of 25% to 37%, with side effects in 16% to 25% of patients.^14-16 Akathisia, the major side effect, can be reduced by lowering the starting dose to 2.5 mg.¹⁷

Olanzapine, combined with fluoxetine in a fixed-dose pill (Symbyax), is the other FDA-approved agent for treatment-resistant depression. Trivedi and colleagues showed a remission rate of 25.5% and side effects ranging from 10% to 40%.¹⁸

Risperidone (Risperdal) rivals the efficacy of the medications previously discussed, but starting doses and titration schedules vary widely, making it difficult to determine which treatment course would be most efficacious.^19-21

Quetiapine (Seroquel) has produced mixed results in treatment-resistant patients. That may have been because some studies used lower daily doses—25 to 100 mg—vs 150 to 600 mg in other trials.²² Bauer found higher remission rates compared to placebo (36% vs 24%)²³ while Garakani²² did not (49% for quetiapine vs. 63% for placebo when using an intention-to-treat analysis; a similar lack of efficacy was found if those who dropped out of the study were excluded; P<0.29). Garakani also found that dry mouth, sedation, and other side effects occurred in up to 76% of patients. (Of note: Most studies of atypical antipsychotics are industry funded.)

Mirtazapine has not been well studied
Unfortunately, there are very little data on mirtazapine (Remeron). When the drug was added to ongoing antidepressant therapy, a single double-blind, randomized controlled trial found significantly better response rates compared with placebo.²⁴ One advantage of the drug was that it helped relieve the sexual side effects of ongoing selective serotonin reuptake inhibitor (SSRI) therapy.²⁵

Folic acid and SAMe also haven’t been well studied

Up to 50% of Americans have low levels of central nervous system L-methylfolate, which is a key co-factor in monoamine neurotransmitter production. Although lower plasma folate has been linked to depression, folate supplementation as a primary treatment for major depression has not been well studied and its use in treatment-resistant depression is limited to 1 study by Alpert and colleagues.²⁶ Using an open-label, nonplacebo-controlled design in which folinic acid—an activated form of folic acid—was compared with placebo, researchers found remission rates of 18%, which is not significantly higher than the placebo response seen in other studies.

Similarly, there’s limited research on S-adenosyl-L-methionine (SAMe). Using a similar open-label, nonplacebo-controlled design, Alpert found a 43% remission rate in patients with treatment-resistant depression. Side effects occurred in up to half of patients, prompting 6.6% of patients to leave the study.²⁷

Omega-3 fatty acids: The news is mixed
Data are contradictory on the value of omega-3 fatty acid for major depression^28,29 and the evidence to support its use in treatment-resistant depression is likewise limited and contradictory.³⁰ Until the data are more consistent and robust, it’s unclear whether omega-3 fatty acid supplementation can benefit patients.

Exercise helps
Some studies suggest that exercise can have a dose-responsive effect on clinical depression.³¹ As a result, the Treatment with Exercise Augmentation for Depression (TREAD) trial is underway to examine whether it can augment drug therapy. Preliminary evidence suggests that 30 minutes of aerobic exercise most days of the week can be effective.^32,33 (For more on exercise, see “Does exercise alleviate symptoms of depression?”)

Cognitive therapy helps patients, and there are no side effects

The STAR*D study showed equivalent efficacy in achieving remission (23% vs. 33%) when comparing augmentation with cognitive therapy and augmentation with medications, with a low rate of side effects seen for cognitive therapy.³⁴ Notably, it took longer to achieve remission (55 vs 40 days) when cognitive therapy was added to antidepressants. (For more on the STAR*D trial, see “What we learned from the STAR*D trial”) It’s also worth mentioning that other researchers did not find similar results for augmentation when they compared cognitive behavioral and brief supportive psychotherapy with medication therapy only.³⁵

What we learned from the STAR*D trial

The Sequenced Treatment Alternatives to Relieve Depression (STAR*D) study is noteworthy because of its size, duration, design, and impact on the treatment of depression.⁵ This was a multicenter study of 3671 patients with major unipolar depression. Remission was determined by 2 rating scales—HRSD17 and QIDS-SR16*—and side effects were measured at each visit. There was no placebo control, no consideration for atypical antipsychotic medications or other therapies, and randomization was limited.

The FIGURE below shows how STAR*D was structured. In Level 1, patients were treated with citalopram (Celexa). If they didn’t go into remission, they were encouraged to proceed to Level 2, which involved 3 arms: switching agents, augmenting agents, or cognitive therapy. Patients chose which arm they wanted and were randomized to the specific medications used in that level. Switch agents included bupropion SR (Wellbutrin SR), sertraline (Zoloft), and venlafaxine XR (Effexor XR). Augmentation was with buspirone (BuSpar) and bupropion. Cognitive therapy was with or without continued citalopram with a second augmentation level of bupropion SR or venlafaxine XR. Level 3 also involved switch and augmentation arms using nortriptyline (Aventyl, Pamelor) or mirtazapine (Remeron) for switch agents and lithium or thyroxine for augmentation. Level 4 involved switching to tranylcypromine (Parnate) or venlafaxine XR plus mirtazapine.

Remission rates were 36.8% for those in Level 1, 30.6% in Level 2, 13.7% in Level 3, and 13% in Level 4. The cumulative remission rate was 67%. While remission rates for switch strategies in Levels 2 and 3 appeared to be lower for pharmacologic agents compared with augmentation strategies (27% vs 35% in Level 2 and 10.7% vs 20.5% in Level 3), the sample sizes were too small for the differences to be statistically significant, and the study was underpowered to establish the superiority of switch vs augmentation strategies.

Finally, it’s important to mention that patients of primary care physicians had similar remission rates, when compared with psychiatrists’ patients.⁴² Keep in mind, however, that the care provided in the STAR*D trial was structured and protocol-driven. After randomization into 1 medication group or another, dose adjustments were standardized at set intervals and based on inadequate response to validated depression severity assessment tools. Outcomes were patient-based and uniformly applied. This study structure may explain why remission rates were identical regardless of the specialty of the treating physician.

*The 17-item Hamilton Rating Scale of Depression and the 16-item Quick Inventory of Depressive Symptomatology–Self Report.

FIGURE
Structure of the Sequenced Treatment Alternatives to Relieve Depression (STAR*D)

These agents don’t appear helpful in augmentation efforts

Several other agents have been studied with no effect on remission rates. These include pindolol,³⁶ modafinil,³⁷ buspirone, lamotrigine,³⁸ stimulants, and estrogen replacement.

Consider side effects, cost, and patient preference

Because most studies showed that the efficacy of the tested drugs was similar, how do you decide which augmenting agent to prescribe? The usual standard is lithium, which offers rates of remission that are high, but not statistically significantly better when compared with thyroid supplementation and cognitive therapy. Quetiapine, while not more efficacious, may lower scores on depression rating scales more quickly than lithium.³⁹ Additionally, the STAR*D trial suggests that many agents may be used in augmentation with similar results.

In the final analysis, the family physician has to consider factors other than efficacy. You also have to factor in the costs of medicines and lab testing, patient preference, and side effects. Lithium is most likely to cause side effects. Atypical antipsychotics seem to have lower short-term side effect profiles with efficacy similar to cognitive therapy. However, the potential drawbacks of antipsychotics, including aberrations in glucose metabolism, weight, and lipid profiles, are not typically seen in short-term studies. Obviously, more long-term studies are necessary before 1 agent can be deemed superior.

When should you stop therapy?

That question still doesn’t have a definitive answer. The STAR*D trial found that patients who achieved full remission were less likely to relapse/worsen than those who had only a partial response. The time to relapse ranged from 2.5 to 4.5 months and was shorter for those patients requiring 2 or more levels of treatment. As there was no control of the therapeutic interventions used during this time, we can’t be certain about what caused the relapse, but burden of disease, income levels, and ethnicity all played roles in symptom severity, decreasing remission rates, and increasing relapse rates.^40,41

While we cannot identify the ideal or even preferred duration of augmentation in patients with treatment-resistant depression, it seems clear that relapse is extremely common and tends to occur relatively early after achieving remission.

CORRESPONDENCE Paul Hicks, MD, Department of Family and Community Medicine, 1450 North Cherry Avenue, Tucson, AZ 85719; [email protected]

The sobering truth about major depression is that too often it goes unrecognized or undertreated.^1,2 And even when it is treated correctly, up to 34% of patients fail to respond to treatment.³ In the United States alone, the lifetime prevalence of the disease is 16.2%, and more than 6% of adults experience symptoms of major depression in any given year.² Obviously, we need to do more for these patients.

Treatment-resistant depression has been defined as the failure to achieve remission after continuous therapy for about 6 to 12 weeks with an adequate dose of a single antidepressant.⁴ Remission is typically defined as a 50% reduction of scores on depression severity scales, with the 17-item Hamilton Rating Scale of Depression (HRSD17) and the 16-item Quick Inventory of Depressive Symptomatology–Self-Report (QIDS-SR16) being the most often used. An adequate dose is the lowest effective dose that doesn’t cause intolerable side effects.

What are our best options for treatment-resistant depression?

To answer this question, we reviewed all English language studies in PubMed or Medline that were performed among adults using the search terms “augmentation, antidepressants, major depression.” We excluded studies involving patients with comorbid anxiety, bipolar disorder, or other major mental illnesses.

Based on our review of the literature, we found support for several augmentative treatments for patients with treatment-resistant depression (TABLE). Of note, though: Most of these studies were randomized trials dating back nearly 2 decades and had limitations. Most were not blinded, nor did they have consistent placebo controls. The studies were typically small (albeit frequently still showing efficacy for the agent despite lower statistical power) and of relatively short duration (typically 6-14 weeks). There were few studies that looked at treatment approaches over longer periods or that considered indications and timelines for tapering of medications once remission had been achieved. A major exception to the rule was the Sequenced Treatment Alternatives to Relieve Depression (STAR*D) study, which we’ll discuss below.⁵

The research points to several viable options: Clinicians can switch antidepressants, augment these agents by adding others—usually nonantidepressants—or completely change the therapeutic approach. Since most of the research discusses augmentation, we’ll focus on that here. Because the decision to use electroconvulsive therapy, vagus nerve stimulators, or other nonpharmaceutical approaches is made in consultation with a psychiatrist, we will not discuss these options.

TABLE
How these agents compare for treatment-resistant depression

Lithium: Good results, but with side effects
Lithium is one of the oldest and most well studied agents.^6-9 Most studies used 900 mg divided into 3 daily doses and titrated to lithium plasma levels of 0.5 to 1 mmol/L. Unfortunately, though, lithium is more likely to cause side effects than many other psychotropic agents (>25%).^5-9 And while the research has found that lithium produces significant remission rates, it’s not very effective in patients who have failed multiple antidepressant trials, with only 12.5% to 15.9% of patients in that cohort achieving remission.^10,11

Thyroid hormone has only mild side effects
Low-dose thyroid supplementation has been used for decades in euthyroid patients with treatment-resistant depression; remission rates range from 25% to 59%.^10-13 Most studies used a low fixed-dose therapy between 25 and 50 mcg daily; side effects were mild and occurred at rates similar to placebo.¹³

Atypical antipsychotics: Many pluses, but weight gain is an issue
Atypical antipsychotics are an attractive alternative to typical antipsychotics for treatment-resistant depression. They are much less likely to cause extrapyramidal symptoms, tardive dyskinesia, and other motor symptoms, but as a trade-off they often cause weight gain, abnormal glucose metabolism, dyslipidemia, and hyperprolactinemia.

Aripiprazole (Abilify) is 1 of 2 medications approved by the Food and Drug Administration (FDA) for treatment-resistant depression. Studies have shown remission rates of 25% to 37%, with side effects in 16% to 25% of patients.^14-16 Akathisia, the major side effect, can be reduced by lowering the starting dose to 2.5 mg.¹⁷

Olanzapine, combined with fluoxetine in a fixed-dose pill (Symbyax), is the other FDA-approved agent for treatment-resistant depression. Trivedi and colleagues showed a remission rate of 25.5% and side effects ranging from 10% to 40%.¹⁸

Risperidone (Risperdal) rivals the efficacy of the medications previously discussed, but starting doses and titration schedules vary widely, making it difficult to determine which treatment course would be most efficacious.^19-21

Quetiapine (Seroquel) has produced mixed results in treatment-resistant patients. That may have been because some studies used lower daily doses—25 to 100 mg—vs 150 to 600 mg in other trials.²² Bauer found higher remission rates compared to placebo (36% vs 24%)²³ while Garakani²² did not (49% for quetiapine vs. 63% for placebo when using an intention-to-treat analysis; a similar lack of efficacy was found if those who dropped out of the study were excluded; P<0.29). Garakani also found that dry mouth, sedation, and other side effects occurred in up to 76% of patients. (Of note: Most studies of atypical antipsychotics are industry funded.)

Mirtazapine has not been well studied
Unfortunately, there are very little data on mirtazapine (Remeron). When the drug was added to ongoing antidepressant therapy, a single double-blind, randomized controlled trial found significantly better response rates compared with placebo.²⁴ One advantage of the drug was that it helped relieve the sexual side effects of ongoing selective serotonin reuptake inhibitor (SSRI) therapy.²⁵

Folic acid and SAMe also haven’t been well studied

Up to 50% of Americans have low levels of central nervous system L-methylfolate, which is a key co-factor in monoamine neurotransmitter production. Although lower plasma folate has been linked to depression, folate supplementation as a primary treatment for major depression has not been well studied and its use in treatment-resistant depression is limited to 1 study by Alpert and colleagues.²⁶ Using an open-label, nonplacebo-controlled design in which folinic acid—an activated form of folic acid—was compared with placebo, researchers found remission rates of 18%, which is not significantly higher than the placebo response seen in other studies.

Similarly, there’s limited research on S-adenosyl-L-methionine (SAMe). Using a similar open-label, nonplacebo-controlled design, Alpert found a 43% remission rate in patients with treatment-resistant depression. Side effects occurred in up to half of patients, prompting 6.6% of patients to leave the study.²⁷

Omega-3 fatty acids: The news is mixed
Data are contradictory on the value of omega-3 fatty acid for major depression^28,29 and the evidence to support its use in treatment-resistant depression is likewise limited and contradictory.³⁰ Until the data are more consistent and robust, it’s unclear whether omega-3 fatty acid supplementation can benefit patients.

Exercise helps
Some studies suggest that exercise can have a dose-responsive effect on clinical depression.³¹ As a result, the Treatment with Exercise Augmentation for Depression (TREAD) trial is underway to examine whether it can augment drug therapy. Preliminary evidence suggests that 30 minutes of aerobic exercise most days of the week can be effective.^32,33 (For more on exercise, see “Does exercise alleviate symptoms of depression?”)

Cognitive therapy helps patients, and there are no side effects

The STAR*D study showed equivalent efficacy in achieving remission (23% vs. 33%) when comparing augmentation with cognitive therapy and augmentation with medications, with a low rate of side effects seen for cognitive therapy.³⁴ Notably, it took longer to achieve remission (55 vs 40 days) when cognitive therapy was added to antidepressants. (For more on the STAR*D trial, see “What we learned from the STAR*D trial”) It’s also worth mentioning that other researchers did not find similar results for augmentation when they compared cognitive behavioral and brief supportive psychotherapy with medication therapy only.³⁵

What we learned from the STAR*D trial

The Sequenced Treatment Alternatives to Relieve Depression (STAR*D) study is noteworthy because of its size, duration, design, and impact on the treatment of depression.⁵ This was a multicenter study of 3671 patients with major unipolar depression. Remission was determined by 2 rating scales—HRSD17 and QIDS-SR16*—and side effects were measured at each visit. There was no placebo control, no consideration for atypical antipsychotic medications or other therapies, and randomization was limited.

The FIGURE below shows how STAR*D was structured. In Level 1, patients were treated with citalopram (Celexa). If they didn’t go into remission, they were encouraged to proceed to Level 2, which involved 3 arms: switching agents, augmenting agents, or cognitive therapy. Patients chose which arm they wanted and were randomized to the specific medications used in that level. Switch agents included bupropion SR (Wellbutrin SR), sertraline (Zoloft), and venlafaxine XR (Effexor XR). Augmentation was with buspirone (BuSpar) and bupropion. Cognitive therapy was with or without continued citalopram with a second augmentation level of bupropion SR or venlafaxine XR. Level 3 also involved switch and augmentation arms using nortriptyline (Aventyl, Pamelor) or mirtazapine (Remeron) for switch agents and lithium or thyroxine for augmentation. Level 4 involved switching to tranylcypromine (Parnate) or venlafaxine XR plus mirtazapine.

Remission rates were 36.8% for those in Level 1, 30.6% in Level 2, 13.7% in Level 3, and 13% in Level 4. The cumulative remission rate was 67%. While remission rates for switch strategies in Levels 2 and 3 appeared to be lower for pharmacologic agents compared with augmentation strategies (27% vs 35% in Level 2 and 10.7% vs 20.5% in Level 3), the sample sizes were too small for the differences to be statistically significant, and the study was underpowered to establish the superiority of switch vs augmentation strategies.

Finally, it’s important to mention that patients of primary care physicians had similar remission rates, when compared with psychiatrists’ patients.⁴² Keep in mind, however, that the care provided in the STAR*D trial was structured and protocol-driven. After randomization into 1 medication group or another, dose adjustments were standardized at set intervals and based on inadequate response to validated depression severity assessment tools. Outcomes were patient-based and uniformly applied. This study structure may explain why remission rates were identical regardless of the specialty of the treating physician.

*The 17-item Hamilton Rating Scale of Depression and the 16-item Quick Inventory of Depressive Symptomatology–Self Report.

FIGURE
Structure of the Sequenced Treatment Alternatives to Relieve Depression (STAR*D)

These agents don’t appear helpful in augmentation efforts

Several other agents have been studied with no effect on remission rates. These include pindolol,³⁶ modafinil,³⁷ buspirone, lamotrigine,³⁸ stimulants, and estrogen replacement.

Consider side effects, cost, and patient preference

Because most studies showed that the efficacy of the tested drugs was similar, how do you decide which augmenting agent to prescribe? The usual standard is lithium, which offers rates of remission that are high, but not statistically significantly better when compared with thyroid supplementation and cognitive therapy. Quetiapine, while not more efficacious, may lower scores on depression rating scales more quickly than lithium.³⁹ Additionally, the STAR*D trial suggests that many agents may be used in augmentation with similar results.

In the final analysis, the family physician has to consider factors other than efficacy. You also have to factor in the costs of medicines and lab testing, patient preference, and side effects. Lithium is most likely to cause side effects. Atypical antipsychotics seem to have lower short-term side effect profiles with efficacy similar to cognitive therapy. However, the potential drawbacks of antipsychotics, including aberrations in glucose metabolism, weight, and lipid profiles, are not typically seen in short-term studies. Obviously, more long-term studies are necessary before 1 agent can be deemed superior.

When should you stop therapy?

That question still doesn’t have a definitive answer. The STAR*D trial found that patients who achieved full remission were less likely to relapse/worsen than those who had only a partial response. The time to relapse ranged from 2.5 to 4.5 months and was shorter for those patients requiring 2 or more levels of treatment. As there was no control of the therapeutic interventions used during this time, we can’t be certain about what caused the relapse, but burden of disease, income levels, and ethnicity all played roles in symptom severity, decreasing remission rates, and increasing relapse rates.^40,41

While we cannot identify the ideal or even preferred duration of augmentation in patients with treatment-resistant depression, it seems clear that relapse is extremely common and tends to occur relatively early after achieving remission.

