Applied Evidence

Shortly after the US Food and Drug Administration (FDA) approved fluoxetine for the treatment of major depression nearly a quarter of a century ago, Prozac became a household name. And, as a handful of additional selective serotonin reuptake inhibitors (SSRIs) were approved, the popularity of this category of antidepressant continued to grow.¹ In 2008, 5 of the 6 SSRIs on the US market (fluvoxamine was the exception) were among the 200 most frequently dispensed prescription drugs.²

While much has been written about the properties of SSRIs, uncertainty about many features of particular agents and what should be considered in selecting an antidepressant for a particular patient remains. To help clear up the confusion, we’ve addressed 7 questions about SSRIs that we are often asked. Although there are no simple answers, the evidence presented, both in the answers and in the tables that follow, will help primary care physicians make more informed choices for patients who require antidepressant therapy.

1. Which is the best SSRI to start a patient on?

A recent meta-analysis of 117 head-to-head studies assessed the efficacy and acceptability of 12 newer antidepressants, including all 6 SSRIs (citalopram, escitalopram, fluoxetine, fluvoxamine, paroxetine, and sertraline).³ The researchers found 2 SSRIs, sertraline and escitalopram, to be superior to the other medications studied on both counts. A choice between sertraline or escitalopram may be a reasonable starting point in many cases, but it is impossible to recommend 1 or 2 SSRIs that are effective for all patients. There are several reasons for this.

The first is addressed by the Cipriani meta-analysis. The researchers assessed the efficacy of initial antidepressant therapy at 8 weeks, so the results cannot be extrapolated to long-term response rates or acceptability. (For a detailed review of the meta-analysis, see “Try these 2 drugs first for depression,” J Fam Pract. 2009;58:365-369.)

A second reason is the substantial publication bias associated with studies of antidepressants. Turner et al assessed 74 studies registered with the FDA, determined whether the results were positive or negative, and categorized the studies based on publication status.⁴ Their conclusions: 97% of studies with positive findings (n=38) were published.

The majority of the remaining studies, all of which were determined to have negative findings, were either not published or published in a manner that suggested a positive outcome. When these additional studies are taken into account, the percentage of published studies with a positive response drops to 51%.⁴ In addition, the effect size of each agent is reduced when all the studies are included.

There are many confounding variables associated with publication, so Turner and his colleagues were unable to definitively determine the reason for the disparity. Nonetheless, their evaluation raises questions about the reported effectiveness, not only of SSRIs, but of antidepressants in general.

A recent Cochrane review of SSRIs for treatment of depression in children and adolescents (a topic covered in greater detail in the answer to Question 6) raised similar concerns. The reviewers cited study methodology and recruitment methods as potential sources of the conflicting results they found.⁵

Finally, individual differences, such as age, comorbidities, and medication history, require an individualized approach to SSRI treatment.

2. Which side effects are common to all SSRIs, and which can be resolved by switching agents?

As a class, SSRIs are well tolerated, but they do have some common adverse effects, most notably gastrointestinal (GI) problems, sexual dysfunction, and sleep disturbances. There are also considerable differences in SSRI profiles and a few adverse effects that a switch to another SSRI may alleviate or resolve.

Compared with other SSRIs, for example, sertraline has a higher risk of diarrhea, but this adverse event does not usually lead to medication discontinuation. If it is bothersome to the patient or does not resolve with continued therapy, a change to another agent might eliminate this adverse effect.

Weight gain has been found to be more significant with paroxetine than with fluoxetine or sertraline, which may be due to the anticholinergic action of the drug.⁶ For someone who is particularly concerned about extra weight, avoiding paroxetine in the first place, or changing a patient who’s already taking it to fluoxetine or sertraline, might be an effective treatment strategy.

Overall, paroxetine has more anticholinergic side effects than other SSRIs, the likely result of its higher affinity for muscarinic receptors ( TABLE 1 ). Its potential sedating effect,⁷ which can be extremely disturbing to some patients, may be a desired feature for others. A patient with insomnia, for example, might benefit from taking paroxetine at bedtime or switching to a less activating SSRI, such as citalopram or escitalopram. Excessive activation, such as significant insomnia, may also be a warning sign of undiagnosed bipolar disorder. Thus, careful screening is needed prior to switching medications or adding a hypnotic agent. (See the box below “Treating depression in primary care: Tips from a psychiatrist”.)

A switch to another SSRI is not the only way to alleviate a troublesome side effect, of course. Insomnia can also be managed by adding a short course of trazodone to the drug regimen or by switching the patient to mirtazapine or a tricyclic antidepressant (TCA).⁸

Augmentation with either bupropion or mirtazapine may alleviate sexual side effects and should be tried prior to switching the patient to a different antidepressant; however, the positive effects of augmentation should be balanced with any additional adverse events either agent may cause. For male sexual dysfunction, a trial of a phosphodiesterase inhibitor such as sildenafil is another alternative.

TABLE 1
SSRIs: Neurotransmitter affinity and side effects⁹

3. What drug interactions with SSRIs should I be most concerned about?

Like other psychotropic medications, SSRIs interact with drugs in a number of ways. There are interactions that occur at the cytochrome (CYP) 450 level and can result in toxicity or loss of effect, interactions that increase the likelihood of bleeding, and interactions that can lead to serotonin syndrome.

CYP 450 interactions. Depending on the CYP substrates that the SSRI and the other medication act upon, the result could be an increased concentration of the other agent and increased or decreased concentrations of the SSRI. The additive toxicity that could result has the potential to result in rare SSRI-associated adverse events, such as seizures and syndrome of inappropriate antidiuretic hormone (SIADH). Three exceptions to the increased concentration interaction are codeine, tamoxifen, and clopidogrel. Codeine, which relies on CYP metabolism to morphine, may have less analgesic effect if given with a CYP inhibitor. Tamoxifen may not be converted to endoxifen if given with a CYP inhibitor, resulting in a potentially lower therapeutic effect. Theoretically, a similar interaction could occur with clopidogrel when a CYP inhibitor is administered concurrently.

Fluoxetine, fluvoxamine, and paroxetine are the SSRIs with the greatest likelihood of having a significant CYP 450 interaction by inhibiting the metabolism of medications mediated by CYP 2D6, CYP 1A2, and CYP 2C19.^9,10 (A partial list of medications and drug classes mediated by these substrates appears in TABLE 2 .)

Risk of bleeding. Combining any SSRI with a nonsteroidal anti-inflammatory drug (NSAID) without the addition of an acid-suppressing agent would cause 1 in 250 patients to experience an upper GI bleed, a recent study found.¹¹ One in 500 patients treated with an SSRI and an antiplatelet agent developed an upper GI bleed. If a patient taking an SSRI requires antiplatelet or anticoagulant therapy, it is crucial to alert him or her to the risk and to carefully review signs and symptoms of bleeding. Hepatitis C, cirrhosis, hepatic failure, and portal hypertension are independent causes of coagulopathy, so patients with any of these conditions face an elevated risk of bleeding and would need to be monitored more closely.¹¹ Avoid prescribing fluoxetine for patients with severe liver disease; an SSRI with a shorter half-life would be a more appropriate choice.

Serotonin syndrome. Combining an SSRI with a drug that affects serotonin (venlafaxine, mirtazapine, and serotonin receptor agonists such as sumatriptan, TCAs, St. John’s wort, meperidine, and tryptophan) or a drug that exhibits monoamine oxidase inhibition properties (isocarboxazid, linezolid, phenelzine, phentermine, selegiline, and tranylcypromine) may lead to serotonin syndrome. This toxidrome is identified by autonomic instability, neuromuscular changes, and altered mental status in a patient who has ingested a substance that could elevate serotonin levels, but typically resolves within 24 hours after the serotonergic agent is discontinued.¹²

Because of the high risk of serotonin syndrome when a monoamine oxidase inhibitor (MAOI) is combined with an SSRI, do not prescribe a drug in this class until the patient has been off the SSRI for at least 2 weeks. Fluoxetine has a longer half-life, so a patient should be off of it for 5 weeks before taking an MAOI.¹³

TABLE 2
CYP 450 interactions: Beware of these drug pairings^9,10,13-17

4. What precautions are necessary when starting a patient on an SSRI or modifying therapy?

Dosing is the initial concern, with adjustments made based on specific patient factors. Elderly patients should be started on a low dose and titrated up more slowly than younger patients, for example. Low starting doses are also recommended for patients with hepatic dysfunction.^14-17

“Start low and go slow” is a good rule to follow when prescribing an SSRI to anyone whom you suspect may be intolerant to common side effects—a patient with GI symptoms associated with depression, for example.

Patient comorbidities affect choice of agent as well as dose. For a patient with a creatinine clearance <20 mL/min, citalopram and escitalopram should be used with caution.^14,15 Paroxetine should be initiated at lower doses for patients with a creatinine clearance <30 mL/min. While citalopram and escitalopram may not be ideal SSRIs for patients with renal impairment because of the potential for accumulation, they lack the substantial drug interactions and marked discontinuation syndrome seen with SSRIs such as paroxetine.

Two key concerns when changing from an SSRI to another class of antidepressant, or vice versa, are the increased risk of adverse events and a reduction in symptom control. A cross-titration strategy is appropriate for most such changes, provided the other drug is not an MAOI.

Discontinuation syndrome, which can be remembered by the mnemonic FInISH (Flu-like symptoms, Insomnia, Imbalance, Sensory disturbances, and Hyperarousal),¹⁸ is also a concern when antidepressant therapy is modified. The likelihood that a patient will develop discontinuation syndrome appears to be related to dose and agent, but not to the duration of treatment.¹⁹

While discontinuation syndrome is self-limiting, it is prudent to taper SSRI therapy whenever possible to minimize the risk of this adverse event, especially with paroxetine. A sample taper would be to reduce paroxetine by 10 mg per day every 5 to 7 days until the dose is down to 5 to 10 mg daily, then to discontinue the drug completely.²⁰ Cross-titration to a different medication will also prevent withdrawal symptoms and minimize the risk that a patient who was taking the maximum dose of an SSRI will develop serotonin syndrome.²¹ Because of the long half-life of fluoxetine and its metabolite, norfluoxetine, fluoxetine is less likely than other SSRIs to cause discontinuation syndrome. Basically, it self-tapers.

5. What should I tell pregnant patients about the risks of SSRIs?

Be upfront with them that depression in pregnancy presents a dilemma.

Tell them that on the one hand, untreated depression has been found to increase the risk of preterm labor, low birth weight, decreased fetal growth, preeclampsia, and a worsening psychiatric condition after childbirth.^22,23 In a 2006 study of 201 pregnant women with a previous diagnosis of depression, 43% relapsed during pregnancy. Those who were not taking antidepressants were 2.6 times more likely to relapse than women being treated for depression.²⁴

Patients also need to be informed that antidepressant therapy during pregnancy carries its own set of risks. Five SSRIs are pregnancy category C,²⁵ indicating either animal studies have found the drug to be harmful to fetuses and there are no well-done studies in pregnant women or that no animal studies and no human studies have tested its safety during pregnancy (the data were gathered after pregnancy). The sole exception is paroxetine, which has a D rating.²⁵ Studies have linked paroxetine to an increased risk of cardiovascular malformations in babies who were exposed to it during the first trimester.²⁶ These ratings may change shortly, however, as they are under FDA review. In May 2008, a new classification system for medication use in pregnancy was proposed.²⁷ While this system would have great clinical utility, no target date for its release has been identified.

Use of SSRIs during the second and third trimester increases the risk of neonatal pulmonary hypertension. One study found that exposed newborns were 6 times more likely to experience persistent pulmonary hypertension, compared with newborns who were not exposed to SSRIs in the second and third trimesters.²⁸

In addition, a derivative of the discontinuation syndrome is associated with neonatal withdrawal after in utero exposure, especially during the third trimester. Up to 30% of infants exposed to an SSRI may experience withdrawal symptoms, including increased or decreased muscle tone, jitteriness, feeding problems, irritability, sleep disturbance, and respiratory distress.²⁹

Under the circumstances, the best you can do is to provide the patient with as much information as possible about the benefits and risks of each strategy. In any case, pregnant women suffering from depression should receive frequent follow-up and a referral to a mental health professional. Emphasize the importance of discussing their current medications and symptoms of depression with their obstetrician and psychiatrist or psychotherapist.

6. What can I tell adolescents and their parents about SSRI safety?

Explain that 4 SSRIs—escitalopram, fluoxetine, fluvoxamine, and sertraline—are approved for use in this age group, for specific indications. Fluoxetine and escitalopram are approved for the treatment of depression in children ≥8 and ≥12 years of age, respectively. Fluvoxamine, fluoxetine, and sertraline are approved for obsessive-compulsive disorder in children ≥8, ≥7, and ≥6 years, respectively.

You can also tell patients and parents that adolescents typically fare better when they receive a combination of medication and psychotherapy, compared with medications or therapy alone.

The FDA issued an initial warning about antidepressant use in pediatric and adolescent patients in 2003, based on data from 23 randomized controlled trials submitted by 8 different drug manufacturers.³⁰ Most of the studies reported roughly twice the risk for suicidal ideation in adolescents taking SSRIs, compared with placebo. It is noteworthy, however, that there were no reports of completed suicides in the submitted trials.³⁰ In fact, data suggest that despite some increased suicidal ideation when SSRIs are initiated, these antidepressants result in improved symptom control. In 2007, after a data review, the FDA issued an advisory warning physicians about increased suicidality in young patients.³¹

The FDA has recommended increased monitoring of adolescents taking SSRIs, with office visits once a week for the first month of treatment and every 2 weeks for the second month, followed by 1 visit every 3 months. This stringent schedule has proven difficult to adhere to. One study showed that only 5% of adolescent patients received this level of attention.³² The American Academy of Child and Adolescent Psychiatry and the American Psychiatric Association advocate an individualized treatment plan instead.³³

If you prescribe SSRIs for depressed adolescents, educate patients and parents about the atypical presentation of depression that is common in patients of this age group. Advise them to watch closely for, and promptly report, increases in agitation, anxiety, impulsiveness, and restlessness, and symptoms of mania or hypomania.³³

7. When should I refer a patient to a mental health professional?

Refer patients to a mental health specialist when the optimal treatment calls for a combination of psychotherapy and medication, as is the case with depressed adolescents. Referral is recommended, too, for any complex patients. Examples include elderly individuals who are taking multiple medications or have comorbidities that can interfere with optimal treatment, and pregnant women who need additional help weighing the benefits and risks of antidepressant therapy vs nonpharmacologic treatments.

Finally, referral is critical for any patient who does not respond to treatment, even after dose adjustments, for patients who need cross-tapering that may be better handled by a specialist, and certainly for any patient who you suspect may have suicidal ideation.

CORRESPONDENCE
Patricia R. Wigle, PharmD, BCPS, University of Cincinnati, The James L. Winkle College of Pharmacy, 3225 Eden Ave., Cincinnati, OH 45267; [email protected]

Shortly after the US Food and Drug Administration (FDA) approved fluoxetine for the treatment of major depression nearly a quarter of a century ago, Prozac became a household name. And, as a handful of additional selective serotonin reuptake inhibitors (SSRIs) were approved, the popularity of this category of antidepressant continued to grow.¹ In 2008, 5 of the 6 SSRIs on the US market (fluvoxamine was the exception) were among the 200 most frequently dispensed prescription drugs.²

While much has been written about the properties of SSRIs, uncertainty about many features of particular agents and what should be considered in selecting an antidepressant for a particular patient remains. To help clear up the confusion, we’ve addressed 7 questions about SSRIs that we are often asked. Although there are no simple answers, the evidence presented, both in the answers and in the tables that follow, will help primary care physicians make more informed choices for patients who require antidepressant therapy.

1. Which is the best SSRI to start a patient on?

A recent meta-analysis of 117 head-to-head studies assessed the efficacy and acceptability of 12 newer antidepressants, including all 6 SSRIs (citalopram, escitalopram, fluoxetine, fluvoxamine, paroxetine, and sertraline).³ The researchers found 2 SSRIs, sertraline and escitalopram, to be superior to the other medications studied on both counts. A choice between sertraline or escitalopram may be a reasonable starting point in many cases, but it is impossible to recommend 1 or 2 SSRIs that are effective for all patients. There are several reasons for this.

The first is addressed by the Cipriani meta-analysis. The researchers assessed the efficacy of initial antidepressant therapy at 8 weeks, so the results cannot be extrapolated to long-term response rates or acceptability. (For a detailed review of the meta-analysis, see “Try these 2 drugs first for depression,” J Fam Pract. 2009;58:365-369.)

A second reason is the substantial publication bias associated with studies of antidepressants. Turner et al assessed 74 studies registered with the FDA, determined whether the results were positive or negative, and categorized the studies based on publication status.⁴ Their conclusions: 97% of studies with positive findings (n=38) were published.

The majority of the remaining studies, all of which were determined to have negative findings, were either not published or published in a manner that suggested a positive outcome. When these additional studies are taken into account, the percentage of published studies with a positive response drops to 51%.⁴ In addition, the effect size of each agent is reduced when all the studies are included.

There are many confounding variables associated with publication, so Turner and his colleagues were unable to definitively determine the reason for the disparity. Nonetheless, their evaluation raises questions about the reported effectiveness, not only of SSRIs, but of antidepressants in general.

A recent Cochrane review of SSRIs for treatment of depression in children and adolescents (a topic covered in greater detail in the answer to Question 6) raised similar concerns. The reviewers cited study methodology and recruitment methods as potential sources of the conflicting results they found.⁵

Finally, individual differences, such as age, comorbidities, and medication history, require an individualized approach to SSRI treatment.

2. Which side effects are common to all SSRIs, and which can be resolved by switching agents?

As a class, SSRIs are well tolerated, but they do have some common adverse effects, most notably gastrointestinal (GI) problems, sexual dysfunction, and sleep disturbances. There are also considerable differences in SSRI profiles and a few adverse effects that a switch to another SSRI may alleviate or resolve.

