Patricia R. Wigle, PharmD, BCPS

Arterial and venous thromboses are major causes of morbidity and mortality in the United States. Each year, about 100 out of 100,000 Americans (0.1%) experience a venous thromboembolism (VTE), and the incidence is considerably higher among hospitalized patients.¹ Incidence and early mortality after a first-time event increase with age. Mortality and the potential for a pulmonary embolism (PE) to occur after a deep vein thrombosis (DVT) depend on the location of the DVT and how well the DVT is managed. Proximal DVTs are more likely to develop into a PE. Mortality rates for patients with PE are as high as 17% 3 months after diagnosis.²

Anticoagulant therapy is the foundation for prevention and treatment of thromboembolic disease, and family physicians are on the front line of management when patients with DVT are discharged from the hospital. There are many therapeutic options to choose from: unfractionated heparin (UFH), low-molecular-weight heparins (LMWHs), the factor Xa inhibitor fondaparinux, direct thrombin inhibitors, or vitamin K antagonists (VKAs). All of these agents are effective, but you’ll need to keep clinical considerations and drug limitations in mind to use them properly.

The salient details for optimal use of these agents are set out in the 8th edition of the American College of Chest Physicians (ACCP) Evidence-based Clinical Practice Guidelines, released in 2008.³ But following these complex guidelines to maximize patient safety and minimize both cost and inconvenience raises many questions for the busy family physician. This article provides the answers you need to maximize your care.

1. What therapies can be used in the outpatient setting to treat acute DVT or PE?

You can manage DVT with LMWH—dalteparin (Fragmin), enoxaparin (Lovenox), or tinzaparin (Innohep)—or the factor Xa inhibitor fondaparinux (Arixtra) overlapped with warfarin (Coumadin). UFH is generally not recommended in the outpatient setting. Patients who are obese or have a creatinine clearance <30 mL/min will need inpatient treatment with UFH in most cases.

Outpatient management of PE based on clinical prediction rules that stratify patients by risk factors has been attempted, although the safety and efficacy of this practice have not been conclusively demonstrated. Prediction rules are available at http://www.medicalcriteria.com/criteria/car_thrombosis.htm. Note that LMWH and fondaparinux are not approved by the US Food and Drug Administration (FDA) for the outpatient treatment of PE.

Dosing guidelines for LMWH agents and fondaparinux are given in TABLE 1, and recommendations for treatment of DVT are summarized in TABLE 2.

TABLE 1
Low-molecular-weight heparins and fondaparinux dosing for DVT

TABLE 2
Treating DVT: Recommended options^3,4,21-24

2. When, and at what dosage, should i initiate warfarin?

With a medically stable patient, you can start warfarin shortly after the first dose of LMWH or fondaparinux, and overlap both therapies for at least 5 days until the patient’s international normalized ratio (INR) is ≥2 for 24 hours. If the INR does not reach 2 within 5 days, LMWH or fondaparinux should be continued. The target INR for DVT is 2.5.

The initial dose of warfarin for most patients should be between 5 and 10 mg per day for the first 2 doses, with 10-mg doses reserved for younger patients without significant drug interactions or comorbidities.⁴ Consider a starting dose ≤5 mg in elderly patients, those with certain medical conditions (eg, liver disease or heart failure), and patients taking medications known to significantly inhibit warfarin metabolism.^3,4TABLE 3 provides a suggested method for initiation of warfarin in ambulatory patients.

Continue warfarin for at least 3 months, and possibly longer, depending on the cause of DVT/PE and underlying or ongoing risk factors. Evaluate the risk vs benefit of continued therapy 3 months after the initial thromboembolic event. Patients with cancer, whose risk for VTE is greater, should receive LMWH for the first 3 to 6 months, followed by long-term therapy with warfarin or LMWH until the cancer is resolved.^3,4

TABLE 3
Average warfarin daily dosing for INR goal 2-3

3. Is it time to customize anticoagulant therapy based on genetic testing?

No. Currently, FDA and ACCP guidelines do not recommend genetic testing before initiating warfarin.^5,6 Theoretically, genetic testing should be helpful in predicting an individual’s optimal starting warfarin dose. At present, however, no good clinical data support this practice.⁵ If randomized trials show improved clinical outcomes with pharmacogenetic dosing of warfarin, genotyping may become part of clinical practice in the future.