CORRESPONDENCE Paul Hicks, MD, Department of Family and Community Medicine, 1450 North Cherry Avenue, Tucson, AZ 85719; [email protected]

Though we have the knowledge and the means to reduce complications of type 2 diabetes mellitus (T2DM), most patients may not be reaching all the glycemic goals necessary to achieve optimal risk reduction.¹ Maintaining an acceptable level of glycosylated hemoglobin (HbA1c) is one of the important glycemic goals. But that measurement is an average of glucose levels occurring over the prior 3 months. Regardless of a given HbA1c measurement, an emerging body of evidence supports the presumption that glycemic variability over each 24-hour cycle is an independent risk factor for vascular complications.^2-15

In this article, I review the literature pertaining to the risk associated with glycemic variability and to the benefit in correcting it. I also review the comparative outcomes achievable with normal human insulin and insulin analogs, as well as the advisability of starting insulin earlier in the management process.

Glycemic variability increases vascular risk independently

HbA1c, considered the gold standard for monitoring glycemic control in patients with T2DM, is an average of the full range of glucose values in the preceding 3 months, including fasting plasma glucose (FPG) and 2-hour postprandial glucose (PPG) levels. Studies have linked lowering HbA1c to reducing the risk and progression of micro- and macrovascular complications associated with diabetes.^16,17 But evidence shows that other glycemic values are also important.

The Diabetes Control and Complications Trial (DCCT) was a landmark study in which patients with type 1 diabetes mellitus who received targeted intensive insulin therapy experienced delayed onset and slowed progression of micro-vascular complications compared with those who received conventional insulin treatment.¹⁶ Interestingly, this study also reported that patients randomized to receive conventional insulin treatment did not exhibit a reduction in the risk of progression of microvascular disease despite having HbA1c values comparable to those in the intensive-treatment group. One hypothesis is that glucose excursions occurred more frequently in the conventionally treated group, which received fewer daily insulin injections.⁵

Acute glucose fluctuations during the postprandial period trigger oxidative stress and are more predictive of atherosclerosis development than are FPG or HbA1c^6,7 (see “Implications of glycemic variability” below). This suggests that therapy for patients with T2DM should not only target HbA1c as a long-term goal, but also aim to avoid acute glucose fluctuations as an immediate goal. Several studies have shown that postprandial hyperglycemia is an independent risk factor for vascular complications in patients with T2DM.^{2,7-9,12,14,15}

Evidence of increased vascular risk with glycemic variability. The Diabetes Epidemiology: COllaborative analysis of Diagnostic criteria in Europe study (DECODE) followed more than 25,000 patients for more than 7 years and found that increased mortality was more closely associated with increased 2-hour PPG levels than with FPG.¹⁴ In the Framingham Offspring Study, Meigs et al⁹ reported that, in nondiabetic subjects, an elevated glucose level 2 hours after an oral challenge increased the relative risk for cardiovascular disease by up to 40%, independent of fasting hyperglycemia.

Mixed outcomes with Hba1c reduction only. Macrovascular risk reduction with intensive HbA1c management was not apparent in 3 recent studies—Action to Control Cardiovascular Risk in Diabetes (ACCORD),¹⁸ Action in Diabetes and Vascular Disease: Preterax and Diamicron MR Controlled Evaluation (ADVANCE),¹⁹ and Veterans Affairs Diabetes Trial (VADT).²⁰ The ACCORD study, in fact, showed an increase in cardiovascular events in the intensively managed group (HbA1c target <6.0%). Indeed, previous studies had suggested an association between fasting hypoglycemia and poor cardiovascular outcomes.^3,4 Retrospective subanalysis of the ACCORD study suggested that patients with poorer glycemic control had a greater risk of hypoglycemia independent of HbA1c values, and that patients who had difficulty reaching lower HbA1c levels may have had poorer cardiovascular outcomes.²¹

The apparent absence of a reduction in macrovascular events in the ACCORD, ADVANCE, and VADT studies also suggests an additive effect of nonglycemic risk factors that frequently accompany diabetes—ie, hypertension, hyperlipidemia, and hypercoagulability/pro-inflammatory states.

Long-term follow-up in the United Kingdom Prospective Diabetes Study (UKPDS) showed ongoing risk reduction for both microvascular and macrovascular complications.²² A separate meta-analysis showed a significant 10% reduction in cardiovascular events with intensive glycemic control when data were combined from the ACCORD trial, ADVANCE trial, VADT, and the UKPDS.²³

An improvement in long-term outcomes for patients with T2DM might be expected when initiating a targeted, intensified, multi-factorial interventional regimen to reduce not only HbA1c, but also glucose variability. The STENO-2 trial showed that a targeted multifactorial treatment regimen in patients with T2DM could decrease long-term vascular complications.²⁴

Consider assessing true variability in your patients. Because postprandial glucose levels alone may not equate to overall glycemic variability, you may want to ask select patients to take readings with their glucose meters at various times of the day across several days to get a more accurate picture.⁵

FIGURE 1
How oxidative stress secondary to hyperglycemia leads to vascular complications in diabetes^66-68

Following through with targeted, intensified management

Consider the following treatment goals for patients with T2DM: (1) lowering HbA1c levels; (2) lowering fasting blood glucose levels; (3) minimizing glycemic variability, including postprandial glucose excursions. TABLE 1 lists the values that the American Diabetes Association (ADA) and the American Association of Clinical Endocrinologists (AACE) have assigned to these glycemic-control goals.

In addition to managing glycemic levels, reducing risk of cardiovascular disease in T2DM involves aggressive interventions, as needed, to correct blood pressure and lipid levels.^24,25

TABLE 1
Aim to reach 3 glycemic goals in treating type 2 diabetes mellitus

Challenges to achieving glycemic control
Despite current recommendations for more aggressive management of patients with T2DM,²⁵ estimates are that as many as 60% of patients with T2DM do not achieve glycemic targets, and, as the disease progresses, many of the available treatment options fail to sustain levels previously reached.^1,26,27

A shortcoming of older treatment strategies still in use is the slow transition to more effective therapy, resulting in long periods of inadequate glycemic control.¹ Brown et al²⁷ found that patients receiving monotherapy with either a sulfonylurea or metformin had HbA1c levels >8% for a mean of 20 months and 14 months, respectively, before treatment was changed. Current recommendations call for treatment changes within 2 to 3 months of initiation of therapy if the HbA1c goal is not reached.^27-29

Turning to insulin earlier. Insulin is most effective for lowering HbA1c and delaying subsequent complications related to diabetes; however, there is often reluctance to using it early in diabetes management. Consequently, by the time insulin therapy is started, many patients will have had unacceptable glycemic levels for 10 years or more and may already be developing complications.²⁷ And, as noted, the HbA1c level is an average measurement that does not detect glycemic variability. Continuous glucose monitoring will likely lead to more responsive adjustments in treatment regimens and to improved quality of care for patients with T2DM.

Insulin has many beneficial effects

Insulin exerts an anti-inflammatory effect by reducing the increase in C-reactive protein and serum amyloid A.³⁰ It also partially restores insulin-stimulated endothelial function,³¹ facilitates vasodilation by increasing nitric oxide production,³² and improves fibrinolytic profiles.³³ Early initiation of insulin therapy can increase peripheral insulin sensitivity and preserve beta cell function.^34-36

When oral agents have failed, insulin can significantly improve patients’ beta cell function,^34,35,37 and short periods of insulin therapy in patients newly diagnosed with T2DM may even set the foundation for better long-term control.^38,39

But not all insulin is alike
Ideally, insulin therapy should mimic physiologic insulin secretion. However, conventional human insulin products fail to do so because of their suboptimal pharmacodynamic profiles. With recombinant DNA technology, molecular modifications of the human insulin molecule have overcome some of the limitations of conventional human insulin products.

Unfortunately, many practitioners still hold insulin in reserve until combination therapy with oral agents has failed, possibly resulting in years of suboptimal glycemic control. Newer strategies recommend earlier initiation of insulin—ie, once diet and exercise fail, or when treatment with 1 oral agent fails. The development of insulin analogs is a significant milestone on the road to achieving improved outcomes for patients with T2DM.

Rapid-acting agents
Compared with regular human insulin, newer rapid-acting insulin analogs may improve glycemic control when used at mealtimes. However, due to their shorter half-lives, these insulin analogs require augmentation with basal insulin to control hyperglycemia between meals and during the night.

Insulin lispro was the first commercially available rapid-acting insulin analog, introduced in 1996. This agent differs from human insulin by an inversion of amino acid residues in positions 28 and 29 of the insulin B-chain. Inversion prevents the formation of hexamers and dimers that tend to diffuse more slowly, thereby facilitating a rapid uptake of the insulin analog into blood and tissues.^40,41 The second such agent, marketed in 2000, was insulin aspart, in which aspartic acid replaces proline at position 28 of the B-chain of human insulin.^41,42 The most recent rapid-acting analog is insulin glulisine, in which lysine replaces asparagine near the N-terminus of the B-chain, and glutamic acid replaces lysine near the C-terminus of human insulin.

The molecular changes made in creating these analogs allows them to dissociate quickly into monomers that are absorbed rapidly and achieve faster peak levels compared with regular human insulin.^41,42 These changes do not, however, interfere with the analogs’ ability to bind to the insulin receptor.^43,44

Dosing considerations. Absorption of regular human insulin is not sufficiently rapid at mealtimes to control prandial glucose levels.⁴⁵ Therefore, it is essential to give regular insulin 30 to 60 minutes before meals. For patients who have erratic daily schedules, adhering to this sort of routine can be difficult. But even if scheduling is not a problem, the prolonged duration of action of human insulin can predispose patients to hypoglycemia. Moreover, absorption of regular insulin can vary dramatically from day to day.^46,47

The insulin analogs correct the pharmacokinetic and pharmacodynamic deficiencies of regular insulin, producing plasma profiles that more closely simulate normal, physiologic meal-stimulated insulin release.^48-50 The 3 rapid-acting agents (aspart, glulisine, lispro) have very similar onset and duration of action, with peak effect occurring close to injection time (TABLE 2 and FIGURE 2).^48-50

Advantages of rapid-acting agents. These agents can be administered closer to meals, giving patients more flexibility and likely tighter postprandial glucose control, with reductions in glycemic excursions. Another advantage is the ability to better match insulin dose to anticipated carbohydrate in-take, affording better postprandial control.^51-53 Rapid-acting analogs also result in fewer episodes of hypoglycemia. In a meta-analysis of 2576 patients, hypoglycemic events occurred 25% less often with insulin lispro compared with regular human insulin in patients with type 1 diabetes mellitus (T1DM).⁵² In clinical trials, insulin aspart and insulin glulisine have also caused fewer hypoglycemic events compared with regular human insulin.^51-53

Long-acting agents
Basal, or long-acting, insulins are important for maintaining normoglycemia over 24 hours. Neutral protamine Hagedorn (NPH) insulin reaches its peak effect 4 to 10 hours after injection, and its total effect lasts only 12 to 18 hours. NPH is therefore often dosed twice daily. Absorption of NPH can vary significantly, causing day-to-day blood glucose fluctuations.^46,47 Therefore, this agent’s activity does not closely resemble normal physiologic basal insulin secretion.

The newer long-acting insulin analogs—insulin detemir and insulin glargine—were designed to more closely replicate normal physiologic basal insulin secretion. Insulin glargine was first to reach the market, in 2001. It contains glycine instead of asparagine in the alpha-chain and 2 arginine residues at the C-terminus, and the addition of zinc enhances the aggregation and slow release at a neutral pH. Insulin glargine precipitates in the subcutaneous tissue, which slows its absorption and results in a relatively flat insulin plasma profile and extended action.^54,55 Insulin detemir is a combination of the original insulin molecule and a saturated fatty acid (myristic acid). Insulin detemir is designed to bind albumin (98% albumin-bound in circulation) through this fatty acid chain in the plasma after injection, resulting in an extended plasma profile.^54,56 Insulin glargine and NPH form crystalline depots, but detemir is soluble and the subcutaneous depot remains in a liquid state; this may account for differences in absorption variability.⁵⁶

Advantages of the long-acting insulin analogs. Compared with conventional basal insulin such as NPH, the analogs have a prolonged duration of action (up to 24 hours) without pronounced peaks, permitting once-daily dosing in many patients (TABLE 2 and FIGURE 2).^46,47,50,55 The pharmacodynamic and pharmacokinetic properties of the long- acting agents make them less likely than NPH to cause nocturnal and overall hypoglycemia, a benefit that has been observed in several clinical trials.^57-61

Comparative clinical trials evaluating glycemic variability. Both of the long-acting analogs have shown lower within-subject variability in blood and plasma glucose measurements when compared with NPH.^62-64 In head-to-head comparisons of the analogs in glucose clamp studies, insulin detemir has demonstrated less within-subject variability of blood glucose levels than insulin glargine, in patients with T1DM or T2DM.^56,64 In clinical practice, different patients may have better results with one of these basal insulins as opposed to the other, and treatment choices will need to be tailored to the individual patient.

TABLE 2
Pharmacokinetic properties giving insulin analogs an advantage over regular insulin^46-50

FIGURE 2
Pharmacokinetic profiles of human insulin and insulin analogs

CORRESPONDENCE Eric L. Johnson, MD, University of North Dakota School of Medicine and Health Sciences, Department of Family & Community Medicine, 501 North Columbia Road, Grand Forks, ND 58202-9037; [email protected]

Though we have the knowledge and the means to reduce complications of type 2 diabetes mellitus (T2DM), most patients may not be reaching all the glycemic goals necessary to achieve optimal risk reduction.¹ Maintaining an acceptable level of glycosylated hemoglobin (HbA1c) is one of the important glycemic goals. But that measurement is an average of glucose levels occurring over the prior 3 months. Regardless of a given HbA1c measurement, an emerging body of evidence supports the presumption that glycemic variability over each 24-hour cycle is an independent risk factor for vascular complications.^2-15

In this article, I review the literature pertaining to the risk associated with glycemic variability and to the benefit in correcting it. I also review the comparative outcomes achievable with normal human insulin and insulin analogs, as well as the advisability of starting insulin earlier in the management process.

Glycemic variability increases vascular risk independently

HbA1c, considered the gold standard for monitoring glycemic control in patients with T2DM, is an average of the full range of glucose values in the preceding 3 months, including fasting plasma glucose (FPG) and 2-hour postprandial glucose (PPG) levels. Studies have linked lowering HbA1c to reducing the risk and progression of micro- and macrovascular complications associated with diabetes.^16,17 But evidence shows that other glycemic values are also important.

The Diabetes Control and Complications Trial (DCCT) was a landmark study in which patients with type 1 diabetes mellitus who received targeted intensive insulin therapy experienced delayed onset and slowed progression of micro-vascular complications compared with those who received conventional insulin treatment.¹⁶ Interestingly, this study also reported that patients randomized to receive conventional insulin treatment did not exhibit a reduction in the risk of progression of microvascular disease despite having HbA1c values comparable to those in the intensive-treatment group. One hypothesis is that glucose excursions occurred more frequently in the conventionally treated group, which received fewer daily insulin injections.⁵

Acute glucose fluctuations during the postprandial period trigger oxidative stress and are more predictive of atherosclerosis development than are FPG or HbA1c^6,7 (see “Implications of glycemic variability” below). This suggests that therapy for patients with T2DM should not only target HbA1c as a long-term goal, but also aim to avoid acute glucose fluctuations as an immediate goal. Several studies have shown that postprandial hyperglycemia is an independent risk factor for vascular complications in patients with T2DM.^{2,7-9,12,14,15}

Evidence of increased vascular risk with glycemic variability. The Diabetes Epidemiology: COllaborative analysis of Diagnostic criteria in Europe study (DECODE) followed more than 25,000 patients for more than 7 years and found that increased mortality was more closely associated with increased 2-hour PPG levels than with FPG.¹⁴ In the Framingham Offspring Study, Meigs et al⁹ reported that, in nondiabetic subjects, an elevated glucose level 2 hours after an oral challenge increased the relative risk for cardiovascular disease by up to 40%, independent of fasting hyperglycemia.

Mixed outcomes with Hba1c reduction only. Macrovascular risk reduction with intensive HbA1c management was not apparent in 3 recent studies—Action to Control Cardiovascular Risk in Diabetes (ACCORD),¹⁸ Action in Diabetes and Vascular Disease: Preterax and Diamicron MR Controlled Evaluation (ADVANCE),¹⁹ and Veterans Affairs Diabetes Trial (VADT).²⁰ The ACCORD study, in fact, showed an increase in cardiovascular events in the intensively managed group (HbA1c target <6.0%). Indeed, previous studies had suggested an association between fasting hypoglycemia and poor cardiovascular outcomes.^3,4 Retrospective subanalysis of the ACCORD study suggested that patients with poorer glycemic control had a greater risk of hypoglycemia independent of HbA1c values, and that patients who had difficulty reaching lower HbA1c levels may have had poorer cardiovascular outcomes.²¹

The apparent absence of a reduction in macrovascular events in the ACCORD, ADVANCE, and VADT studies also suggests an additive effect of nonglycemic risk factors that frequently accompany diabetes—ie, hypertension, hyperlipidemia, and hypercoagulability/pro-inflammatory states.

Long-term follow-up in the United Kingdom Prospective Diabetes Study (UKPDS) showed ongoing risk reduction for both microvascular and macrovascular complications.²² A separate meta-analysis showed a significant 10% reduction in cardiovascular events with intensive glycemic control when data were combined from the ACCORD trial, ADVANCE trial, VADT, and the UKPDS.²³

An improvement in long-term outcomes for patients with T2DM might be expected when initiating a targeted, intensified, multi-factorial interventional regimen to reduce not only HbA1c, but also glucose variability. The STENO-2 trial showed that a targeted multifactorial treatment regimen in patients with T2DM could decrease long-term vascular complications.²⁴

Consider assessing true variability in your patients. Because postprandial glucose levels alone may not equate to overall glycemic variability, you may want to ask select patients to take readings with their glucose meters at various times of the day across several days to get a more accurate picture.⁵

FIGURE 1
How oxidative stress secondary to hyperglycemia leads to vascular complications in diabetes^66-68

Following through with targeted, intensified management

Consider the following treatment goals for patients with T2DM: (1) lowering HbA1c levels; (2) lowering fasting blood glucose levels; (3) minimizing glycemic variability, including postprandial glucose excursions. TABLE 1 lists the values that the American Diabetes Association (ADA) and the American Association of Clinical Endocrinologists (AACE) have assigned to these glycemic-control goals.

In addition to managing glycemic levels, reducing risk of cardiovascular disease in T2DM involves aggressive interventions, as needed, to correct blood pressure and lipid levels.^24,25

TABLE 1
Aim to reach 3 glycemic goals in treating type 2 diabetes mellitus

Challenges to achieving glycemic control
Despite current recommendations for more aggressive management of patients with T2DM,²⁵ estimates are that as many as 60% of patients with T2DM do not achieve glycemic targets, and, as the disease progresses, many of the available treatment options fail to sustain levels previously reached.^1,26,27

A shortcoming of older treatment strategies still in use is the slow transition to more effective therapy, resulting in long periods of inadequate glycemic control.¹ Brown et al²⁷ found that patients receiving monotherapy with either a sulfonylurea or metformin had HbA1c levels >8% for a mean of 20 months and 14 months, respectively, before treatment was changed. Current recommendations call for treatment changes within 2 to 3 months of initiation of therapy if the HbA1c goal is not reached.^27-29

Turning to insulin earlier. Insulin is most effective for lowering HbA1c and delaying subsequent complications related to diabetes; however, there is often reluctance to using it early in diabetes management. Consequently, by the time insulin therapy is started, many patients will have had unacceptable glycemic levels for 10 years or more and may already be developing complications.²⁷ And, as noted, the HbA1c level is an average measurement that does not detect glycemic variability. Continuous glucose monitoring will likely lead to more responsive adjustments in treatment regimens and to improved quality of care for patients with T2DM.

