Applied Evidence

 

CASE John L, a 63-year-old man taking lovastatin (40 mg/d) and ramipril (5 mg/d) for hypercholesterolemia and arterial hypertension was hospitalized with atrial fibrillation. Three days later, he was discharged, with a prescription for amiodarone (200 mg/d). After a month, he was readmitted to the hospital with dark urine and intensifying thigh weakness and achiness. Laboratory testing revealed aspartate aminotransferase and alanine aminotransferase levels 10 times the upper limit of normal, and elevated urine and serum myoglobin.

Drug-drug interactions (DDIs) like the one John experienced between lovastatin and amiodarone are a common cause of readmissions, as well as emergency department visits and hospitalizations, for everything from myopathy to electrolyte imbalance, gastrointestinal (GI) bleeding, hepatotoxicity, renal dysfunction, and changes in blood pressure and heart rate.^1-3

Yet many, if not most, DDIs can be avoided. With diligence and the right tools, you can do much to reduce the incidence of such interactions and adverse outcomes.

Polypharmacy and age pose the highest risks

The more medications a patient is taking, of course, the greater the likelihood of a clinically significant DDI. According to 1 study, 13% of patients taking 2 drugs develop a DDI; the incidence approaches 40% for patients taking 5 drugs, and exceeds 80% for patients taking 7 or more medications.⁴

In addition to polypharmacy, age alone is a key risk factor for DDIs.⁵ Pharmacokinetics and pharmacodynamics are frequently altered in older people, who may have slower intestinal transit time; diminished absorption capacity; decreased liver metabolism, mitochondrial function, and renal excretion; and alterations in volemia and body fat distribution.⁶ Although the speed at which these changes occur varies, aging is associated with a progressive deficiency in the regulation of most homeo-static mechanisms and an altered response to receptor stimulation.⁷

Whether age, multiple medications, or both are to blame, the impact on the elderly is striking. One recent retrospective study found 25% of elderly outpatients to be at risk for DDIs.⁸

Very young patients (<5 years) are also at risk for DDIs because of the immaturity of their enzymatic metabolic system.^5,9,10 Additional risk factors, detailed in TABLE 1, include the presence of an infection or other acute medical condition, a metabolic or endocrine disorder, and taking 1 or more drugs with a narrow therapeutic range. Women are also at higher risk for DDIs than their male counterparts, the result of a slower metabolic capacity and interference with sex hormones.^3-5,8,9 Pharmacogenetics may also play a big part in DDIs, and more and more studies are focusing on identifying patients at greatest risk.

TABLE 1
Risk factors for drug-drug interactions^3-5,8,9

 

Risk factor	Potential result
Acute medical condition (eg, dehydration, infection, alcoholism)	Augmented risk of elevated plasma drug concentration, increased catabolism, inhibition of hepatic drug metabolism
Age (very young [<5 years] and elderly)	Reduced metabolic capacity (greater accumulation of drugs)
Decreased renal and/or hepatic function	Decreased drug clearance/elimination; greater accumulation of drugs or their metabolites
Drug(s) with narrow therapeutic range	Increased risk for dose-related side effects
Female sex	Reduced metabolic capacity, interference with sex hormones
Metabolic or endocrine conditions (eg, fatty liver, obesity, hypothyroidism)	Altered hepatic metabolism, increased body distribution volumes, augmented risk of accumulation for hydrophobic molecules
Polypharmacy (≥3 medications)	Increased risk of metabolic and/or pharmacodynamic interference
Pharmacogenetics	Altered metabolic capacity (greater accumulation of drugs or their metabolites)

Lack of coordinated care also increases risk
Another risk factor involves the use of multiple providers.⁹ A woman may be treated by—and receive prescriptions from—an endocrinolo-gist, a gynecologist, and a family physician (FP), for instance, and get medications from a local pharmacy, a nationwide discount chain, and a mail order pharmacy. As with the number of medications being taken, the greater the number of health care professionals a patient sees, the higher the risk.