An estimated one-third of patients on warfarin therapy may be at higher risk for adverse outcomes because they carry genes that make them more or less sensitive to warfarin.⁵ Variants of 2 genes—cytochrome P450 2C9 (CYP2C9) and the vitamin K oxide reductase complex 1 (VKORC1)—are thought to be responsible for this variance in warfarin response.⁵

Patients with variations of CYP2C9 may need lower starting doses of warfarin. Mutations in the VKORC1 gene affect the enzymes that activate vitamin K, which are the target for warfarin’s inhibitory effect on clotting. Mutations in this gene therefore result in varying sensitivities to warfarin and may be the cause of hereditary warfarin resistance in some individuals. Genetic variations in VKORC1 are estimated to occur in 14% to 37% of Caucasians and African Americans and may exist in as many as 89% of Asians.⁵ Several tests to detect some variants in these genes have been approved by the FDA.

In August 2007, a labeling change for Coumadin and its generics detailed the influence of gene variations on warfarin sensitivity.⁷ A report from the American Enterprise Institute-Brookings Joint Center for Regulatory Studies estimated that genetic testing could prevent 85,000 serious bleeding events and 17,000 strokes per year, resulting in a $1.1 billion reduction in warfarin-related health care spending. Costs of genetic testing for the 2 million Americans who begin warfarin therapy each year would be approximately $1 billion.⁶

4. Which warfarin–drug interactions are clinically important?

Drugs, supplements, and foods that potentiate or inhibit warfarin’s anticoagulant effect or increase the risk of bleeding are clinically important. The list of such interactions has been referred to as the 8 “As”: antibiotics, antifungals, antidepressants, antiplatelets, amiodarone, anti-inflammatories, high-dose acetaminophen, and alternative remedies.⁸ (For details on common warfarin interactions, see TABLE W1.)

These and other medications can affect how warfarin is absorbed, distributed, and metabolized. For example, sucralfate and bile-acid sequestrants such as cholestyramine can inhibit absorption. You can minimize this interaction by staggering the time each medication is ingested. Drugs that induce cytochrome P450 enzymes (eg, rifampin, carbamazepine) enhance warfarin clearance, while drugs that inhibit CYP enzymes (amiodarone or itraconazole) decrease warfarin clearance.² Most clinically relevant interactions affect warfarin metabolism.

5. How should i proceed when a patient taking warfarin also needs antiplatelet medications?

Monitor warfarin more frequently in such patients and target the lower end of the INR therapeutic range (2-2.5).⁹ Keep an eye on your patient’s overall medication regimen and avoid medications like nonsteroidal anti-inflammatory drugs (NSAIDs) that increase bleeding risk. If NSAIDs must be used, avoid chronic use, high doses, and NSAIDs with a long half-life. You may also want to consider referral to an anticoagulation clinic.

Many of these patients have cardiac conditions for which dual antiplatelet therapy is recommended. For example, patients with coronary stents may need aspirin and clopidogrel for a specified period of time. They may have underlying atrial fibrillation or valve replacement requiring warfarin therapy. Data examining triple therapy (aspirin, clopidogrel, and warfarin) are primarily limited to patients with acute coronary syndrome or those who have had percutaneous coronary intervention. Unfortunately, the data are also retrospective, based on a small sample, and inconsistent.¹⁰ In these patients, you need to weigh the increased risk of bleeding against the proven preventive value of each of these modalities.

For patients with stents, current guidelines recommend a lower dose of aspirin and discontinuation of clopidogrel after a certain length of time, depending on the type of stent.^10,11 However, 1 study showed that aspirin dose and INR values did not influence bleeding risk in patients on triple therapy.¹²

It is imperative that you counsel patients on triple therapy to report the first sign of bleeding.

6. What is the best approach when a patient’s INR is elevated?

You’ll need to minimize the risk of bleeding while at the same time ensuring adequate levels of anticoagulation. You can use oral vitamin K (phytonadione [Mephyton]) to reverse the effects of warfarin without inducing warfarin resistance. Avoid subcutaneous administration; the effects are unpredictable and response is delayed.⁴

Send patients with active or life-threatening bleeding to the emergency department. Reserve intravenous vitamin K administration for patients who are bleeding or have an INR >20. The ACCP guidelines provide recommendations on managing elevated INRs in patients receiving warfarin (TABLE 4).⁴

TABLE 4
Managing elevated INR

7. What new anticoagulants are on the horizon?

Several alternative treatments for DVT are currently in clinical trials, and 1 recently received FDA approval.

Ximelagatran, a direct thrombin inhibitor, appeared to hold promise as an oral anticoagulant, but was denied FDA approval and eventually withdrawn by its manufacturer when reports of hepatotoxicity and possible myocardial ischemia surfaced.^13,14 Other oral treatment options to be aware of include another direct thrombin inhibitor, dabigatran, and the factor Xa inhibitors apixaban and rivaroxaban.