Insulin has many beneficial effects

Insulin exerts an anti-inflammatory effect by reducing the increase in C-reactive protein and serum amyloid A.³⁰ It also partially restores insulin-stimulated endothelial function,³¹ facilitates vasodilation by increasing nitric oxide production,³² and improves fibrinolytic profiles.³³ Early initiation of insulin therapy can increase peripheral insulin sensitivity and preserve beta cell function.^34-36

When oral agents have failed, insulin can significantly improve patients’ beta cell function,^34,35,37 and short periods of insulin therapy in patients newly diagnosed with T2DM may even set the foundation for better long-term control.^38,39

But not all insulin is alike
Ideally, insulin therapy should mimic physiologic insulin secretion. However, conventional human insulin products fail to do so because of their suboptimal pharmacodynamic profiles. With recombinant DNA technology, molecular modifications of the human insulin molecule have overcome some of the limitations of conventional human insulin products.

Unfortunately, many practitioners still hold insulin in reserve until combination therapy with oral agents has failed, possibly resulting in years of suboptimal glycemic control. Newer strategies recommend earlier initiation of insulin—ie, once diet and exercise fail, or when treatment with 1 oral agent fails. The development of insulin analogs is a significant milestone on the road to achieving improved outcomes for patients with T2DM.

Rapid-acting agents
Compared with regular human insulin, newer rapid-acting insulin analogs may improve glycemic control when used at mealtimes. However, due to their shorter half-lives, these insulin analogs require augmentation with basal insulin to control hyperglycemia between meals and during the night.

Insulin lispro was the first commercially available rapid-acting insulin analog, introduced in 1996. This agent differs from human insulin by an inversion of amino acid residues in positions 28 and 29 of the insulin B-chain. Inversion prevents the formation of hexamers and dimers that tend to diffuse more slowly, thereby facilitating a rapid uptake of the insulin analog into blood and tissues.^40,41 The second such agent, marketed in 2000, was insulin aspart, in which aspartic acid replaces proline at position 28 of the B-chain of human insulin.^41,42 The most recent rapid-acting analog is insulin glulisine, in which lysine replaces asparagine near the N-terminus of the B-chain, and glutamic acid replaces lysine near the C-terminus of human insulin.

The molecular changes made in creating these analogs allows them to dissociate quickly into monomers that are absorbed rapidly and achieve faster peak levels compared with regular human insulin.^41,42 These changes do not, however, interfere with the analogs’ ability to bind to the insulin receptor.^43,44

Dosing considerations. Absorption of regular human insulin is not sufficiently rapid at mealtimes to control prandial glucose levels.⁴⁵ Therefore, it is essential to give regular insulin 30 to 60 minutes before meals. For patients who have erratic daily schedules, adhering to this sort of routine can be difficult. But even if scheduling is not a problem, the prolonged duration of action of human insulin can predispose patients to hypoglycemia. Moreover, absorption of regular insulin can vary dramatically from day to day.^46,47

The insulin analogs correct the pharmacokinetic and pharmacodynamic deficiencies of regular insulin, producing plasma profiles that more closely simulate normal, physiologic meal-stimulated insulin release.^48-50 The 3 rapid-acting agents (aspart, glulisine, lispro) have very similar onset and duration of action, with peak effect occurring close to injection time (TABLE 2 and FIGURE 2).^48-50

Advantages of rapid-acting agents. These agents can be administered closer to meals, giving patients more flexibility and likely tighter postprandial glucose control, with reductions in glycemic excursions. Another advantage is the ability to better match insulin dose to anticipated carbohydrate in-take, affording better postprandial control.^51-53 Rapid-acting analogs also result in fewer episodes of hypoglycemia. In a meta-analysis of 2576 patients, hypoglycemic events occurred 25% less often with insulin lispro compared with regular human insulin in patients with type 1 diabetes mellitus (T1DM).⁵² In clinical trials, insulin aspart and insulin glulisine have also caused fewer hypoglycemic events compared with regular human insulin.^51-53

Long-acting agents
Basal, or long-acting, insulins are important for maintaining normoglycemia over 24 hours. Neutral protamine Hagedorn (NPH) insulin reaches its peak effect 4 to 10 hours after injection, and its total effect lasts only 12 to 18 hours. NPH is therefore often dosed twice daily. Absorption of NPH can vary significantly, causing day-to-day blood glucose fluctuations.^46,47 Therefore, this agent’s activity does not closely resemble normal physiologic basal insulin secretion.

The newer long-acting insulin analogs—insulin detemir and insulin glargine—were designed to more closely replicate normal physiologic basal insulin secretion. Insulin glargine was first to reach the market, in 2001. It contains glycine instead of asparagine in the alpha-chain and 2 arginine residues at the C-terminus, and the addition of zinc enhances the aggregation and slow release at a neutral pH. Insulin glargine precipitates in the subcutaneous tissue, which slows its absorption and results in a relatively flat insulin plasma profile and extended action.^54,55 Insulin detemir is a combination of the original insulin molecule and a saturated fatty acid (myristic acid). Insulin detemir is designed to bind albumin (98% albumin-bound in circulation) through this fatty acid chain in the plasma after injection, resulting in an extended plasma profile.^54,56 Insulin glargine and NPH form crystalline depots, but detemir is soluble and the subcutaneous depot remains in a liquid state; this may account for differences in absorption variability.⁵⁶

Advantages of the long-acting insulin analogs. Compared with conventional basal insulin such as NPH, the analogs have a prolonged duration of action (up to 24 hours) without pronounced peaks, permitting once-daily dosing in many patients (TABLE 2 and FIGURE 2).^46,47,50,55 The pharmacodynamic and pharmacokinetic properties of the long- acting agents make them less likely than NPH to cause nocturnal and overall hypoglycemia, a benefit that has been observed in several clinical trials.^57-61

Comparative clinical trials evaluating glycemic variability. Both of the long-acting analogs have shown lower within-subject variability in blood and plasma glucose measurements when compared with NPH.^62-64 In head-to-head comparisons of the analogs in glucose clamp studies, insulin detemir has demonstrated less within-subject variability of blood glucose levels than insulin glargine, in patients with T1DM or T2DM.^56,64 In clinical practice, different patients may have better results with one of these basal insulins as opposed to the other, and treatment choices will need to be tailored to the individual patient.

TABLE 2
Pharmacokinetic properties giving insulin analogs an advantage over regular insulin^46-50

FIGURE 2
Pharmacokinetic profiles of human insulin and insulin analogs

CORRESPONDENCE Eric L. Johnson, MD, University of North Dakota School of Medicine and Health Sciences, Department of Family & Community Medicine, 501 North Columbia Road, Grand Forks, ND 58202-9037; [email protected]

Though we have the knowledge and the means to reduce complications of type 2 diabetes mellitus (T2DM), most patients may not be reaching all the glycemic goals necessary to achieve optimal risk reduction.¹ Maintaining an acceptable level of glycosylated hemoglobin (HbA1c) is one of the important glycemic goals. But that measurement is an average of glucose levels occurring over the prior 3 months. Regardless of a given HbA1c measurement, an emerging body of evidence supports the presumption that glycemic variability over each 24-hour cycle is an independent risk factor for vascular complications.^2-15

In this article, I review the literature pertaining to the risk associated with glycemic variability and to the benefit in correcting it. I also review the comparative outcomes achievable with normal human insulin and insulin analogs, as well as the advisability of starting insulin earlier in the management process.

Glycemic variability increases vascular risk independently

HbA1c, considered the gold standard for monitoring glycemic control in patients with T2DM, is an average of the full range of glucose values in the preceding 3 months, including fasting plasma glucose (FPG) and 2-hour postprandial glucose (PPG) levels. Studies have linked lowering HbA1c to reducing the risk and progression of micro- and macrovascular complications associated with diabetes.^16,17 But evidence shows that other glycemic values are also important.

The Diabetes Control and Complications Trial (DCCT) was a landmark study in which patients with type 1 diabetes mellitus who received targeted intensive insulin therapy experienced delayed onset and slowed progression of micro-vascular complications compared with those who received conventional insulin treatment.¹⁶ Interestingly, this study also reported that patients randomized to receive conventional insulin treatment did not exhibit a reduction in the risk of progression of microvascular disease despite having HbA1c values comparable to those in the intensive-treatment group. One hypothesis is that glucose excursions occurred more frequently in the conventionally treated group, which received fewer daily insulin injections.⁵

Acute glucose fluctuations during the postprandial period trigger oxidative stress and are more predictive of atherosclerosis development than are FPG or HbA1c^6,7 (see “Implications of glycemic variability” below). This suggests that therapy for patients with T2DM should not only target HbA1c as a long-term goal, but also aim to avoid acute glucose fluctuations as an immediate goal. Several studies have shown that postprandial hyperglycemia is an independent risk factor for vascular complications in patients with T2DM.^{2,7-9,12,14,15}

Evidence of increased vascular risk with glycemic variability. The Diabetes Epidemiology: COllaborative analysis of Diagnostic criteria in Europe study (DECODE) followed more than 25,000 patients for more than 7 years and found that increased mortality was more closely associated with increased 2-hour PPG levels than with FPG.¹⁴ In the Framingham Offspring Study, Meigs et al⁹ reported that, in nondiabetic subjects, an elevated glucose level 2 hours after an oral challenge increased the relative risk for cardiovascular disease by up to 40%, independent of fasting hyperglycemia.

Mixed outcomes with Hba1c reduction only. Macrovascular risk reduction with intensive HbA1c management was not apparent in 3 recent studies—Action to Control Cardiovascular Risk in Diabetes (ACCORD),¹⁸ Action in Diabetes and Vascular Disease: Preterax and Diamicron MR Controlled Evaluation (ADVANCE),¹⁹ and Veterans Affairs Diabetes Trial (VADT).²⁰ The ACCORD study, in fact, showed an increase in cardiovascular events in the intensively managed group (HbA1c target <6.0%). Indeed, previous studies had suggested an association between fasting hypoglycemia and poor cardiovascular outcomes.^3,4 Retrospective subanalysis of the ACCORD study suggested that patients with poorer glycemic control had a greater risk of hypoglycemia independent of HbA1c values, and that patients who had difficulty reaching lower HbA1c levels may have had poorer cardiovascular outcomes.²¹

The apparent absence of a reduction in macrovascular events in the ACCORD, ADVANCE, and VADT studies also suggests an additive effect of nonglycemic risk factors that frequently accompany diabetes—ie, hypertension, hyperlipidemia, and hypercoagulability/pro-inflammatory states.

Long-term follow-up in the United Kingdom Prospective Diabetes Study (UKPDS) showed ongoing risk reduction for both microvascular and macrovascular complications.²² A separate meta-analysis showed a significant 10% reduction in cardiovascular events with intensive glycemic control when data were combined from the ACCORD trial, ADVANCE trial, VADT, and the UKPDS.²³

An improvement in long-term outcomes for patients with T2DM might be expected when initiating a targeted, intensified, multi-factorial interventional regimen to reduce not only HbA1c, but also glucose variability. The STENO-2 trial showed that a targeted multifactorial treatment regimen in patients with T2DM could decrease long-term vascular complications.²⁴

Consider assessing true variability in your patients. Because postprandial glucose levels alone may not equate to overall glycemic variability, you may want to ask select patients to take readings with their glucose meters at various times of the day across several days to get a more accurate picture.⁵

FIGURE 1
How oxidative stress secondary to hyperglycemia leads to vascular complications in diabetes^66-68

Following through with targeted, intensified management

Consider the following treatment goals for patients with T2DM: (1) lowering HbA1c levels; (2) lowering fasting blood glucose levels; (3) minimizing glycemic variability, including postprandial glucose excursions. TABLE 1 lists the values that the American Diabetes Association (ADA) and the American Association of Clinical Endocrinologists (AACE) have assigned to these glycemic-control goals.

In addition to managing glycemic levels, reducing risk of cardiovascular disease in T2DM involves aggressive interventions, as needed, to correct blood pressure and lipid levels.^24,25

TABLE 1
Aim to reach 3 glycemic goals in treating type 2 diabetes mellitus

Challenges to achieving glycemic control
Despite current recommendations for more aggressive management of patients with T2DM,²⁵ estimates are that as many as 60% of patients with T2DM do not achieve glycemic targets, and, as the disease progresses, many of the available treatment options fail to sustain levels previously reached.^1,26,27

A shortcoming of older treatment strategies still in use is the slow transition to more effective therapy, resulting in long periods of inadequate glycemic control.¹ Brown et al²⁷ found that patients receiving monotherapy with either a sulfonylurea or metformin had HbA1c levels >8% for a mean of 20 months and 14 months, respectively, before treatment was changed. Current recommendations call for treatment changes within 2 to 3 months of initiation of therapy if the HbA1c goal is not reached.^27-29

Turning to insulin earlier. Insulin is most effective for lowering HbA1c and delaying subsequent complications related to diabetes; however, there is often reluctance to using it early in diabetes management. Consequently, by the time insulin therapy is started, many patients will have had unacceptable glycemic levels for 10 years or more and may already be developing complications.²⁷ And, as noted, the HbA1c level is an average measurement that does not detect glycemic variability. Continuous glucose monitoring will likely lead to more responsive adjustments in treatment regimens and to improved quality of care for patients with T2DM.

Insulin has many beneficial effects

Insulin exerts an anti-inflammatory effect by reducing the increase in C-reactive protein and serum amyloid A.³⁰ It also partially restores insulin-stimulated endothelial function,³¹ facilitates vasodilation by increasing nitric oxide production,³² and improves fibrinolytic profiles.³³ Early initiation of insulin therapy can increase peripheral insulin sensitivity and preserve beta cell function.^34-36

When oral agents have failed, insulin can significantly improve patients’ beta cell function,^34,35,37 and short periods of insulin therapy in patients newly diagnosed with T2DM may even set the foundation for better long-term control.^38,39

But not all insulin is alike
Ideally, insulin therapy should mimic physiologic insulin secretion. However, conventional human insulin products fail to do so because of their suboptimal pharmacodynamic profiles. With recombinant DNA technology, molecular modifications of the human insulin molecule have overcome some of the limitations of conventional human insulin products.

Unfortunately, many practitioners still hold insulin in reserve until combination therapy with oral agents has failed, possibly resulting in years of suboptimal glycemic control. Newer strategies recommend earlier initiation of insulin—ie, once diet and exercise fail, or when treatment with 1 oral agent fails. The development of insulin analogs is a significant milestone on the road to achieving improved outcomes for patients with T2DM.

Rapid-acting agents
Compared with regular human insulin, newer rapid-acting insulin analogs may improve glycemic control when used at mealtimes. However, due to their shorter half-lives, these insulin analogs require augmentation with basal insulin to control hyperglycemia between meals and during the night.

Insulin lispro was the first commercially available rapid-acting insulin analog, introduced in 1996. This agent differs from human insulin by an inversion of amino acid residues in positions 28 and 29 of the insulin B-chain. Inversion prevents the formation of hexamers and dimers that tend to diffuse more slowly, thereby facilitating a rapid uptake of the insulin analog into blood and tissues.^40,41 The second such agent, marketed in 2000, was insulin aspart, in which aspartic acid replaces proline at position 28 of the B-chain of human insulin.^41,42 The most recent rapid-acting analog is insulin glulisine, in which lysine replaces asparagine near the N-terminus of the B-chain, and glutamic acid replaces lysine near the C-terminus of human insulin.

The molecular changes made in creating these analogs allows them to dissociate quickly into monomers that are absorbed rapidly and achieve faster peak levels compared with regular human insulin.^41,42 These changes do not, however, interfere with the analogs’ ability to bind to the insulin receptor.^43,44

Dosing considerations. Absorption of regular human insulin is not sufficiently rapid at mealtimes to control prandial glucose levels.⁴⁵ Therefore, it is essential to give regular insulin 30 to 60 minutes before meals. For patients who have erratic daily schedules, adhering to this sort of routine can be difficult. But even if scheduling is not a problem, the prolonged duration of action of human insulin can predispose patients to hypoglycemia. Moreover, absorption of regular insulin can vary dramatically from day to day.^46,47

The insulin analogs correct the pharmacokinetic and pharmacodynamic deficiencies of regular insulin, producing plasma profiles that more closely simulate normal, physiologic meal-stimulated insulin release.^48-50 The 3 rapid-acting agents (aspart, glulisine, lispro) have very similar onset and duration of action, with peak effect occurring close to injection time (TABLE 2 and FIGURE 2).^48-50

Advantages of rapid-acting agents. These agents can be administered closer to meals, giving patients more flexibility and likely tighter postprandial glucose control, with reductions in glycemic excursions. Another advantage is the ability to better match insulin dose to anticipated carbohydrate in-take, affording better postprandial control.^51-53 Rapid-acting analogs also result in fewer episodes of hypoglycemia. In a meta-analysis of 2576 patients, hypoglycemic events occurred 25% less often with insulin lispro compared with regular human insulin in patients with type 1 diabetes mellitus (T1DM).⁵² In clinical trials, insulin aspart and insulin glulisine have also caused fewer hypoglycemic events compared with regular human insulin.^51-53

Long-acting agents
Basal, or long-acting, insulins are important for maintaining normoglycemia over 24 hours. Neutral protamine Hagedorn (NPH) insulin reaches its peak effect 4 to 10 hours after injection, and its total effect lasts only 12 to 18 hours. NPH is therefore often dosed twice daily. Absorption of NPH can vary significantly, causing day-to-day blood glucose fluctuations.^46,47 Therefore, this agent’s activity does not closely resemble normal physiologic basal insulin secretion.

The newer long-acting insulin analogs—insulin detemir and insulin glargine—were designed to more closely replicate normal physiologic basal insulin secretion. Insulin glargine was first to reach the market, in 2001. It contains glycine instead of asparagine in the alpha-chain and 2 arginine residues at the C-terminus, and the addition of zinc enhances the aggregation and slow release at a neutral pH. Insulin glargine precipitates in the subcutaneous tissue, which slows its absorption and results in a relatively flat insulin plasma profile and extended action.^54,55 Insulin detemir is a combination of the original insulin molecule and a saturated fatty acid (myristic acid). Insulin detemir is designed to bind albumin (98% albumin-bound in circulation) through this fatty acid chain in the plasma after injection, resulting in an extended plasma profile.^54,56 Insulin glargine and NPH form crystalline depots, but detemir is soluble and the subcutaneous depot remains in a liquid state; this may account for differences in absorption variability.⁵⁶

Advantages of the long-acting insulin analogs. Compared with conventional basal insulin such as NPH, the analogs have a prolonged duration of action (up to 24 hours) without pronounced peaks, permitting once-daily dosing in many patients (TABLE 2 and FIGURE 2).^46,47,50,55 The pharmacodynamic and pharmacokinetic properties of the long- acting agents make them less likely than NPH to cause nocturnal and overall hypoglycemia, a benefit that has been observed in several clinical trials.^57-61

Comparative clinical trials evaluating glycemic variability. Both of the long-acting analogs have shown lower within-subject variability in blood and plasma glucose measurements when compared with NPH.^62-64 In head-to-head comparisons of the analogs in glucose clamp studies, insulin detemir has demonstrated less within-subject variability of blood glucose levels than insulin glargine, in patients with T1DM or T2DM.^56,64 In clinical practice, different patients may have better results with one of these basal insulins as opposed to the other, and treatment choices will need to be tailored to the individual patient.

TABLE 2
Pharmacokinetic properties giving insulin analogs an advantage over regular insulin^46-50

FIGURE 2
Pharmacokinetic profiles of human insulin and insulin analogs

CORRESPONDENCE Eric L. Johnson, MD, University of North Dakota School of Medicine and Health Sciences, Department of Family & Community Medicine, 501 North Columbia Road, Grand Forks, ND 58202-9037; [email protected]

Test your skills: How would you treat these 3 patients?