To mitigate the risk, encourage your patients to fill all their prescriptions at the same pharmacy—and for your part, take a complete medication history before writing a new prescription.

Medication history in doubt? Schedule a “brown bag review”
Ask patients to provide the name and dose of every medication they’re taking. Inquire specifically about over-the-counter (OTC) cough and cold remedies and complementary and alternative medicines, including herbal remedies, vitamins, and supplements. Patients often neglect to mention nonprescription remedies and may not even think of them as medicine, but OTC products with the potential to interact adversely with prescription drugs may otherwise remain undetected.^11-13

 

Consider a “brown bag review” for patients who don’t know what dosage they’re taking or have difficulty identifying the drugs other physicians have prescribed. Ask them to put all their medications in a brown bag and bring them in on their next visit.^14,15

 

Steps to take to reduce risk

Software systems. A number of free and low-cost software systems identify potential DDIs (See “Check for drug interactions: Software programs to consider”). While such electronic programs can indeed lower the risk,^16-18 they cannot be counted on to detect or avert every possible adverse interaction.

The downside. One problem is that some software programs fail to distinguish between clinically significant and nonsignificant interactions, causing some prescribers to override system alerts—and possibly miss an important warning.¹⁹ Another problem: While most systems do an excellent job of checking to see whether 2 drugs can be safely taken together, few are capable of checking for all potential interactions among multiple medications. What’s more, many drugs have not been evaluated for their potential to interact with other agents, so the absence of reported interactions is no guarantee of a lack of DDIs.

 

Other strategies to consider:

Minimize the number of prescriptions. While it may not be possible to avoid prescribing a new agent for a patient who is already taking multiple medications, limiting the number of new drugs to those that are absolutely essential will help to minimize DDIs. Whenever possible, select a compound with the desired effects. Prescribe a single agent with antihypertensive as well as uricosuric effects for a patient with elevated blood pressure and uric acid levels rather than 2 different drugs (eg, losartan instead of an anti-hypertensive agent plus allopurinol).

Alter the dosing regimen. Several active molecules may cause DDIs by interfering with intestinal absorption or GI transit time if they’re taken closely together. For example, a quinolone should not be administered at the same time as a cation because of possible chelation in the GI tract. If a patient needs both, however, you may be able to avert a DDI by advising the patient to take them at least 2 hours apart.²⁰ Another possibility is to temporarily discontinue a maintenance medication if it has the potential to interact adversely with a drug that is needed for only a short duration.

Choose a different drug (or drug class). Some drug classes should never be mixed—nitrates and phosphodiesterase type-5 inhibitors, taken together, greatly increase the risk of vasodilation and may result in severe hypotension, for example. There are also drug classes with a low potential for DDIs, including cholinesterase inhibitors and anti-hypertensives (angiotensin-converting enzyme inhibitors, angiotensin receptor blockers, beta-blockers, and thiazides).²¹

 

Frequently, though, medications within the same drug class do not share the same potential for DDIs. In such cases, an adverse outcome can often be averted by being aware of combinations likely to result in clinically significant DDIs (TABLE 2) and, whenever possible, prescribing another agent. If a patient taking carbamazepine needs a macro-lide antibiotic, for instance, azithromycin is a better choice than erythromycin. That’s because erythromycin inhibits the hepatic metabolism of the anticonvulsant, increasing the serum level of carbamazepine, while azithromycin does not interfere with carbamazepine metabolism.²²

TABLE 2
Clinically significant drug-drug interactions^21,22

 