Dabigatran (Pradaxa), an oral direct thrombin inhibitor similar to ximelagatran, received FDA approval last month for stroke prevention in atrial fibrillation. One study, RE-COVER, studied dabigatran vs warfarin for the treatment of acute VTE—both lower extremity DVT and PE. This noninferiority trial compared dabigatran 150 mg twice daily with daily warfarin adjusted to achieve an INR of 2.0 to 3.0, with the 6-month recurrence of VTE as the primary outcome. Dabigatran was found to be as effective as warfarin at preventing recurrent or fatal VTE. There was no difference in major bleeding between the dabigatran and warfarin groups, although the dabigatran group did show more major or clinically relevant nonmajor bleeding. No differences in other adverse events were observed between the 2 groups.¹⁵

Three studies, RE-MOBILIZE, RENOVATE, and RE-MODEL, compared dabigatran’s efficacy and safety with enoxaparin for the prevention of VTE after knee and hip replacement surgery. In the RE-MOBILIZE trial, dabigatran was effective compared with enoxaparin once daily, but not effective compared with twice-daily enoxaparin.¹⁶ The RE-NOVATE and RE-MODEL studies also showed dabigatran’s efficacy compared with once-daily enoxaparin.^17-19 Major bleeding occurred in approximately 1% of patients in both the dabigatran and enoxaparin treatment groups, and the incidence of hepatotoxicity was similar.^17-19

The RE-LY trial studied warfarin vs 2 different doses of dabigatran in atrial fibrillation for the prevention of stroke or systemic embolism. Both doses of dabigatran (110 mg twice daily or 150 mg once daily) were similar to warfarin for the study’s primary outcome, and dabigatran at the 110-mg dose had a significantly lower incidence of hemorrhagic stroke.¹⁹ Studies on the use of dabigatran in acute coronary syndrome are ongoing.

Apixaban and rivaroxaban are oral inhibitors of both free and fibrin-bound factor Xa. They are similar in activity to the currently available, injectable fondaparinux. In the RECORD 1, 2, 3, and 4 trials, rivaroxaban was compared with once- or twice-daily enoxaparin in patients undergoing hip and knee replacement surgery.²⁰ Rivaroxaban was significantly better in preventing VTE and it had a comparable rate of major bleeding (approximately 0.2%). Rivaroxaban has been approved in Canada and Europe for thromboprophylaxis after major orthopedic surgery. Rivaroxaban was recommended for approval by an FDA advisory panel, but the FDA has not issued an approval as yet.

Phase III trials for other indications of rivaroxaban and apixaban are currently underway. The long-term safety and adverse event profiles are as yet unclear. If and when these new medications are approved, they should be used judiciously while issues related to reversibility, long-term adverse events, and monitoring are still unresolved. For some patients, warfarin may continue to be the most appropriate oral anticoagulant medication.

TABLE W1
Important warfarin interactions*

CORRESPONDENCE Anne H. Metzger, PharmD, BCPS, University of Cincinnati, The James L. Winkle College of Pharmacy, 3225 Eden Avenue, Cincinnati, OH 45267; [email protected]

Arterial and venous thromboses are major causes of morbidity and mortality in the United States. Each year, about 100 out of 100,000 Americans (0.1%) experience a venous thromboembolism (VTE), and the incidence is considerably higher among hospitalized patients.¹ Incidence and early mortality after a first-time event increase with age. Mortality and the potential for a pulmonary embolism (PE) to occur after a deep vein thrombosis (DVT) depend on the location of the DVT and how well the DVT is managed. Proximal DVTs are more likely to develop into a PE. Mortality rates for patients with PE are as high as 17% 3 months after diagnosis.²

Anticoagulant therapy is the foundation for prevention and treatment of thromboembolic disease, and family physicians are on the front line of management when patients with DVT are discharged from the hospital. There are many therapeutic options to choose from: unfractionated heparin (UFH), low-molecular-weight heparins (LMWHs), the factor Xa inhibitor fondaparinux, direct thrombin inhibitors, or vitamin K antagonists (VKAs). All of these agents are effective, but you’ll need to keep clinical considerations and drug limitations in mind to use them properly.

The salient details for optimal use of these agents are set out in the 8th edition of the American College of Chest Physicians (ACCP) Evidence-based Clinical Practice Guidelines, released in 2008.³ But following these complex guidelines to maximize patient safety and minimize both cost and inconvenience raises many questions for the busy family physician. This article provides the answers you need to maximize your care.