CASE 1: Casey is a 14-year-old with a normal body mass index who has had heavy vaginal bleeding for 10 days. For the last 3 days, the bleeding has been so heavy she has been soaking more than 15 pads a day. She feels tired and is light-headed and dizzy when she stands up. Casey had her first period 13 months ago. Since then, her periods have varied in length from 18 to 40 days, with heavy bleeding for 7 to 14 days. She tells you she is not taking any prescription or herbal medications or over-the-counter supplements, and does not have any other medical problems. She is not sexually active. Her physical examination was remarkable only for pale skin and a positive tilt test. She feels frustrated and wants something done immediately.

CASE 2: Sarah is a 35-year-old obese woman whose chief complaint is irregular periods. For the past 3 years, she has had only 3 to 6 periods per year, each period lasting for 3 to 10 days. Her most recent period was 4 months ago. Sarah has a moderate amount of acne and facial hair.

CASE 3: Joan is a 53-year-old postmenopausal woman who has never been pregnant. She has a history of type 2 diabetes mellitus, hypertension, and obesity. She has come to your office for a routine physical examination. She tells you that her periods were regular until she stopped menstruating 4 years ago. But for the past 6 to 8 months she says she’s had irregular bleeding every 30 to 45 days, each period of bleeding lasting for 3 to 7 days. Her previous Pap smears and mammograms were normal. She has no family history of breast, gastrointestinal, or genital tract cancer. Her physical examination, including her pelvic examination, is negative.

These 3 women are fairly typical patients in a family medicine practice. Most women experience episodes of abnormal uterine bleeding (AUB) at some point in their reproductive lives. The condition occurs in approximately 1 in every 3 women of reproductive age and 1 in 10 postmenopausal women, and the impact on quality of life is often substantial.^1,2 Abnormal bleeding can be divided into 4 major categories: genital tract pathology, systemic disease, exposure to medication or radiation, and dysfunctional uterine bleeding (DUB). Specific conditions within each category are listed in TABLE 1. The focus of this article will be on DUB, the category that remains after the other possibilities are excluded.

First, find out what your patient means by “abnormal”
A normal menstrual cycle varies in length between 24 and 35 days, with menstrual flow lasting 2 to 7 days. Blood loss of 30 to 80 cc per cycle is considered normal.^3-5 To quantify blood loss, ask the patient how many pads or tampons she uses each period (<21 would be normal), how often she has to change pads (every 3 hours is usual), the size of clots (less than 1 cm is normal), and whether she has to get up at night to change pads.⁶ If blood loss is sufficient to cause anemia, the condition is always considered abnormal and requires further evaluation. When your patient’s description leaves you in doubt about whether her bleeding is abnormal, base your evaluation and treatment on her perception of a change in her menstrual cycle.

Stages of the reproductive life cycle
The meaning of abnormal bleeding varies with your patient’s stage in her reproductive life cycle. Uterine bleeding in a premenarchal child or a postmenopausal woman is always abnormal and must be evaluated.^7-9

Premenarchal children with vaginal bleeding should be evaluated for trauma, sexual or physical abuse, foreign bodies, signs of precocious puberty, and possible infectious etiologies.⁷ If the cause of the bleeding is not obvious, these patients should be immediately referred to a pediatric gynecologist.

Postmenopausal bleeding is defined as any bleeding that occurs more than 1 year after the last menstrual period.⁸ Cancer is the primary concern in these women and must always be excluded. (More on that, in a bit.)

History may reveal underlying pathology

The initial approach to evaluating abnormal bleeding is a thorough history and physical. Ask about stress, dietary habits, exercise, medications, radiation exposure, visual disturbances, headache, weight loss or gain, galactorrhea, palpitations, abdominal symptoms, jaundice, or excessive hair growth. Your patient’s answers to these questions may point to pathologies that underlie the abnormal bleeding, as listed in TABLE 1.

TABLE 1
Abnormal uterine bleeding: A typology^3,5,17,18,22-^24,30

In postmenopausal women, rule out cancer
The initial work-up for a postmenopausal patient should begin with a pelvic examination, followed by an assessment of her endometrial cavity by transvaginal ultrasound (TVUS), saline infusion sonohysterogram (SIS), or biopsy. An SIS, in particular, is often superior to TVUS in screening for anatomic anomalies.¹⁰ If a sonogram shows an endometrial thickness greater than 5 mm or the patient has risk factors for endometrial neoplasia, an endometrial biopsy for histologic diagnosis will be needed.¹¹ Risk factors for endometrial cancer include age older than 40, infertility, diabetes mellitus, hypertension, obesity, and estrogen medication. Repeat the sampling if the biopsy is inadequate. If the patient continues to have uterine bleeding, further evaluation with hysteroscopy by a gynecologist should be considered.^12-14

Evaluating bleeding in women of child-bearing age
Bleeding in this age group is most often related to pregnancy, so the diagnostic work-up should begin with a urine pregnancy test.^3,15 If pregnancy is ruled out, most etiologies in these women are benign, respond to conservative therapy, and can often be managed exclusively by family physicians.

After excluding pregnancy, look for genital tract pathology, including infection, polyps, uterine fibroids, and signs of cancer; iatrogenic causes such as medications or radiation; and systemic illnesses. In teenagers, look for inherited clotting disorders such as Von Willebrand’s disorder. If cervical cancer screening is not up to date, do a Pap smear.¹⁶ Cervical dysplasia generally does not cause heavy vaginal bleeding, but can cause postcoital bleeding.¹⁷ A complete blood count and a thyroid-stimulating hormone (TSH) level will allow you to rule out anemia, leukemia, thrombocytopenia, and thyroid disorders.¹¹

Dysfunctional uterine bleeding (DUB): A diagnosis of exclusion

Once you have ruled out genital tract pathology, systemic disease, and iatrogenic causes, you are left with a diagnosis of dysfunctional uterine bleeding. DUB occurs most commonly at the onset of regular menstrual cycles or when menstruation is coming to an end during menopause.

The menstrual cycle in a woman with DUB may be ovulatory or anovulatory. Women who have ovulatory cycles usually know the characteristics of their menses and are often aware of minor variations in the timing or flow. A patient with an anatomic problem who has ovulatory cycles will usually present with complaints of menorrhagia.

Anovulatory cycles are more typical, occurring in 90% of patients with DUB.¹⁸ In anovulatory cycles, the corpus luteum is not produced and the ovaries do not secrete progesterone. In the absence of progesterone, constant estrogen stimulation produces a proliferative endometrium that is not sustainable. As 1 area of bleeding heals, another site begins to slough, and the result is an irregular and prolonged bleeding pattern that is unpredictable. The clinical result in this scenario is varying cycle lengths and differing amounts of menstrual blood loss.

Treatment depends on the etiology

Cervical and endometrial cancer should be ruled out early, because early diagnosis and treatment may improve survival. If the source of abnormal bleeding is an anatomic abnormality such as an endometrial polyp, removing the polyp under hysteroscopic guidance should alleviate the problem. If the bleeding is due to medication exposure or a systemic disease such as hypothyroidism, withdrawing the off ending agent or treating the systemic disorder will generally alleviate the problem.

For patients with DUB, hormonal (medroxyprogesterone acetate, oral contraceptive pills [OCPs], levonorgestrel intrauterine device [IUD]) and nonhormonal treatment (nonsteroidal anti-inflammatory drugs, tranexamic acid, danazol) decisions are based on the age of the patient, severity of bleeding and symptoms, and the patient’s hematocrit.¹⁹ Patients with persistent bleeding despite medical treatment require a complete reevaluation and referral to a gynecologist if an explanation is not found or if surgical treatment is required.

CASE 1: Casey

Where does Casey fit in this typology? She is in the perimenarchal stage of her reproductive life cycle, when anovulatory bleeding is common. For the first 18 to 24 months after menarche, the immature hypothalamic-pituitary-ovarian axis may fail to respond to estrogen and progesterone stimulation, resulting in anovulation and irregular, often heavy bleeding. Her urine test rules out pregnancy. Blood tests confirm the anemia her pallor and fatigue suggest. Her initial, empiric treatment would be iron supplementation for anemia and cyclic medroxyprogesterone acetate or OCPs (TABLE 2) to regulate her periods. If these conservative measures are not sufficient, further evaluation would be indicated.

Blood dyscrasias (5%-20% incidence in teenagers) and systemic disorders, including Von Willebrand’s disease, idiopathic thrombocytopenic purpura, and leukemia, are the major diseases to consider.^20-23 An endometrial biopsy is not indicated, because the incidence of endometrial cancer in Casey’s age group is less than 1 in 100,000. ²⁴

CASE 2: Sarah

How can you explain Sarah’s irregular periods? A negative urine test rules out pregnancy, and her responses to questions about diet, exercise, and stress rule out hypothalamic suppression. She doesn’t complain of headaches, visual field changes, or galactorrhea, which would exclude a pituitary microadenoma. She does not exhibit symptoms of a thyroid disorder and her TSH is normal. Complaints of frequent urination, thirst, or weight loss could be indications of diabetes mellitus, but Sarah does not present with these symptoms. Her facial hair and acne suggest androgen excess originating from the adrenal glands or ovaries.

Sarah’s history of infrequent and heavy menses, as well as an absence of breast tenderness, bloating, or mittelschmerz, indicate she is not ovulating. The most likely explanation for her failure to ovulate is polycystic ovarian syndrome (PCOS), and you can initiate treatment immediately. The major treatment options for this disorder are observation, medroxyprogesterone acetate, and OCPs (TABLE 2).

If Sarah does not respond to hormonal therapy, a thorough reevaluation is indicated, including additional laboratory tests and a pelvic sonogram to evaluate the uterus and ovaries. Other tests to consider include prolactin, fasting blood sugar, early morning 17-hydroxy-progesterone, dehydroepiandrosterone (DHEA) and dehydroepiandrosterone sulfate (DHEA-S), testosterone, and cortisol. For an extensive review of PCOS and its relationship with endocrine, metabolic, and reproductive disorders, as well as cardiovascular disease and obstructive sleep apnea, see the excellent review by Ehrmann.²⁵ If hormonal therapy is unsuccessful, a hysteroscopy with endometrial ablation could then be offered. In refractory cases, a hysterectomy can be performed.

Although Sarah is only 35, her prolonged exposure to unopposed estrogen (>3 years, according to her history) warrants an endometrial biopsy. The presence of other endometrial cancer risk factors (obesity, chronic anovulation, nulliparity) supports this decision. The incidence of endometrial cancer is 2.3 cases per 100,000 patients in 30- to 34-year-old women, 6.1 cases per 100,000 patients in 35- to 39-year-old women, increasing to 36.5 cases per 100,000 in women ages 40 to 49 years. ²⁴

If Sarah is troubled by her infertility, consider referring her to a specialist. Treatment options for her infertility would include weight loss, insulin-lowering medications, and clomiphene citrate to induce ovulation.

CASE 3: Joan

Uterine bleeding in a postmenopausal patient like Joan is always abnormal. In 5% to 10% of cases, such bleeding indicates endometrial cancer. ^26,27 An endometrial biopsy to rule out cancer is the first order of business. If the biopsy is nondiagnostic or reveals endometrial polyps or submucosal fibroids, the next step would be a diagnostic hysteroscopy. Alternatively, Joan’s endometrium could first be evaluated with a TVUS. If the sonogram showed an endometrium 5 mm in thickness or more, an endometrial biopsy could be performed then.^26-29

If these tests rule out a cancer diagnosis, your next step would be to try low-dose cyclic OCPs or medroxyprogesterone acetate (TABLE 2) to control the bleeding. If hormonal therapy is not effective or Joan doesn’t want to try it, an endometrial ablation in conjunction with a hysteroscopy performed by a gynecologist is another option. But if Joan’s bleeding is light, it may be due simply to her postmenopausal hypoestrogenic state, and can be left untreated as long as Joan is comfortable with this option.

TABLE 2
Medical treatment for dysfunctional uterine bleeding

Lessons learned

Patients like Casey, Sarah, and Joan can be successfully managed by the family physician. A thorough history, physical examination, and basic laboratory tests will usually suffice to rule out anatomic, systemic, or iatrogenic explanations. Pregnancy, the most common explanation for abnormal uterine bleeding, can be ruled out with a urine pregnancy test. Patients like Sarah and Joan, who have some of the risk factors for endometrial cancer, require an evaluation of the endometrium to rule out that possibility. When none of these etiologies is the culprit, your working diagnosis is DUB, and medical treatment for it is well within your competence.

CORRESPONDENCE David L. Maness, DO, MSS, Department of Family Medicine, University of Tennessee Health Science Center, College of Medicine, 1301 Primacy Parkway, Memphis, TN 38119; [email protected]

Test your skills: How would you treat these 3 patients?

CASE 1: Casey is a 14-year-old with a normal body mass index who has had heavy vaginal bleeding for 10 days. For the last 3 days, the bleeding has been so heavy she has been soaking more than 15 pads a day. She feels tired and is light-headed and dizzy when she stands up. Casey had her first period 13 months ago. Since then, her periods have varied in length from 18 to 40 days, with heavy bleeding for 7 to 14 days. She tells you she is not taking any prescription or herbal medications or over-the-counter supplements, and does not have any other medical problems. She is not sexually active. Her physical examination was remarkable only for pale skin and a positive tilt test. She feels frustrated and wants something done immediately.

CASE 2: Sarah is a 35-year-old obese woman whose chief complaint is irregular periods. For the past 3 years, she has had only 3 to 6 periods per year, each period lasting for 3 to 10 days. Her most recent period was 4 months ago. Sarah has a moderate amount of acne and facial hair.

CASE 3: Joan is a 53-year-old postmenopausal woman who has never been pregnant. She has a history of type 2 diabetes mellitus, hypertension, and obesity. She has come to your office for a routine physical examination. She tells you that her periods were regular until she stopped menstruating 4 years ago. But for the past 6 to 8 months she says she’s had irregular bleeding every 30 to 45 days, each period of bleeding lasting for 3 to 7 days. Her previous Pap smears and mammograms were normal. She has no family history of breast, gastrointestinal, or genital tract cancer. Her physical examination, including her pelvic examination, is negative.

These 3 women are fairly typical patients in a family medicine practice. Most women experience episodes of abnormal uterine bleeding (AUB) at some point in their reproductive lives. The condition occurs in approximately 1 in every 3 women of reproductive age and 1 in 10 postmenopausal women, and the impact on quality of life is often substantial.^1,2 Abnormal bleeding can be divided into 4 major categories: genital tract pathology, systemic disease, exposure to medication or radiation, and dysfunctional uterine bleeding (DUB). Specific conditions within each category are listed in TABLE 1. The focus of this article will be on DUB, the category that remains after the other possibilities are excluded.

First, find out what your patient means by “abnormal”
A normal menstrual cycle varies in length between 24 and 35 days, with menstrual flow lasting 2 to 7 days. Blood loss of 30 to 80 cc per cycle is considered normal.^3-5 To quantify blood loss, ask the patient how many pads or tampons she uses each period (<21 would be normal), how often she has to change pads (every 3 hours is usual), the size of clots (less than 1 cm is normal), and whether she has to get up at night to change pads.⁶ If blood loss is sufficient to cause anemia, the condition is always considered abnormal and requires further evaluation. When your patient’s description leaves you in doubt about whether her bleeding is abnormal, base your evaluation and treatment on her perception of a change in her menstrual cycle.

Stages of the reproductive life cycle
The meaning of abnormal bleeding varies with your patient’s stage in her reproductive life cycle. Uterine bleeding in a premenarchal child or a postmenopausal woman is always abnormal and must be evaluated.^7-9

Premenarchal children with vaginal bleeding should be evaluated for trauma, sexual or physical abuse, foreign bodies, signs of precocious puberty, and possible infectious etiologies.⁷ If the cause of the bleeding is not obvious, these patients should be immediately referred to a pediatric gynecologist.

Postmenopausal bleeding is defined as any bleeding that occurs more than 1 year after the last menstrual period.⁸ Cancer is the primary concern in these women and must always be excluded. (More on that, in a bit.)

History may reveal underlying pathology

The initial approach to evaluating abnormal bleeding is a thorough history and physical. Ask about stress, dietary habits, exercise, medications, radiation exposure, visual disturbances, headache, weight loss or gain, galactorrhea, palpitations, abdominal symptoms, jaundice, or excessive hair growth. Your patient’s answers to these questions may point to pathologies that underlie the abnormal bleeding, as listed in TABLE 1.

TABLE 1
Abnormal uterine bleeding: A typology^3,5,17,18,22-^24,30

In postmenopausal women, rule out cancer
The initial work-up for a postmenopausal patient should begin with a pelvic examination, followed by an assessment of her endometrial cavity by transvaginal ultrasound (TVUS), saline infusion sonohysterogram (SIS), or biopsy. An SIS, in particular, is often superior to TVUS in screening for anatomic anomalies.¹⁰ If a sonogram shows an endometrial thickness greater than 5 mm or the patient has risk factors for endometrial neoplasia, an endometrial biopsy for histologic diagnosis will be needed.¹¹ Risk factors for endometrial cancer include age older than 40, infertility, diabetes mellitus, hypertension, obesity, and estrogen medication. Repeat the sampling if the biopsy is inadequate. If the patient continues to have uterine bleeding, further evaluation with hysteroscopy by a gynecologist should be considered.^12-14

Evaluating bleeding in women of child-bearing age
Bleeding in this age group is most often related to pregnancy, so the diagnostic work-up should begin with a urine pregnancy test.^3,15 If pregnancy is ruled out, most etiologies in these women are benign, respond to conservative therapy, and can often be managed exclusively by family physicians.

After excluding pregnancy, look for genital tract pathology, including infection, polyps, uterine fibroids, and signs of cancer; iatrogenic causes such as medications or radiation; and systemic illnesses. In teenagers, look for inherited clotting disorders such as Von Willebrand’s disorder. If cervical cancer screening is not up to date, do a Pap smear.¹⁶ Cervical dysplasia generally does not cause heavy vaginal bleeding, but can cause postcoital bleeding.¹⁷ A complete blood count and a thyroid-stimulating hormone (TSH) level will allow you to rule out anemia, leukemia, thrombocytopenia, and thyroid disorders.¹¹

Dysfunctional uterine bleeding (DUB): A diagnosis of exclusion

Once you have ruled out genital tract pathology, systemic disease, and iatrogenic causes, you are left with a diagnosis of dysfunctional uterine bleeding. DUB occurs most commonly at the onset of regular menstrual cycles or when menstruation is coming to an end during menopause.

The menstrual cycle in a woman with DUB may be ovulatory or anovulatory. Women who have ovulatory cycles usually know the characteristics of their menses and are often aware of minor variations in the timing or flow. A patient with an anatomic problem who has ovulatory cycles will usually present with complaints of menorrhagia.

Anovulatory cycles are more typical, occurring in 90% of patients with DUB.¹⁸ In anovulatory cycles, the corpus luteum is not produced and the ovaries do not secrete progesterone. In the absence of progesterone, constant estrogen stimulation produces a proliferative endometrium that is not sustainable. As 1 area of bleeding heals, another site begins to slough, and the result is an irregular and prolonged bleeding pattern that is unpredictable. The clinical result in this scenario is varying cycle lengths and differing amounts of menstrual blood loss.

Treatment depends on the etiology

Cervical and endometrial cancer should be ruled out early, because early diagnosis and treatment may improve survival. If the source of abnormal bleeding is an anatomic abnormality such as an endometrial polyp, removing the polyp under hysteroscopic guidance should alleviate the problem. If the bleeding is due to medication exposure or a systemic disease such as hypothyroidism, withdrawing the off ending agent or treating the systemic disorder will generally alleviate the problem.