Compared with other SSRIs, for example, sertraline has a higher risk of diarrhea, but this adverse event does not usually lead to medication discontinuation. If it is bothersome to the patient or does not resolve with continued therapy, a change to another agent might eliminate this adverse effect.

Weight gain has been found to be more significant with paroxetine than with fluoxetine or sertraline, which may be due to the anticholinergic action of the drug.⁶ For someone who is particularly concerned about extra weight, avoiding paroxetine in the first place, or changing a patient who’s already taking it to fluoxetine or sertraline, might be an effective treatment strategy.

Overall, paroxetine has more anticholinergic side effects than other SSRIs, the likely result of its higher affinity for muscarinic receptors ( TABLE 1 ). Its potential sedating effect,⁷ which can be extremely disturbing to some patients, may be a desired feature for others. A patient with insomnia, for example, might benefit from taking paroxetine at bedtime or switching to a less activating SSRI, such as citalopram or escitalopram. Excessive activation, such as significant insomnia, may also be a warning sign of undiagnosed bipolar disorder. Thus, careful screening is needed prior to switching medications or adding a hypnotic agent. (See the box below “Treating depression in primary care: Tips from a psychiatrist”.)

A switch to another SSRI is not the only way to alleviate a troublesome side effect, of course. Insomnia can also be managed by adding a short course of trazodone to the drug regimen or by switching the patient to mirtazapine or a tricyclic antidepressant (TCA).⁸

Augmentation with either bupropion or mirtazapine may alleviate sexual side effects and should be tried prior to switching the patient to a different antidepressant; however, the positive effects of augmentation should be balanced with any additional adverse events either agent may cause. For male sexual dysfunction, a trial of a phosphodiesterase inhibitor such as sildenafil is another alternative.

TABLE 1
SSRIs: Neurotransmitter affinity and side effects⁹

3. What drug interactions with SSRIs should I be most concerned about?

Like other psychotropic medications, SSRIs interact with drugs in a number of ways. There are interactions that occur at the cytochrome (CYP) 450 level and can result in toxicity or loss of effect, interactions that increase the likelihood of bleeding, and interactions that can lead to serotonin syndrome.

CYP 450 interactions. Depending on the CYP substrates that the SSRI and the other medication act upon, the result could be an increased concentration of the other agent and increased or decreased concentrations of the SSRI. The additive toxicity that could result has the potential to result in rare SSRI-associated adverse events, such as seizures and syndrome of inappropriate antidiuretic hormone (SIADH). Three exceptions to the increased concentration interaction are codeine, tamoxifen, and clopidogrel. Codeine, which relies on CYP metabolism to morphine, may have less analgesic effect if given with a CYP inhibitor. Tamoxifen may not be converted to endoxifen if given with a CYP inhibitor, resulting in a potentially lower therapeutic effect. Theoretically, a similar interaction could occur with clopidogrel when a CYP inhibitor is administered concurrently.

Fluoxetine, fluvoxamine, and paroxetine are the SSRIs with the greatest likelihood of having a significant CYP 450 interaction by inhibiting the metabolism of medications mediated by CYP 2D6, CYP 1A2, and CYP 2C19.^9,10 (A partial list of medications and drug classes mediated by these substrates appears in TABLE 2 .)

Risk of bleeding. Combining any SSRI with a nonsteroidal anti-inflammatory drug (NSAID) without the addition of an acid-suppressing agent would cause 1 in 250 patients to experience an upper GI bleed, a recent study found.¹¹ One in 500 patients treated with an SSRI and an antiplatelet agent developed an upper GI bleed. If a patient taking an SSRI requires antiplatelet or anticoagulant therapy, it is crucial to alert him or her to the risk and to carefully review signs and symptoms of bleeding. Hepatitis C, cirrhosis, hepatic failure, and portal hypertension are independent causes of coagulopathy, so patients with any of these conditions face an elevated risk of bleeding and would need to be monitored more closely.¹¹ Avoid prescribing fluoxetine for patients with severe liver disease; an SSRI with a shorter half-life would be a more appropriate choice.

Serotonin syndrome. Combining an SSRI with a drug that affects serotonin (venlafaxine, mirtazapine, and serotonin receptor agonists such as sumatriptan, TCAs, St. John’s wort, meperidine, and tryptophan) or a drug that exhibits monoamine oxidase inhibition properties (isocarboxazid, linezolid, phenelzine, phentermine, selegiline, and tranylcypromine) may lead to serotonin syndrome. This toxidrome is identified by autonomic instability, neuromuscular changes, and altered mental status in a patient who has ingested a substance that could elevate serotonin levels, but typically resolves within 24 hours after the serotonergic agent is discontinued.¹²

Because of the high risk of serotonin syndrome when a monoamine oxidase inhibitor (MAOI) is combined with an SSRI, do not prescribe a drug in this class until the patient has been off the SSRI for at least 2 weeks. Fluoxetine has a longer half-life, so a patient should be off of it for 5 weeks before taking an MAOI.¹³

TABLE 2
CYP 450 interactions: Beware of these drug pairings^9,10,13-17

4. What precautions are necessary when starting a patient on an SSRI or modifying therapy?

Dosing is the initial concern, with adjustments made based on specific patient factors. Elderly patients should be started on a low dose and titrated up more slowly than younger patients, for example. Low starting doses are also recommended for patients with hepatic dysfunction.^14-17

“Start low and go slow” is a good rule to follow when prescribing an SSRI to anyone whom you suspect may be intolerant to common side effects—a patient with GI symptoms associated with depression, for example.

Patient comorbidities affect choice of agent as well as dose. For a patient with a creatinine clearance <20 mL/min, citalopram and escitalopram should be used with caution.^14,15 Paroxetine should be initiated at lower doses for patients with a creatinine clearance <30 mL/min. While citalopram and escitalopram may not be ideal SSRIs for patients with renal impairment because of the potential for accumulation, they lack the substantial drug interactions and marked discontinuation syndrome seen with SSRIs such as paroxetine.

Two key concerns when changing from an SSRI to another class of antidepressant, or vice versa, are the increased risk of adverse events and a reduction in symptom control. A cross-titration strategy is appropriate for most such changes, provided the other drug is not an MAOI.

Discontinuation syndrome, which can be remembered by the mnemonic FInISH (Flu-like symptoms, Insomnia, Imbalance, Sensory disturbances, and Hyperarousal),¹⁸ is also a concern when antidepressant therapy is modified. The likelihood that a patient will develop discontinuation syndrome appears to be related to dose and agent, but not to the duration of treatment.¹⁹

While discontinuation syndrome is self-limiting, it is prudent to taper SSRI therapy whenever possible to minimize the risk of this adverse event, especially with paroxetine. A sample taper would be to reduce paroxetine by 10 mg per day every 5 to 7 days until the dose is down to 5 to 10 mg daily, then to discontinue the drug completely.²⁰ Cross-titration to a different medication will also prevent withdrawal symptoms and minimize the risk that a patient who was taking the maximum dose of an SSRI will develop serotonin syndrome.²¹ Because of the long half-life of fluoxetine and its metabolite, norfluoxetine, fluoxetine is less likely than other SSRIs to cause discontinuation syndrome. Basically, it self-tapers.

5. What should I tell pregnant patients about the risks of SSRIs?

Be upfront with them that depression in pregnancy presents a dilemma.

Tell them that on the one hand, untreated depression has been found to increase the risk of preterm labor, low birth weight, decreased fetal growth, preeclampsia, and a worsening psychiatric condition after childbirth.^22,23 In a 2006 study of 201 pregnant women with a previous diagnosis of depression, 43% relapsed during pregnancy. Those who were not taking antidepressants were 2.6 times more likely to relapse than women being treated for depression.²⁴

Patients also need to be informed that antidepressant therapy during pregnancy carries its own set of risks. Five SSRIs are pregnancy category C,²⁵ indicating either animal studies have found the drug to be harmful to fetuses and there are no well-done studies in pregnant women or that no animal studies and no human studies have tested its safety during pregnancy (the data were gathered after pregnancy). The sole exception is paroxetine, which has a D rating.²⁵ Studies have linked paroxetine to an increased risk of cardiovascular malformations in babies who were exposed to it during the first trimester.²⁶ These ratings may change shortly, however, as they are under FDA review. In May 2008, a new classification system for medication use in pregnancy was proposed.²⁷ While this system would have great clinical utility, no target date for its release has been identified.

Use of SSRIs during the second and third trimester increases the risk of neonatal pulmonary hypertension. One study found that exposed newborns were 6 times more likely to experience persistent pulmonary hypertension, compared with newborns who were not exposed to SSRIs in the second and third trimesters.²⁸

In addition, a derivative of the discontinuation syndrome is associated with neonatal withdrawal after in utero exposure, especially during the third trimester. Up to 30% of infants exposed to an SSRI may experience withdrawal symptoms, including increased or decreased muscle tone, jitteriness, feeding problems, irritability, sleep disturbance, and respiratory distress.²⁹

Under the circumstances, the best you can do is to provide the patient with as much information as possible about the benefits and risks of each strategy. In any case, pregnant women suffering from depression should receive frequent follow-up and a referral to a mental health professional. Emphasize the importance of discussing their current medications and symptoms of depression with their obstetrician and psychiatrist or psychotherapist.

6. What can I tell adolescents and their parents about SSRI safety?

Explain that 4 SSRIs—escitalopram, fluoxetine, fluvoxamine, and sertraline—are approved for use in this age group, for specific indications. Fluoxetine and escitalopram are approved for the treatment of depression in children ≥8 and ≥12 years of age, respectively. Fluvoxamine, fluoxetine, and sertraline are approved for obsessive-compulsive disorder in children ≥8, ≥7, and ≥6 years, respectively.

You can also tell patients and parents that adolescents typically fare better when they receive a combination of medication and psychotherapy, compared with medications or therapy alone.

The FDA issued an initial warning about antidepressant use in pediatric and adolescent patients in 2003, based on data from 23 randomized controlled trials submitted by 8 different drug manufacturers.³⁰ Most of the studies reported roughly twice the risk for suicidal ideation in adolescents taking SSRIs, compared with placebo. It is noteworthy, however, that there were no reports of completed suicides in the submitted trials.³⁰ In fact, data suggest that despite some increased suicidal ideation when SSRIs are initiated, these antidepressants result in improved symptom control. In 2007, after a data review, the FDA issued an advisory warning physicians about increased suicidality in young patients.³¹

The FDA has recommended increased monitoring of adolescents taking SSRIs, with office visits once a week for the first month of treatment and every 2 weeks for the second month, followed by 1 visit every 3 months. This stringent schedule has proven difficult to adhere to. One study showed that only 5% of adolescent patients received this level of attention.³² The American Academy of Child and Adolescent Psychiatry and the American Psychiatric Association advocate an individualized treatment plan instead.³³

If you prescribe SSRIs for depressed adolescents, educate patients and parents about the atypical presentation of depression that is common in patients of this age group. Advise them to watch closely for, and promptly report, increases in agitation, anxiety, impulsiveness, and restlessness, and symptoms of mania or hypomania.³³

7. When should I refer a patient to a mental health professional?

Refer patients to a mental health specialist when the optimal treatment calls for a combination of psychotherapy and medication, as is the case with depressed adolescents. Referral is recommended, too, for any complex patients. Examples include elderly individuals who are taking multiple medications or have comorbidities that can interfere with optimal treatment, and pregnant women who need additional help weighing the benefits and risks of antidepressant therapy vs nonpharmacologic treatments.

Finally, referral is critical for any patient who does not respond to treatment, even after dose adjustments, for patients who need cross-tapering that may be better handled by a specialist, and certainly for any patient who you suspect may have suicidal ideation.

CORRESPONDENCE
Patricia R. Wigle, PharmD, BCPS, University of Cincinnati, The James L. Winkle College of Pharmacy, 3225 Eden Ave., Cincinnati, OH 45267; [email protected]

Shortly after the US Food and Drug Administration (FDA) approved fluoxetine for the treatment of major depression nearly a quarter of a century ago, Prozac became a household name. And, as a handful of additional selective serotonin reuptake inhibitors (SSRIs) were approved, the popularity of this category of antidepressant continued to grow.¹ In 2008, 5 of the 6 SSRIs on the US market (fluvoxamine was the exception) were among the 200 most frequently dispensed prescription drugs.²

While much has been written about the properties of SSRIs, uncertainty about many features of particular agents and what should be considered in selecting an antidepressant for a particular patient remains. To help clear up the confusion, we’ve addressed 7 questions about SSRIs that we are often asked. Although there are no simple answers, the evidence presented, both in the answers and in the tables that follow, will help primary care physicians make more informed choices for patients who require antidepressant therapy.

1. Which is the best SSRI to start a patient on?

A recent meta-analysis of 117 head-to-head studies assessed the efficacy and acceptability of 12 newer antidepressants, including all 6 SSRIs (citalopram, escitalopram, fluoxetine, fluvoxamine, paroxetine, and sertraline).³ The researchers found 2 SSRIs, sertraline and escitalopram, to be superior to the other medications studied on both counts. A choice between sertraline or escitalopram may be a reasonable starting point in many cases, but it is impossible to recommend 1 or 2 SSRIs that are effective for all patients. There are several reasons for this.

The first is addressed by the Cipriani meta-analysis. The researchers assessed the efficacy of initial antidepressant therapy at 8 weeks, so the results cannot be extrapolated to long-term response rates or acceptability. (For a detailed review of the meta-analysis, see “Try these 2 drugs first for depression,” J Fam Pract. 2009;58:365-369.)

A second reason is the substantial publication bias associated with studies of antidepressants. Turner et al assessed 74 studies registered with the FDA, determined whether the results were positive or negative, and categorized the studies based on publication status.⁴ Their conclusions: 97% of studies with positive findings (n=38) were published.

The majority of the remaining studies, all of which were determined to have negative findings, were either not published or published in a manner that suggested a positive outcome. When these additional studies are taken into account, the percentage of published studies with a positive response drops to 51%.⁴ In addition, the effect size of each agent is reduced when all the studies are included.

There are many confounding variables associated with publication, so Turner and his colleagues were unable to definitively determine the reason for the disparity. Nonetheless, their evaluation raises questions about the reported effectiveness, not only of SSRIs, but of antidepressants in general.

A recent Cochrane review of SSRIs for treatment of depression in children and adolescents (a topic covered in greater detail in the answer to Question 6) raised similar concerns. The reviewers cited study methodology and recruitment methods as potential sources of the conflicting results they found.⁵

Finally, individual differences, such as age, comorbidities, and medication history, require an individualized approach to SSRI treatment.

2. Which side effects are common to all SSRIs, and which can be resolved by switching agents?

As a class, SSRIs are well tolerated, but they do have some common adverse effects, most notably gastrointestinal (GI) problems, sexual dysfunction, and sleep disturbances. There are also considerable differences in SSRI profiles and a few adverse effects that a switch to another SSRI may alleviate or resolve.

Compared with other SSRIs, for example, sertraline has a higher risk of diarrhea, but this adverse event does not usually lead to medication discontinuation. If it is bothersome to the patient or does not resolve with continued therapy, a change to another agent might eliminate this adverse effect.

Weight gain has been found to be more significant with paroxetine than with fluoxetine or sertraline, which may be due to the anticholinergic action of the drug.⁶ For someone who is particularly concerned about extra weight, avoiding paroxetine in the first place, or changing a patient who’s already taking it to fluoxetine or sertraline, might be an effective treatment strategy.

Overall, paroxetine has more anticholinergic side effects than other SSRIs, the likely result of its higher affinity for muscarinic receptors ( TABLE 1 ). Its potential sedating effect,⁷ which can be extremely disturbing to some patients, may be a desired feature for others. A patient with insomnia, for example, might benefit from taking paroxetine at bedtime or switching to a less activating SSRI, such as citalopram or escitalopram. Excessive activation, such as significant insomnia, may also be a warning sign of undiagnosed bipolar disorder. Thus, careful screening is needed prior to switching medications or adding a hypnotic agent. (See the box below “Treating depression in primary care: Tips from a psychiatrist”.)

A switch to another SSRI is not the only way to alleviate a troublesome side effect, of course. Insomnia can also be managed by adding a short course of trazodone to the drug regimen or by switching the patient to mirtazapine or a tricyclic antidepressant (TCA).⁸

Augmentation with either bupropion or mirtazapine may alleviate sexual side effects and should be tried prior to switching the patient to a different antidepressant; however, the positive effects of augmentation should be balanced with any additional adverse events either agent may cause. For male sexual dysfunction, a trial of a phosphodiesterase inhibitor such as sildenafil is another alternative.

TABLE 1
SSRIs: Neurotransmitter affinity and side effects⁹

3. What drug interactions with SSRIs should I be most concerned about?

Like other psychotropic medications, SSRIs interact with drugs in a number of ways. There are interactions that occur at the cytochrome (CYP) 450 level and can result in toxicity or loss of effect, interactions that increase the likelihood of bleeding, and interactions that can lead to serotonin syndrome.

CYP 450 interactions. Depending on the CYP substrates that the SSRI and the other medication act upon, the result could be an increased concentration of the other agent and increased or decreased concentrations of the SSRI. The additive toxicity that could result has the potential to result in rare SSRI-associated adverse events, such as seizures and syndrome of inappropriate antidiuretic hormone (SIADH). Three exceptions to the increased concentration interaction are codeine, tamoxifen, and clopidogrel. Codeine, which relies on CYP metabolism to morphine, may have less analgesic effect if given with a CYP inhibitor. Tamoxifen may not be converted to endoxifen if given with a CYP inhibitor, resulting in a potentially lower therapeutic effect. Theoretically, a similar interaction could occur with clopidogrel when a CYP inhibitor is administered concurrently.

Fluoxetine, fluvoxamine, and paroxetine are the SSRIs with the greatest likelihood of having a significant CYP 450 interaction by inhibiting the metabolism of medications mediated by CYP 2D6, CYP 1A2, and CYP 2C19.^9,10 (A partial list of medications and drug classes mediated by these substrates appears in TABLE 2 .)