1. What therapies can be used in the outpatient setting to treat acute DVT or PE?

You can manage DVT with LMWH—dalteparin (Fragmin), enoxaparin (Lovenox), or tinzaparin (Innohep)—or the factor Xa inhibitor fondaparinux (Arixtra) overlapped with warfarin (Coumadin). UFH is generally not recommended in the outpatient setting. Patients who are obese or have a creatinine clearance <30 mL/min will need inpatient treatment with UFH in most cases.

Outpatient management of PE based on clinical prediction rules that stratify patients by risk factors has been attempted, although the safety and efficacy of this practice have not been conclusively demonstrated. Prediction rules are available at http://www.medicalcriteria.com/criteria/car_thrombosis.htm. Note that LMWH and fondaparinux are not approved by the US Food and Drug Administration (FDA) for the outpatient treatment of PE.

Dosing guidelines for LMWH agents and fondaparinux are given in TABLE 1, and recommendations for treatment of DVT are summarized in TABLE 2.

TABLE 1
Low-molecular-weight heparins and fondaparinux dosing for DVT

TABLE 2
Treating DVT: Recommended options^3,4,21-24

2. When, and at what dosage, should i initiate warfarin?

With a medically stable patient, you can start warfarin shortly after the first dose of LMWH or fondaparinux, and overlap both therapies for at least 5 days until the patient’s international normalized ratio (INR) is ≥2 for 24 hours. If the INR does not reach 2 within 5 days, LMWH or fondaparinux should be continued. The target INR for DVT is 2.5.

The initial dose of warfarin for most patients should be between 5 and 10 mg per day for the first 2 doses, with 10-mg doses reserved for younger patients without significant drug interactions or comorbidities.⁴ Consider a starting dose ≤5 mg in elderly patients, those with certain medical conditions (eg, liver disease or heart failure), and patients taking medications known to significantly inhibit warfarin metabolism.^3,4TABLE 3 provides a suggested method for initiation of warfarin in ambulatory patients.

Continue warfarin for at least 3 months, and possibly longer, depending on the cause of DVT/PE and underlying or ongoing risk factors. Evaluate the risk vs benefit of continued therapy 3 months after the initial thromboembolic event. Patients with cancer, whose risk for VTE is greater, should receive LMWH for the first 3 to 6 months, followed by long-term therapy with warfarin or LMWH until the cancer is resolved.^3,4

TABLE 3
Average warfarin daily dosing for INR goal 2-3

3. Is it time to customize anticoagulant therapy based on genetic testing?

No. Currently, FDA and ACCP guidelines do not recommend genetic testing before initiating warfarin.^5,6 Theoretically, genetic testing should be helpful in predicting an individual’s optimal starting warfarin dose. At present, however, no good clinical data support this practice.⁵ If randomized trials show improved clinical outcomes with pharmacogenetic dosing of warfarin, genotyping may become part of clinical practice in the future.

An estimated one-third of patients on warfarin therapy may be at higher risk for adverse outcomes because they carry genes that make them more or less sensitive to warfarin.⁵ Variants of 2 genes—cytochrome P450 2C9 (CYP2C9) and the vitamin K oxide reductase complex 1 (VKORC1)—are thought to be responsible for this variance in warfarin response.⁵

Patients with variations of CYP2C9 may need lower starting doses of warfarin. Mutations in the VKORC1 gene affect the enzymes that activate vitamin K, which are the target for warfarin’s inhibitory effect on clotting. Mutations in this gene therefore result in varying sensitivities to warfarin and may be the cause of hereditary warfarin resistance in some individuals. Genetic variations in VKORC1 are estimated to occur in 14% to 37% of Caucasians and African Americans and may exist in as many as 89% of Asians.⁵ Several tests to detect some variants in these genes have been approved by the FDA.

In August 2007, a labeling change for Coumadin and its generics detailed the influence of gene variations on warfarin sensitivity.⁷ A report from the American Enterprise Institute-Brookings Joint Center for Regulatory Studies estimated that genetic testing could prevent 85,000 serious bleeding events and 17,000 strokes per year, resulting in a $1.1 billion reduction in warfarin-related health care spending. Costs of genetic testing for the 2 million Americans who begin warfarin therapy each year would be approximately $1 billion.⁶

4. Which warfarin–drug interactions are clinically important?

Drugs, supplements, and foods that potentiate or inhibit warfarin’s anticoagulant effect or increase the risk of bleeding are clinically important. The list of such interactions has been referred to as the 8 “As”: antibiotics, antifungals, antidepressants, antiplatelets, amiodarone, anti-inflammatories, high-dose acetaminophen, and alternative remedies.⁸ (For details on common warfarin interactions, see TABLE W1.)