For patients with DUB, hormonal (medroxyprogesterone acetate, oral contraceptive pills [OCPs], levonorgestrel intrauterine device [IUD]) and nonhormonal treatment (nonsteroidal anti-inflammatory drugs, tranexamic acid, danazol) decisions are based on the age of the patient, severity of bleeding and symptoms, and the patient’s hematocrit.¹⁹ Patients with persistent bleeding despite medical treatment require a complete reevaluation and referral to a gynecologist if an explanation is not found or if surgical treatment is required.

CASE 1: Casey

Where does Casey fit in this typology? She is in the perimenarchal stage of her reproductive life cycle, when anovulatory bleeding is common. For the first 18 to 24 months after menarche, the immature hypothalamic-pituitary-ovarian axis may fail to respond to estrogen and progesterone stimulation, resulting in anovulation and irregular, often heavy bleeding. Her urine test rules out pregnancy. Blood tests confirm the anemia her pallor and fatigue suggest. Her initial, empiric treatment would be iron supplementation for anemia and cyclic medroxyprogesterone acetate or OCPs (TABLE 2) to regulate her periods. If these conservative measures are not sufficient, further evaluation would be indicated.

Blood dyscrasias (5%-20% incidence in teenagers) and systemic disorders, including Von Willebrand’s disease, idiopathic thrombocytopenic purpura, and leukemia, are the major diseases to consider.^20-23 An endometrial biopsy is not indicated, because the incidence of endometrial cancer in Casey’s age group is less than 1 in 100,000. ²⁴

CASE 2: Sarah

How can you explain Sarah’s irregular periods? A negative urine test rules out pregnancy, and her responses to questions about diet, exercise, and stress rule out hypothalamic suppression. She doesn’t complain of headaches, visual field changes, or galactorrhea, which would exclude a pituitary microadenoma. She does not exhibit symptoms of a thyroid disorder and her TSH is normal. Complaints of frequent urination, thirst, or weight loss could be indications of diabetes mellitus, but Sarah does not present with these symptoms. Her facial hair and acne suggest androgen excess originating from the adrenal glands or ovaries.

Sarah’s history of infrequent and heavy menses, as well as an absence of breast tenderness, bloating, or mittelschmerz, indicate she is not ovulating. The most likely explanation for her failure to ovulate is polycystic ovarian syndrome (PCOS), and you can initiate treatment immediately. The major treatment options for this disorder are observation, medroxyprogesterone acetate, and OCPs (TABLE 2).

If Sarah does not respond to hormonal therapy, a thorough reevaluation is indicated, including additional laboratory tests and a pelvic sonogram to evaluate the uterus and ovaries. Other tests to consider include prolactin, fasting blood sugar, early morning 17-hydroxy-progesterone, dehydroepiandrosterone (DHEA) and dehydroepiandrosterone sulfate (DHEA-S), testosterone, and cortisol. For an extensive review of PCOS and its relationship with endocrine, metabolic, and reproductive disorders, as well as cardiovascular disease and obstructive sleep apnea, see the excellent review by Ehrmann.²⁵ If hormonal therapy is unsuccessful, a hysteroscopy with endometrial ablation could then be offered. In refractory cases, a hysterectomy can be performed.

Although Sarah is only 35, her prolonged exposure to unopposed estrogen (>3 years, according to her history) warrants an endometrial biopsy. The presence of other endometrial cancer risk factors (obesity, chronic anovulation, nulliparity) supports this decision. The incidence of endometrial cancer is 2.3 cases per 100,000 patients in 30- to 34-year-old women, 6.1 cases per 100,000 patients in 35- to 39-year-old women, increasing to 36.5 cases per 100,000 in women ages 40 to 49 years. ²⁴

If Sarah is troubled by her infertility, consider referring her to a specialist. Treatment options for her infertility would include weight loss, insulin-lowering medications, and clomiphene citrate to induce ovulation.

CASE 3: Joan

Uterine bleeding in a postmenopausal patient like Joan is always abnormal. In 5% to 10% of cases, such bleeding indicates endometrial cancer. ^26,27 An endometrial biopsy to rule out cancer is the first order of business. If the biopsy is nondiagnostic or reveals endometrial polyps or submucosal fibroids, the next step would be a diagnostic hysteroscopy. Alternatively, Joan’s endometrium could first be evaluated with a TVUS. If the sonogram showed an endometrium 5 mm in thickness or more, an endometrial biopsy could be performed then.^26-29

If these tests rule out a cancer diagnosis, your next step would be to try low-dose cyclic OCPs or medroxyprogesterone acetate (TABLE 2) to control the bleeding. If hormonal therapy is not effective or Joan doesn’t want to try it, an endometrial ablation in conjunction with a hysteroscopy performed by a gynecologist is another option. But if Joan’s bleeding is light, it may be due simply to her postmenopausal hypoestrogenic state, and can be left untreated as long as Joan is comfortable with this option.

TABLE 2
Medical treatment for dysfunctional uterine bleeding

Lessons learned

Patients like Casey, Sarah, and Joan can be successfully managed by the family physician. A thorough history, physical examination, and basic laboratory tests will usually suffice to rule out anatomic, systemic, or iatrogenic explanations. Pregnancy, the most common explanation for abnormal uterine bleeding, can be ruled out with a urine pregnancy test. Patients like Sarah and Joan, who have some of the risk factors for endometrial cancer, require an evaluation of the endometrium to rule out that possibility. When none of these etiologies is the culprit, your working diagnosis is DUB, and medical treatment for it is well within your competence.

CORRESPONDENCE David L. Maness, DO, MSS, Department of Family Medicine, University of Tennessee Health Science Center, College of Medicine, 1301 Primacy Parkway, Memphis, TN 38119; [email protected]

Test your skills: How would you treat these 3 patients?

CASE 1: Casey is a 14-year-old with a normal body mass index who has had heavy vaginal bleeding for 10 days. For the last 3 days, the bleeding has been so heavy she has been soaking more than 15 pads a day. She feels tired and is light-headed and dizzy when she stands up. Casey had her first period 13 months ago. Since then, her periods have varied in length from 18 to 40 days, with heavy bleeding for 7 to 14 days. She tells you she is not taking any prescription or herbal medications or over-the-counter supplements, and does not have any other medical problems. She is not sexually active. Her physical examination was remarkable only for pale skin and a positive tilt test. She feels frustrated and wants something done immediately.

CASE 2: Sarah is a 35-year-old obese woman whose chief complaint is irregular periods. For the past 3 years, she has had only 3 to 6 periods per year, each period lasting for 3 to 10 days. Her most recent period was 4 months ago. Sarah has a moderate amount of acne and facial hair.

CASE 3: Joan is a 53-year-old postmenopausal woman who has never been pregnant. She has a history of type 2 diabetes mellitus, hypertension, and obesity. She has come to your office for a routine physical examination. She tells you that her periods were regular until she stopped menstruating 4 years ago. But for the past 6 to 8 months she says she’s had irregular bleeding every 30 to 45 days, each period of bleeding lasting for 3 to 7 days. Her previous Pap smears and mammograms were normal. She has no family history of breast, gastrointestinal, or genital tract cancer. Her physical examination, including her pelvic examination, is negative.

These 3 women are fairly typical patients in a family medicine practice. Most women experience episodes of abnormal uterine bleeding (AUB) at some point in their reproductive lives. The condition occurs in approximately 1 in every 3 women of reproductive age and 1 in 10 postmenopausal women, and the impact on quality of life is often substantial.^1,2 Abnormal bleeding can be divided into 4 major categories: genital tract pathology, systemic disease, exposure to medication or radiation, and dysfunctional uterine bleeding (DUB). Specific conditions within each category are listed in TABLE 1. The focus of this article will be on DUB, the category that remains after the other possibilities are excluded.

First, find out what your patient means by “abnormal”
A normal menstrual cycle varies in length between 24 and 35 days, with menstrual flow lasting 2 to 7 days. Blood loss of 30 to 80 cc per cycle is considered normal.^3-5 To quantify blood loss, ask the patient how many pads or tampons she uses each period (<21 would be normal), how often she has to change pads (every 3 hours is usual), the size of clots (less than 1 cm is normal), and whether she has to get up at night to change pads.⁶ If blood loss is sufficient to cause anemia, the condition is always considered abnormal and requires further evaluation. When your patient’s description leaves you in doubt about whether her bleeding is abnormal, base your evaluation and treatment on her perception of a change in her menstrual cycle.

Stages of the reproductive life cycle
The meaning of abnormal bleeding varies with your patient’s stage in her reproductive life cycle. Uterine bleeding in a premenarchal child or a postmenopausal woman is always abnormal and must be evaluated.^7-9

Premenarchal children with vaginal bleeding should be evaluated for trauma, sexual or physical abuse, foreign bodies, signs of precocious puberty, and possible infectious etiologies.⁷ If the cause of the bleeding is not obvious, these patients should be immediately referred to a pediatric gynecologist.

Postmenopausal bleeding is defined as any bleeding that occurs more than 1 year after the last menstrual period.⁸ Cancer is the primary concern in these women and must always be excluded. (More on that, in a bit.)

History may reveal underlying pathology

The initial approach to evaluating abnormal bleeding is a thorough history and physical. Ask about stress, dietary habits, exercise, medications, radiation exposure, visual disturbances, headache, weight loss or gain, galactorrhea, palpitations, abdominal symptoms, jaundice, or excessive hair growth. Your patient’s answers to these questions may point to pathologies that underlie the abnormal bleeding, as listed in TABLE 1.

TABLE 1
Abnormal uterine bleeding: A typology^3,5,17,18,22-^24,30

In postmenopausal women, rule out cancer
The initial work-up for a postmenopausal patient should begin with a pelvic examination, followed by an assessment of her endometrial cavity by transvaginal ultrasound (TVUS), saline infusion sonohysterogram (SIS), or biopsy. An SIS, in particular, is often superior to TVUS in screening for anatomic anomalies.¹⁰ If a sonogram shows an endometrial thickness greater than 5 mm or the patient has risk factors for endometrial neoplasia, an endometrial biopsy for histologic diagnosis will be needed.¹¹ Risk factors for endometrial cancer include age older than 40, infertility, diabetes mellitus, hypertension, obesity, and estrogen medication. Repeat the sampling if the biopsy is inadequate. If the patient continues to have uterine bleeding, further evaluation with hysteroscopy by a gynecologist should be considered.^12-14

Evaluating bleeding in women of child-bearing age
Bleeding in this age group is most often related to pregnancy, so the diagnostic work-up should begin with a urine pregnancy test.^3,15 If pregnancy is ruled out, most etiologies in these women are benign, respond to conservative therapy, and can often be managed exclusively by family physicians.

After excluding pregnancy, look for genital tract pathology, including infection, polyps, uterine fibroids, and signs of cancer; iatrogenic causes such as medications or radiation; and systemic illnesses. In teenagers, look for inherited clotting disorders such as Von Willebrand’s disorder. If cervical cancer screening is not up to date, do a Pap smear.¹⁶ Cervical dysplasia generally does not cause heavy vaginal bleeding, but can cause postcoital bleeding.¹⁷ A complete blood count and a thyroid-stimulating hormone (TSH) level will allow you to rule out anemia, leukemia, thrombocytopenia, and thyroid disorders.¹¹

Dysfunctional uterine bleeding (DUB): A diagnosis of exclusion

Once you have ruled out genital tract pathology, systemic disease, and iatrogenic causes, you are left with a diagnosis of dysfunctional uterine bleeding. DUB occurs most commonly at the onset of regular menstrual cycles or when menstruation is coming to an end during menopause.

The menstrual cycle in a woman with DUB may be ovulatory or anovulatory. Women who have ovulatory cycles usually know the characteristics of their menses and are often aware of minor variations in the timing or flow. A patient with an anatomic problem who has ovulatory cycles will usually present with complaints of menorrhagia.

Anovulatory cycles are more typical, occurring in 90% of patients with DUB.¹⁸ In anovulatory cycles, the corpus luteum is not produced and the ovaries do not secrete progesterone. In the absence of progesterone, constant estrogen stimulation produces a proliferative endometrium that is not sustainable. As 1 area of bleeding heals, another site begins to slough, and the result is an irregular and prolonged bleeding pattern that is unpredictable. The clinical result in this scenario is varying cycle lengths and differing amounts of menstrual blood loss.

Treatment depends on the etiology

Cervical and endometrial cancer should be ruled out early, because early diagnosis and treatment may improve survival. If the source of abnormal bleeding is an anatomic abnormality such as an endometrial polyp, removing the polyp under hysteroscopic guidance should alleviate the problem. If the bleeding is due to medication exposure or a systemic disease such as hypothyroidism, withdrawing the off ending agent or treating the systemic disorder will generally alleviate the problem.

For patients with DUB, hormonal (medroxyprogesterone acetate, oral contraceptive pills [OCPs], levonorgestrel intrauterine device [IUD]) and nonhormonal treatment (nonsteroidal anti-inflammatory drugs, tranexamic acid, danazol) decisions are based on the age of the patient, severity of bleeding and symptoms, and the patient’s hematocrit.¹⁹ Patients with persistent bleeding despite medical treatment require a complete reevaluation and referral to a gynecologist if an explanation is not found or if surgical treatment is required.

CASE 1: Casey

Where does Casey fit in this typology? She is in the perimenarchal stage of her reproductive life cycle, when anovulatory bleeding is common. For the first 18 to 24 months after menarche, the immature hypothalamic-pituitary-ovarian axis may fail to respond to estrogen and progesterone stimulation, resulting in anovulation and irregular, often heavy bleeding. Her urine test rules out pregnancy. Blood tests confirm the anemia her pallor and fatigue suggest. Her initial, empiric treatment would be iron supplementation for anemia and cyclic medroxyprogesterone acetate or OCPs (TABLE 2) to regulate her periods. If these conservative measures are not sufficient, further evaluation would be indicated.

Blood dyscrasias (5%-20% incidence in teenagers) and systemic disorders, including Von Willebrand’s disease, idiopathic thrombocytopenic purpura, and leukemia, are the major diseases to consider.^20-23 An endometrial biopsy is not indicated, because the incidence of endometrial cancer in Casey’s age group is less than 1 in 100,000. ²⁴

CASE 2: Sarah

How can you explain Sarah’s irregular periods? A negative urine test rules out pregnancy, and her responses to questions about diet, exercise, and stress rule out hypothalamic suppression. She doesn’t complain of headaches, visual field changes, or galactorrhea, which would exclude a pituitary microadenoma. She does not exhibit symptoms of a thyroid disorder and her TSH is normal. Complaints of frequent urination, thirst, or weight loss could be indications of diabetes mellitus, but Sarah does not present with these symptoms. Her facial hair and acne suggest androgen excess originating from the adrenal glands or ovaries.

Sarah’s history of infrequent and heavy menses, as well as an absence of breast tenderness, bloating, or mittelschmerz, indicate she is not ovulating. The most likely explanation for her failure to ovulate is polycystic ovarian syndrome (PCOS), and you can initiate treatment immediately. The major treatment options for this disorder are observation, medroxyprogesterone acetate, and OCPs (TABLE 2).

If Sarah does not respond to hormonal therapy, a thorough reevaluation is indicated, including additional laboratory tests and a pelvic sonogram to evaluate the uterus and ovaries. Other tests to consider include prolactin, fasting blood sugar, early morning 17-hydroxy-progesterone, dehydroepiandrosterone (DHEA) and dehydroepiandrosterone sulfate (DHEA-S), testosterone, and cortisol. For an extensive review of PCOS and its relationship with endocrine, metabolic, and reproductive disorders, as well as cardiovascular disease and obstructive sleep apnea, see the excellent review by Ehrmann.²⁵ If hormonal therapy is unsuccessful, a hysteroscopy with endometrial ablation could then be offered. In refractory cases, a hysterectomy can be performed.

Although Sarah is only 35, her prolonged exposure to unopposed estrogen (>3 years, according to her history) warrants an endometrial biopsy. The presence of other endometrial cancer risk factors (obesity, chronic anovulation, nulliparity) supports this decision. The incidence of endometrial cancer is 2.3 cases per 100,000 patients in 30- to 34-year-old women, 6.1 cases per 100,000 patients in 35- to 39-year-old women, increasing to 36.5 cases per 100,000 in women ages 40 to 49 years. ²⁴

If Sarah is troubled by her infertility, consider referring her to a specialist. Treatment options for her infertility would include weight loss, insulin-lowering medications, and clomiphene citrate to induce ovulation.

CASE 3: Joan

Uterine bleeding in a postmenopausal patient like Joan is always abnormal. In 5% to 10% of cases, such bleeding indicates endometrial cancer. ^26,27 An endometrial biopsy to rule out cancer is the first order of business. If the biopsy is nondiagnostic or reveals endometrial polyps or submucosal fibroids, the next step would be a diagnostic hysteroscopy. Alternatively, Joan’s endometrium could first be evaluated with a TVUS. If the sonogram showed an endometrium 5 mm in thickness or more, an endometrial biopsy could be performed then.^26-29

If these tests rule out a cancer diagnosis, your next step would be to try low-dose cyclic OCPs or medroxyprogesterone acetate (TABLE 2) to control the bleeding. If hormonal therapy is not effective or Joan doesn’t want to try it, an endometrial ablation in conjunction with a hysteroscopy performed by a gynecologist is another option. But if Joan’s bleeding is light, it may be due simply to her postmenopausal hypoestrogenic state, and can be left untreated as long as Joan is comfortable with this option.

TABLE 2
Medical treatment for dysfunctional uterine bleeding

Lessons learned

Patients like Casey, Sarah, and Joan can be successfully managed by the family physician. A thorough history, physical examination, and basic laboratory tests will usually suffice to rule out anatomic, systemic, or iatrogenic explanations. Pregnancy, the most common explanation for abnormal uterine bleeding, can be ruled out with a urine pregnancy test. Patients like Sarah and Joan, who have some of the risk factors for endometrial cancer, require an evaluation of the endometrium to rule out that possibility. When none of these etiologies is the culprit, your working diagnosis is DUB, and medical treatment for it is well within your competence.

CORRESPONDENCE David L. Maness, DO, MSS, Department of Family Medicine, University of Tennessee Health Science Center, College of Medicine, 1301 Primacy Parkway, Memphis, TN 38119; [email protected]

CASE A 50-year-old construction worker comes in for an office visit because he’s been experiencing intermittent low back pain that’s been occurring more frequently. He says he has not been injured and that he always takes care on the job to minimize physical risk. He reports no symptoms other than the back pain.

How would you proceed with this patient’s care? Would you order a plain radiograph to be sure nothing dire is causing the patient’s pain—or would you skip it? And would you know how to match your patient’s history and exam findings with specific physical therapy interventions?

The following review brings the latest guidelines and research to bear on these questions—and others—as you care for patients with nonspecific low back pain.

Categorizing low back pain to direct your investigation

The 2007 Joint Clinical Practice Guideline issued by the American College of Physicians and the American Pain Society encourages clinicians to perform a focused history and physical exam to classify patients into 1 of 3 broad categories: nonspecific low back pain (LBP), LBP potentially associated with radiculopathy or spinal stenosis, or LBP potentially associated with another specific spinal cause.¹

Patients in the last category often exhibit findings in the history and physical examination suggestive of severe or progressive neurologic defi cits. Refer these patients for further diagnostic testing.

Patients presenting with persistent (>4 weeks) LBP and signs and symptoms of radiculopathy or spinal stenosis are best referred for, preferably, magnetic resonance imaging (MRI) or for computed tomography (CT)—but only if the patient is a candidate for surgery or epidural steroid injection.