Combination (effect)
Allopurinol and captopril (augments allopurinol’s effect)
Antidepressants (SSRIs, MAOIs) and antiepileptics (augments antidepressant effect)
Clopidogrel and omeprazole or esomeprazole (reduces clopidogrel’s effect)
Erythromycin and carbamazepine (augments carbamazepine’s effect)
Erythromycin and terfenadine (augments terfenadine’s effect)
Ketoconazole and PPIs (reduces ketoconazole’s absorption)
Levodopa and metoclopramide (augments levodopa’s effect)
MAOIs and narcotic analgesics (augments effects of both drugs)
Nitrates and phosphodiesterase type-5 inhibitors (augments effects of both drugs)
OCs and penicillins, phenobarbital, or tetracycline (reduces OCs’ effect)
Phenobarbital and simvastatin* (reduces simvastatin’s effect)
Quinolone and cation (reduces quinolone’s absorption and effect)
Repaglinide and diltiazem (augments repaglinide’s effect)
Simvastatin* or lovastatin and amiodarone or itraconazole (augments statin’s effect)
Theophylline and cimetidine or ciprofloxacin (augments theophylline’s effect)
MAOIs, monoamine oxidase inhibitors; OCs, oral contraceptives; PPIs, proton pump inhibitors; SSRIs, selective serotonin reuptake inhibitors.
*For a complete list of drugs that may interact with simvastatin, see US Food and Drug Administration.28

 

 

How is the drug metabolized?

DDIs may occur as a result of pharmaco-dynamic interaction (when 2 drugs act on the same receptor, site of action, or physiologic system) or pharmacokinetic changes (interference with absorption, albumin binding, distribution, metabolism, or elimination).²³ As already noted, age-related changes in pharmacokinetics and pharmacodynamics contribute to the high prevalence of DDIs in elderly patients.

In the liver, drug metabolism, particularly via the cytochrome P450 (CYP450) system, is the cornerstone of drug transformation.²³ Al-though the CYP system consists of “superfami-lies” with more than 100 types of enzymes, only a few are responsible for the majority of biotransformation.²³ The CYP system is also subject to genetic polymorphism, making some patients especially prone to DDIs.

P-glycoproteins (PGPs), which regulate drug absorption by transporting the drugs across cell membranes, also play a key role. PGP inhibitors or inducers help determine whether the accumulation of the molecule or the increased delivery of toxic metabolites leads to adverse effects.¹⁰

Reviewing the mechanism of action of any drug you prescribe for a patient taking other medications may alert you to a potential DDI—and the need to either switch the newly prescribed agent or alter the individual’s drug regimen in some other way.

CASE When John was readmitted to the hospital, he was taken off both the lovastatin and amiodarone and hydrated with forced alkaline diuresis. After a week, his symptoms resolved, and he was discharged soon after. His blood tests normalized 1 month later. The severe DDI he experienced occurred because lovastatin (which is metabolized primarily by CYP3A4) and amiodarone (a CYP3A4 inhibitor) were taken together. (Statins that are substrates of CYP3A4 have the greatest potential for interacting with drugs known to inhibit the CYP450 system [eg, cyclosporine, morphine derivatives, ketoconazole, and amiodarone].)

This adverse interaction could have been avoided if the physician who started John on amiodarone had been aware of the potential DDI—and switched him to an HMG-CoA inhibitor other than lovastatin. Pravastatin, which is not metabolized via CYP450, would have been an excellent choice.

Warfarin warrants special attention
Medications that have a particularly high potential for adverse interactions require special attention and patient monitoring, warfarin foremost among them. Warfarin metabolism and its anticoagulant effects can be dramatically changed if it is administered with a drug with a higher affinity for PGPs or an agent that competes with it within the CYP450 system.²⁴ Because of warfarin’s narrow therapeutic range, there are many drugs and drug classes that patients on warfarin should avoid (TABLE 3)—a fact that patients as well as their physicians need to be aware of.²⁴ Indeed, warfarin is often involved in drug-related hospital admissions for DDIs, especially in elderly patients and in those who are also taking nonsteroidal anti-inflamma-tory drugs (NSAIDs) or macrolides—2 of the many drug classes that patients taking warfarin should avoid.²⁴

TABLE 3
Patient on warfarin? Steer clear of these drugs^10,20,23

 