These and other medications can affect how warfarin is absorbed, distributed, and metabolized. For example, sucralfate and bile-acid sequestrants such as cholestyramine can inhibit absorption. You can minimize this interaction by staggering the time each medication is ingested. Drugs that induce cytochrome P450 enzymes (eg, rifampin, carbamazepine) enhance warfarin clearance, while drugs that inhibit CYP enzymes (amiodarone or itraconazole) decrease warfarin clearance.² Most clinically relevant interactions affect warfarin metabolism.

5. How should i proceed when a patient taking warfarin also needs antiplatelet medications?

Monitor warfarin more frequently in such patients and target the lower end of the INR therapeutic range (2-2.5).⁹ Keep an eye on your patient’s overall medication regimen and avoid medications like nonsteroidal anti-inflammatory drugs (NSAIDs) that increase bleeding risk. If NSAIDs must be used, avoid chronic use, high doses, and NSAIDs with a long half-life. You may also want to consider referral to an anticoagulation clinic.

Many of these patients have cardiac conditions for which dual antiplatelet therapy is recommended. For example, patients with coronary stents may need aspirin and clopidogrel for a specified period of time. They may have underlying atrial fibrillation or valve replacement requiring warfarin therapy. Data examining triple therapy (aspirin, clopidogrel, and warfarin) are primarily limited to patients with acute coronary syndrome or those who have had percutaneous coronary intervention. Unfortunately, the data are also retrospective, based on a small sample, and inconsistent.¹⁰ In these patients, you need to weigh the increased risk of bleeding against the proven preventive value of each of these modalities.

For patients with stents, current guidelines recommend a lower dose of aspirin and discontinuation of clopidogrel after a certain length of time, depending on the type of stent.^10,11 However, 1 study showed that aspirin dose and INR values did not influence bleeding risk in patients on triple therapy.¹²

It is imperative that you counsel patients on triple therapy to report the first sign of bleeding.

6. What is the best approach when a patient’s INR is elevated?

You’ll need to minimize the risk of bleeding while at the same time ensuring adequate levels of anticoagulation. You can use oral vitamin K (phytonadione [Mephyton]) to reverse the effects of warfarin without inducing warfarin resistance. Avoid subcutaneous administration; the effects are unpredictable and response is delayed.⁴

Send patients with active or life-threatening bleeding to the emergency department. Reserve intravenous vitamin K administration for patients who are bleeding or have an INR >20. The ACCP guidelines provide recommendations on managing elevated INRs in patients receiving warfarin (TABLE 4).⁴

TABLE 4
Managing elevated INR

7. What new anticoagulants are on the horizon?

Several alternative treatments for DVT are currently in clinical trials, and 1 recently received FDA approval.

Ximelagatran, a direct thrombin inhibitor, appeared to hold promise as an oral anticoagulant, but was denied FDA approval and eventually withdrawn by its manufacturer when reports of hepatotoxicity and possible myocardial ischemia surfaced.^13,14 Other oral treatment options to be aware of include another direct thrombin inhibitor, dabigatran, and the factor Xa inhibitors apixaban and rivaroxaban.

Dabigatran (Pradaxa), an oral direct thrombin inhibitor similar to ximelagatran, received FDA approval last month for stroke prevention in atrial fibrillation. One study, RE-COVER, studied dabigatran vs warfarin for the treatment of acute VTE—both lower extremity DVT and PE. This noninferiority trial compared dabigatran 150 mg twice daily with daily warfarin adjusted to achieve an INR of 2.0 to 3.0, with the 6-month recurrence of VTE as the primary outcome. Dabigatran was found to be as effective as warfarin at preventing recurrent or fatal VTE. There was no difference in major bleeding between the dabigatran and warfarin groups, although the dabigatran group did show more major or clinically relevant nonmajor bleeding. No differences in other adverse events were observed between the 2 groups.¹⁵

Three studies, RE-MOBILIZE, RENOVATE, and RE-MODEL, compared dabigatran’s efficacy and safety with enoxaparin for the prevention of VTE after knee and hip replacement surgery. In the RE-MOBILIZE trial, dabigatran was effective compared with enoxaparin once daily, but not effective compared with twice-daily enoxaparin.¹⁶ The RE-NOVATE and RE-MODEL studies also showed dabigatran’s efficacy compared with once-daily enoxaparin.^17-19 Major bleeding occurred in approximately 1% of patients in both the dabigatran and enoxaparin treatment groups, and the incidence of hepatotoxicity was similar.^17-19

The RE-LY trial studied warfarin vs 2 different doses of dabigatran in atrial fibrillation for the prevention of stroke or systemic embolism. Both doses of dabigatran (110 mg twice daily or 150 mg once daily) were similar to warfarin for the study’s primary outcome, and dabigatran at the 110-mg dose had a significantly lower incidence of hemorrhagic stroke.¹⁹ Studies on the use of dabigatran in acute coronary syndrome are ongoing.