Risk of bleeding. Combining any SSRI with a nonsteroidal anti-inflammatory drug (NSAID) without the addition of an acid-suppressing agent would cause 1 in 250 patients to experience an upper GI bleed, a recent study found.¹¹ One in 500 patients treated with an SSRI and an antiplatelet agent developed an upper GI bleed. If a patient taking an SSRI requires antiplatelet or anticoagulant therapy, it is crucial to alert him or her to the risk and to carefully review signs and symptoms of bleeding. Hepatitis C, cirrhosis, hepatic failure, and portal hypertension are independent causes of coagulopathy, so patients with any of these conditions face an elevated risk of bleeding and would need to be monitored more closely.¹¹ Avoid prescribing fluoxetine for patients with severe liver disease; an SSRI with a shorter half-life would be a more appropriate choice.

Serotonin syndrome. Combining an SSRI with a drug that affects serotonin (venlafaxine, mirtazapine, and serotonin receptor agonists such as sumatriptan, TCAs, St. John’s wort, meperidine, and tryptophan) or a drug that exhibits monoamine oxidase inhibition properties (isocarboxazid, linezolid, phenelzine, phentermine, selegiline, and tranylcypromine) may lead to serotonin syndrome. This toxidrome is identified by autonomic instability, neuromuscular changes, and altered mental status in a patient who has ingested a substance that could elevate serotonin levels, but typically resolves within 24 hours after the serotonergic agent is discontinued.¹²

Because of the high risk of serotonin syndrome when a monoamine oxidase inhibitor (MAOI) is combined with an SSRI, do not prescribe a drug in this class until the patient has been off the SSRI for at least 2 weeks. Fluoxetine has a longer half-life, so a patient should be off of it for 5 weeks before taking an MAOI.¹³

TABLE 2
CYP 450 interactions: Beware of these drug pairings^9,10,13-17

4. What precautions are necessary when starting a patient on an SSRI or modifying therapy?

Dosing is the initial concern, with adjustments made based on specific patient factors. Elderly patients should be started on a low dose and titrated up more slowly than younger patients, for example. Low starting doses are also recommended for patients with hepatic dysfunction.^14-17

“Start low and go slow” is a good rule to follow when prescribing an SSRI to anyone whom you suspect may be intolerant to common side effects—a patient with GI symptoms associated with depression, for example.

Patient comorbidities affect choice of agent as well as dose. For a patient with a creatinine clearance <20 mL/min, citalopram and escitalopram should be used with caution.^14,15 Paroxetine should be initiated at lower doses for patients with a creatinine clearance <30 mL/min. While citalopram and escitalopram may not be ideal SSRIs for patients with renal impairment because of the potential for accumulation, they lack the substantial drug interactions and marked discontinuation syndrome seen with SSRIs such as paroxetine.

Two key concerns when changing from an SSRI to another class of antidepressant, or vice versa, are the increased risk of adverse events and a reduction in symptom control. A cross-titration strategy is appropriate for most such changes, provided the other drug is not an MAOI.

Discontinuation syndrome, which can be remembered by the mnemonic FInISH (Flu-like symptoms, Insomnia, Imbalance, Sensory disturbances, and Hyperarousal),¹⁸ is also a concern when antidepressant therapy is modified. The likelihood that a patient will develop discontinuation syndrome appears to be related to dose and agent, but not to the duration of treatment.¹⁹

While discontinuation syndrome is self-limiting, it is prudent to taper SSRI therapy whenever possible to minimize the risk of this adverse event, especially with paroxetine. A sample taper would be to reduce paroxetine by 10 mg per day every 5 to 7 days until the dose is down to 5 to 10 mg daily, then to discontinue the drug completely.²⁰ Cross-titration to a different medication will also prevent withdrawal symptoms and minimize the risk that a patient who was taking the maximum dose of an SSRI will develop serotonin syndrome.²¹ Because of the long half-life of fluoxetine and its metabolite, norfluoxetine, fluoxetine is less likely than other SSRIs to cause discontinuation syndrome. Basically, it self-tapers.

5. What should I tell pregnant patients about the risks of SSRIs?

Be upfront with them that depression in pregnancy presents a dilemma.

Tell them that on the one hand, untreated depression has been found to increase the risk of preterm labor, low birth weight, decreased fetal growth, preeclampsia, and a worsening psychiatric condition after childbirth.^22,23 In a 2006 study of 201 pregnant women with a previous diagnosis of depression, 43% relapsed during pregnancy. Those who were not taking antidepressants were 2.6 times more likely to relapse than women being treated for depression.²⁴

Patients also need to be informed that antidepressant therapy during pregnancy carries its own set of risks. Five SSRIs are pregnancy category C,²⁵ indicating either animal studies have found the drug to be harmful to fetuses and there are no well-done studies in pregnant women or that no animal studies and no human studies have tested its safety during pregnancy (the data were gathered after pregnancy). The sole exception is paroxetine, which has a D rating.²⁵ Studies have linked paroxetine to an increased risk of cardiovascular malformations in babies who were exposed to it during the first trimester.²⁶ These ratings may change shortly, however, as they are under FDA review. In May 2008, a new classification system for medication use in pregnancy was proposed.²⁷ While this system would have great clinical utility, no target date for its release has been identified.

Use of SSRIs during the second and third trimester increases the risk of neonatal pulmonary hypertension. One study found that exposed newborns were 6 times more likely to experience persistent pulmonary hypertension, compared with newborns who were not exposed to SSRIs in the second and third trimesters.²⁸

In addition, a derivative of the discontinuation syndrome is associated with neonatal withdrawal after in utero exposure, especially during the third trimester. Up to 30% of infants exposed to an SSRI may experience withdrawal symptoms, including increased or decreased muscle tone, jitteriness, feeding problems, irritability, sleep disturbance, and respiratory distress.²⁹

Under the circumstances, the best you can do is to provide the patient with as much information as possible about the benefits and risks of each strategy. In any case, pregnant women suffering from depression should receive frequent follow-up and a referral to a mental health professional. Emphasize the importance of discussing their current medications and symptoms of depression with their obstetrician and psychiatrist or psychotherapist.

6. What can I tell adolescents and their parents about SSRI safety?

Explain that 4 SSRIs—escitalopram, fluoxetine, fluvoxamine, and sertraline—are approved for use in this age group, for specific indications. Fluoxetine and escitalopram are approved for the treatment of depression in children ≥8 and ≥12 years of age, respectively. Fluvoxamine, fluoxetine, and sertraline are approved for obsessive-compulsive disorder in children ≥8, ≥7, and ≥6 years, respectively.

You can also tell patients and parents that adolescents typically fare better when they receive a combination of medication and psychotherapy, compared with medications or therapy alone.

The FDA issued an initial warning about antidepressant use in pediatric and adolescent patients in 2003, based on data from 23 randomized controlled trials submitted by 8 different drug manufacturers.³⁰ Most of the studies reported roughly twice the risk for suicidal ideation in adolescents taking SSRIs, compared with placebo. It is noteworthy, however, that there were no reports of completed suicides in the submitted trials.³⁰ In fact, data suggest that despite some increased suicidal ideation when SSRIs are initiated, these antidepressants result in improved symptom control. In 2007, after a data review, the FDA issued an advisory warning physicians about increased suicidality in young patients.³¹

The FDA has recommended increased monitoring of adolescents taking SSRIs, with office visits once a week for the first month of treatment and every 2 weeks for the second month, followed by 1 visit every 3 months. This stringent schedule has proven difficult to adhere to. One study showed that only 5% of adolescent patients received this level of attention.³² The American Academy of Child and Adolescent Psychiatry and the American Psychiatric Association advocate an individualized treatment plan instead.³³

If you prescribe SSRIs for depressed adolescents, educate patients and parents about the atypical presentation of depression that is common in patients of this age group. Advise them to watch closely for, and promptly report, increases in agitation, anxiety, impulsiveness, and restlessness, and symptoms of mania or hypomania.³³

7. When should I refer a patient to a mental health professional?

Refer patients to a mental health specialist when the optimal treatment calls for a combination of psychotherapy and medication, as is the case with depressed adolescents. Referral is recommended, too, for any complex patients. Examples include elderly individuals who are taking multiple medications or have comorbidities that can interfere with optimal treatment, and pregnant women who need additional help weighing the benefits and risks of antidepressant therapy vs nonpharmacologic treatments.

Finally, referral is critical for any patient who does not respond to treatment, even after dose adjustments, for patients who need cross-tapering that may be better handled by a specialist, and certainly for any patient who you suspect may have suicidal ideation.

CORRESPONDENCE
Patricia R. Wigle, PharmD, BCPS, University of Cincinnati, The James L. Winkle College of Pharmacy, 3225 Eden Ave., Cincinnati, OH 45267; [email protected]

CASE: AN OBESE PATIENT WITH BURNING FOOT PAIN

Mr. F., an obese 47-year-old with hypertension and type 2 diabetes, is trying to lose weight, and comes in to discuss a new exercise program. He recently had a negative cardiac workup for chest pain, which was ultimately diagnosed as gastroesophageal reflux disease.

The patient, whose most recent glycosylated hemoglobin (HbA1c) is 7.5, reports painful burning in his feet at night. A foot exam reveals no ulcers, lesions, or calluses; 2+ dorsalis pedis pulses bilaterally; and loss of sensation to 10-g monofilament testing at 3 sites on the bottom of each foot.

Based on Mr. F.’s presentation, it seems safe from a cardiac standpoint for him to embark on an exercise program, but what about his risk for a foot ulcer? An accurate assessment would enable you to tailor your recommendations and take the appropriate steps to modify the patient’s risk now—and to minimize complications down the road.

The incidence of diabetic foot ulcers—like that of diabetes itself—has risen in recent years.¹ More than 15% of the approximately 23.5 million US adults with diabetes are expected to develop a foot ulcer at some point in their lives.² Improvements in wound care and increased use of revascularization techniques have led to a decline in the number of ulcer-related amputations.³ But for 7% of those who develop diabetic foot ulcers, amputation is still the end result.⁴

As a family physician, you can play a key role in guarding against that outcome. This review—and the risk classification, ulcer grading, and treatment mnemonic tools that are detailed in the pages that follow—will help you optimize the foot care you provide to patients with diabetes.

Simple classification system accurately predicts risk

The longer an individual lives with diabetes and the poorer the level of blood glucose control, the greater the risk of ulceration.^1,5 Other major risk factors include neuropathy, peripheral arterial disease, and a previous foot ulcer or amputation. All patients with diabetes should undergo an annual risk assessment for foot ulcers, which can be easily incorporated into their yearly physical. Although the causes of ulceration are numerous and complex, classification of risk based on simple measures has been found to accurately predict risk.⁶

Risk stratification tool. Medical history and in-office assessment of pulses, sensation, and foot deformities form the basis for a handy assessment tool (TABLE 1). In a study of 3256 patients, these simple criteria successfully predicted ulcer risk. During the nearly 2 years of the study, fewer than 1% of those categorized as low risk (n=2253) developed ulcers, while more than 29% of the high-risk group (n=477) did.⁶

To assess a patient’s risk, visually examine both feet, inspecting the skin, nails, and structure. Autonomic neuropathy may cause decreased sweating, leading to dry, breakable skin, while motor neuropathy can cause deformities such as hammer or claw toes, which also increase ulcer risk. Palpate for dorsalis pedis and posterior tibialis pulses. Although posterior tibialis pulses may be congenitally absent, the absence of a dorsalis pedis pulse is associated with a 6-fold increased risk of foot ulcer.¹

Compared with patients with diabetes who have never had a foot ulcer, patients with a history of foot ulcer have 13 times the risk.⁷ As already noted, glucose control is also a key risk factor. For every 2-point increase in HbA1c, the odds ratio for ulcer development increases by 1.6,⁵ which is likely the result of hyperglycemia’s contribution to microvascular disease and peripheral neuropathy.

TABLE 1
Diabetic foot ulcer: Risk stratification

Screening for neuropathy, ulceration’s most consistent risk factor

Forty percent of patients with diabetes develop distal peripheral neuropathy.⁸ Damage to sensory nerves often results in the burning foot pain described by Mr. F., which can significantly affect quality of life. It may also present simply as decreased sensation or vibration sense. While decreased sensation is not painful and may not be troubling to the patient, it substantially heightens the risk of foot injury.

To screen for neuropathy, place a 10-g monofilament on noncallused plantar surfaces of the distal hallux and metatarsal heads, with enough pressure to slightly bend the filament. Instructing patients to close their eyes and report any sensation is more effective than prompting them for a response.

Recent guidelines from the American Diabetes Association and American Association of Clinical Endocrinologists recommend 1 additional screening test for neuropathy,⁹ such as pinprick or vibration sense.

Patients with neuropathy or other risk factors require more frequent follow-up to check for early signs of ulceration. Take the opportunity to provide education about the importance of self-examination of the feet, among other preventive measures (See “Preventing ulcers in high-risk feet”). Be alert to evidence of Charcot neuroarthropathy (CN). This progressive and irreversible condition of bone and joint slippage, dislocation, or fracture can affect any part of the foot, although it is typically found in the midfoot.

Acute CN is a clinical emergency characterized by pain, warmth, and erythema—making it clinically indistinguishable from cellulitis, osteomyelitis, or gout. Blood tests, including a white blood cell count, sedimentary rate, and uric acid level, may be necessary for diagnosis, as well as radiographs or magnetic resonance imaging. Treatment is long-term (4-6 months) immobilization of the joint to prevent further destruction,¹⁰ and bisphosphonates for pain during the acute phase.¹¹ Untreated CN typically results in a rocker-bottom foot deformity (FIGURE) that puts patients at greater risk for plantar ulceration.

FIGURE
Rocker-bottom foot deformity

CASE: MR. F.’S RISK BECOMES A REALITY

Mr. F. has a moderate risk of foot ulceration, based on evidence of neuropathy with 6 insensate sites (TABLE 1). You emphasize the importance of foot care, including appropriate footwear, and refer him to a podiatrist. You strongly support his decision to begin an exercise program to improve his glycemic control, decrease his cardiovascular mortality risk, and possibly help him lose weight.

Non-weight-bearing activity can be safely recommended to patients with diabetic neuropathy. So can daily weight-bearing activity, which actually decreases the risk of foot ulcer by maintaining leg muscle tone and plantar tissue tolerance to stress.^12,13 Recent studies suggest that increasing weight-bearing activity slowly–walking daily and adding a total of 100 additional steps every 2 weeks, for example—in a carefully monitored program is not associated with increased risk of foot ulceration.¹⁴ You report these findings to Mr. F.

When he returns to your office in 2 months, Mr. F. has lost 7 pounds and his HbA1c has fallen to 6.5. Despite these positive developments, an examination of his feet reveals a full thickness ulcer on the left metatarsal head, which the patient had not noticed.

You recognize the clinical urgency of effectively treating Mr. F.’s diabetic foot ulcer, as size, duration, and grade are the greatest predictors of healing. Ulcers that are larger than 2 cm, have been present for more than 6 months,¹⁵ or have a higher grade on a scale such as the Wagner Foot Classification System¹⁶ (TABLE 2) are far less likely to heal than smaller, low-grade ulcers of shorter duration.

AIM DOC mnemonic guides ulcer care

Comprehensive, coordinated care improves outcomes for diabetic ulcers and has repeatedly been shown to reduce amputation rates.^17-19 Large clinical trials have not evaluated each aspect of diabetic ulcer care, however, so the recommendations that follow are based on expert opinion and available evidence. These include assessing the limb’s arterial supply and ensuring that the patient undergoes revascularization, as needed; promptly treating infection; and providing optimal wound care, including debridement of callused and necrotic tissue, off-loading pressure, and applying moist wound dressings.²⁰

AIM DOC, developed by 1 of the authors (JE), is a handy treatment tool. The mnemonic represents both the elements of treatment and the order in which they should be carried out. The letters stand for:

• Arterial disease
• Infection
• Measure
• Debride
• Off-load
• Cover
  Here’s how to use AIM DOC, step by step:

Step 1: Assess for arterial disease

Start by assessing the vascular supply to the affected limb, which can be presumed to be adequate if pulses are palpable. If pulses cannot be palpated, the patient should undergo an ankle-brachial index (ABI) test and, if necessary, referred to a specialist to be evaluated for angioplasty or vascular bypass surgery. An ABI <0.9 is abnormal; 0.5 is considered the threshold for healing without such intervention.³ Keep in mind, however, that the ABI is falsely elevated in approximately 15% of patients with diabetes. If classic signs and symptoms of arterial disease are present, further evaluation is needed even if the ABI is normal.

Classic signs and symptoms of arterial disease include cool, hairless feet with shiny skin, and claudication. Location may provide another clue to etiology: Ulcers located on the heel, the outside of the foot, or between the toes tend to be associated with vascular disease, while ulcers with surrounding callus, such as the classic mal perforans ulcer on the metatarsal head, are neuropathic.

Step 2: Treat—or rule out—infection

While patients with diabetes have a 5-fold increase in risk of infection²¹ compared with individuals without the disease, there is no value in treating an uninfected ulcer with antibiotics. Diagnosing infection can be challenging because patients with diabetes may be less likely to demonstrate evidence of infection.²¹

An elevated white blood cell count, purulent drainage, foul odor, and/or erythema >2 cm around the wound clearly indicates a need for systemic antibiotics.^22,23 Tissue necrosis, often assumed to represent ischemia, may result from neutrophilic vasculitis from soft tissue infection.²⁴ Superficial cultures reflect colonization and should not be used to diagnose infection. In inspecting the wound for signs of infection, probe the ulcer and evaluate for osteomyelitis if it reaches bone.