Drug class: agent(s)
Antiarrhythmics: amiodarone, propafenone
Antibiotics: ciprofloxacin, metronidazole, rifampin, trimethoprim/sulfamethoxazole
Anticonvulsants: carbamazepine, valproate
Antidepressants: fluoxetine, fluvoxamine, paroxetine, sertraline, trazodone
Antidiabetics: chlorpropamide
Antifungals: danazol, fluconazole, itraconazole, miconazole
Antimalarial agents: quinidine
Antineoplastics: azathioprine, fluorouracil, flutamide, ifosfamide, tamoxifen
Antiplatelet agents: ticlopidine
Antipsychotics: clozapine
Diuretics: spironolactone
GI drugs: cimetidine, esomeprazole, lansoprazole, omeprazole, pantoprazole, rabeprazole, ranitidine
Gout treatment: allopurinol
Hypolipidemics: atorvastatin, cholestyramine, ezetimibe, fenofibrate, fluvastatin, gemfibrozil, lovastatin, pravastatin, simvastatin
NSAIDs: aspirin, celecoxib, diclofenac, ibuprofen, indomethacin, ketoprofen, ketorolac, naproxen, piroxicam, sulindac
Thrombolytics: heparin, tissue plasminogen activator
Thyroid drugs: methimazole, propylthiouracil
Uricosuric agents: sulfinpyrazone
GI, gastrointestinal; NSAIDs, nonsteroidal anti-inflammatory drugs.

Keep an eye on these drug combinations, as well

Among the many combinations likely to result in clinically significant DDIs (TABLE 2), the following are worth mentioning:

 

Clopidogrel + certain proton pump inhibitors. The addition of a PPI to clopidogrel has been associated with a significant increase of recurrent infarction.²⁵ This may occur because clopidogrel is a prodrug and is converted in the liver to its active form by CYP2C19, an enzyme specifically inhibited by various PPIs—thereby altering the effectiveness of the antiplatelet agent. However, a recent analysis suggests that there is no need to avoid the concomitant use of a PPI and clopidogrel—and that the interference appears to be limited to omeprazole and esomeprazole.²⁶

Oral contraceptives (OCs) + penicillins, phenobarbital, or tetracycline. Each of these drugs reduces the effect of OCs, and women who are taking them concomitantly need to be advised to use another means of contraception.

Phenobarbital + simvastatin. Pheno-barbital (a CYP3A4 inducer) may reduce the efficacy of simvastatin.

Repaglinide + diltiazem. Diltiazem inhibits the metabolism of repaglinide (a CYP3A4 substrate), thus increasing the risk of hypoglycemia.

 

Simvastatin + amiodarone or itraconazole. Either of these antiarrhythmic agents decreases simvastatin metabolism, raising the risk of myopathy; with amiodarone, however, the likelihood of an adverse outcome is especially high. In 2008, the US Food and Drug Administration (FDA) issued a warning to healthcare professionals of the increased risk for rhabdomyolysis when simvastatin doses greater than 20 mg are administered together with amiodarone.²⁷ The agency issued a safety review of simvastatin, warning of its potential for DDIs with amiodarone and numerous other medications, earlier this year.²⁸

 

As John’s case illustrates, use of lovastatin with amiodarone should be avoided, as well.

Keep others safe: Report adverse events
When a DDI occurs despite your best efforts, you can help ensure that other patients do not experience the same adverse outcome by reporting it to MedWatch, the FDA’s voluntary safety information and adverse event reporting program. Go to https://www.accessdata.fda.gov/scripts/medwatch/medwatch-online.htm to file a report online.