Apixaban and rivaroxaban are oral inhibitors of both free and fibrin-bound factor Xa. They are similar in activity to the currently available, injectable fondaparinux. In the RECORD 1, 2, 3, and 4 trials, rivaroxaban was compared with once- or twice-daily enoxaparin in patients undergoing hip and knee replacement surgery.²⁰ Rivaroxaban was significantly better in preventing VTE and it had a comparable rate of major bleeding (approximately 0.2%). Rivaroxaban has been approved in Canada and Europe for thromboprophylaxis after major orthopedic surgery. Rivaroxaban was recommended for approval by an FDA advisory panel, but the FDA has not issued an approval as yet.

Phase III trials for other indications of rivaroxaban and apixaban are currently underway. The long-term safety and adverse event profiles are as yet unclear. If and when these new medications are approved, they should be used judiciously while issues related to reversibility, long-term adverse events, and monitoring are still unresolved. For some patients, warfarin may continue to be the most appropriate oral anticoagulant medication.

TABLE W1
Important warfarin interactions*

CORRESPONDENCE Anne H. Metzger, PharmD, BCPS, University of Cincinnati, The James L. Winkle College of Pharmacy, 3225 Eden Avenue, Cincinnati, OH 45267; [email protected]

Arterial and venous thromboses are major causes of morbidity and mortality in the United States. Each year, about 100 out of 100,000 Americans (0.1%) experience a venous thromboembolism (VTE), and the incidence is considerably higher among hospitalized patients.¹ Incidence and early mortality after a first-time event increase with age. Mortality and the potential for a pulmonary embolism (PE) to occur after a deep vein thrombosis (DVT) depend on the location of the DVT and how well the DVT is managed. Proximal DVTs are more likely to develop into a PE. Mortality rates for patients with PE are as high as 17% 3 months after diagnosis.²

Anticoagulant therapy is the foundation for prevention and treatment of thromboembolic disease, and family physicians are on the front line of management when patients with DVT are discharged from the hospital. There are many therapeutic options to choose from: unfractionated heparin (UFH), low-molecular-weight heparins (LMWHs), the factor Xa inhibitor fondaparinux, direct thrombin inhibitors, or vitamin K antagonists (VKAs). All of these agents are effective, but you’ll need to keep clinical considerations and drug limitations in mind to use them properly.

The salient details for optimal use of these agents are set out in the 8th edition of the American College of Chest Physicians (ACCP) Evidence-based Clinical Practice Guidelines, released in 2008.³ But following these complex guidelines to maximize patient safety and minimize both cost and inconvenience raises many questions for the busy family physician. This article provides the answers you need to maximize your care.

1. What therapies can be used in the outpatient setting to treat acute DVT or PE?

You can manage DVT with LMWH—dalteparin (Fragmin), enoxaparin (Lovenox), or tinzaparin (Innohep)—or the factor Xa inhibitor fondaparinux (Arixtra) overlapped with warfarin (Coumadin). UFH is generally not recommended in the outpatient setting. Patients who are obese or have a creatinine clearance <30 mL/min will need inpatient treatment with UFH in most cases.

Outpatient management of PE based on clinical prediction rules that stratify patients by risk factors has been attempted, although the safety and efficacy of this practice have not been conclusively demonstrated. Prediction rules are available at http://www.medicalcriteria.com/criteria/car_thrombosis.htm. Note that LMWH and fondaparinux are not approved by the US Food and Drug Administration (FDA) for the outpatient treatment of PE.

Dosing guidelines for LMWH agents and fondaparinux are given in TABLE 1, and recommendations for treatment of DVT are summarized in TABLE 2.

TABLE 1
Low-molecular-weight heparins and fondaparinux dosing for DVT

TABLE 2
Treating DVT: Recommended options^3,4,21-24

2. When, and at what dosage, should i initiate warfarin?

With a medically stable patient, you can start warfarin shortly after the first dose of LMWH or fondaparinux, and overlap both therapies for at least 5 days until the patient’s international normalized ratio (INR) is ≥2 for 24 hours. If the INR does not reach 2 within 5 days, LMWH or fondaparinux should be continued. The target INR for DVT is 2.5.