Gram-positive cocci, especially Staphylococcus aureus, are the predominant pathogens in diabetic foot infections, and antibiotics effective against them may be sufficient for mild-to-moderate infections in patients who have not recently received antibiotic therapy. Gram-negative rods may be found in chronic wounds, however, and anaerobic pathogens may be present in patients with foot ischemia or gangrene; in both cases, broader-spectrum antimicrobials are required.²² Highly bioavailable oral antibiotics are indicated for infection, including some cases of osteomyelitis. Silver dressings may be helpful as a topical antimicrobial; there have been reports of successful treatment with honey, as well.^25,26

Step 3: Measure (and grade) the wound

Accurate measurement of ulcer size is critical, both when you initially detect it and at each subsequent visit. To get an accurate measure, simply multiply the greatest length by the greatest width. The percent change in wound size after 4 weeks of treatment is a significant predictor of healing.²⁵

Some clinicians also use photographs to track wound size, but these can be misleading. A better approach is to trace the wound on a sheet of acetate to document progress over time. If you see no improvement within 2 weeks, treatment should be modified.

Use a grading tool. There are a number of systems used for grading ulcers, none of which is universally accepted. One well-known tool is the Wagner Classification System¹⁶ (TABLE 2) referred to earlier. The higher the grade, the lower the likelihood that the ulcer will heal.

TABLE 2
Grading the ulcer: The Wagner system

Step 4: Debride the ulcer

Frequent sharp debridement—to remove necrotic, callused, infected, and hypergranulation tissue—has long been considered essential in the treatment of neuropathic ulcers,^17,21 especially for chronic or infected wounds. Debridement is thought to aid in healing by reducing pressure on the ulcer, decreasing bacterial contamination, enhancing platelet activation, releasing growth factors, and stimulating granulation tissue.

While some physicians are hesitant to perform debridement in the office, the process can actually be carried out without difficulty in an outpatient setting. Because of the neuropathy associated with most diabetic ulcers, no anesthesia is required. While the procedure is not typically painful, you will need to provide patient education to prepare the patient for the possibility of bleeding. Debride the wound to the outer edge of the hyperkeratotic tissue. If bleeding occurs, simply apply pressure until it stops.

When not to debride. Debridement is contraindicated under certain circumstances—if the limb has poor circulation, for example. Similarly, avoid debriding heel ulcers covered by eschar if there is no fluctuance in the underlying tissue, as the eschar provides a protective barrier.²⁴ When sharp debridement is not possible, consider topical hydrogel or maggot therapy, an adjuvant treatment we’ll discuss in a bit.

Step 5: Off-load the wound

Mechanical load relief is vital for treating neuropathic ulcers, both to redistribute plantar pressures and protect granulation tissue. Total contact casting (TCC) is the gold standard, healing 90% of ulcers within 6 to 8 weeks.²⁸ TCC is costly when applied weekly, however, and should only be done by a specialist, as an incorrectly applied cast can lead to the creation of new ulcers.

Because of the heaviness and inconvenience of the casts, many patients prefer removable devices, but these devices are much less effective. One study found that the average removable off-loading device is worn no more than 30% of the time that the patient is walking.²⁹ Removable devices can be temporarily secured with plaster of Paris (a process that is sometimes referred to as instant contact casting) to ensure compliance.

Acceptable removable devices include a heel pressure relief shoe for heel ulcers and a CAM (controlled ankle motion) walker for metatarsal ulcers. Be sure off-loading devices are applied securely so no slippage can occur.

Step 6: Cover with moist dressings

The purpose of any topical dressing is to keep wounds moist, absorb exudate, and prevent contamination. A variety of moist dressings have been successfully used to treat ulcers, although evidence to recommend any particular dressing is insufficient.¹⁹ While wound vacuum-assisted closure (VAC) devices are widely used, there is little support for their use. A review of 7 trials comparing VAC devices with moistened gauze dressings or other topical agents found no evidence that topical negative pressure increases chronic wound healing.³⁰

When foot ulcers do not heal

AIM DOC highlights the steps of diabetic ulcer care that are most likely to result in healing. When ulcers are slow to heal, review the 6 steps of treatment, paying particular attention to off-loading. If you establish that these have been appropriately applied and the wound is still not responding, consider alternative diagnoses such as venous insufficiency, vasculitis, or malignancy. Venous insufficiency ulcers may be difficult to differentiate from neuropathic ulcers, but they won’t heal without compression dressings. Diagnosis of vasculitis and malignancy can be made by biopsying the ulcer edge.

2 adjunctive therapies to consider

If there is no evidence to support an alternative diagnosis, consider adjunctive treatments, with hyperbaric oxygen therapy (HBOT) foremost among them.

A Cochrane review of HBOT found that it reduced the number of major amputations in diabetes patients with chronic foot ulcers, and improved healing at 1 year.³¹ Both Medicare and Medicaid cover HBOT for patients with diabetic ulcers classified as Wagner Grade 3 or higher that have not responded to 30 days of standard treatment.

Maggots, scientifically known as Lucilia sericata (Greenbottle) fly larvae, secrete proteolytic enzymes that debride necrotic tissue but are inactivated by living tissue. One meta-analysis found that neither surgical debridement nor larval therapy showed significant benefit over hydrogel.³² A subsequent small study did show statistically improved healing in ulcers debrided with maggots, compared with surgical debridement.³³

CASE: MR. F.’S ULCER HEALS

Because Mr. F.’s ulcer is small and shallow and has been present for a short time, it has an excellent chance of healing if you follow the AIM DOC steps. You determine that he has adequate arterial supply and that the wound is uninfected (there is a strong dorsalis pedis pulse and no warmth, exudate, or erythema around the wound). Using a #15 blade, you pare away the callus surrounding the ulcer and document the length and width of the wound. You cover the ulcer with a moist dressing and instruct Mr. F to replace it twice a day, cautioning him against using alcohol or hydrogen peroxide, which could harm the healing skin. You discuss the importance of avoiding all weight bearing on the ulcerated foot, prescribe a CAM walker to wear at all times except while he’s sleeping, and schedule weekly follow-up visits to track progress.

In 2 weeks, the wound has resolved. You educate Mr. F. about his risk of ulcer recurrence and outline appropriate preventive steps. You also refer him for fitted extra-depth diabetic shoes and ongoing podiatry follow-up.

ACKNOWLEDGMENT

The authors wish to thank Gregory Mack, DPM, University of Wisconsin School of Medicine and Public Health, for his teaching and collaboration.

CORRESPONDENCE
Jennifer Eddy, MD, University of Wisconsin School of Medicine and Public Health, 617 W. Clairemont Avenue, Eau Claire, WI 54701; [email protected]

CASE: AN OBESE PATIENT WITH BURNING FOOT PAIN

Mr. F., an obese 47-year-old with hypertension and type 2 diabetes, is trying to lose weight, and comes in to discuss a new exercise program. He recently had a negative cardiac workup for chest pain, which was ultimately diagnosed as gastroesophageal reflux disease.

The patient, whose most recent glycosylated hemoglobin (HbA1c) is 7.5, reports painful burning in his feet at night. A foot exam reveals no ulcers, lesions, or calluses; 2+ dorsalis pedis pulses bilaterally; and loss of sensation to 10-g monofilament testing at 3 sites on the bottom of each foot.

Based on Mr. F.’s presentation, it seems safe from a cardiac standpoint for him to embark on an exercise program, but what about his risk for a foot ulcer? An accurate assessment would enable you to tailor your recommendations and take the appropriate steps to modify the patient’s risk now—and to minimize complications down the road.

The incidence of diabetic foot ulcers—like that of diabetes itself—has risen in recent years.¹ More than 15% of the approximately 23.5 million US adults with diabetes are expected to develop a foot ulcer at some point in their lives.² Improvements in wound care and increased use of revascularization techniques have led to a decline in the number of ulcer-related amputations.³ But for 7% of those who develop diabetic foot ulcers, amputation is still the end result.⁴

As a family physician, you can play a key role in guarding against that outcome. This review—and the risk classification, ulcer grading, and treatment mnemonic tools that are detailed in the pages that follow—will help you optimize the foot care you provide to patients with diabetes.

Simple classification system accurately predicts risk

The longer an individual lives with diabetes and the poorer the level of blood glucose control, the greater the risk of ulceration.^1,5 Other major risk factors include neuropathy, peripheral arterial disease, and a previous foot ulcer or amputation. All patients with diabetes should undergo an annual risk assessment for foot ulcers, which can be easily incorporated into their yearly physical. Although the causes of ulceration are numerous and complex, classification of risk based on simple measures has been found to accurately predict risk.⁶

Risk stratification tool. Medical history and in-office assessment of pulses, sensation, and foot deformities form the basis for a handy assessment tool (TABLE 1). In a study of 3256 patients, these simple criteria successfully predicted ulcer risk. During the nearly 2 years of the study, fewer than 1% of those categorized as low risk (n=2253) developed ulcers, while more than 29% of the high-risk group (n=477) did.⁶

To assess a patient’s risk, visually examine both feet, inspecting the skin, nails, and structure. Autonomic neuropathy may cause decreased sweating, leading to dry, breakable skin, while motor neuropathy can cause deformities such as hammer or claw toes, which also increase ulcer risk. Palpate for dorsalis pedis and posterior tibialis pulses. Although posterior tibialis pulses may be congenitally absent, the absence of a dorsalis pedis pulse is associated with a 6-fold increased risk of foot ulcer.¹

Compared with patients with diabetes who have never had a foot ulcer, patients with a history of foot ulcer have 13 times the risk.⁷ As already noted, glucose control is also a key risk factor. For every 2-point increase in HbA1c, the odds ratio for ulcer development increases by 1.6,⁵ which is likely the result of hyperglycemia’s contribution to microvascular disease and peripheral neuropathy.

TABLE 1
Diabetic foot ulcer: Risk stratification

Screening for neuropathy, ulceration’s most consistent risk factor

Forty percent of patients with diabetes develop distal peripheral neuropathy.⁸ Damage to sensory nerves often results in the burning foot pain described by Mr. F., which can significantly affect quality of life. It may also present simply as decreased sensation or vibration sense. While decreased sensation is not painful and may not be troubling to the patient, it substantially heightens the risk of foot injury.

To screen for neuropathy, place a 10-g monofilament on noncallused plantar surfaces of the distal hallux and metatarsal heads, with enough pressure to slightly bend the filament. Instructing patients to close their eyes and report any sensation is more effective than prompting them for a response.

Recent guidelines from the American Diabetes Association and American Association of Clinical Endocrinologists recommend 1 additional screening test for neuropathy,⁹ such as pinprick or vibration sense.

Patients with neuropathy or other risk factors require more frequent follow-up to check for early signs of ulceration. Take the opportunity to provide education about the importance of self-examination of the feet, among other preventive measures (See “Preventing ulcers in high-risk feet”). Be alert to evidence of Charcot neuroarthropathy (CN). This progressive and irreversible condition of bone and joint slippage, dislocation, or fracture can affect any part of the foot, although it is typically found in the midfoot.

Acute CN is a clinical emergency characterized by pain, warmth, and erythema—making it clinically indistinguishable from cellulitis, osteomyelitis, or gout. Blood tests, including a white blood cell count, sedimentary rate, and uric acid level, may be necessary for diagnosis, as well as radiographs or magnetic resonance imaging. Treatment is long-term (4-6 months) immobilization of the joint to prevent further destruction,¹⁰ and bisphosphonates for pain during the acute phase.¹¹ Untreated CN typically results in a rocker-bottom foot deformity (FIGURE) that puts patients at greater risk for plantar ulceration.

FIGURE
Rocker-bottom foot deformity

CASE: MR. F.’S RISK BECOMES A REALITY

Mr. F. has a moderate risk of foot ulceration, based on evidence of neuropathy with 6 insensate sites (TABLE 1). You emphasize the importance of foot care, including appropriate footwear, and refer him to a podiatrist. You strongly support his decision to begin an exercise program to improve his glycemic control, decrease his cardiovascular mortality risk, and possibly help him lose weight.

Non-weight-bearing activity can be safely recommended to patients with diabetic neuropathy. So can daily weight-bearing activity, which actually decreases the risk of foot ulcer by maintaining leg muscle tone and plantar tissue tolerance to stress.^12,13 Recent studies suggest that increasing weight-bearing activity slowly–walking daily and adding a total of 100 additional steps every 2 weeks, for example—in a carefully monitored program is not associated with increased risk of foot ulceration.¹⁴ You report these findings to Mr. F.

When he returns to your office in 2 months, Mr. F. has lost 7 pounds and his HbA1c has fallen to 6.5. Despite these positive developments, an examination of his feet reveals a full thickness ulcer on the left metatarsal head, which the patient had not noticed.

You recognize the clinical urgency of effectively treating Mr. F.’s diabetic foot ulcer, as size, duration, and grade are the greatest predictors of healing. Ulcers that are larger than 2 cm, have been present for more than 6 months,¹⁵ or have a higher grade on a scale such as the Wagner Foot Classification System¹⁶ (TABLE 2) are far less likely to heal than smaller, low-grade ulcers of shorter duration.

AIM DOC mnemonic guides ulcer care

Comprehensive, coordinated care improves outcomes for diabetic ulcers and has repeatedly been shown to reduce amputation rates.^17-19 Large clinical trials have not evaluated each aspect of diabetic ulcer care, however, so the recommendations that follow are based on expert opinion and available evidence. These include assessing the limb’s arterial supply and ensuring that the patient undergoes revascularization, as needed; promptly treating infection; and providing optimal wound care, including debridement of callused and necrotic tissue, off-loading pressure, and applying moist wound dressings.²⁰

AIM DOC, developed by 1 of the authors (JE), is a handy treatment tool. The mnemonic represents both the elements of treatment and the order in which they should be carried out. The letters stand for:

• Arterial disease
• Infection
• Measure
• Debride
• Off-load
• Cover
  Here’s how to use AIM DOC, step by step:

Step 1: Assess for arterial disease

Start by assessing the vascular supply to the affected limb, which can be presumed to be adequate if pulses are palpable. If pulses cannot be palpated, the patient should undergo an ankle-brachial index (ABI) test and, if necessary, referred to a specialist to be evaluated for angioplasty or vascular bypass surgery. An ABI <0.9 is abnormal; 0.5 is considered the threshold for healing without such intervention.³ Keep in mind, however, that the ABI is falsely elevated in approximately 15% of patients with diabetes. If classic signs and symptoms of arterial disease are present, further evaluation is needed even if the ABI is normal.

Classic signs and symptoms of arterial disease include cool, hairless feet with shiny skin, and claudication. Location may provide another clue to etiology: Ulcers located on the heel, the outside of the foot, or between the toes tend to be associated with vascular disease, while ulcers with surrounding callus, such as the classic mal perforans ulcer on the metatarsal head, are neuropathic.

Step 2: Treat—or rule out—infection

While patients with diabetes have a 5-fold increase in risk of infection²¹ compared with individuals without the disease, there is no value in treating an uninfected ulcer with antibiotics. Diagnosing infection can be challenging because patients with diabetes may be less likely to demonstrate evidence of infection.²¹

An elevated white blood cell count, purulent drainage, foul odor, and/or erythema >2 cm around the wound clearly indicates a need for systemic antibiotics.^22,23 Tissue necrosis, often assumed to represent ischemia, may result from neutrophilic vasculitis from soft tissue infection.²⁴ Superficial cultures reflect colonization and should not be used to diagnose infection. In inspecting the wound for signs of infection, probe the ulcer and evaluate for osteomyelitis if it reaches bone.

Gram-positive cocci, especially Staphylococcus aureus, are the predominant pathogens in diabetic foot infections, and antibiotics effective against them may be sufficient for mild-to-moderate infections in patients who have not recently received antibiotic therapy. Gram-negative rods may be found in chronic wounds, however, and anaerobic pathogens may be present in patients with foot ischemia or gangrene; in both cases, broader-spectrum antimicrobials are required.²² Highly bioavailable oral antibiotics are indicated for infection, including some cases of osteomyelitis. Silver dressings may be helpful as a topical antimicrobial; there have been reports of successful treatment with honey, as well.^25,26

Step 3: Measure (and grade) the wound

Accurate measurement of ulcer size is critical, both when you initially detect it and at each subsequent visit. To get an accurate measure, simply multiply the greatest length by the greatest width. The percent change in wound size after 4 weeks of treatment is a significant predictor of healing.²⁵

Some clinicians also use photographs to track wound size, but these can be misleading. A better approach is to trace the wound on a sheet of acetate to document progress over time. If you see no improvement within 2 weeks, treatment should be modified.

Use a grading tool. There are a number of systems used for grading ulcers, none of which is universally accepted. One well-known tool is the Wagner Classification System¹⁶ (TABLE 2) referred to earlier. The higher the grade, the lower the likelihood that the ulcer will heal.

TABLE 2
Grading the ulcer: The Wagner system

Step 4: Debride the ulcer

Frequent sharp debridement—to remove necrotic, callused, infected, and hypergranulation tissue—has long been considered essential in the treatment of neuropathic ulcers,^17,21 especially for chronic or infected wounds. Debridement is thought to aid in healing by reducing pressure on the ulcer, decreasing bacterial contamination, enhancing platelet activation, releasing growth factors, and stimulating granulation tissue.

While some physicians are hesitant to perform debridement in the office, the process can actually be carried out without difficulty in an outpatient setting. Because of the neuropathy associated with most diabetic ulcers, no anesthesia is required. While the procedure is not typically painful, you will need to provide patient education to prepare the patient for the possibility of bleeding. Debride the wound to the outer edge of the hyperkeratotic tissue. If bleeding occurs, simply apply pressure until it stops.

When not to debride. Debridement is contraindicated under certain circumstances—if the limb has poor circulation, for example. Similarly, avoid debriding heel ulcers covered by eschar if there is no fluctuance in the underlying tissue, as the eschar provides a protective barrier.²⁴ When sharp debridement is not possible, consider topical hydrogel or maggot therapy, an adjuvant treatment we’ll discuss in a bit.