For patients with nonspecific LBP, which accounts for most cases, practice guidelines recommend against routine use of imaging or other diagnostic procedures.¹

Unfortunately, however, some clinicians still use routine imaging in the absence of significant findings or without clear indication.² One argument used to justify this action—particularly by some who consider nonspecific LBP to be a diagnosis of exclusion —is the desire to rule out a serious underlying spinal condition.

What the research tells us about routine imaging

A recently published systematic review and meta-analysis compiled data relevant to more than 1800 subjects from 6 randomized controlled trials (RCTs). The authors examined early, routine use of lumbar imaging (radiography, MRI, or CT) in patients with acute or subacute LBP. They found that, without clear indication from findings in the history and physical examination, immediate imaging does not improve clinical outcomes (ie, diminish pain or improve daily function).³

Another study took a closer look at advanced imaging for LBP. In an RCT including more than 300 patients with a mean age of 53 years, investigators compared outcomes for patients receiving either plain radiographs or rapid MRI. No differences were noted in outcomes for back-related disability, pain, health survey results, preference scores, satisfaction, or costs at 12 months.

Furthermore, patients receiving rapid MRI were more likely to undergo surgery, which also failed to improve outcomes. As a result, the authors cautioned against unnecessary use of advanced imaging, as it could increase costs of care and possibly increase surgical intervention without improved outcomes.⁴ These studies substantiate practice guidelines regarding the use of imaging for patients with nonspecific LBP.

Not helpful, and perhaps harmful?
When imaging is unwarranted, it unnecessarily exposes patients to ionizing radiation, especially objectionable for younger women.¹ Imaging can also lead to the identifi cation of pathology unrelated to a patient’s LBP.^1,5

As mentioned above, patients receiving rapid MRI were more likely to receive surgical intervention that did not improve outcomes.⁴ This observation may reflect, in part, findings of pathoanatomical abnormalities that have little or no correlation with patient symptoms. In a random sample of 148 subjects ages 36 to 71 years—nearly half of whom had never experienced back pain—Jarvik and colleagues⁵ found MRI evidence of annular tears, disc bulges, disc protrusions, facet joint degeneration, end plate changes, and mild spondylolisthesis. The authors concluded that such MRI findings are therefore of limited diagnostic value.⁵

Labeling can be harmful, too. Identification of pathology that could well be unrelated to LBP can result in a specific, presumed diagnosis, possibly inducing a phenomenon known as the labeling effect. The search for a specific diagnosis or label for patients with nonspecific LBP could cause them to perceive their low back condition as being more serious than it actually is. Patients may then develop distorted beliefs regarding the true nature of their health status. The labeling effect could even alter the natural course of an otherwise benign condition.^6,7

It’s time to treat: Tell patients to remain active

Practice guidelines for nonspecific LBP recommend providing patients with evidence-based education that emphasizes the favorable course of this condition and that encourages them to remain active.¹ This recommendation was assessed retrospectively in a study of nearly 1200 patients receiving physical therapy for acute LBP. The authors found that adherence to the recommendation for activity and exercise yielded significant reductions in disability and pain, and resulted in significantly fewer visits and lower charges.⁸

For acute cases of nonspecific LBP, good evidence supports the short-term effectiveness of acetaminophen and nonsteroidal anti-inflammatory drugs, as well as skeletal muscle relaxants.^1,9 For chronic cases, good evidence exists for prescribing tricyclic antidepressants.⁹

Nonpharmacologic interventions include, in acute cases, active care, spinal manipulation, and superficial heat (eg, hot packs). For subacute and chronic cases, think about intensive interdisciplinary rehabilitation interventions—therapeutic exercise, soft-tissue manual techniques, acupuncture, movement re-education techniques, spinal manipulation, cognitive-behavioral therapy, or progressive relaxation.^1,10

Further customize your Tx approach

While recent data suggest that some front-line physicians who treat patients with LBP may have insufficient knowledge to do so^11-13 (see “Are we out of step?” ), there are promising developments, as well. Many primary care clinicians and researchers believe that nonspecific LBP resembles a heterogeneous condition and that intervention is more effective when treatment is matched to the patient’s history and examination findings.^14,15 In a survey of more than 600 primary care clinicians from multiple disciplines, including physical therapy, chiropractic, and medicine, 93% of the participants reported that they treat nonspecific LBP cases differently, depending on signs and symptoms, with 74% believing there are recognizable subgroups to guide management.¹⁴

An example of subgrouping patients with nonspecific LBP is the idea of treatment-based classification, which evidence has found to be reliable, effective, and cost-efficient for patients with LBP (TABLE).¹¹ In an RCT, Brennan and colleagues¹⁵ examined 123 patients with acute LBP (ie, back pain lasting <90 days) referred to a physical therapist for treatment. Patients were examined and then classified into 1 of 3 subgroups according to the type of treatment expected to work best for them: manipulation, stabilization, or specific exercise. Each patient was then randomly assigned to receive 1 of the 3 treatments. Post-treatment analysis compared outcomes between patients who received treatment matched to their classification and those whose treatment did not match their classification.

At 4 weeks, the matched-treatment group had significantly greater reductions in disability compared with the unmatched-treatment group; this difference was still evident at 1 year. The authors agreed that LBP should not be thought of as a homogenous condition, and found that outcomes can be improved with subgrouping to guide intervention selection.¹⁵

TABLE
Matching physical therapy to low back pain attributes can improve outcomes

CORRESPONDENCE William R. VanWye, PT, DPT, 1600 Albany Street, Beech Grove, IN 46107; [email protected]

CASE A 50-year-old construction worker comes in for an office visit because he’s been experiencing intermittent low back pain that’s been occurring more frequently. He says he has not been injured and that he always takes care on the job to minimize physical risk. He reports no symptoms other than the back pain.

How would you proceed with this patient’s care? Would you order a plain radiograph to be sure nothing dire is causing the patient’s pain—or would you skip it? And would you know how to match your patient’s history and exam findings with specific physical therapy interventions?

The following review brings the latest guidelines and research to bear on these questions—and others—as you care for patients with nonspecific low back pain.

Categorizing low back pain to direct your investigation

The 2007 Joint Clinical Practice Guideline issued by the American College of Physicians and the American Pain Society encourages clinicians to perform a focused history and physical exam to classify patients into 1 of 3 broad categories: nonspecific low back pain (LBP), LBP potentially associated with radiculopathy or spinal stenosis, or LBP potentially associated with another specific spinal cause.¹

Patients in the last category often exhibit findings in the history and physical examination suggestive of severe or progressive neurologic defi cits. Refer these patients for further diagnostic testing.

Patients presenting with persistent (>4 weeks) LBP and signs and symptoms of radiculopathy or spinal stenosis are best referred for, preferably, magnetic resonance imaging (MRI) or for computed tomography (CT)—but only if the patient is a candidate for surgery or epidural steroid injection.

For patients with nonspecific LBP, which accounts for most cases, practice guidelines recommend against routine use of imaging or other diagnostic procedures.¹

Unfortunately, however, some clinicians still use routine imaging in the absence of significant findings or without clear indication.² One argument used to justify this action—particularly by some who consider nonspecific LBP to be a diagnosis of exclusion —is the desire to rule out a serious underlying spinal condition.

What the research tells us about routine imaging

A recently published systematic review and meta-analysis compiled data relevant to more than 1800 subjects from 6 randomized controlled trials (RCTs). The authors examined early, routine use of lumbar imaging (radiography, MRI, or CT) in patients with acute or subacute LBP. They found that, without clear indication from findings in the history and physical examination, immediate imaging does not improve clinical outcomes (ie, diminish pain or improve daily function).³

Another study took a closer look at advanced imaging for LBP. In an RCT including more than 300 patients with a mean age of 53 years, investigators compared outcomes for patients receiving either plain radiographs or rapid MRI. No differences were noted in outcomes for back-related disability, pain, health survey results, preference scores, satisfaction, or costs at 12 months.

Furthermore, patients receiving rapid MRI were more likely to undergo surgery, which also failed to improve outcomes. As a result, the authors cautioned against unnecessary use of advanced imaging, as it could increase costs of care and possibly increase surgical intervention without improved outcomes.⁴ These studies substantiate practice guidelines regarding the use of imaging for patients with nonspecific LBP.

Not helpful, and perhaps harmful?
When imaging is unwarranted, it unnecessarily exposes patients to ionizing radiation, especially objectionable for younger women.¹ Imaging can also lead to the identifi cation of pathology unrelated to a patient’s LBP.^1,5

As mentioned above, patients receiving rapid MRI were more likely to receive surgical intervention that did not improve outcomes.⁴ This observation may reflect, in part, findings of pathoanatomical abnormalities that have little or no correlation with patient symptoms. In a random sample of 148 subjects ages 36 to 71 years—nearly half of whom had never experienced back pain—Jarvik and colleagues⁵ found MRI evidence of annular tears, disc bulges, disc protrusions, facet joint degeneration, end plate changes, and mild spondylolisthesis. The authors concluded that such MRI findings are therefore of limited diagnostic value.⁵

Labeling can be harmful, too. Identification of pathology that could well be unrelated to LBP can result in a specific, presumed diagnosis, possibly inducing a phenomenon known as the labeling effect. The search for a specific diagnosis or label for patients with nonspecific LBP could cause them to perceive their low back condition as being more serious than it actually is. Patients may then develop distorted beliefs regarding the true nature of their health status. The labeling effect could even alter the natural course of an otherwise benign condition.^6,7

It’s time to treat: Tell patients to remain active

Practice guidelines for nonspecific LBP recommend providing patients with evidence-based education that emphasizes the favorable course of this condition and that encourages them to remain active.¹ This recommendation was assessed retrospectively in a study of nearly 1200 patients receiving physical therapy for acute LBP. The authors found that adherence to the recommendation for activity and exercise yielded significant reductions in disability and pain, and resulted in significantly fewer visits and lower charges.⁸

For acute cases of nonspecific LBP, good evidence supports the short-term effectiveness of acetaminophen and nonsteroidal anti-inflammatory drugs, as well as skeletal muscle relaxants.^1,9 For chronic cases, good evidence exists for prescribing tricyclic antidepressants.⁹

Nonpharmacologic interventions include, in acute cases, active care, spinal manipulation, and superficial heat (eg, hot packs). For subacute and chronic cases, think about intensive interdisciplinary rehabilitation interventions—therapeutic exercise, soft-tissue manual techniques, acupuncture, movement re-education techniques, spinal manipulation, cognitive-behavioral therapy, or progressive relaxation.^1,10

Further customize your Tx approach

While recent data suggest that some front-line physicians who treat patients with LBP may have insufficient knowledge to do so^11-13 (see “Are we out of step?” ), there are promising developments, as well. Many primary care clinicians and researchers believe that nonspecific LBP resembles a heterogeneous condition and that intervention is more effective when treatment is matched to the patient’s history and examination findings.^14,15 In a survey of more than 600 primary care clinicians from multiple disciplines, including physical therapy, chiropractic, and medicine, 93% of the participants reported that they treat nonspecific LBP cases differently, depending on signs and symptoms, with 74% believing there are recognizable subgroups to guide management.¹⁴

An example of subgrouping patients with nonspecific LBP is the idea of treatment-based classification, which evidence has found to be reliable, effective, and cost-efficient for patients with LBP (TABLE).¹¹ In an RCT, Brennan and colleagues¹⁵ examined 123 patients with acute LBP (ie, back pain lasting <90 days) referred to a physical therapist for treatment. Patients were examined and then classified into 1 of 3 subgroups according to the type of treatment expected to work best for them: manipulation, stabilization, or specific exercise. Each patient was then randomly assigned to receive 1 of the 3 treatments. Post-treatment analysis compared outcomes between patients who received treatment matched to their classification and those whose treatment did not match their classification.

At 4 weeks, the matched-treatment group had significantly greater reductions in disability compared with the unmatched-treatment group; this difference was still evident at 1 year. The authors agreed that LBP should not be thought of as a homogenous condition, and found that outcomes can be improved with subgrouping to guide intervention selection.¹⁵

TABLE
Matching physical therapy to low back pain attributes can improve outcomes

CORRESPONDENCE William R. VanWye, PT, DPT, 1600 Albany Street, Beech Grove, IN 46107; [email protected]

CASE A 50-year-old construction worker comes in for an office visit because he’s been experiencing intermittent low back pain that’s been occurring more frequently. He says he has not been injured and that he always takes care on the job to minimize physical risk. He reports no symptoms other than the back pain.

How would you proceed with this patient’s care? Would you order a plain radiograph to be sure nothing dire is causing the patient’s pain—or would you skip it? And would you know how to match your patient’s history and exam findings with specific physical therapy interventions?

The following review brings the latest guidelines and research to bear on these questions—and others—as you care for patients with nonspecific low back pain.

Categorizing low back pain to direct your investigation

The 2007 Joint Clinical Practice Guideline issued by the American College of Physicians and the American Pain Society encourages clinicians to perform a focused history and physical exam to classify patients into 1 of 3 broad categories: nonspecific low back pain (LBP), LBP potentially associated with radiculopathy or spinal stenosis, or LBP potentially associated with another specific spinal cause.¹

Patients in the last category often exhibit findings in the history and physical examination suggestive of severe or progressive neurologic defi cits. Refer these patients for further diagnostic testing.

Patients presenting with persistent (>4 weeks) LBP and signs and symptoms of radiculopathy or spinal stenosis are best referred for, preferably, magnetic resonance imaging (MRI) or for computed tomography (CT)—but only if the patient is a candidate for surgery or epidural steroid injection.

For patients with nonspecific LBP, which accounts for most cases, practice guidelines recommend against routine use of imaging or other diagnostic procedures.¹

Unfortunately, however, some clinicians still use routine imaging in the absence of significant findings or without clear indication.² One argument used to justify this action—particularly by some who consider nonspecific LBP to be a diagnosis of exclusion —is the desire to rule out a serious underlying spinal condition.

What the research tells us about routine imaging

A recently published systematic review and meta-analysis compiled data relevant to more than 1800 subjects from 6 randomized controlled trials (RCTs). The authors examined early, routine use of lumbar imaging (radiography, MRI, or CT) in patients with acute or subacute LBP. They found that, without clear indication from findings in the history and physical examination, immediate imaging does not improve clinical outcomes (ie, diminish pain or improve daily function).³

Another study took a closer look at advanced imaging for LBP. In an RCT including more than 300 patients with a mean age of 53 years, investigators compared outcomes for patients receiving either plain radiographs or rapid MRI. No differences were noted in outcomes for back-related disability, pain, health survey results, preference scores, satisfaction, or costs at 12 months.

Furthermore, patients receiving rapid MRI were more likely to undergo surgery, which also failed to improve outcomes. As a result, the authors cautioned against unnecessary use of advanced imaging, as it could increase costs of care and possibly increase surgical intervention without improved outcomes.⁴ These studies substantiate practice guidelines regarding the use of imaging for patients with nonspecific LBP.

Not helpful, and perhaps harmful?
When imaging is unwarranted, it unnecessarily exposes patients to ionizing radiation, especially objectionable for younger women.¹ Imaging can also lead to the identifi cation of pathology unrelated to a patient’s LBP.^1,5

As mentioned above, patients receiving rapid MRI were more likely to receive surgical intervention that did not improve outcomes.⁴ This observation may reflect, in part, findings of pathoanatomical abnormalities that have little or no correlation with patient symptoms. In a random sample of 148 subjects ages 36 to 71 years—nearly half of whom had never experienced back pain—Jarvik and colleagues⁵ found MRI evidence of annular tears, disc bulges, disc protrusions, facet joint degeneration, end plate changes, and mild spondylolisthesis. The authors concluded that such MRI findings are therefore of limited diagnostic value.⁵

Labeling can be harmful, too. Identification of pathology that could well be unrelated to LBP can result in a specific, presumed diagnosis, possibly inducing a phenomenon known as the labeling effect. The search for a specific diagnosis or label for patients with nonspecific LBP could cause them to perceive their low back condition as being more serious than it actually is. Patients may then develop distorted beliefs regarding the true nature of their health status. The labeling effect could even alter the natural course of an otherwise benign condition.^6,7

It’s time to treat: Tell patients to remain active

Practice guidelines for nonspecific LBP recommend providing patients with evidence-based education that emphasizes the favorable course of this condition and that encourages them to remain active.¹ This recommendation was assessed retrospectively in a study of nearly 1200 patients receiving physical therapy for acute LBP. The authors found that adherence to the recommendation for activity and exercise yielded significant reductions in disability and pain, and resulted in significantly fewer visits and lower charges.⁸

For acute cases of nonspecific LBP, good evidence supports the short-term effectiveness of acetaminophen and nonsteroidal anti-inflammatory drugs, as well as skeletal muscle relaxants.^1,9 For chronic cases, good evidence exists for prescribing tricyclic antidepressants.⁹

Nonpharmacologic interventions include, in acute cases, active care, spinal manipulation, and superficial heat (eg, hot packs). For subacute and chronic cases, think about intensive interdisciplinary rehabilitation interventions—therapeutic exercise, soft-tissue manual techniques, acupuncture, movement re-education techniques, spinal manipulation, cognitive-behavioral therapy, or progressive relaxation.^1,10

Further customize your Tx approach

While recent data suggest that some front-line physicians who treat patients with LBP may have insufficient knowledge to do so^11-13 (see “Are we out of step?” ), there are promising developments, as well. Many primary care clinicians and researchers believe that nonspecific LBP resembles a heterogeneous condition and that intervention is more effective when treatment is matched to the patient’s history and examination findings.^14,15 In a survey of more than 600 primary care clinicians from multiple disciplines, including physical therapy, chiropractic, and medicine, 93% of the participants reported that they treat nonspecific LBP cases differently, depending on signs and symptoms, with 74% believing there are recognizable subgroups to guide management.¹⁴

An example of subgrouping patients with nonspecific LBP is the idea of treatment-based classification, which evidence has found to be reliable, effective, and cost-efficient for patients with LBP (TABLE).¹¹ In an RCT, Brennan and colleagues¹⁵ examined 123 patients with acute LBP (ie, back pain lasting <90 days) referred to a physical therapist for treatment. Patients were examined and then classified into 1 of 3 subgroups according to the type of treatment expected to work best for them: manipulation, stabilization, or specific exercise. Each patient was then randomly assigned to receive 1 of the 3 treatments. Post-treatment analysis compared outcomes between patients who received treatment matched to their classification and those whose treatment did not match their classification.

At 4 weeks, the matched-treatment group had significantly greater reductions in disability compared with the unmatched-treatment group; this difference was still evident at 1 year. The authors agreed that LBP should not be thought of as a homogenous condition, and found that outcomes can be improved with subgrouping to guide intervention selection.¹⁵

TABLE
Matching physical therapy to low back pain attributes can improve outcomes

CORRESPONDENCE William R. VanWye, PT, DPT, 1600 Albany Street, Beech Grove, IN 46107; [email protected]

With preparticipation physical evaluations (PPEs) required for competitive athletic activities at colleges nationwide¹ and at high schools and middle schools in the vast majority of states,² the start of a new school year often brings a barrage of student visits. Yet despite this almost universal requirement, there is no universal standard for the PPE.

There is, however, a new evidence-based guideline. Preparticipation Physical Evaluation, 4th edition,³ released earlier this year, was created by the American Academy of Family Physicians, American Academy of Pediatrics, American College of Sports Medicine, American Medical Society for Sports Medicine, American Orthopaedic Society for Sports Medicine, and American Osteopathic Academy of Sports Medicine.

This guideline describes how to conduct a thorough medical history and a targeted physical exam, with a focus on activity-related risks to various organ systems. Regardless of the specific activity the student is interested in pursuing, however, emphasis on cardiac, neurological, and musculoskeletal systems is crucial because of the frequency and gravity of complications.