The initial dose of warfarin for most patients should be between 5 and 10 mg per day for the first 2 doses, with 10-mg doses reserved for younger patients without significant drug interactions or comorbidities.⁴ Consider a starting dose ≤5 mg in elderly patients, those with certain medical conditions (eg, liver disease or heart failure), and patients taking medications known to significantly inhibit warfarin metabolism.^3,4TABLE 3 provides a suggested method for initiation of warfarin in ambulatory patients.

Continue warfarin for at least 3 months, and possibly longer, depending on the cause of DVT/PE and underlying or ongoing risk factors. Evaluate the risk vs benefit of continued therapy 3 months after the initial thromboembolic event. Patients with cancer, whose risk for VTE is greater, should receive LMWH for the first 3 to 6 months, followed by long-term therapy with warfarin or LMWH until the cancer is resolved.^3,4

TABLE 3
Average warfarin daily dosing for INR goal 2-3

3. Is it time to customize anticoagulant therapy based on genetic testing?

No. Currently, FDA and ACCP guidelines do not recommend genetic testing before initiating warfarin.^5,6 Theoretically, genetic testing should be helpful in predicting an individual’s optimal starting warfarin dose. At present, however, no good clinical data support this practice.⁵ If randomized trials show improved clinical outcomes with pharmacogenetic dosing of warfarin, genotyping may become part of clinical practice in the future.

An estimated one-third of patients on warfarin therapy may be at higher risk for adverse outcomes because they carry genes that make them more or less sensitive to warfarin.⁵ Variants of 2 genes—cytochrome P450 2C9 (CYP2C9) and the vitamin K oxide reductase complex 1 (VKORC1)—are thought to be responsible for this variance in warfarin response.⁵

Patients with variations of CYP2C9 may need lower starting doses of warfarin. Mutations in the VKORC1 gene affect the enzymes that activate vitamin K, which are the target for warfarin’s inhibitory effect on clotting. Mutations in this gene therefore result in varying sensitivities to warfarin and may be the cause of hereditary warfarin resistance in some individuals. Genetic variations in VKORC1 are estimated to occur in 14% to 37% of Caucasians and African Americans and may exist in as many as 89% of Asians.⁵ Several tests to detect some variants in these genes have been approved by the FDA.

In August 2007, a labeling change for Coumadin and its generics detailed the influence of gene variations on warfarin sensitivity.⁷ A report from the American Enterprise Institute-Brookings Joint Center for Regulatory Studies estimated that genetic testing could prevent 85,000 serious bleeding events and 17,000 strokes per year, resulting in a $1.1 billion reduction in warfarin-related health care spending. Costs of genetic testing for the 2 million Americans who begin warfarin therapy each year would be approximately $1 billion.⁶

4. Which warfarin–drug interactions are clinically important?

Drugs, supplements, and foods that potentiate or inhibit warfarin’s anticoagulant effect or increase the risk of bleeding are clinically important. The list of such interactions has been referred to as the 8 “As”: antibiotics, antifungals, antidepressants, antiplatelets, amiodarone, anti-inflammatories, high-dose acetaminophen, and alternative remedies.⁸ (For details on common warfarin interactions, see TABLE W1.)

These and other medications can affect how warfarin is absorbed, distributed, and metabolized. For example, sucralfate and bile-acid sequestrants such as cholestyramine can inhibit absorption. You can minimize this interaction by staggering the time each medication is ingested. Drugs that induce cytochrome P450 enzymes (eg, rifampin, carbamazepine) enhance warfarin clearance, while drugs that inhibit CYP enzymes (amiodarone or itraconazole) decrease warfarin clearance.² Most clinically relevant interactions affect warfarin metabolism.

5. How should i proceed when a patient taking warfarin also needs antiplatelet medications?

Monitor warfarin more frequently in such patients and target the lower end of the INR therapeutic range (2-2.5).⁹ Keep an eye on your patient’s overall medication regimen and avoid medications like nonsteroidal anti-inflammatory drugs (NSAIDs) that increase bleeding risk. If NSAIDs must be used, avoid chronic use, high doses, and NSAIDs with a long half-life. You may also want to consider referral to an anticoagulation clinic.

Many of these patients have cardiac conditions for which dual antiplatelet therapy is recommended. For example, patients with coronary stents may need aspirin and clopidogrel for a specified period of time. They may have underlying atrial fibrillation or valve replacement requiring warfarin therapy. Data examining triple therapy (aspirin, clopidogrel, and warfarin) are primarily limited to patients with acute coronary syndrome or those who have had percutaneous coronary intervention. Unfortunately, the data are also retrospective, based on a small sample, and inconsistent.¹⁰ In these patients, you need to weigh the increased risk of bleeding against the proven preventive value of each of these modalities.