Step 5: Off-load the wound

Mechanical load relief is vital for treating neuropathic ulcers, both to redistribute plantar pressures and protect granulation tissue. Total contact casting (TCC) is the gold standard, healing 90% of ulcers within 6 to 8 weeks.²⁸ TCC is costly when applied weekly, however, and should only be done by a specialist, as an incorrectly applied cast can lead to the creation of new ulcers.

Because of the heaviness and inconvenience of the casts, many patients prefer removable devices, but these devices are much less effective. One study found that the average removable off-loading device is worn no more than 30% of the time that the patient is walking.²⁹ Removable devices can be temporarily secured with plaster of Paris (a process that is sometimes referred to as instant contact casting) to ensure compliance.

Acceptable removable devices include a heel pressure relief shoe for heel ulcers and a CAM (controlled ankle motion) walker for metatarsal ulcers. Be sure off-loading devices are applied securely so no slippage can occur.

Step 6: Cover with moist dressings

The purpose of any topical dressing is to keep wounds moist, absorb exudate, and prevent contamination. A variety of moist dressings have been successfully used to treat ulcers, although evidence to recommend any particular dressing is insufficient.¹⁹ While wound vacuum-assisted closure (VAC) devices are widely used, there is little support for their use. A review of 7 trials comparing VAC devices with moistened gauze dressings or other topical agents found no evidence that topical negative pressure increases chronic wound healing.³⁰

When foot ulcers do not heal

AIM DOC highlights the steps of diabetic ulcer care that are most likely to result in healing. When ulcers are slow to heal, review the 6 steps of treatment, paying particular attention to off-loading. If you establish that these have been appropriately applied and the wound is still not responding, consider alternative diagnoses such as venous insufficiency, vasculitis, or malignancy. Venous insufficiency ulcers may be difficult to differentiate from neuropathic ulcers, but they won’t heal without compression dressings. Diagnosis of vasculitis and malignancy can be made by biopsying the ulcer edge.

2 adjunctive therapies to consider

If there is no evidence to support an alternative diagnosis, consider adjunctive treatments, with hyperbaric oxygen therapy (HBOT) foremost among them.

A Cochrane review of HBOT found that it reduced the number of major amputations in diabetes patients with chronic foot ulcers, and improved healing at 1 year.³¹ Both Medicare and Medicaid cover HBOT for patients with diabetic ulcers classified as Wagner Grade 3 or higher that have not responded to 30 days of standard treatment.

Maggots, scientifically known as Lucilia sericata (Greenbottle) fly larvae, secrete proteolytic enzymes that debride necrotic tissue but are inactivated by living tissue. One meta-analysis found that neither surgical debridement nor larval therapy showed significant benefit over hydrogel.³² A subsequent small study did show statistically improved healing in ulcers debrided with maggots, compared with surgical debridement.³³

CASE: MR. F.’S ULCER HEALS

Because Mr. F.’s ulcer is small and shallow and has been present for a short time, it has an excellent chance of healing if you follow the AIM DOC steps. You determine that he has adequate arterial supply and that the wound is uninfected (there is a strong dorsalis pedis pulse and no warmth, exudate, or erythema around the wound). Using a #15 blade, you pare away the callus surrounding the ulcer and document the length and width of the wound. You cover the ulcer with a moist dressing and instruct Mr. F to replace it twice a day, cautioning him against using alcohol or hydrogen peroxide, which could harm the healing skin. You discuss the importance of avoiding all weight bearing on the ulcerated foot, prescribe a CAM walker to wear at all times except while he’s sleeping, and schedule weekly follow-up visits to track progress.

In 2 weeks, the wound has resolved. You educate Mr. F. about his risk of ulcer recurrence and outline appropriate preventive steps. You also refer him for fitted extra-depth diabetic shoes and ongoing podiatry follow-up.

ACKNOWLEDGMENT

The authors wish to thank Gregory Mack, DPM, University of Wisconsin School of Medicine and Public Health, for his teaching and collaboration.

CORRESPONDENCE
Jennifer Eddy, MD, University of Wisconsin School of Medicine and Public Health, 617 W. Clairemont Avenue, Eau Claire, WI 54701; [email protected]

CASE: AN OBESE PATIENT WITH BURNING FOOT PAIN

Mr. F., an obese 47-year-old with hypertension and type 2 diabetes, is trying to lose weight, and comes in to discuss a new exercise program. He recently had a negative cardiac workup for chest pain, which was ultimately diagnosed as gastroesophageal reflux disease.

The patient, whose most recent glycosylated hemoglobin (HbA1c) is 7.5, reports painful burning in his feet at night. A foot exam reveals no ulcers, lesions, or calluses; 2+ dorsalis pedis pulses bilaterally; and loss of sensation to 10-g monofilament testing at 3 sites on the bottom of each foot.

Based on Mr. F.’s presentation, it seems safe from a cardiac standpoint for him to embark on an exercise program, but what about his risk for a foot ulcer? An accurate assessment would enable you to tailor your recommendations and take the appropriate steps to modify the patient’s risk now—and to minimize complications down the road.

The incidence of diabetic foot ulcers—like that of diabetes itself—has risen in recent years.¹ More than 15% of the approximately 23.5 million US adults with diabetes are expected to develop a foot ulcer at some point in their lives.² Improvements in wound care and increased use of revascularization techniques have led to a decline in the number of ulcer-related amputations.³ But for 7% of those who develop diabetic foot ulcers, amputation is still the end result.⁴

As a family physician, you can play a key role in guarding against that outcome. This review—and the risk classification, ulcer grading, and treatment mnemonic tools that are detailed in the pages that follow—will help you optimize the foot care you provide to patients with diabetes.

Simple classification system accurately predicts risk

The longer an individual lives with diabetes and the poorer the level of blood glucose control, the greater the risk of ulceration.^1,5 Other major risk factors include neuropathy, peripheral arterial disease, and a previous foot ulcer or amputation. All patients with diabetes should undergo an annual risk assessment for foot ulcers, which can be easily incorporated into their yearly physical. Although the causes of ulceration are numerous and complex, classification of risk based on simple measures has been found to accurately predict risk.⁶

Risk stratification tool. Medical history and in-office assessment of pulses, sensation, and foot deformities form the basis for a handy assessment tool (TABLE 1). In a study of 3256 patients, these simple criteria successfully predicted ulcer risk. During the nearly 2 years of the study, fewer than 1% of those categorized as low risk (n=2253) developed ulcers, while more than 29% of the high-risk group (n=477) did.⁶

To assess a patient’s risk, visually examine both feet, inspecting the skin, nails, and structure. Autonomic neuropathy may cause decreased sweating, leading to dry, breakable skin, while motor neuropathy can cause deformities such as hammer or claw toes, which also increase ulcer risk. Palpate for dorsalis pedis and posterior tibialis pulses. Although posterior tibialis pulses may be congenitally absent, the absence of a dorsalis pedis pulse is associated with a 6-fold increased risk of foot ulcer.¹

Compared with patients with diabetes who have never had a foot ulcer, patients with a history of foot ulcer have 13 times the risk.⁷ As already noted, glucose control is also a key risk factor. For every 2-point increase in HbA1c, the odds ratio for ulcer development increases by 1.6,⁵ which is likely the result of hyperglycemia’s contribution to microvascular disease and peripheral neuropathy.

TABLE 1
Diabetic foot ulcer: Risk stratification

Screening for neuropathy, ulceration’s most consistent risk factor

Forty percent of patients with diabetes develop distal peripheral neuropathy.⁸ Damage to sensory nerves often results in the burning foot pain described by Mr. F., which can significantly affect quality of life. It may also present simply as decreased sensation or vibration sense. While decreased sensation is not painful and may not be troubling to the patient, it substantially heightens the risk of foot injury.

To screen for neuropathy, place a 10-g monofilament on noncallused plantar surfaces of the distal hallux and metatarsal heads, with enough pressure to slightly bend the filament. Instructing patients to close their eyes and report any sensation is more effective than prompting them for a response.

Recent guidelines from the American Diabetes Association and American Association of Clinical Endocrinologists recommend 1 additional screening test for neuropathy,⁹ such as pinprick or vibration sense.

Patients with neuropathy or other risk factors require more frequent follow-up to check for early signs of ulceration. Take the opportunity to provide education about the importance of self-examination of the feet, among other preventive measures (See “Preventing ulcers in high-risk feet”). Be alert to evidence of Charcot neuroarthropathy (CN). This progressive and irreversible condition of bone and joint slippage, dislocation, or fracture can affect any part of the foot, although it is typically found in the midfoot.

Acute CN is a clinical emergency characterized by pain, warmth, and erythema—making it clinically indistinguishable from cellulitis, osteomyelitis, or gout. Blood tests, including a white blood cell count, sedimentary rate, and uric acid level, may be necessary for diagnosis, as well as radiographs or magnetic resonance imaging. Treatment is long-term (4-6 months) immobilization of the joint to prevent further destruction,¹⁰ and bisphosphonates for pain during the acute phase.¹¹ Untreated CN typically results in a rocker-bottom foot deformity (FIGURE) that puts patients at greater risk for plantar ulceration.

FIGURE
Rocker-bottom foot deformity

CASE: MR. F.’S RISK BECOMES A REALITY

Mr. F. has a moderate risk of foot ulceration, based on evidence of neuropathy with 6 insensate sites (TABLE 1). You emphasize the importance of foot care, including appropriate footwear, and refer him to a podiatrist. You strongly support his decision to begin an exercise program to improve his glycemic control, decrease his cardiovascular mortality risk, and possibly help him lose weight.

Non-weight-bearing activity can be safely recommended to patients with diabetic neuropathy. So can daily weight-bearing activity, which actually decreases the risk of foot ulcer by maintaining leg muscle tone and plantar tissue tolerance to stress.^12,13 Recent studies suggest that increasing weight-bearing activity slowly–walking daily and adding a total of 100 additional steps every 2 weeks, for example—in a carefully monitored program is not associated with increased risk of foot ulceration.¹⁴ You report these findings to Mr. F.

When he returns to your office in 2 months, Mr. F. has lost 7 pounds and his HbA1c has fallen to 6.5. Despite these positive developments, an examination of his feet reveals a full thickness ulcer on the left metatarsal head, which the patient had not noticed.

You recognize the clinical urgency of effectively treating Mr. F.’s diabetic foot ulcer, as size, duration, and grade are the greatest predictors of healing. Ulcers that are larger than 2 cm, have been present for more than 6 months,¹⁵ or have a higher grade on a scale such as the Wagner Foot Classification System¹⁶ (TABLE 2) are far less likely to heal than smaller, low-grade ulcers of shorter duration.

AIM DOC mnemonic guides ulcer care

Comprehensive, coordinated care improves outcomes for diabetic ulcers and has repeatedly been shown to reduce amputation rates.^17-19 Large clinical trials have not evaluated each aspect of diabetic ulcer care, however, so the recommendations that follow are based on expert opinion and available evidence. These include assessing the limb’s arterial supply and ensuring that the patient undergoes revascularization, as needed; promptly treating infection; and providing optimal wound care, including debridement of callused and necrotic tissue, off-loading pressure, and applying moist wound dressings.²⁰

AIM DOC, developed by 1 of the authors (JE), is a handy treatment tool. The mnemonic represents both the elements of treatment and the order in which they should be carried out. The letters stand for:

• Arterial disease
• Infection
• Measure
• Debride
• Off-load
• Cover
  Here’s how to use AIM DOC, step by step:

Step 1: Assess for arterial disease

Start by assessing the vascular supply to the affected limb, which can be presumed to be adequate if pulses are palpable. If pulses cannot be palpated, the patient should undergo an ankle-brachial index (ABI) test and, if necessary, referred to a specialist to be evaluated for angioplasty or vascular bypass surgery. An ABI <0.9 is abnormal; 0.5 is considered the threshold for healing without such intervention.³ Keep in mind, however, that the ABI is falsely elevated in approximately 15% of patients with diabetes. If classic signs and symptoms of arterial disease are present, further evaluation is needed even if the ABI is normal.

Classic signs and symptoms of arterial disease include cool, hairless feet with shiny skin, and claudication. Location may provide another clue to etiology: Ulcers located on the heel, the outside of the foot, or between the toes tend to be associated with vascular disease, while ulcers with surrounding callus, such as the classic mal perforans ulcer on the metatarsal head, are neuropathic.

Step 2: Treat—or rule out—infection

While patients with diabetes have a 5-fold increase in risk of infection²¹ compared with individuals without the disease, there is no value in treating an uninfected ulcer with antibiotics. Diagnosing infection can be challenging because patients with diabetes may be less likely to demonstrate evidence of infection.²¹

An elevated white blood cell count, purulent drainage, foul odor, and/or erythema >2 cm around the wound clearly indicates a need for systemic antibiotics.^22,23 Tissue necrosis, often assumed to represent ischemia, may result from neutrophilic vasculitis from soft tissue infection.²⁴ Superficial cultures reflect colonization and should not be used to diagnose infection. In inspecting the wound for signs of infection, probe the ulcer and evaluate for osteomyelitis if it reaches bone.

Gram-positive cocci, especially Staphylococcus aureus, are the predominant pathogens in diabetic foot infections, and antibiotics effective against them may be sufficient for mild-to-moderate infections in patients who have not recently received antibiotic therapy. Gram-negative rods may be found in chronic wounds, however, and anaerobic pathogens may be present in patients with foot ischemia or gangrene; in both cases, broader-spectrum antimicrobials are required.²² Highly bioavailable oral antibiotics are indicated for infection, including some cases of osteomyelitis. Silver dressings may be helpful as a topical antimicrobial; there have been reports of successful treatment with honey, as well.^25,26

Step 3: Measure (and grade) the wound

Accurate measurement of ulcer size is critical, both when you initially detect it and at each subsequent visit. To get an accurate measure, simply multiply the greatest length by the greatest width. The percent change in wound size after 4 weeks of treatment is a significant predictor of healing.²⁵

Some clinicians also use photographs to track wound size, but these can be misleading. A better approach is to trace the wound on a sheet of acetate to document progress over time. If you see no improvement within 2 weeks, treatment should be modified.

Use a grading tool. There are a number of systems used for grading ulcers, none of which is universally accepted. One well-known tool is the Wagner Classification System¹⁶ (TABLE 2) referred to earlier. The higher the grade, the lower the likelihood that the ulcer will heal.

TABLE 2
Grading the ulcer: The Wagner system

Step 4: Debride the ulcer

Frequent sharp debridement—to remove necrotic, callused, infected, and hypergranulation tissue—has long been considered essential in the treatment of neuropathic ulcers,^17,21 especially for chronic or infected wounds. Debridement is thought to aid in healing by reducing pressure on the ulcer, decreasing bacterial contamination, enhancing platelet activation, releasing growth factors, and stimulating granulation tissue.

While some physicians are hesitant to perform debridement in the office, the process can actually be carried out without difficulty in an outpatient setting. Because of the neuropathy associated with most diabetic ulcers, no anesthesia is required. While the procedure is not typically painful, you will need to provide patient education to prepare the patient for the possibility of bleeding. Debride the wound to the outer edge of the hyperkeratotic tissue. If bleeding occurs, simply apply pressure until it stops.

When not to debride. Debridement is contraindicated under certain circumstances—if the limb has poor circulation, for example. Similarly, avoid debriding heel ulcers covered by eschar if there is no fluctuance in the underlying tissue, as the eschar provides a protective barrier.²⁴ When sharp debridement is not possible, consider topical hydrogel or maggot therapy, an adjuvant treatment we’ll discuss in a bit.

Step 5: Off-load the wound

Mechanical load relief is vital for treating neuropathic ulcers, both to redistribute plantar pressures and protect granulation tissue. Total contact casting (TCC) is the gold standard, healing 90% of ulcers within 6 to 8 weeks.²⁸ TCC is costly when applied weekly, however, and should only be done by a specialist, as an incorrectly applied cast can lead to the creation of new ulcers.

Because of the heaviness and inconvenience of the casts, many patients prefer removable devices, but these devices are much less effective. One study found that the average removable off-loading device is worn no more than 30% of the time that the patient is walking.²⁹ Removable devices can be temporarily secured with plaster of Paris (a process that is sometimes referred to as instant contact casting) to ensure compliance.

Acceptable removable devices include a heel pressure relief shoe for heel ulcers and a CAM (controlled ankle motion) walker for metatarsal ulcers. Be sure off-loading devices are applied securely so no slippage can occur.

Step 6: Cover with moist dressings

The purpose of any topical dressing is to keep wounds moist, absorb exudate, and prevent contamination. A variety of moist dressings have been successfully used to treat ulcers, although evidence to recommend any particular dressing is insufficient.¹⁹ While wound vacuum-assisted closure (VAC) devices are widely used, there is little support for their use. A review of 7 trials comparing VAC devices with moistened gauze dressings or other topical agents found no evidence that topical negative pressure increases chronic wound healing.³⁰

When foot ulcers do not heal

AIM DOC highlights the steps of diabetic ulcer care that are most likely to result in healing. When ulcers are slow to heal, review the 6 steps of treatment, paying particular attention to off-loading. If you establish that these have been appropriately applied and the wound is still not responding, consider alternative diagnoses such as venous insufficiency, vasculitis, or malignancy. Venous insufficiency ulcers may be difficult to differentiate from neuropathic ulcers, but they won’t heal without compression dressings. Diagnosis of vasculitis and malignancy can be made by biopsying the ulcer edge.

2 adjunctive therapies to consider

If there is no evidence to support an alternative diagnosis, consider adjunctive treatments, with hyperbaric oxygen therapy (HBOT) foremost among them.

A Cochrane review of HBOT found that it reduced the number of major amputations in diabetes patients with chronic foot ulcers, and improved healing at 1 year.³¹ Both Medicare and Medicaid cover HBOT for patients with diabetic ulcers classified as Wagner Grade 3 or higher that have not responded to 30 days of standard treatment.