Help with the medical history: A downloadable form

The history is the most important part of the PPE, which should be scheduled at least 6 weeks before the sports season starts to allow time for follow-up testing or consultation. A form (which can be downloaded along with a physical exam form from the American College of Sports Medicine’s [ACSM] Web site at http://www.ppesportsevaluation.org/body.html) features 54 questions, which cover a range of organ systems and highlight areas of common sports injury and disability. The answers to these questions alone can identify 75% of problems affecting athletes,³ including chronic conditions and medications that may require adjustment or closer monitoring.

Start with the cardiovascular system
As many as 85% of sudden deaths in young athletes are related to underlying cardiac abnormalities, according to a 10-year study of 150 such cases.⁴ Indeed, sudden cardiac death occurs in approximately 1 in 200,000 high school athletes.^5,6 Not surprisingly, those engaged in high-intensity activity are at highest risk.⁷

Thus, screening school-aged athletes for potential causes of sudden cardiac death is a primary objective of the PPE.³ The American Heart Association (AHA) recommends a 12-part evaluation (TABLE 1) of the heart, with 8 history questions and 4 physical exam components, to properly screen for cardiac disease in the young athlete.⁸ Syncope, chest pain, and dyspnea, particularly if associated with exertion, may be signs of underlying cardiac disorders that warrant confirmatory testing;⁵ a family history of premature death or disability from cardiac disease indicate a need for additional testing, as well. Further evaluation normally entails a combination of cardiac testing and consultation with a cardiologist.

Ask about asthma, including exercise-induced asthma (EIA). Identifying any respiratory conditions is vital to ensure adequate treatment and optimal performance. Patients with EIA are normally asymptomatic at rest, with no disturbance of peak expiratory flow. The presence of cough, chest tightness, wheezing, dyspnea, or loss of endurance during exercise suggests an EIA diagnosis.⁹ If a patient reports any such symptoms, order pulmonary function tests for confirmation.

EIA affects between 10% and 50% of athletes, depending on the sport.¹⁰ In up to 80% of athletes with EIA, use of inhaled short-acting beta-agonists prior to participation can help to prevent symptoms.⁹ Any athlete who reports symptoms of asthma, whether or not it is exercise-induced, requires treatment to prevent serious respiratory sequelae.

TABLE 1
Ask these 12 questions during cardiovascular screening⁵

Take a neurologic history
The neurologic system is highly susceptible to injury during sports activity. Identifying a history of concussion, nerve injury, or neurologic deficit is important both to prevent future injury and to avoid worsening of a current disability.

Always ask about transient neuropraxia, also known as a “stinger” or “burner”—a traction or compression injury to the nerves of the brachial plexus or cervical nerve roots that is sustained by 50% to 65% of college football players.¹¹ The injury causes numbness, weakness, or both, in an upper extremity. Although symptoms are commonly transient, the injury may be severe enough to cause more prolonged symptoms; about 5% to 10% of the time, symptoms related to transient neuropraxia persist for hours to weeks.¹¹

A history of numbness or weakness that occurs simultaneously in more than 1 extremity is suggestive of underlying cervical cord neuropraxia.¹² Patients with such symptoms require further evaluation for the presence of spinal stenosis, cervical ligamentous injury, and spinal cord injury before being cleared to participate in sports.

Concussion is a common neurologic insult among competitive athletes (See “Update on concussion: Here’s what the experts have to say”), accounting for anywhere from 6.5% to 18.5% of injuries in collegiate contact sports.^1,13 Identifying students who have had head injuries and are therefore at increased risk for future concussion is a key function of the PPE.^13,14

An individual who suffers even a mild head injury prior to full resolution of an initial concussion is at risk for second impact syndrome—a devastating brain injury that causes a loss of autoregulation of cerebral blood flow, leading to rapid swelling, herniation, and death. Thus, any athlete with a recent history of concussion requires further evaluation—and disqualification from play until symptoms fully resolve. In fact, anyone with a history of concussion—recent or not—may need neuropsychologic testing to assess baseline level of cognitive function. Thorough documentation is critical in such cases. In the event of a repeat concussion, having this information may assist with treatment and return-to-play decisions.

Review other significant findings noted on the history form. Follow up, as needed, with questions about any missing or dysfunctional organ or bodily system (The ACSM Web site includes a supplemental history form for athletes with special needs to document additional details, if necessary).

In some cases, protective equipment may be sufficient (a student with defective vision in 1 eye can wear protective eyewear to prevent injury to the other eye, for instance). In other cases, a conversation with the patient and his or her family regarding the risk of serious injury may be in order. Parents of a child with only 1 kidney, for example, should be advised that contact sports pose a small, but real risk of damage to that kidney, potentially resulting in the need for dialysis.

Take the opportunity to discuss body image, mental health

For many adolescents—up to 50%, by some counts¹⁵—the PPE is their only interaction with a clinician. Thus, it provides a good opportunity for you to talk to young athletes about such sensitive topics as high-risk behaviors, mental health issues, body image, and personal safety. To help ensure that students feel free to talk openly, the new PPE forms have removed key questions about high-risk behaviors from the patient history (which requires a parent’s signature). Instead, a series of questions is listed under the heading “physician reminders” at the top of the form for the physical exam,³ which is typically conducted in a private setting.

Overweight? Underweight? Height and weight, a standard part of the physical along with blood pressure and pulse rate, may clue you in to the existence of an eating disorder. Calculate body mass index (BMI), looking for students who are underweight (BMI <19 kg/m²) or overweight (BMI >25 kg/m²). Overweight athletes are at increased risk for heat illness and may need a preseason conditioning program.¹⁶ Underweight athletes—particularly young women—may be at additional risk.

Female athlete triad. When examining young women, be alert to signs and symptoms of the female athlete triad, a syndrome of disordered eating, amenorrhea, and osteopenia or osteoporosis. This related spectrum of medical problems can pose a significant health risk, as it involves a cycle of low energy intake that “turns off” the reproductive cycle and creates a hypoestrogenic state. The lack of estrogen has a devastating effect on bone mineral resorption and can lead to osteopenia.¹⁷ While the prevalence of the female athlete triad is low in the general population, 1 study noted that nearly 6% of young female athletes met the criteria for 2 of the 3 components of the triad, and as many as 20% had at least 1.¹⁸

Is eye protection needed? Ear plugs?
Check vision in each eye, both with and without corrective lenses. A student whose corrected vision is 20/50 or worse in 1 eye qualifies as functionally one-eyed—and must wear full eye protection during athletic activities.¹⁹ Document pupil size for all patients to ensure that anisocoria is not confused with a neurological insult in the event of a sports injury.

Water sports? Examine the nares, inspecting the septum for signs of deviation or perforation. Th is portion of the exam, however, can be tailored to the type of sport in which the patient plans to participate. Students who expect to be on a swim team or participate in other water sports require an evaluation of the ears, including the tympanic membranes and external auditory canals. Swimming and humid environments are risk factors for otitis externa, and repeated cold water exposure is a risk for the development of external auditory exostoses.^20,21 A swimmer with perforated tympanic membranes should be advised to wear ear plugs to protect the middle ear.

Is the patient a wrestler? Pay close attention to the ears of any student who plans to join a wrestling team, documenting the absence—or presence—of cauliflower ear. A notation about the presence of this ear-deforming condition in the medical record ensures that any new auricular hematoma can be evaluated and treated with knowledge of the prior injury.

Cardiac findings that warrant further work-up
Elevated blood pressure, the most common cardiovascular abnormality in people participating in competitive sports,²² is categorized in stages as defined by the Second Task Force on Blood Pressure Control in Children²³ and the Joint National Commission VII for adults ages 18 and older.²⁴ Children with significant hypertension (TABLE 2) and adults with stage I hypertension (140-159/90-99 mm Hg) may participate as long as there is no indication of end organ disease. A child with severe hypertension or an adult with stage II hypertension (>159/99 mm Hg), however, should not be cleared for participation until he or she has had further evaluation and treatment.^22,25

Similarly, any patient with an elevated heart rate needs a medical work-up prior to participation to determine the underlying cause of tachycardia.²² This may be a sign of an underlying cardiac arrhythmia or another medical condition that must be treated prior to athletic competition.

Auscultate the heart to screen for underlying cardiac disease. Listen to all 4 standard regions, with the patient in both a supine and a standing position. If you detect a murmur, perform auscultation while the patient squats and while performing the Valsalva maneuver. The murmur of hypertrophic cardiomyopathy (HCM)—the key cause of sudden cardiac death—is systolic, increasing with standing and the Valsalva maneuver and decreasing with squatting and a supine position. Any murmur that is 3/6 or greater in sound or has characteristics of HCM needs further evaluation before the patient can be cleared for sports activities. (See “Hypertrophic cardiomyopathy: Ask athletes these 9 questions,” J Fam Pract. 2009;58:576-584).

Palpate the femoral pulses. Absence of, or decreased, femoral pulses compared with brachial pulses may suggest coarctation of the aorta.

TABLE 2
Hypertension in children: Is it significant or severe?²⁴

Check lungs, abdomen, and skin
Auscultate the lungs, which should be clear and absent of adventitious or diminished breath sounds. Any patient with abnormal breath sounds requires further evaluation and/or treatment before being cleared for sports participation.

Examine the abdomen to assess the presence of organomegaly. Enlargement of the liver and spleen may be a sign of an underlying disease process. Mononucleosis, which causes splenomegaly, is associated with a 0.1% to 0.2% rate of rupture that may be related to trauma or the Valsalva maneuver.²⁶ Athletes with organomegaly should not be cleared for athletic activity without further evaluation.³

Perform a genitourinary examination on young men. Check for the presence of both testicles and palpate for masses and inguinal hernias. A patient with a solitary testicle will require protective gear during certain sporting events to prevent injury.²⁷

Assess the condition of the skin. In this case, the activity the student plans to pursue will determine the extent of the examination. Skin condition is particularly important for wrestlers, as the close contact involved predisposes the athletes to infectious skin conditions. Molluscum contagiosum, tinea corporis, and herpes gladiatorum, in particular, should be treated before the student is cleared to participate in wrestling. Regardless of the sport involved, however, identify chronic skin problems and institute prophylaxis, as needed.

Conduct a general musculoskeletal exam
A brief evaluation of strength and mobility of each joint and muscle group is sufficient to determine if a student has adequate baseline musculoskeletal function to compete.³ A joint-specific exam will not only help to assess the particular function and stability of each joint, but may reveal subtle deficits of particular joints or muscle groups that may be amenable to rehabilitation or other treatments prior to the start of the sport season.

The emphasis here, too, may vary based on the sport the patient plans to participate in. If you examine the ankles of a volleyball or soccer player and find ligamentous laxity from a prior injury, for example, consider prescribing a brace or physical therapy.

Follow up with a neurologic exam, which is often paired with the musculoskeletal exam and is considered adequate if the patient possesses full strength in all muscle groups. A patient with a history of multiple stingers may warrant a more detailed examination of strength, reflex, and sensation in the upper extremities to screen for signs of residual nerve injury. Similarly, a patient with a history of multiple concussions may need a more detailed exam that includes the cranial nerves, evaluation of balance, and possibly baseline neuropsychological testing.

Screen for Marfan syndrome

This genetic disorder involves mutations of the fibrillin gene that lead to a diverse presentation of abnormalities in multiple organ systems. Primarily because of the effects of Marfan syndrome on the cardiovascular system, it is important to identify it as part of the PPE. Individuals with this disorder have an increased risk for valvular disorders and for aortic dilation that can lead to dissection or rupture. If the history reveals that the patient has a family member with this disorder or has a history of spontaneous pneumothorax or mitral valve prolapse, look closely for the following skeletal and cardiac abnormalities:

• pectus carinatum or excavatum
• arm span-to-height ratio >1.05
• arachnodactyly
• pes planus
• scoliosis
• reduced elbow extension
• highly arched palate
• murmur of mitral valve prolapse or regurgitation.

If you find any of these abnormalities, postpone sports clearance until the patient undergoes further evaluation.²⁸ To establish a diagnosis of Marfan syndrome, the patient must have major criteria in 2 organ systems, with involvement of a third system.

Make a determination: Should the patient be cleared?

The final part of the PPE, of course, is your decision as to whether the patient should be cleared to engage in competitive activities. Clearance falls into 4 categories: (1) Clearance without restriction; (2) clearance with recommendations for further evaluation or treatment; (3) not cleared; restricted until completion of further testing/consultation; (4) complete restriction from certain or all sports.

The goal of the PPE is to provide safe participation in sports for all athletes, not to disqualify anyone. Fortunately, the vast majority of student athletes qualify for clearance without restriction. About 3% to 10% of those who undergo preparticipation screening require further evaluation prior to sports clearance, and less than 1% are disqualified.³

Being prevented from participating in sports can be a major stressor that interferes with the student’s sense of well-being. The inability to exercise or participate in organized sports poses a risk to the patient’s overall health, and may contribute to social isolation. In fact, being restricted from playing organized sports has been reported to be as stressful as the death of a close friend.²⁹

Help in making the determination. Nonetheless, there are circumstances when restriction is necessary. In addition to the new PPE, there are 2 sources you can turn to for help in making that determination. The Bethesda guidelines—Eligibility Recommendations for Competitive Athletes with Cardiovascular Abnormalities, published by the American College of Cardiology and available at http://content.onlinejacc.org/cgi/content/full/45/8/1318, call for disqualifying students only when the activity would pose a clear danger to their health.²³

The American Academy of Pediatrics (AAP) has published a set of sports participation guidelines of its own (available at http://www.guideline.gov/summary/summary.aspx?doc_id=13439) for children with conditions that range from diabetes mellitus and human immunodeficiency virus to mitral valve prolapse, rheumatoid arthritis, and bleeding disorders. There are, of course, some disorders for which the AAP would disqualify students from participation, but in many cases it recommends a “qualified Yes.”^27,30

CORRESPONDENCE Jason Womack, MD, UMDNJ - Robert Wood Johnson Medical School, 1 Robert Wood Johnson Place, Department of Family Medicine, MEB 2nd floor, New Brunswick, NJ 08903; [email protected]

With preparticipation physical evaluations (PPEs) required for competitive athletic activities at colleges nationwide¹ and at high schools and middle schools in the vast majority of states,² the start of a new school year often brings a barrage of student visits. Yet despite this almost universal requirement, there is no universal standard for the PPE.

There is, however, a new evidence-based guideline. Preparticipation Physical Evaluation, 4th edition,³ released earlier this year, was created by the American Academy of Family Physicians, American Academy of Pediatrics, American College of Sports Medicine, American Medical Society for Sports Medicine, American Orthopaedic Society for Sports Medicine, and American Osteopathic Academy of Sports Medicine.

This guideline describes how to conduct a thorough medical history and a targeted physical exam, with a focus on activity-related risks to various organ systems. Regardless of the specific activity the student is interested in pursuing, however, emphasis on cardiac, neurological, and musculoskeletal systems is crucial because of the frequency and gravity of complications.

Help with the medical history: A downloadable form

The history is the most important part of the PPE, which should be scheduled at least 6 weeks before the sports season starts to allow time for follow-up testing or consultation. A form (which can be downloaded along with a physical exam form from the American College of Sports Medicine’s [ACSM] Web site at http://www.ppesportsevaluation.org/body.html) features 54 questions, which cover a range of organ systems and highlight areas of common sports injury and disability. The answers to these questions alone can identify 75% of problems affecting athletes,³ including chronic conditions and medications that may require adjustment or closer monitoring.

Start with the cardiovascular system
As many as 85% of sudden deaths in young athletes are related to underlying cardiac abnormalities, according to a 10-year study of 150 such cases.⁴ Indeed, sudden cardiac death occurs in approximately 1 in 200,000 high school athletes.^5,6 Not surprisingly, those engaged in high-intensity activity are at highest risk.⁷

Thus, screening school-aged athletes for potential causes of sudden cardiac death is a primary objective of the PPE.³ The American Heart Association (AHA) recommends a 12-part evaluation (TABLE 1) of the heart, with 8 history questions and 4 physical exam components, to properly screen for cardiac disease in the young athlete.⁸ Syncope, chest pain, and dyspnea, particularly if associated with exertion, may be signs of underlying cardiac disorders that warrant confirmatory testing;⁵ a family history of premature death or disability from cardiac disease indicate a need for additional testing, as well. Further evaluation normally entails a combination of cardiac testing and consultation with a cardiologist.

Ask about asthma, including exercise-induced asthma (EIA). Identifying any respiratory conditions is vital to ensure adequate treatment and optimal performance. Patients with EIA are normally asymptomatic at rest, with no disturbance of peak expiratory flow. The presence of cough, chest tightness, wheezing, dyspnea, or loss of endurance during exercise suggests an EIA diagnosis.⁹ If a patient reports any such symptoms, order pulmonary function tests for confirmation.

EIA affects between 10% and 50% of athletes, depending on the sport.¹⁰ In up to 80% of athletes with EIA, use of inhaled short-acting beta-agonists prior to participation can help to prevent symptoms.⁹ Any athlete who reports symptoms of asthma, whether or not it is exercise-induced, requires treatment to prevent serious respiratory sequelae.

TABLE 1
Ask these 12 questions during cardiovascular screening⁵

Take a neurologic history
The neurologic system is highly susceptible to injury during sports activity. Identifying a history of concussion, nerve injury, or neurologic deficit is important both to prevent future injury and to avoid worsening of a current disability.

Always ask about transient neuropraxia, also known as a “stinger” or “burner”—a traction or compression injury to the nerves of the brachial plexus or cervical nerve roots that is sustained by 50% to 65% of college football players.¹¹ The injury causes numbness, weakness, or both, in an upper extremity. Although symptoms are commonly transient, the injury may be severe enough to cause more prolonged symptoms; about 5% to 10% of the time, symptoms related to transient neuropraxia persist for hours to weeks.¹¹

A history of numbness or weakness that occurs simultaneously in more than 1 extremity is suggestive of underlying cervical cord neuropraxia.¹² Patients with such symptoms require further evaluation for the presence of spinal stenosis, cervical ligamentous injury, and spinal cord injury before being cleared to participate in sports.

Concussion is a common neurologic insult among competitive athletes (See “Update on concussion: Here’s what the experts have to say”), accounting for anywhere from 6.5% to 18.5% of injuries in collegiate contact sports.^1,13 Identifying students who have had head injuries and are therefore at increased risk for future concussion is a key function of the PPE.^13,14

An individual who suffers even a mild head injury prior to full resolution of an initial concussion is at risk for second impact syndrome—a devastating brain injury that causes a loss of autoregulation of cerebral blood flow, leading to rapid swelling, herniation, and death. Thus, any athlete with a recent history of concussion requires further evaluation—and disqualification from play until symptoms fully resolve. In fact, anyone with a history of concussion—recent or not—may need neuropsychologic testing to assess baseline level of cognitive function. Thorough documentation is critical in such cases. In the event of a repeat concussion, having this information may assist with treatment and return-to-play decisions.

Review other significant findings noted on the history form. Follow up, as needed, with questions about any missing or dysfunctional organ or bodily system (The ACSM Web site includes a supplemental history form for athletes with special needs to document additional details, if necessary).

In some cases, protective equipment may be sufficient (a student with defective vision in 1 eye can wear protective eyewear to prevent injury to the other eye, for instance). In other cases, a conversation with the patient and his or her family regarding the risk of serious injury may be in order. Parents of a child with only 1 kidney, for example, should be advised that contact sports pose a small, but real risk of damage to that kidney, potentially resulting in the need for dialysis.