For patients with stents, current guidelines recommend a lower dose of aspirin and discontinuation of clopidogrel after a certain length of time, depending on the type of stent.^10,11 However, 1 study showed that aspirin dose and INR values did not influence bleeding risk in patients on triple therapy.¹²

It is imperative that you counsel patients on triple therapy to report the first sign of bleeding.

6. What is the best approach when a patient’s INR is elevated?

You’ll need to minimize the risk of bleeding while at the same time ensuring adequate levels of anticoagulation. You can use oral vitamin K (phytonadione [Mephyton]) to reverse the effects of warfarin without inducing warfarin resistance. Avoid subcutaneous administration; the effects are unpredictable and response is delayed.⁴

Send patients with active or life-threatening bleeding to the emergency department. Reserve intravenous vitamin K administration for patients who are bleeding or have an INR >20. The ACCP guidelines provide recommendations on managing elevated INRs in patients receiving warfarin (TABLE 4).⁴

TABLE 4
Managing elevated INR

7. What new anticoagulants are on the horizon?

Several alternative treatments for DVT are currently in clinical trials, and 1 recently received FDA approval.

Ximelagatran, a direct thrombin inhibitor, appeared to hold promise as an oral anticoagulant, but was denied FDA approval and eventually withdrawn by its manufacturer when reports of hepatotoxicity and possible myocardial ischemia surfaced.^13,14 Other oral treatment options to be aware of include another direct thrombin inhibitor, dabigatran, and the factor Xa inhibitors apixaban and rivaroxaban.

Dabigatran (Pradaxa), an oral direct thrombin inhibitor similar to ximelagatran, received FDA approval last month for stroke prevention in atrial fibrillation. One study, RE-COVER, studied dabigatran vs warfarin for the treatment of acute VTE—both lower extremity DVT and PE. This noninferiority trial compared dabigatran 150 mg twice daily with daily warfarin adjusted to achieve an INR of 2.0 to 3.0, with the 6-month recurrence of VTE as the primary outcome. Dabigatran was found to be as effective as warfarin at preventing recurrent or fatal VTE. There was no difference in major bleeding between the dabigatran and warfarin groups, although the dabigatran group did show more major or clinically relevant nonmajor bleeding. No differences in other adverse events were observed between the 2 groups.¹⁵

Three studies, RE-MOBILIZE, RENOVATE, and RE-MODEL, compared dabigatran’s efficacy and safety with enoxaparin for the prevention of VTE after knee and hip replacement surgery. In the RE-MOBILIZE trial, dabigatran was effective compared with enoxaparin once daily, but not effective compared with twice-daily enoxaparin.¹⁶ The RE-NOVATE and RE-MODEL studies also showed dabigatran’s efficacy compared with once-daily enoxaparin.^17-19 Major bleeding occurred in approximately 1% of patients in both the dabigatran and enoxaparin treatment groups, and the incidence of hepatotoxicity was similar.^17-19

The RE-LY trial studied warfarin vs 2 different doses of dabigatran in atrial fibrillation for the prevention of stroke or systemic embolism. Both doses of dabigatran (110 mg twice daily or 150 mg once daily) were similar to warfarin for the study’s primary outcome, and dabigatran at the 110-mg dose had a significantly lower incidence of hemorrhagic stroke.¹⁹ Studies on the use of dabigatran in acute coronary syndrome are ongoing.

Apixaban and rivaroxaban are oral inhibitors of both free and fibrin-bound factor Xa. They are similar in activity to the currently available, injectable fondaparinux. In the RECORD 1, 2, 3, and 4 trials, rivaroxaban was compared with once- or twice-daily enoxaparin in patients undergoing hip and knee replacement surgery.²⁰ Rivaroxaban was significantly better in preventing VTE and it had a comparable rate of major bleeding (approximately 0.2%). Rivaroxaban has been approved in Canada and Europe for thromboprophylaxis after major orthopedic surgery. Rivaroxaban was recommended for approval by an FDA advisory panel, but the FDA has not issued an approval as yet.

Phase III trials for other indications of rivaroxaban and apixaban are currently underway. The long-term safety and adverse event profiles are as yet unclear. If and when these new medications are approved, they should be used judiciously while issues related to reversibility, long-term adverse events, and monitoring are still unresolved. For some patients, warfarin may continue to be the most appropriate oral anticoagulant medication.

TABLE W1
Important warfarin interactions*

CORRESPONDENCE Anne H. Metzger, PharmD, BCPS, University of Cincinnati, The James L. Winkle College of Pharmacy, 3225 Eden Avenue, 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]

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]