Maggots, scientifically known as Lucilia sericata (Greenbottle) fly larvae, secrete proteolytic enzymes that debride necrotic tissue but are inactivated by living tissue. One meta-analysis found that neither surgical debridement nor larval therapy showed significant benefit over hydrogel.³² A subsequent small study did show statistically improved healing in ulcers debrided with maggots, compared with surgical debridement.³³

CASE: MR. F.’S ULCER HEALS

Because Mr. F.’s ulcer is small and shallow and has been present for a short time, it has an excellent chance of healing if you follow the AIM DOC steps. You determine that he has adequate arterial supply and that the wound is uninfected (there is a strong dorsalis pedis pulse and no warmth, exudate, or erythema around the wound). Using a #15 blade, you pare away the callus surrounding the ulcer and document the length and width of the wound. You cover the ulcer with a moist dressing and instruct Mr. F to replace it twice a day, cautioning him against using alcohol or hydrogen peroxide, which could harm the healing skin. You discuss the importance of avoiding all weight bearing on the ulcerated foot, prescribe a CAM walker to wear at all times except while he’s sleeping, and schedule weekly follow-up visits to track progress.

In 2 weeks, the wound has resolved. You educate Mr. F. about his risk of ulcer recurrence and outline appropriate preventive steps. You also refer him for fitted extra-depth diabetic shoes and ongoing podiatry follow-up.

ACKNOWLEDGMENT

The authors wish to thank Gregory Mack, DPM, University of Wisconsin School of Medicine and Public Health, for his teaching and collaboration.

CORRESPONDENCE
Jennifer Eddy, MD, University of Wisconsin School of Medicine and Public Health, 617 W. Clairemont Avenue, Eau Claire, WI 54701; [email protected]

When The Journal of Family Practice asked family physicians across the country whether access to primary care was a serious problem in their communities, the answer was a unanimous Yes. Each of the physicians we interviewed is taking steps to increase access. But all agree that the problems contributing to the growing crisis—including a shrinking pool of primary care physicians (PCPs)—are too big to be solved without a national policy shift.

“Physicians owe so much by the time they finish medical school that unless reimbursement is substantially increased, deciding on primary care will be a very difficult decision to make,” says KAROL DAVIS, MD, a family physician (FP) at Heartland Primary Care in Kansas City, Kan.

STANLEY KOZAKOWSKI, MD, director of the Hunterdon Medical Center Family Medicine Residency in Flemington, NJ, echoes Davis’s concern. “We need to address the inequities between the earnings of primary care and non-primary care physicians,” he says, adding that by some accounts, what’s needed is an increase in PCP salaries in the order of 60%.

But money is neither the sole problem, nor the sole solution. An advocate of the patient-centered medical home, Kozakowski would end the “hamster wheel model” of practice—in which PCPs squeeze more and more patients into shorter and shorter time slots, but can never catch up. To enable FPs to partner with and coordinate care for their patients, he maintains, “We need to redesign our system to support an advanced primary care model.” In the model he envisions, a PCP’s panel might be closer to 1000 than 2300.

That’s in sharp contrast to the 3300-patient panel of DAVID SWITZER, MD. An FP with Page Healthcare Associates in rural Luray, Va, Switzer has spent the last 12 years in this community, a designated primary care shortage area. “As the population ages and fewer doctors go into primary care,” he says, “the problem is only going to get worse.”

Improving access, 1 practice at a time

While Switzer agrees that “as a society, we need to figure out ways to make it more attractive for physicians to enter primary care disciplines,” he also believes—as do the other FPs we spoke to—that increasing access is the responsibility of individual physicians, too.

With that in mind, his rural practice added 1 full-time and 1 half-time physician a year ago. That increased access in terms of “warm bodies,” Switzer notes, “but we were hampered by space constraints because we had no additional exam rooms.” The solution? The practice instituted staggered schedules, which resulted in opening up earlier in the morning and remaining open a little later in the day.

To further improve access, Switzer and his colleagues implemented an open-access scheduling system. “That has certainly made us more accessible than the conventional scheduling system we previously used,” he acknowledges. “But when you’re limited in terms of capacity, as we are, open access can only go so far.”

If Switzer is not counting on health care reform to fix the PCP access problem, he is hopeful that the recent affliation of his employer, Page Memorial Hospital, with a larger regional health system will make the group more attractive to recruits over the next year or 2. “We now have access to capital to expand the office space, which will help increase our capacity,” he says. Equally helpful is the hospital’s new “convenient care” clinic, adjacent to the emergency department (ED), where patients can go for after-hours care at primary care—not ED—prices.

Heartland Primary Care in Kansas City runs its own urgent care clinic, which is open until 8 pm during the week and until 2 pm on Saturdays and Sundays. Although Davis rotates with several other physicians and nurse practitioners (NPs) in the practice (she works 1 evening a week and 1 weekend a month), after-hours care outside of the hospital is not the norm in their area.

“We routinely see patients at the clinic who have their own PCP, but their doctors close their doors at the end of the day and tell them if they’re sick to go to the ED or to our urgent care clinic, “Davis says. Those who choose the clinic, according to Davis, fare considerably better, in terms of convenience as well as costs: “We’ve gotten the wait down to a point where it’s usually less than 20 minutes,” she says. “At the ED down the street, I’m told, patients can wait as long as 3 hours.”

Davis has taken additional steps to improve patient access. After having a closed panel for 2 years, she hired an NP to help her work down the backlog and see more patients. Now she’s able to accommodate 2 new patients every workday. “I love meeting new people and would like to accept more new patients,” says Davis. But her panel size—now at 2700—is approaching the point where she may have to close it again.

While some FPs are frustrated by the focus of reform efforts, Davis sees the proposed legislation as an important first step. Another step in the right direction would be to institute reimbursement for electronic patient interactions, she adds, such as those supported by the patient portal Heartland is about to launch. Another key step: Require public and private health plans alike to pay higher fees for after-hours care.

“Plumbers get paid extra for working on Sundays,” Davis points out. “Why shouldn’t we?”

KATE ROWLAND, MD, an FP at Advocate Illinois Masonic Family Medicine Residency in Chicago, says that the residency clinic where she works has taken another step toward increasing access: The group has instituted formal walk-in hours. Between 8 and 10 am, established patients with acute care needs can come in without an appointment and be guaranteed that they’ll be seen by noon.

The new approach—a modified version of same-day scheduling—has been tremendously successful, she says, and relieved a good deal of pressure. Still, “We see some 3000 patients per month, and have just gotten busier and busier.”

What else can be done?

Rowland favors taking small steps, 1 at a time: extending clinic hours (it is now open until 7 pm 2 evenings a week and from 8 am until noon on Saturdays); reallocating physician duties, as needed; and establishing classes for patients with similar issues, such as asthma and diabetes, among other innovations.

She maintains, however, that the US health system is devoted to curing disease and prolonging life rather than on keeping the population healthy. A major refocus is needed for lasting change to occur, Rowland says. But will governmental policy changes and physician culture shift in the right direction?

“Medicine is incredibly stagnant,” the 31-year-old Rowland says. “I wonder whether we’ll still have the same chaotic system when I’m ready to retire.”

When The Journal of Family Practice asked family physicians across the country whether access to primary care was a serious problem in their communities, the answer was a unanimous Yes. Each of the physicians we interviewed is taking steps to increase access. But all agree that the problems contributing to the growing crisis—including a shrinking pool of primary care physicians (PCPs)—are too big to be solved without a national policy shift.

“Physicians owe so much by the time they finish medical school that unless reimbursement is substantially increased, deciding on primary care will be a very difficult decision to make,” says KAROL DAVIS, MD, a family physician (FP) at Heartland Primary Care in Kansas City, Kan.

STANLEY KOZAKOWSKI, MD, director of the Hunterdon Medical Center Family Medicine Residency in Flemington, NJ, echoes Davis’s concern. “We need to address the inequities between the earnings of primary care and non-primary care physicians,” he says, adding that by some accounts, what’s needed is an increase in PCP salaries in the order of 60%.

But money is neither the sole problem, nor the sole solution. An advocate of the patient-centered medical home, Kozakowski would end the “hamster wheel model” of practice—in which PCPs squeeze more and more patients into shorter and shorter time slots, but can never catch up. To enable FPs to partner with and coordinate care for their patients, he maintains, “We need to redesign our system to support an advanced primary care model.” In the model he envisions, a PCP’s panel might be closer to 1000 than 2300.

That’s in sharp contrast to the 3300-patient panel of DAVID SWITZER, MD. An FP with Page Healthcare Associates in rural Luray, Va, Switzer has spent the last 12 years in this community, a designated primary care shortage area. “As the population ages and fewer doctors go into primary care,” he says, “the problem is only going to get worse.”

Improving access, 1 practice at a time

While Switzer agrees that “as a society, we need to figure out ways to make it more attractive for physicians to enter primary care disciplines,” he also believes—as do the other FPs we spoke to—that increasing access is the responsibility of individual physicians, too.

With that in mind, his rural practice added 1 full-time and 1 half-time physician a year ago. That increased access in terms of “warm bodies,” Switzer notes, “but we were hampered by space constraints because we had no additional exam rooms.” The solution? The practice instituted staggered schedules, which resulted in opening up earlier in the morning and remaining open a little later in the day.

To further improve access, Switzer and his colleagues implemented an open-access scheduling system. “That has certainly made us more accessible than the conventional scheduling system we previously used,” he acknowledges. “But when you’re limited in terms of capacity, as we are, open access can only go so far.”

If Switzer is not counting on health care reform to fix the PCP access problem, he is hopeful that the recent affliation of his employer, Page Memorial Hospital, with a larger regional health system will make the group more attractive to recruits over the next year or 2. “We now have access to capital to expand the office space, which will help increase our capacity,” he says. Equally helpful is the hospital’s new “convenient care” clinic, adjacent to the emergency department (ED), where patients can go for after-hours care at primary care—not ED—prices.

Heartland Primary Care in Kansas City runs its own urgent care clinic, which is open until 8 pm during the week and until 2 pm on Saturdays and Sundays. Although Davis rotates with several other physicians and nurse practitioners (NPs) in the practice (she works 1 evening a week and 1 weekend a month), after-hours care outside of the hospital is not the norm in their area.

“We routinely see patients at the clinic who have their own PCP, but their doctors close their doors at the end of the day and tell them if they’re sick to go to the ED or to our urgent care clinic, “Davis says. Those who choose the clinic, according to Davis, fare considerably better, in terms of convenience as well as costs: “We’ve gotten the wait down to a point where it’s usually less than 20 minutes,” she says. “At the ED down the street, I’m told, patients can wait as long as 3 hours.”

Davis has taken additional steps to improve patient access. After having a closed panel for 2 years, she hired an NP to help her work down the backlog and see more patients. Now she’s able to accommodate 2 new patients every workday. “I love meeting new people and would like to accept more new patients,” says Davis. But her panel size—now at 2700—is approaching the point where she may have to close it again.

While some FPs are frustrated by the focus of reform efforts, Davis sees the proposed legislation as an important first step. Another step in the right direction would be to institute reimbursement for electronic patient interactions, she adds, such as those supported by the patient portal Heartland is about to launch. Another key step: Require public and private health plans alike to pay higher fees for after-hours care.

“Plumbers get paid extra for working on Sundays,” Davis points out. “Why shouldn’t we?”

KATE ROWLAND, MD, an FP at Advocate Illinois Masonic Family Medicine Residency in Chicago, says that the residency clinic where she works has taken another step toward increasing access: The group has instituted formal walk-in hours. Between 8 and 10 am, established patients with acute care needs can come in without an appointment and be guaranteed that they’ll be seen by noon.

The new approach—a modified version of same-day scheduling—has been tremendously successful, she says, and relieved a good deal of pressure. Still, “We see some 3000 patients per month, and have just gotten busier and busier.”

What else can be done?

Rowland favors taking small steps, 1 at a time: extending clinic hours (it is now open until 7 pm 2 evenings a week and from 8 am until noon on Saturdays); reallocating physician duties, as needed; and establishing classes for patients with similar issues, such as asthma and diabetes, among other innovations.

She maintains, however, that the US health system is devoted to curing disease and prolonging life rather than on keeping the population healthy. A major refocus is needed for lasting change to occur, Rowland says. But will governmental policy changes and physician culture shift in the right direction?

“Medicine is incredibly stagnant,” the 31-year-old Rowland says. “I wonder whether we’ll still have the same chaotic system when I’m ready to retire.”

When The Journal of Family Practice asked family physicians across the country whether access to primary care was a serious problem in their communities, the answer was a unanimous Yes. Each of the physicians we interviewed is taking steps to increase access. But all agree that the problems contributing to the growing crisis—including a shrinking pool of primary care physicians (PCPs)—are too big to be solved without a national policy shift.

“Physicians owe so much by the time they finish medical school that unless reimbursement is substantially increased, deciding on primary care will be a very difficult decision to make,” says KAROL DAVIS, MD, a family physician (FP) at Heartland Primary Care in Kansas City, Kan.

STANLEY KOZAKOWSKI, MD, director of the Hunterdon Medical Center Family Medicine Residency in Flemington, NJ, echoes Davis’s concern. “We need to address the inequities between the earnings of primary care and non-primary care physicians,” he says, adding that by some accounts, what’s needed is an increase in PCP salaries in the order of 60%.

But money is neither the sole problem, nor the sole solution. An advocate of the patient-centered medical home, Kozakowski would end the “hamster wheel model” of practice—in which PCPs squeeze more and more patients into shorter and shorter time slots, but can never catch up. To enable FPs to partner with and coordinate care for their patients, he maintains, “We need to redesign our system to support an advanced primary care model.” In the model he envisions, a PCP’s panel might be closer to 1000 than 2300.

That’s in sharp contrast to the 3300-patient panel of DAVID SWITZER, MD. An FP with Page Healthcare Associates in rural Luray, Va, Switzer has spent the last 12 years in this community, a designated primary care shortage area. “As the population ages and fewer doctors go into primary care,” he says, “the problem is only going to get worse.”

Improving access, 1 practice at a time

While Switzer agrees that “as a society, we need to figure out ways to make it more attractive for physicians to enter primary care disciplines,” he also believes—as do the other FPs we spoke to—that increasing access is the responsibility of individual physicians, too.

With that in mind, his rural practice added 1 full-time and 1 half-time physician a year ago. That increased access in terms of “warm bodies,” Switzer notes, “but we were hampered by space constraints because we had no additional exam rooms.” The solution? The practice instituted staggered schedules, which resulted in opening up earlier in the morning and remaining open a little later in the day.

To further improve access, Switzer and his colleagues implemented an open-access scheduling system. “That has certainly made us more accessible than the conventional scheduling system we previously used,” he acknowledges. “But when you’re limited in terms of capacity, as we are, open access can only go so far.”

If Switzer is not counting on health care reform to fix the PCP access problem, he is hopeful that the recent affliation of his employer, Page Memorial Hospital, with a larger regional health system will make the group more attractive to recruits over the next year or 2. “We now have access to capital to expand the office space, which will help increase our capacity,” he says. Equally helpful is the hospital’s new “convenient care” clinic, adjacent to the emergency department (ED), where patients can go for after-hours care at primary care—not ED—prices.

Heartland Primary Care in Kansas City runs its own urgent care clinic, which is open until 8 pm during the week and until 2 pm on Saturdays and Sundays. Although Davis rotates with several other physicians and nurse practitioners (NPs) in the practice (she works 1 evening a week and 1 weekend a month), after-hours care outside of the hospital is not the norm in their area.

“We routinely see patients at the clinic who have their own PCP, but their doctors close their doors at the end of the day and tell them if they’re sick to go to the ED or to our urgent care clinic, “Davis says. Those who choose the clinic, according to Davis, fare considerably better, in terms of convenience as well as costs: “We’ve gotten the wait down to a point where it’s usually less than 20 minutes,” she says. “At the ED down the street, I’m told, patients can wait as long as 3 hours.”

Davis has taken additional steps to improve patient access. After having a closed panel for 2 years, she hired an NP to help her work down the backlog and see more patients. Now she’s able to accommodate 2 new patients every workday. “I love meeting new people and would like to accept more new patients,” says Davis. But her panel size—now at 2700—is approaching the point where she may have to close it again.

While some FPs are frustrated by the focus of reform efforts, Davis sees the proposed legislation as an important first step. Another step in the right direction would be to institute reimbursement for electronic patient interactions, she adds, such as those supported by the patient portal Heartland is about to launch. Another key step: Require public and private health plans alike to pay higher fees for after-hours care.

“Plumbers get paid extra for working on Sundays,” Davis points out. “Why shouldn’t we?”

KATE ROWLAND, MD, an FP at Advocate Illinois Masonic Family Medicine Residency in Chicago, says that the residency clinic where she works has taken another step toward increasing access: The group has instituted formal walk-in hours. Between 8 and 10 am, established patients with acute care needs can come in without an appointment and be guaranteed that they’ll be seen by noon.

The new approach—a modified version of same-day scheduling—has been tremendously successful, she says, and relieved a good deal of pressure. Still, “We see some 3000 patients per month, and have just gotten busier and busier.”

What else can be done?

Rowland favors taking small steps, 1 at a time: extending clinic hours (it is now open until 7 pm 2 evenings a week and from 8 am until noon on Saturdays); reallocating physician duties, as needed; and establishing classes for patients with similar issues, such as asthma and diabetes, among other innovations.

She maintains, however, that the US health system is devoted to curing disease and prolonging life rather than on keeping the population healthy. A major refocus is needed for lasting change to occur, Rowland says. But will governmental policy changes and physician culture shift in the right direction?

“Medicine is incredibly stagnant,” the 31-year-old Rowland says. “I wonder whether we’ll still have the same chaotic system when I’m ready to retire.”

As the nation moves closer to a health reform bill designed to extend coverage to the vast majority of the uninsured, a key concern is being overlooked: Simply having health insurance—crucial as that is—is not enough. People also need access to care, particularly to primary care. Yet a growing body of evidence suggests that even among insured Americans, access to primary care is on the decline.