Take the opportunity to discuss body image, mental health

For many adolescents—up to 50%, by some counts¹⁵—the PPE is their only interaction with a clinician. Thus, it provides a good opportunity for you to talk to young athletes about such sensitive topics as high-risk behaviors, mental health issues, body image, and personal safety. To help ensure that students feel free to talk openly, the new PPE forms have removed key questions about high-risk behaviors from the patient history (which requires a parent’s signature). Instead, a series of questions is listed under the heading “physician reminders” at the top of the form for the physical exam,³ which is typically conducted in a private setting.

Overweight? Underweight? Height and weight, a standard part of the physical along with blood pressure and pulse rate, may clue you in to the existence of an eating disorder. Calculate body mass index (BMI), looking for students who are underweight (BMI <19 kg/m²) or overweight (BMI >25 kg/m²). Overweight athletes are at increased risk for heat illness and may need a preseason conditioning program.¹⁶ Underweight athletes—particularly young women—may be at additional risk.

Female athlete triad. When examining young women, be alert to signs and symptoms of the female athlete triad, a syndrome of disordered eating, amenorrhea, and osteopenia or osteoporosis. This related spectrum of medical problems can pose a significant health risk, as it involves a cycle of low energy intake that “turns off” the reproductive cycle and creates a hypoestrogenic state. The lack of estrogen has a devastating effect on bone mineral resorption and can lead to osteopenia.¹⁷ While the prevalence of the female athlete triad is low in the general population, 1 study noted that nearly 6% of young female athletes met the criteria for 2 of the 3 components of the triad, and as many as 20% had at least 1.¹⁸

Is eye protection needed? Ear plugs?
Check vision in each eye, both with and without corrective lenses. A student whose corrected vision is 20/50 or worse in 1 eye qualifies as functionally one-eyed—and must wear full eye protection during athletic activities.¹⁹ Document pupil size for all patients to ensure that anisocoria is not confused with a neurological insult in the event of a sports injury.

Water sports? Examine the nares, inspecting the septum for signs of deviation or perforation. Th is portion of the exam, however, can be tailored to the type of sport in which the patient plans to participate. Students who expect to be on a swim team or participate in other water sports require an evaluation of the ears, including the tympanic membranes and external auditory canals. Swimming and humid environments are risk factors for otitis externa, and repeated cold water exposure is a risk for the development of external auditory exostoses.^20,21 A swimmer with perforated tympanic membranes should be advised to wear ear plugs to protect the middle ear.

Is the patient a wrestler? Pay close attention to the ears of any student who plans to join a wrestling team, documenting the absence—or presence—of cauliflower ear. A notation about the presence of this ear-deforming condition in the medical record ensures that any new auricular hematoma can be evaluated and treated with knowledge of the prior injury.

Cardiac findings that warrant further work-up
Elevated blood pressure, the most common cardiovascular abnormality in people participating in competitive sports,²² is categorized in stages as defined by the Second Task Force on Blood Pressure Control in Children²³ and the Joint National Commission VII for adults ages 18 and older.²⁴ Children with significant hypertension (TABLE 2) and adults with stage I hypertension (140-159/90-99 mm Hg) may participate as long as there is no indication of end organ disease. A child with severe hypertension or an adult with stage II hypertension (>159/99 mm Hg), however, should not be cleared for participation until he or she has had further evaluation and treatment.^22,25

Similarly, any patient with an elevated heart rate needs a medical work-up prior to participation to determine the underlying cause of tachycardia.²² This may be a sign of an underlying cardiac arrhythmia or another medical condition that must be treated prior to athletic competition.

Auscultate the heart to screen for underlying cardiac disease. Listen to all 4 standard regions, with the patient in both a supine and a standing position. If you detect a murmur, perform auscultation while the patient squats and while performing the Valsalva maneuver. The murmur of hypertrophic cardiomyopathy (HCM)—the key cause of sudden cardiac death—is systolic, increasing with standing and the Valsalva maneuver and decreasing with squatting and a supine position. Any murmur that is 3/6 or greater in sound or has characteristics of HCM needs further evaluation before the patient can be cleared for sports activities. (See “Hypertrophic cardiomyopathy: Ask athletes these 9 questions,” J Fam Pract. 2009;58:576-584).

Palpate the femoral pulses. Absence of, or decreased, femoral pulses compared with brachial pulses may suggest coarctation of the aorta.

TABLE 2
Hypertension in children: Is it significant or severe?²⁴

Check lungs, abdomen, and skin
Auscultate the lungs, which should be clear and absent of adventitious or diminished breath sounds. Any patient with abnormal breath sounds requires further evaluation and/or treatment before being cleared for sports participation.

Examine the abdomen to assess the presence of organomegaly. Enlargement of the liver and spleen may be a sign of an underlying disease process. Mononucleosis, which causes splenomegaly, is associated with a 0.1% to 0.2% rate of rupture that may be related to trauma or the Valsalva maneuver.²⁶ Athletes with organomegaly should not be cleared for athletic activity without further evaluation.³

Perform a genitourinary examination on young men. Check for the presence of both testicles and palpate for masses and inguinal hernias. A patient with a solitary testicle will require protective gear during certain sporting events to prevent injury.²⁷

Assess the condition of the skin. In this case, the activity the student plans to pursue will determine the extent of the examination. Skin condition is particularly important for wrestlers, as the close contact involved predisposes the athletes to infectious skin conditions. Molluscum contagiosum, tinea corporis, and herpes gladiatorum, in particular, should be treated before the student is cleared to participate in wrestling. Regardless of the sport involved, however, identify chronic skin problems and institute prophylaxis, as needed.

Conduct a general musculoskeletal exam
A brief evaluation of strength and mobility of each joint and muscle group is sufficient to determine if a student has adequate baseline musculoskeletal function to compete.³ A joint-specific exam will not only help to assess the particular function and stability of each joint, but may reveal subtle deficits of particular joints or muscle groups that may be amenable to rehabilitation or other treatments prior to the start of the sport season.

The emphasis here, too, may vary based on the sport the patient plans to participate in. If you examine the ankles of a volleyball or soccer player and find ligamentous laxity from a prior injury, for example, consider prescribing a brace or physical therapy.

Follow up with a neurologic exam, which is often paired with the musculoskeletal exam and is considered adequate if the patient possesses full strength in all muscle groups. A patient with a history of multiple stingers may warrant a more detailed examination of strength, reflex, and sensation in the upper extremities to screen for signs of residual nerve injury. Similarly, a patient with a history of multiple concussions may need a more detailed exam that includes the cranial nerves, evaluation of balance, and possibly baseline neuropsychological testing.

Screen for Marfan syndrome

This genetic disorder involves mutations of the fibrillin gene that lead to a diverse presentation of abnormalities in multiple organ systems. Primarily because of the effects of Marfan syndrome on the cardiovascular system, it is important to identify it as part of the PPE. Individuals with this disorder have an increased risk for valvular disorders and for aortic dilation that can lead to dissection or rupture. If the history reveals that the patient has a family member with this disorder or has a history of spontaneous pneumothorax or mitral valve prolapse, look closely for the following skeletal and cardiac abnormalities:

• pectus carinatum or excavatum
• arm span-to-height ratio >1.05
• arachnodactyly
• pes planus
• scoliosis
• reduced elbow extension
• highly arched palate
• murmur of mitral valve prolapse or regurgitation.

If you find any of these abnormalities, postpone sports clearance until the patient undergoes further evaluation.²⁸ To establish a diagnosis of Marfan syndrome, the patient must have major criteria in 2 organ systems, with involvement of a third system.

Make a determination: Should the patient be cleared?

The final part of the PPE, of course, is your decision as to whether the patient should be cleared to engage in competitive activities. Clearance falls into 4 categories: (1) Clearance without restriction; (2) clearance with recommendations for further evaluation or treatment; (3) not cleared; restricted until completion of further testing/consultation; (4) complete restriction from certain or all sports.

The goal of the PPE is to provide safe participation in sports for all athletes, not to disqualify anyone. Fortunately, the vast majority of student athletes qualify for clearance without restriction. About 3% to 10% of those who undergo preparticipation screening require further evaluation prior to sports clearance, and less than 1% are disqualified.³

Being prevented from participating in sports can be a major stressor that interferes with the student’s sense of well-being. The inability to exercise or participate in organized sports poses a risk to the patient’s overall health, and may contribute to social isolation. In fact, being restricted from playing organized sports has been reported to be as stressful as the death of a close friend.²⁹

Help in making the determination. Nonetheless, there are circumstances when restriction is necessary. In addition to the new PPE, there are 2 sources you can turn to for help in making that determination. The Bethesda guidelines—Eligibility Recommendations for Competitive Athletes with Cardiovascular Abnormalities, published by the American College of Cardiology and available at http://content.onlinejacc.org/cgi/content/full/45/8/1318, call for disqualifying students only when the activity would pose a clear danger to their health.²³

The American Academy of Pediatrics (AAP) has published a set of sports participation guidelines of its own (available at http://www.guideline.gov/summary/summary.aspx?doc_id=13439) for children with conditions that range from diabetes mellitus and human immunodeficiency virus to mitral valve prolapse, rheumatoid arthritis, and bleeding disorders. There are, of course, some disorders for which the AAP would disqualify students from participation, but in many cases it recommends a “qualified Yes.”^27,30

CORRESPONDENCE Jason Womack, MD, UMDNJ - Robert Wood Johnson Medical School, 1 Robert Wood Johnson Place, Department of Family Medicine, MEB 2nd floor, New Brunswick, NJ 08903; [email protected]

With preparticipation physical evaluations (PPEs) required for competitive athletic activities at colleges nationwide¹ and at high schools and middle schools in the vast majority of states,² the start of a new school year often brings a barrage of student visits. Yet despite this almost universal requirement, there is no universal standard for the PPE.

There is, however, a new evidence-based guideline. Preparticipation Physical Evaluation, 4th edition,³ released earlier this year, was created by the American Academy of Family Physicians, American Academy of Pediatrics, American College of Sports Medicine, American Medical Society for Sports Medicine, American Orthopaedic Society for Sports Medicine, and American Osteopathic Academy of Sports Medicine.

This guideline describes how to conduct a thorough medical history and a targeted physical exam, with a focus on activity-related risks to various organ systems. Regardless of the specific activity the student is interested in pursuing, however, emphasis on cardiac, neurological, and musculoskeletal systems is crucial because of the frequency and gravity of complications.

Help with the medical history: A downloadable form

The history is the most important part of the PPE, which should be scheduled at least 6 weeks before the sports season starts to allow time for follow-up testing or consultation. A form (which can be downloaded along with a physical exam form from the American College of Sports Medicine’s [ACSM] Web site at http://www.ppesportsevaluation.org/body.html) features 54 questions, which cover a range of organ systems and highlight areas of common sports injury and disability. The answers to these questions alone can identify 75% of problems affecting athletes,³ including chronic conditions and medications that may require adjustment or closer monitoring.

Start with the cardiovascular system
As many as 85% of sudden deaths in young athletes are related to underlying cardiac abnormalities, according to a 10-year study of 150 such cases.⁴ Indeed, sudden cardiac death occurs in approximately 1 in 200,000 high school athletes.^5,6 Not surprisingly, those engaged in high-intensity activity are at highest risk.⁷

Thus, screening school-aged athletes for potential causes of sudden cardiac death is a primary objective of the PPE.³ The American Heart Association (AHA) recommends a 12-part evaluation (TABLE 1) of the heart, with 8 history questions and 4 physical exam components, to properly screen for cardiac disease in the young athlete.⁸ Syncope, chest pain, and dyspnea, particularly if associated with exertion, may be signs of underlying cardiac disorders that warrant confirmatory testing;⁵ a family history of premature death or disability from cardiac disease indicate a need for additional testing, as well. Further evaluation normally entails a combination of cardiac testing and consultation with a cardiologist.

Ask about asthma, including exercise-induced asthma (EIA). Identifying any respiratory conditions is vital to ensure adequate treatment and optimal performance. Patients with EIA are normally asymptomatic at rest, with no disturbance of peak expiratory flow. The presence of cough, chest tightness, wheezing, dyspnea, or loss of endurance during exercise suggests an EIA diagnosis.⁹ If a patient reports any such symptoms, order pulmonary function tests for confirmation.

EIA affects between 10% and 50% of athletes, depending on the sport.¹⁰ In up to 80% of athletes with EIA, use of inhaled short-acting beta-agonists prior to participation can help to prevent symptoms.⁹ Any athlete who reports symptoms of asthma, whether or not it is exercise-induced, requires treatment to prevent serious respiratory sequelae.

TABLE 1
Ask these 12 questions during cardiovascular screening⁵

Take a neurologic history
The neurologic system is highly susceptible to injury during sports activity. Identifying a history of concussion, nerve injury, or neurologic deficit is important both to prevent future injury and to avoid worsening of a current disability.

Always ask about transient neuropraxia, also known as a “stinger” or “burner”—a traction or compression injury to the nerves of the brachial plexus or cervical nerve roots that is sustained by 50% to 65% of college football players.¹¹ The injury causes numbness, weakness, or both, in an upper extremity. Although symptoms are commonly transient, the injury may be severe enough to cause more prolonged symptoms; about 5% to 10% of the time, symptoms related to transient neuropraxia persist for hours to weeks.¹¹

A history of numbness or weakness that occurs simultaneously in more than 1 extremity is suggestive of underlying cervical cord neuropraxia.¹² Patients with such symptoms require further evaluation for the presence of spinal stenosis, cervical ligamentous injury, and spinal cord injury before being cleared to participate in sports.

Concussion is a common neurologic insult among competitive athletes (See “Update on concussion: Here’s what the experts have to say”), accounting for anywhere from 6.5% to 18.5% of injuries in collegiate contact sports.^1,13 Identifying students who have had head injuries and are therefore at increased risk for future concussion is a key function of the PPE.^13,14

An individual who suffers even a mild head injury prior to full resolution of an initial concussion is at risk for second impact syndrome—a devastating brain injury that causes a loss of autoregulation of cerebral blood flow, leading to rapid swelling, herniation, and death. Thus, any athlete with a recent history of concussion requires further evaluation—and disqualification from play until symptoms fully resolve. In fact, anyone with a history of concussion—recent or not—may need neuropsychologic testing to assess baseline level of cognitive function. Thorough documentation is critical in such cases. In the event of a repeat concussion, having this information may assist with treatment and return-to-play decisions.

Review other significant findings noted on the history form. Follow up, as needed, with questions about any missing or dysfunctional organ or bodily system (The ACSM Web site includes a supplemental history form for athletes with special needs to document additional details, if necessary).

In some cases, protective equipment may be sufficient (a student with defective vision in 1 eye can wear protective eyewear to prevent injury to the other eye, for instance). In other cases, a conversation with the patient and his or her family regarding the risk of serious injury may be in order. Parents of a child with only 1 kidney, for example, should be advised that contact sports pose a small, but real risk of damage to that kidney, potentially resulting in the need for dialysis.

Take the opportunity to discuss body image, mental health

For many adolescents—up to 50%, by some counts¹⁵—the PPE is their only interaction with a clinician. Thus, it provides a good opportunity for you to talk to young athletes about such sensitive topics as high-risk behaviors, mental health issues, body image, and personal safety. To help ensure that students feel free to talk openly, the new PPE forms have removed key questions about high-risk behaviors from the patient history (which requires a parent’s signature). Instead, a series of questions is listed under the heading “physician reminders” at the top of the form for the physical exam,³ which is typically conducted in a private setting.

Overweight? Underweight? Height and weight, a standard part of the physical along with blood pressure and pulse rate, may clue you in to the existence of an eating disorder. Calculate body mass index (BMI), looking for students who are underweight (BMI <19 kg/m²) or overweight (BMI >25 kg/m²). Overweight athletes are at increased risk for heat illness and may need a preseason conditioning program.¹⁶ Underweight athletes—particularly young women—may be at additional risk.

Female athlete triad. When examining young women, be alert to signs and symptoms of the female athlete triad, a syndrome of disordered eating, amenorrhea, and osteopenia or osteoporosis. This related spectrum of medical problems can pose a significant health risk, as it involves a cycle of low energy intake that “turns off” the reproductive cycle and creates a hypoestrogenic state. The lack of estrogen has a devastating effect on bone mineral resorption and can lead to osteopenia.¹⁷ While the prevalence of the female athlete triad is low in the general population, 1 study noted that nearly 6% of young female athletes met the criteria for 2 of the 3 components of the triad, and as many as 20% had at least 1.¹⁸

Is eye protection needed? Ear plugs?
Check vision in each eye, both with and without corrective lenses. A student whose corrected vision is 20/50 or worse in 1 eye qualifies as functionally one-eyed—and must wear full eye protection during athletic activities.¹⁹ Document pupil size for all patients to ensure that anisocoria is not confused with a neurological insult in the event of a sports injury.

Water sports? Examine the nares, inspecting the septum for signs of deviation or perforation. Th is portion of the exam, however, can be tailored to the type of sport in which the patient plans to participate. Students who expect to be on a swim team or participate in other water sports require an evaluation of the ears, including the tympanic membranes and external auditory canals. Swimming and humid environments are risk factors for otitis externa, and repeated cold water exposure is a risk for the development of external auditory exostoses.^20,21 A swimmer with perforated tympanic membranes should be advised to wear ear plugs to protect the middle ear.

Is the patient a wrestler? Pay close attention to the ears of any student who plans to join a wrestling team, documenting the absence—or presence—of cauliflower ear. A notation about the presence of this ear-deforming condition in the medical record ensures that any new auricular hematoma can be evaluated and treated with knowledge of the prior injury.

Cardiac findings that warrant further work-up
Elevated blood pressure, the most common cardiovascular abnormality in people participating in competitive sports,²² is categorized in stages as defined by the Second Task Force on Blood Pressure Control in Children²³ and the Joint National Commission VII for adults ages 18 and older.²⁴ Children with significant hypertension (TABLE 2) and adults with stage I hypertension (140-159/90-99 mm Hg) may participate as long as there is no indication of end organ disease. A child with severe hypertension or an adult with stage II hypertension (>159/99 mm Hg), however, should not be cleared for participation until he or she has had further evaluation and treatment.^22,25

Similarly, any patient with an elevated heart rate needs a medical work-up prior to participation to determine the underlying cause of tachycardia.²² This may be a sign of an underlying cardiac arrhythmia or another medical condition that must be treated prior to athletic competition.

Auscultate the heart to screen for underlying cardiac disease. Listen to all 4 standard regions, with the patient in both a supine and a standing position. If you detect a murmur, perform auscultation while the patient squats and while performing the Valsalva maneuver. The murmur of hypertrophic cardiomyopathy (HCM)—the key cause of sudden cardiac death—is systolic, increasing with standing and the Valsalva maneuver and decreasing with squatting and a supine position. Any murmur that is 3/6 or greater in sound or has characteristics of HCM needs further evaluation before the patient can be cleared for sports activities. (See “Hypertrophic cardiomyopathy: Ask athletes these 9 questions,” J Fam Pract. 2009;58:576-584).

Palpate the femoral pulses. Absence of, or decreased, femoral pulses compared with brachial pulses may suggest coarctation of the aorta.

TABLE 2
Hypertension in children: Is it significant or severe?²⁴

Check lungs, abdomen, and skin
Auscultate the lungs, which should be clear and absent of adventitious or diminished breath sounds. Any patient with abnormal breath sounds requires further evaluation and/or treatment before being cleared for sports participation.

Examine the abdomen to assess the presence of organomegaly. Enlargement of the liver and spleen may be a sign of an underlying disease process. Mononucleosis, which causes splenomegaly, is associated with a 0.1% to 0.2% rate of rupture that may be related to trauma or the Valsalva maneuver.²⁶ Athletes with organomegaly should not be cleared for athletic activity without further evaluation.³