CORRESPONDENCE Ignazio Grattagliano, MD, General Medicine, Department of Internal and Public Medicine, University of Bari, P.zza G. Cesare, 11 – 70124, Bari, Italy; [email protected]

 

CASE John L, a 63-year-old man taking lovastatin (40 mg/d) and ramipril (5 mg/d) for hypercholesterolemia and arterial hypertension was hospitalized with atrial fibrillation. Three days later, he was discharged, with a prescription for amiodarone (200 mg/d). After a month, he was readmitted to the hospital with dark urine and intensifying thigh weakness and achiness. Laboratory testing revealed aspartate aminotransferase and alanine aminotransferase levels 10 times the upper limit of normal, and elevated urine and serum myoglobin.

Drug-drug interactions (DDIs) like the one John experienced between lovastatin and amiodarone are a common cause of readmissions, as well as emergency department visits and hospitalizations, for everything from myopathy to electrolyte imbalance, gastrointestinal (GI) bleeding, hepatotoxicity, renal dysfunction, and changes in blood pressure and heart rate.^1-3

Yet many, if not most, DDIs can be avoided. With diligence and the right tools, you can do much to reduce the incidence of such interactions and adverse outcomes.

Polypharmacy and age pose the highest risks

The more medications a patient is taking, of course, the greater the likelihood of a clinically significant DDI. According to 1 study, 13% of patients taking 2 drugs develop a DDI; the incidence approaches 40% for patients taking 5 drugs, and exceeds 80% for patients taking 7 or more medications.⁴

In addition to polypharmacy, age alone is a key risk factor for DDIs.⁵ Pharmacokinetics and pharmacodynamics are frequently altered in older people, who may have slower intestinal transit time; diminished absorption capacity; decreased liver metabolism, mitochondrial function, and renal excretion; and alterations in volemia and body fat distribution.⁶ Although the speed at which these changes occur varies, aging is associated with a progressive deficiency in the regulation of most homeo-static mechanisms and an altered response to receptor stimulation.⁷

Whether age, multiple medications, or both are to blame, the impact on the elderly is striking. One recent retrospective study found 25% of elderly outpatients to be at risk for DDIs.⁸

Very young patients (<5 years) are also at risk for DDIs because of the immaturity of their enzymatic metabolic system.^5,9,10 Additional risk factors, detailed in TABLE 1, include the presence of an infection or other acute medical condition, a metabolic or endocrine disorder, and taking 1 or more drugs with a narrow therapeutic range. Women are also at higher risk for DDIs than their male counterparts, the result of a slower metabolic capacity and interference with sex hormones.^3-5,8,9 Pharmacogenetics may also play a big part in DDIs, and more and more studies are focusing on identifying patients at greatest risk.

TABLE 1
Risk factors for drug-drug interactions^3-5,8,9

 

Risk factor	Potential result
Acute medical condition (eg, dehydration, infection, alcoholism)	Augmented risk of elevated plasma drug concentration, increased catabolism, inhibition of hepatic drug metabolism
Age (very young [<5 years] and elderly)	Reduced metabolic capacity (greater accumulation of drugs)
Decreased renal and/or hepatic function	Decreased drug clearance/elimination; greater accumulation of drugs or their metabolites
Drug(s) with narrow therapeutic range	Increased risk for dose-related side effects
Female sex	Reduced metabolic capacity, interference with sex hormones
Metabolic or endocrine conditions (eg, fatty liver, obesity, hypothyroidism)	Altered hepatic metabolism, increased body distribution volumes, augmented risk of accumulation for hydrophobic molecules
Polypharmacy (≥3 medications)	Increased risk of metabolic and/or pharmacodynamic interference
Pharmacogenetics	Altered metabolic capacity (greater accumulation of drugs or their metabolites)

Lack of coordinated care also increases risk
Another risk factor involves the use of multiple providers.⁹ A woman may be treated by—and receive prescriptions from—an endocrinolo-gist, a gynecologist, and a family physician (FP), for instance, and get medications from a local pharmacy, a nationwide discount chain, and a mail order pharmacy. As with the number of medications being taken, the greater the number of health care professionals a patient sees, the higher the risk.

To mitigate the risk, encourage your patients to fill all their prescriptions at the same pharmacy—and for your part, take a complete medication history before writing a new prescription.