Consider the following:

• In 2005, more than 56 million Americans (nearly half of whom had some form of coverage) were “medically disenfranchised”—lacking sufficient access to primary care services. Two years later, that number had grown to 60 million.¹
• In a 2006 national survey, little more than 1 in 4 adults (ages 18 to 64) said they could easily reach their doctors by phone, get after-hours care, or schedule timely office visits.²
• The number of medical school graduates in the United States choosing careers in family medicine fell from 2340 in 1997 to 1132 in 2006; during that same period, the percentage of internal medicine residents entering primary care dropped from 54% to 20%.³
• In 2008, nearly 30% of Medicare beneficiaries seeking a new primary care physician (PCP) reported difficulty finding a doctor—a 17% increase since 2006.⁴

A shrinking pool of primary care physicians

Compared with other Western nations, the United States has a smaller proportion of its physician workforce engaged in primary care.⁵ Shortages of PCPs already exist in numerous states, with Alabama, Alaska, Florida, Kansas, Mississippi, Missouri, Oregon, South Carolina, and Utah among them.¹ In the decade ahead, the Council on Graduate Medical Education, among other professional groups, expects the shortages to become more widespread and more severe.

Recruiting PCPs is increasingly difficult—something that comes as no surprise to physician recruiters. Merritt Hawkins, a large physician recruitment firm, reports that in 2005 the number of searches for open positions for PCPs exceeded searches for specialists for the first time.⁶ And in 2006, nearly half of all primary care residents were contacted by recruiters more than 50 times. In a survey of physician groups that same year, 94% of respondents ranked either internists or family physicians as the most difficult to recruit.⁷

Nurse practitioners (NPs) and physician assistants (PAs) have been important contributors to the primary care workforce, but they, too, may soon be in shorter supply. That’s especially true of PAs, given that less than one third (31%) of them are choosing careers in adult primary care.⁸

Quality of care pays the price. Ironically, the shrinking pool of PCPs coincides with the growing recognition of the importance of the patient-centered medical home. There is increasing evidence, too, of the link between lack of access to primary care and higher mortality rates⁹ and poorer outcomes.¹⁰ Lack of access appears to have an adverse effect on health care spending, as well. While the administration searches for ways to lower the cost of care in order to pay for the expansion in coverage, directors of health plans and medical groups expect medical spending to rise as the looming PCP shortage leads to greater use of specialists and more emergency visits.¹¹

A broader look at inadequate access

The PCP shortage alone, however, is not the whole story. A number of other potent factors related to, but not the direct result of, the shortage contribute to the growing inability of insured Americans to have timely access to primary care. Chief among them are a mismatch between demand for appointments and physicians’ capacity to provide them, limited after-hours care, and organizational problems in primary care practices. We’ve identified the following barriers—and the policy changes and shifts in physician culture that are needed to overcome them.

BARRIER #1: Panel size

The average full-time primary care practitioner has an estimated panel of 2300 patients¹²—too many for a single physician to provide adequate patient care for, according to a recent study.^13,14 Some practices have excessively large panels because they’re located in areas with a shortage of primary care providers. (In an area with 25 PCPs per 100,000 population, for instance, the average panel size would be 4000.) Other practices accept too many patients in order to stay afloat financially. In either case, a situation in which the demand for appointments exceeds the available time slots impedes a patient’s ability to get timely care.

BARRIER #2: Capacity

The number of hours per week that a PCP sees patients and the number of patients scheduled per hour determine that physician’s visit capacity. Quality of care is also at stake. Although physicians who schedule 1 patient every 10 or 15 minutes can, of course, accommodate more patients than doctors who spend 20 or 30 minutes per visit, shorter appointments have been found to adversely affect quality.¹⁵

Further complicating things is the increasing number of physicians who are opting for part-time work.¹⁶ Added to the hospital and nursing home responsibilities many PCPs share, working fewer hours imposes further limits on the number of patients they can care for.

BARRIER #3: Distance

The uneven geographic distribution of PCPs makes access difficult for patients living at a distance from the nearest primary care practice, a particular problem for the homebound and those without transportation. Telemedicine could help mitigate this problem, but few primary care practices are equipped to practice “distance” medicine.

BARRIER #4: Medicaid/Medicare issues

Some primary care practices make decisions about which new patients to accept based on the kind of coverage held by the prospective patients. Medicaid patients have an especially difficult time finding a PCP—far harder than privately insured individuals. Also, in areas in which Medicare fees are below the market rates paid by private insurers, many practices limit the number of Medicare patients they accept.

The bottom line: Already stressed by the economy and low fees, some PCPs say they have little choice but to restrict the number of patients whose care costs them more than they’re paid to provide it.

BARRIER #5: After-hours care

Many patients try, unsuccessfully, to reach their PCP in the evening or on the weekend. The dearth of after-hours access has led to an explosion of “convenience clinics” in pharmacies and shopping malls—and to overuse of the emergency department (ED). In a 2007 national survey, 67% of adults said they had difficulty getting care at night or on weekends unless they went to the ED.¹⁷ In another survey, conducted in California in 2006, nearly half of those who sought care in the ED believed their condition could have been handled in a primary care setting, had it been available.¹⁸

BARRIER #6: Scheduling

Many patients call their PCP’s office for an appointment, only to find that the next available opening is 3 weeks away. While some groups have introduced open-access scheduling—also called same-day scheduling or advanced access—such a system can only be sustained if the demand for appointments is in balance with the practice’s capacity to see patients.

Part of the problem appears to be organizational. Some physicians routinely make monthly follow-up appointments for patients with chronic conditions, such as hypertension, diabetes, or arthritis. However, the return visit interval is often based on the habits of the individual physician or provider group, rather than on the medical needs of the patients. Indeed, 1 study found that prolonging the visit interval resulted in an improvement rather than a decline in quality of care.¹⁹

BARRIER #7: Virtual visits

Many chronic care and preventive care issues could be handled in brief patient encounters via telephone or e-mail. In addition to being convenient for many patients, such virtual visits would increase the practice’s capacity for patients who require in-person visits.²⁰ Here, too, the problem is financial: Insurers generally do not provide reimbursement for virtual visits.

BARRIER #8: Troubles with team care

At some medical groups, nonphysician providers, including registered nurses and pharmacists, use doctor-created protocols and standing orders to address routine chronic care issues and preventive measures for individuals with certain conditions—identified via patient registries. Similarly, medical assistants and community health workers may be trained as health coaches to work with patients on behavior change and adherence to medication regimens, thus freeing up physician time.²¹ Despite the benefits of team care, most insurers only reimburse the services of MDs, NPs and PAs, meaning that no incentives exist for primary care practices to hire other team members.

The solutions: Policy shifts and culture change

What will it take to improve access to primary care and tear down these barriers? First and foremost, we believe the following policy changes are needed:

• Increase reimbursement for primary care.
• Increase loan repayment programs for medical students who establish primary care practices in areas with established shortages.
• Standardize fees paid by private insurers, Medicare, and Medicaid plans.
• Provide financial incentives for PCPs to deliver after-hours care.
• Invest in a national program aimed at helping primary care practices implement same-day scheduling, team care, and other access improvements.²²
• Provide reimbursement for e-mail and telephone encounters and team care, including fees for all allied health professionals who assist PCPs in managing chronic disease and preventive care.

These reforms, if they were to truly come to pass, would ease much of the pressure on PCPs. No matter what policy changes are implemented to increase access to primary care, however, it is clear that a substantial culture change is required on the part of PCPs, as well. Physicians can begin to make changes on their own to increase patient access—expanding the interval between follow-up visits for stable patients, for instance, and reorganizing work schedules so that the practice can remain open for more hours.

It’s clear that providing health insurance to the uninsured without guaranteeing access to primary care can turn a potentially positive development into widespread patient frustration. Unless Americans have greater access to primary care, we fear, the US health care system will undergo significant change without substantial improvement.

CORRESPONDENCE
Thomas Bodenheimer MD, MPH, Department of Family and Community Medicine, University of California at San Francisco, Bldg 80-83, SF General Hospital, 995 Potrero Avenue, San Francisco, CA 94110; [email protected]

As the nation moves closer to a health reform bill designed to extend coverage to the vast majority of the uninsured, a key concern is being overlooked: Simply having health insurance—crucial as that is—is not enough. People also need access to care, particularly to primary care. Yet a growing body of evidence suggests that even among insured Americans, access to primary care is on the decline.

Consider the following:

• In 2005, more than 56 million Americans (nearly half of whom had some form of coverage) were “medically disenfranchised”—lacking sufficient access to primary care services. Two years later, that number had grown to 60 million.¹
• In a 2006 national survey, little more than 1 in 4 adults (ages 18 to 64) said they could easily reach their doctors by phone, get after-hours care, or schedule timely office visits.²
• The number of medical school graduates in the United States choosing careers in family medicine fell from 2340 in 1997 to 1132 in 2006; during that same period, the percentage of internal medicine residents entering primary care dropped from 54% to 20%.³
• In 2008, nearly 30% of Medicare beneficiaries seeking a new primary care physician (PCP) reported difficulty finding a doctor—a 17% increase since 2006.⁴

A shrinking pool of primary care physicians

Compared with other Western nations, the United States has a smaller proportion of its physician workforce engaged in primary care.⁵ Shortages of PCPs already exist in numerous states, with Alabama, Alaska, Florida, Kansas, Mississippi, Missouri, Oregon, South Carolina, and Utah among them.¹ In the decade ahead, the Council on Graduate Medical Education, among other professional groups, expects the shortages to become more widespread and more severe.

Recruiting PCPs is increasingly difficult—something that comes as no surprise to physician recruiters. Merritt Hawkins, a large physician recruitment firm, reports that in 2005 the number of searches for open positions for PCPs exceeded searches for specialists for the first time.⁶ And in 2006, nearly half of all primary care residents were contacted by recruiters more than 50 times. In a survey of physician groups that same year, 94% of respondents ranked either internists or family physicians as the most difficult to recruit.⁷

Nurse practitioners (NPs) and physician assistants (PAs) have been important contributors to the primary care workforce, but they, too, may soon be in shorter supply. That’s especially true of PAs, given that less than one third (31%) of them are choosing careers in adult primary care.⁸

Quality of care pays the price. Ironically, the shrinking pool of PCPs coincides with the growing recognition of the importance of the patient-centered medical home. There is increasing evidence, too, of the link between lack of access to primary care and higher mortality rates⁹ and poorer outcomes.¹⁰ Lack of access appears to have an adverse effect on health care spending, as well. While the administration searches for ways to lower the cost of care in order to pay for the expansion in coverage, directors of health plans and medical groups expect medical spending to rise as the looming PCP shortage leads to greater use of specialists and more emergency visits.¹¹

A broader look at inadequate access

The PCP shortage alone, however, is not the whole story. A number of other potent factors related to, but not the direct result of, the shortage contribute to the growing inability of insured Americans to have timely access to primary care. Chief among them are a mismatch between demand for appointments and physicians’ capacity to provide them, limited after-hours care, and organizational problems in primary care practices. We’ve identified the following barriers—and the policy changes and shifts in physician culture that are needed to overcome them.

BARRIER #1: Panel size

The average full-time primary care practitioner has an estimated panel of 2300 patients¹²—too many for a single physician to provide adequate patient care for, according to a recent study.^13,14 Some practices have excessively large panels because they’re located in areas with a shortage of primary care providers. (In an area with 25 PCPs per 100,000 population, for instance, the average panel size would be 4000.) Other practices accept too many patients in order to stay afloat financially. In either case, a situation in which the demand for appointments exceeds the available time slots impedes a patient’s ability to get timely care.

BARRIER #2: Capacity

The number of hours per week that a PCP sees patients and the number of patients scheduled per hour determine that physician’s visit capacity. Quality of care is also at stake. Although physicians who schedule 1 patient every 10 or 15 minutes can, of course, accommodate more patients than doctors who spend 20 or 30 minutes per visit, shorter appointments have been found to adversely affect quality.¹⁵

Further complicating things is the increasing number of physicians who are opting for part-time work.¹⁶ Added to the hospital and nursing home responsibilities many PCPs share, working fewer hours imposes further limits on the number of patients they can care for.

BARRIER #3: Distance

The uneven geographic distribution of PCPs makes access difficult for patients living at a distance from the nearest primary care practice, a particular problem for the homebound and those without transportation. Telemedicine could help mitigate this problem, but few primary care practices are equipped to practice “distance” medicine.

BARRIER #4: Medicaid/Medicare issues

Some primary care practices make decisions about which new patients to accept based on the kind of coverage held by the prospective patients. Medicaid patients have an especially difficult time finding a PCP—far harder than privately insured individuals. Also, in areas in which Medicare fees are below the market rates paid by private insurers, many practices limit the number of Medicare patients they accept.

The bottom line: Already stressed by the economy and low fees, some PCPs say they have little choice but to restrict the number of patients whose care costs them more than they’re paid to provide it.

BARRIER #5: After-hours care

Many patients try, unsuccessfully, to reach their PCP in the evening or on the weekend. The dearth of after-hours access has led to an explosion of “convenience clinics” in pharmacies and shopping malls—and to overuse of the emergency department (ED). In a 2007 national survey, 67% of adults said they had difficulty getting care at night or on weekends unless they went to the ED.¹⁷ In another survey, conducted in California in 2006, nearly half of those who sought care in the ED believed their condition could have been handled in a primary care setting, had it been available.¹⁸

BARRIER #6: Scheduling

Many patients call their PCP’s office for an appointment, only to find that the next available opening is 3 weeks away. While some groups have introduced open-access scheduling—also called same-day scheduling or advanced access—such a system can only be sustained if the demand for appointments is in balance with the practice’s capacity to see patients.

Part of the problem appears to be organizational. Some physicians routinely make monthly follow-up appointments for patients with chronic conditions, such as hypertension, diabetes, or arthritis. However, the return visit interval is often based on the habits of the individual physician or provider group, rather than on the medical needs of the patients. Indeed, 1 study found that prolonging the visit interval resulted in an improvement rather than a decline in quality of care.¹⁹

BARRIER #7: Virtual visits

Many chronic care and preventive care issues could be handled in brief patient encounters via telephone or e-mail. In addition to being convenient for many patients, such virtual visits would increase the practice’s capacity for patients who require in-person visits.²⁰ Here, too, the problem is financial: Insurers generally do not provide reimbursement for virtual visits.

BARRIER #8: Troubles with team care

At some medical groups, nonphysician providers, including registered nurses and pharmacists, use doctor-created protocols and standing orders to address routine chronic care issues and preventive measures for individuals with certain conditions—identified via patient registries. Similarly, medical assistants and community health workers may be trained as health coaches to work with patients on behavior change and adherence to medication regimens, thus freeing up physician time.²¹ Despite the benefits of team care, most insurers only reimburse the services of MDs, NPs and PAs, meaning that no incentives exist for primary care practices to hire other team members.

The solutions: Policy shifts and culture change

What will it take to improve access to primary care and tear down these barriers? First and foremost, we believe the following policy changes are needed:

• Increase reimbursement for primary care.
• Increase loan repayment programs for medical students who establish primary care practices in areas with established shortages.
• Standardize fees paid by private insurers, Medicare, and Medicaid plans.
• Provide financial incentives for PCPs to deliver after-hours care.
• Invest in a national program aimed at helping primary care practices implement same-day scheduling, team care, and other access improvements.²²
• Provide reimbursement for e-mail and telephone encounters and team care, including fees for all allied health professionals who assist PCPs in managing chronic disease and preventive care.

These reforms, if they were to truly come to pass, would ease much of the pressure on PCPs. No matter what policy changes are implemented to increase access to primary care, however, it is clear that a substantial culture change is required on the part of PCPs, as well. Physicians can begin to make changes on their own to increase patient access—expanding the interval between follow-up visits for stable patients, for instance, and reorganizing work schedules so that the practice can remain open for more hours.

It’s clear that providing health insurance to the uninsured without guaranteeing access to primary care can turn a potentially positive development into widespread patient frustration. Unless Americans have greater access to primary care, we fear, the US health care system will undergo significant change without substantial improvement.

CORRESPONDENCE
Thomas Bodenheimer MD, MPH, Department of Family and Community Medicine, University of California at San Francisco, Bldg 80-83, SF General Hospital, 995 Potrero Avenue, San Francisco, CA 94110; [email protected]

As the nation moves closer to a health reform bill designed to extend coverage to the vast majority of the uninsured, a key concern is being overlooked: Simply having health insurance—crucial as that is—is not enough. People also need access to care, particularly to primary care. Yet a growing body of evidence suggests that even among insured Americans, access to primary care is on the decline.

Consider the following:

• In 2005, more than 56 million Americans (nearly half of whom had some form of coverage) were “medically disenfranchised”—lacking sufficient access to primary care services. Two years later, that number had grown to 60 million.¹
• In a 2006 national survey, little more than 1 in 4 adults (ages 18 to 64) said they could easily reach their doctors by phone, get after-hours care, or schedule timely office visits.²
• The number of medical school graduates in the United States choosing careers in family medicine fell from 2340 in 1997 to 1132 in 2006; during that same period, the percentage of internal medicine residents entering primary care dropped from 54% to 20%.³
• In 2008, nearly 30% of Medicare beneficiaries seeking a new primary care physician (PCP) reported difficulty finding a doctor—a 17% increase since 2006.⁴

A shrinking pool of primary care physicians

Compared with other Western nations, the United States has a smaller proportion of its physician workforce engaged in primary care.⁵ Shortages of PCPs already exist in numerous states, with Alabama, Alaska, Florida, Kansas, Mississippi, Missouri, Oregon, South Carolina, and Utah among them.¹ In the decade ahead, the Council on Graduate Medical Education, among other professional groups, expects the shortages to become more widespread and more severe.

Recruiting PCPs is increasingly difficult—something that comes as no surprise to physician recruiters. Merritt Hawkins, a large physician recruitment firm, reports that in 2005 the number of searches for open positions for PCPs exceeded searches for specialists for the first time.⁶ And in 2006, nearly half of all primary care residents were contacted by recruiters more than 50 times. In a survey of physician groups that same year, 94% of respondents ranked either internists or family physicians as the most difficult to recruit.⁷

Nurse practitioners (NPs) and physician assistants (PAs) have been important contributors to the primary care workforce, but they, too, may soon be in shorter supply. That’s especially true of PAs, given that less than one third (31%) of them are choosing careers in adult primary care.⁸

Quality of care pays the price. Ironically, the shrinking pool of PCPs coincides with the growing recognition of the importance of the patient-centered medical home. There is increasing evidence, too, of the link between lack of access to primary care and higher mortality rates⁹ and poorer outcomes.¹⁰ Lack of access appears to have an adverse effect on health care spending, as well. While the administration searches for ways to lower the cost of care in order to pay for the expansion in coverage, directors of health plans and medical groups expect medical spending to rise as the looming PCP shortage leads to greater use of specialists and more emergency visits.¹¹

A broader look at inadequate access

The PCP shortage alone, however, is not the whole story. A number of other potent factors related to, but not the direct result of, the shortage contribute to the growing inability of insured Americans to have timely access to primary care. Chief among them are a mismatch between demand for appointments and physicians’ capacity to provide them, limited after-hours care, and organizational problems in primary care practices. We’ve identified the following barriers—and the policy changes and shifts in physician culture that are needed to overcome them.

BARRIER #1: Panel size

The average full-time primary care practitioner has an estimated panel of 2300 patients¹²—too many for a single physician to provide adequate patient care for, according to a recent study.^13,14 Some practices have excessively large panels because they’re located in areas with a shortage of primary care providers. (In an area with 25 PCPs per 100,000 population, for instance, the average panel size would be 4000.) Other practices accept too many patients in order to stay afloat financially. In either case, a situation in which the demand for appointments exceeds the available time slots impedes a patient’s ability to get timely care.

BARRIER #2: Capacity

The number of hours per week that a PCP sees patients and the number of patients scheduled per hour determine that physician’s visit capacity. Quality of care is also at stake. Although physicians who schedule 1 patient every 10 or 15 minutes can, of course, accommodate more patients than doctors who spend 20 or 30 minutes per visit, shorter appointments have been found to adversely affect quality.¹⁵

Further complicating things is the increasing number of physicians who are opting for part-time work.¹⁶ Added to the hospital and nursing home responsibilities many PCPs share, working fewer hours imposes further limits on the number of patients they can care for.

BARRIER #3: Distance

The uneven geographic distribution of PCPs makes access difficult for patients living at a distance from the nearest primary care practice, a particular problem for the homebound and those without transportation. Telemedicine could help mitigate this problem, but few primary care practices are equipped to practice “distance” medicine.

BARRIER #4: Medicaid/Medicare issues

Some primary care practices make decisions about which new patients to accept based on the kind of coverage held by the prospective patients. Medicaid patients have an especially difficult time finding a PCP—far harder than privately insured individuals. Also, in areas in which Medicare fees are below the market rates paid by private insurers, many practices limit the number of Medicare patients they accept.

The bottom line: Already stressed by the economy and low fees, some PCPs say they have little choice but to restrict the number of patients whose care costs them more than they’re paid to provide it.

BARRIER #5: After-hours care

Many patients try, unsuccessfully, to reach their PCP in the evening or on the weekend. The dearth of after-hours access has led to an explosion of “convenience clinics” in pharmacies and shopping malls—and to overuse of the emergency department (ED). In a 2007 national survey, 67% of adults said they had difficulty getting care at night or on weekends unless they went to the ED.¹⁷ In another survey, conducted in California in 2006, nearly half of those who sought care in the ED believed their condition could have been handled in a primary care setting, had it been available.¹⁸

BARRIER #6: Scheduling

Many patients call their PCP’s office for an appointment, only to find that the next available opening is 3 weeks away. While some groups have introduced open-access scheduling—also called same-day scheduling or advanced access—such a system can only be sustained if the demand for appointments is in balance with the practice’s capacity to see patients.

Part of the problem appears to be organizational. Some physicians routinely make monthly follow-up appointments for patients with chronic conditions, such as hypertension, diabetes, or arthritis. However, the return visit interval is often based on the habits of the individual physician or provider group, rather than on the medical needs of the patients. Indeed, 1 study found that prolonging the visit interval resulted in an improvement rather than a decline in quality of care.¹⁹

BARRIER #7: Virtual visits

Many chronic care and preventive care issues could be handled in brief patient encounters via telephone or e-mail. In addition to being convenient for many patients, such virtual visits would increase the practice’s capacity for patients who require in-person visits.²⁰ Here, too, the problem is financial: Insurers generally do not provide reimbursement for virtual visits.

BARRIER #8: Troubles with team care

At some medical groups, nonphysician providers, including registered nurses and pharmacists, use doctor-created protocols and standing orders to address routine chronic care issues and preventive measures for individuals with certain conditions—identified via patient registries. Similarly, medical assistants and community health workers may be trained as health coaches to work with patients on behavior change and adherence to medication regimens, thus freeing up physician time.²¹ Despite the benefits of team care, most insurers only reimburse the services of MDs, NPs and PAs, meaning that no incentives exist for primary care practices to hire other team members.

The solutions: Policy shifts and culture change

What will it take to improve access to primary care and tear down these barriers? First and foremost, we believe the following policy changes are needed:

• Increase reimbursement for primary care.
• Increase loan repayment programs for medical students who establish primary care practices in areas with established shortages.
• Standardize fees paid by private insurers, Medicare, and Medicaid plans.
• Provide financial incentives for PCPs to deliver after-hours care.
• Invest in a national program aimed at helping primary care practices implement same-day scheduling, team care, and other access improvements.²²
• Provide reimbursement for e-mail and telephone encounters and team care, including fees for all allied health professionals who assist PCPs in managing chronic disease and preventive care.

These reforms, if they were to truly come to pass, would ease much of the pressure on PCPs. No matter what policy changes are implemented to increase access to primary care, however, it is clear that a substantial culture change is required on the part of PCPs, as well. Physicians can begin to make changes on their own to increase patient access—expanding the interval between follow-up visits for stable patients, for instance, and reorganizing work schedules so that the practice can remain open for more hours.

It’s clear that providing health insurance to the uninsured without guaranteeing access to primary care can turn a potentially positive development into widespread patient frustration. Unless Americans have greater access to primary care, we fear, the US health care system will undergo significant change without substantial improvement.

CORRESPONDENCE
Thomas Bodenheimer MD, MPH, Department of Family and Community Medicine, University of California at San Francisco, Bldg 80-83, SF General Hospital, 995 Potrero Avenue, San Francisco, CA 94110; [email protected]

CASE 1 Mr. Brown is a 60-year-old African American man who works as a mid-level executive at a local bank. About a year and a half ago, he was diagnosed with hypertension, joining a number of other family members who also have the condition. Reviewing his chart, you note that at many of his visits—but not all of them—his systolic blood pressure runs close to 150 mm Hg, with diastolic pressure <90 mm Hg. Today his blood pressure is 148/88 mm Hg, numbers that exceed the parameters of the currently accepted definition of hypertension: blood pressure ≥140 mm Hg systolic or ≥90 mm Hg diastolic, taken on 2 separate occasions with the patient sitting down.1

Mr. Brown is one of the more than 65 million American adults suffering from high blood pressure—the No. 1 diagnosis reported in outpatient medical offices.^2,3 Despite such prevalence, blood pressure control in the United States is suboptimal, with only one-third of hypertensive patients under adequate control.¹ In 2007, the total estimated cost of treating high blood pressure in the United States exceeded $66 billion.⁴

When you tell Mr. Brown that his blood pressure doesn’t meet the therapeutic goal of <140/90 mm Hg, he gives you his reasons: He was late, he rushed, the traffic was bad, and the nurse rushed him into the exam room before he even had time to catch his breath. He insists his blood pressure is “normal” at home, and blames the elevated numbers on anxiety. He also tells you he was up working most of the night before, drinking coffee to stay awake to finish an urgent project.

Mr. Brown’s current medication regimen includes daily doses of lisinopril-hydrochlorothiazide (HCTZ) 20/25 mg daily and simvastatin 20 mg. He tells you he has no chest pain, shortness of breath, cough, edema, claudication, paroxysmal nocturnal dyspnea, or orthopnea. When you ask if he takes his lisinopril-HCTZ every day, he says Yes, but you have your suspicions.

Consider pseudo-resistance

Suboptimal blood pressure control can be classified as either pseudo-resistant or resistant hypertension. According to the definition used in the Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure (JNC7), resistance is the “failure to achieve goal BP in patients who are adhering to full doses of an appropriate 3-drug regimen that includes a diuretic.”¹

Hypertension is described as “pseudo-resistant” when persistent elevations in blood pressure are the result of a failure to comply with the medication regimen, “white-coat” syndrome, poor blood pressure technique, or a combination of these.^5,6 Physician failure to prescribe adequate doses of medication, so-called clinical inertia, may also be a factor.^5,7 Causes and remedies for pseudo-resistance are summarized in the algorithm.

In Mr. Brown’s case, he’s told you that he’s taking his lisinopril-HCTZ every day, but when you check your records, you see that the intervals between his refill requests range between 34 and 36 days. So you ask him, again, whether he is taking his lisinopril-HCTZ daily, and this time he says he takes the medication “at least 5 or 6 days a week.”

Encourage compliance. To motivate Mr. Brown, you tell him his blood pressure does not meet the goal of <140/90 mm Hg and that pressure higher than goal is a significant risk factor for cardiovascular disease. You emphasize the importance of taking his medication every day—a mantra you’ve repeated to countless patients over the years. In fact, it’s estimated that up to 40% of patients will discontinue their medication at some point during treatment.⁸ Patients’ reasons vary and may include medication side effects, the cost of treatment, or a patient’s personal philosophy. Cultural differences may also play a role in noncompliance.⁹

You talk with Mr. Brown a bit more and confirm that his poor compliance is due to simple forgetfulness and not another underlying reason that would need to be addressed. You hand him a pill calendar to help him keep track of his medication. You review the dosage and feel confident that you’ve prescribed a regimen simple enough for Mr. Brown to stick to and adequate to bring his pressure down.

When your white coat is the problem. Though not the case with Mr. Brown, a patient’s elevated readings could be a case of white coat hypertension—a conditioned response in some patients that is probably the result of anxiety in the medical setting.¹⁰ Patients with white coat hypertension have significantly less risk of death and reduced target organ damage than patients with truly resistant hypertension.^11,12

If you suspect that a patient of yours has white coat hypertension, you can test your hypothesis by encouraging him or her to buy a blood pressure monitor to use at home, keep a log of the readings, and bring the log in to the next appointment. Improved technology has made home blood pressure monitoring an important tool in the treatment of high blood pressure.^13-15

According to a 2008 statement issued jointly by the American Heart Association, the American Society of Hypertension, and the Preventive Cardiovascular Nurses Association, “There is a rapidly growing literature showing that measurements taken by patients at home are often lower than readings taken in the office and closer to the average blood pressure recorded by 24-hour ambulatory monitors, which is the blood pressure that best predicts cardiovascular risk.”¹⁴

Arm monitors are the most accurate because they measure brachial artery blood pressure.¹⁴ Most wrist monitors have not been validated in studies, but obese patients who cannot find a properly sized cuff may need to use them.¹⁴ Finger devices should never be used.¹⁴

Sometimes technique is at fault. To ensure that you get proper readings when your patient is in the office, advise the nursing staff not to take a patient’s blood pressure for at least 5 minutes after bringing him or her to an exam room. Specifically, the nurse can either make blood pressure measurement the last thing she (or he) does when “rooming” the patient, or she may prefer to go back to the exam room a few minutes after the patient is there to take the reading. The nurse should take care that the cuff is the right size and that the pressure is taken with the patient’s arm at heart level. In hypertensive patients, placing the arm below the horizontal in the dependent position can raise blood pressure 23/10 mm Hg.^1,16,17

Newer blood pressure monitoring devices for the office are available that obviate artificially high readings caused by errors in technique and by white coat syndrome. Researchers have found that readings taken with the BpTRU automatic device, which obtains and records 5 blood pressure readings over a 5-minute period, are lower than those taken by office personnel, and that the white coat effect is eliminated.^18,19

One pseudo-resistant case solved. At Mr. Brown’s next visit, you review his pill calendar with him and note that his blood pressure is now running 138/88 mm Hg and he is within his target goal of less than 140/90 mm Hg.

If Mr. Brown’s blood pressure control had continued to be suboptimal even after you’d taken steps to address cause(s) of pseudo-resistance, you would have had to shift gears and consider his case one of resistant hypertension.¹

Consider resistant hypertension

Causes of resistant hypertension include lifestyle factors, side effects of medications, and secondary causes.^5,6 Tobacco use, obesity, lack of exercise, a high sodium diet, and alcohol consumption can all contribute to hypertension.^1,5,6,20,21 Dietary sodium in particular has significant adverse effects. Reducing dietary sodium not only reduces mortality, but has positive cardiovascular effects separate from blood pressure reduction, such as improved endothelium-dependent vasodilation.^22-24

Medications including nonsteroidal anti-inflammatory drugs (NSAIDs), oral contraceptives, sympathomimetics, glucocorticoids, and black licorice (found in some herbal supplements) may all raise blood pressure.⁵ Medications and other substances associated with increased blood pressure are listed in the TABLE. A search for such products, with changes or discontinuations that are clinically appropriate, should be part of any evaluation for resistant hypertension.

CASE 2 Ms. Stevens is a 30-year-old woman who was diagnosed with high blood pressure 7 months ago. She is currently taking triamterene-HCTZ 37.5/25 mg daily, amlodipine 10 mg daily, and metoprolol succinate 100 mg daily. She is compliant with medication, sticks to a low-sodium diet, and exercises 5 days a week. She does not drink alcohol. A review of her chart tells you her baseline serum chemistries, chest x-ray, and urinalysis are all normal. Today’s blood pressure taken in the sitting position after a 5-minute rest is 160/92 mm Hg. How should you proceed?

TABLE
Medications and other agents that may cause hypertension

Is it resistant hypertension? Look for clues

You go over Ms. Stevens’ history carefully. Her hypertension medications should be adequate, and she is compliant with the prescribed regimen. Lifestyle factors are clearly not to blame: She’s not obese, she follows a low-sodium diet, doesn’t drink or smoke, and gets plenty of exercise. You ask her about other medications or over-the-counter supplements she takes, and nothing on her list raises a red flag. This doesn’t seem to be pseudo-resistance, and though you’re aware that only about 5% of cases of hypertension can be attributed to secondary causes, you think Ms. Stevens may be one of those patients.²⁵

Numerous causes of secondary hypertension exist, and evaluations for such causes should be considered in all patients with resistant hypertension. Renal artery stenosis and pheochromocytoma are the most well-known causes, but less well known but significant contributors are obstructive sleep apnea and hyperaldosteronism.^1,6,26,27

The relationship between obstructive sleep apnea and hypertension is not clear, and most of the studies linking the 2 conditions are population based.²⁸ Nevertheless, because studies show that continuous positive airway pressure (CPAP) does improve blood pressure control in patients with obstructive sleep apnea, it is appropriate to investigate patients with resistant hypertension for this condition.²⁹

Aldosterone’s role in resistant hypertension appears to be greatest in those with insulin resistance. The mechanism appears to be amplified sodium retention resulting in increased volume expansion, negative effects on oxidative stress, and elevated inflammation.²⁷ Many patients with hypertension also have insulin resistance, so testing for hyperaldosteronism should be part of your investigation.

Test, and as necessary, test some more

Taking into consideration that 2 of the more common causes of secondary hypertension are obstructive sleep apnea and renal artery stenosis, you schedule an overnight sleep study for Ms. Stevens and a magnetic resonance arteriogram of the kidneys.^26,30 (For more on the diagnostic tests and treatments for secondary causes of resistant hypertension, see the ALGORITHM.)

Reconsider medications. Ms. Stevens is already receiving the maximum dose of amlodipine, so you increase her metoprolol succinate to 200 mg daily and ask her to return in 1 month after the results of her tests come in.

Change medications and keep testing. At her next visit, Ms. Stevens’ blood pressure is 150/88 mm Hg, and her pulse rate is 64. Results of her overnight sleep study and magnetic resonance arteriogram are normal. At this juncture, with her work-up for common secondary causes of hypertension negative, you consider changes in her medication regimen. Because her pulse rate of 64 rules out any increase in the metoprolol, you decide to add lisinopril-HCTZ 10/12.5 mg to her regimen and discontinue her triamterene-HCTZ. Because her blood pressure is still elevated at this visit, you schedule additional testing to look for other secondary causes.

More tests, higher doses. One month later, Ms. Stevens returns with a blood pressure of 146/88 mm Hg. Appropriate testing for primary aldosteronism, pheochromocytoma, Cushing’s disease, and coarctation are all normal. Her lisinopril-HCTZ dose is increased to 20/12.5 mg, and when she returns 1 month later her blood pressure reading is 138/88 mm Hg.

You conclude that she is a patient with hard-to-control hypertension with no underlying cause who requires a significant amount of medication. That is not an uncommon situation. Although her blood pressure is now controlled to an appropriate level, you make a chart notation to consider referral to a hypertension specialist if her blood pressure become elevated again and treatment changes are not effective.⁶

ALGORITHM
Hard-to-control hypertension? This treatment algorithm can help

Another consideration: Pseudohypertension

This uncommon condition sometimes occurs in elderly patients who suffer from a stiffened or thickened brachial artery.¹ Because the blood pressure cuff cannot compress the artery, blood pressure readings overestimate true arterial pressure. The condition should be considered in patients when increased doses of medication fail to achieve the desired response, particularly in patients who exhibit postural hypotension.¹

This was not a likely explanation for Ms. Stevens, however, given the clinical scenario and her young age.

CORRESPONDENCE Randy Wexler, MD, MPH, B0902B Cramblett Hall, 456 W 10th Avenue, Columbus, OH 43210; [email protected]