Medication history in doubt? Schedule a “brown bag review”
Ask patients to provide the name and dose of every medication they’re taking. Inquire specifically about over-the-counter (OTC) cough and cold remedies and complementary and alternative medicines, including herbal remedies, vitamins, and supplements. Patients often neglect to mention nonprescription remedies and may not even think of them as medicine, but OTC products with the potential to interact adversely with prescription drugs may otherwise remain undetected.^11-13

 

Consider a “brown bag review” for patients who don’t know what dosage they’re taking or have difficulty identifying the drugs other physicians have prescribed. Ask them to put all their medications in a brown bag and bring them in on their next visit.^14,15

 

Steps to take to reduce risk

Software systems. A number of free and low-cost software systems identify potential DDIs (See “Check for drug interactions: Software programs to consider”). While such electronic programs can indeed lower the risk,^16-18 they cannot be counted on to detect or avert every possible adverse interaction.

The downside. One problem is that some software programs fail to distinguish between clinically significant and nonsignificant interactions, causing some prescribers to override system alerts—and possibly miss an important warning.¹⁹ Another problem: While most systems do an excellent job of checking to see whether 2 drugs can be safely taken together, few are capable of checking for all potential interactions among multiple medications. What’s more, many drugs have not been evaluated for their potential to interact with other agents, so the absence of reported interactions is no guarantee of a lack of DDIs.

 

Other strategies to consider:

Minimize the number of prescriptions. While it may not be possible to avoid prescribing a new agent for a patient who is already taking multiple medications, limiting the number of new drugs to those that are absolutely essential will help to minimize DDIs. Whenever possible, select a compound with the desired effects. Prescribe a single agent with antihypertensive as well as uricosuric effects for a patient with elevated blood pressure and uric acid levels rather than 2 different drugs (eg, losartan instead of an anti-hypertensive agent plus allopurinol).

Alter the dosing regimen. Several active molecules may cause DDIs by interfering with intestinal absorption or GI transit time if they’re taken closely together. For example, a quinolone should not be administered at the same time as a cation because of possible chelation in the GI tract. If a patient needs both, however, you may be able to avert a DDI by advising the patient to take them at least 2 hours apart.²⁰ Another possibility is to temporarily discontinue a maintenance medication if it has the potential to interact adversely with a drug that is needed for only a short duration.

Choose a different drug (or drug class). Some drug classes should never be mixed—nitrates and phosphodiesterase type-5 inhibitors, taken together, greatly increase the risk of vasodilation and may result in severe hypotension, for example. There are also drug classes with a low potential for DDIs, including cholinesterase inhibitors and anti-hypertensives (angiotensin-converting enzyme inhibitors, angiotensin receptor blockers, beta-blockers, and thiazides).²¹

 

Frequently, though, medications within the same drug class do not share the same potential for DDIs. In such cases, an adverse outcome can often be averted by being aware of combinations likely to result in clinically significant DDIs (TABLE 2) and, whenever possible, prescribing another agent. If a patient taking carbamazepine needs a macro-lide antibiotic, for instance, azithromycin is a better choice than erythromycin. That’s because erythromycin inhibits the hepatic metabolism of the anticonvulsant, increasing the serum level of carbamazepine, while azithromycin does not interfere with carbamazepine metabolism.²²

TABLE 2
Clinically significant drug-drug interactions^21,22

 

Combination (effect)
Allopurinol and captopril (augments allopurinol’s effect)
Antidepressants (SSRIs, MAOIs) and antiepileptics (augments antidepressant effect)
Clopidogrel and omeprazole or esomeprazole (reduces clopidogrel’s effect)
Erythromycin and carbamazepine (augments carbamazepine’s effect)
Erythromycin and terfenadine (augments terfenadine’s effect)
Ketoconazole and PPIs (reduces ketoconazole’s absorption)
Levodopa and metoclopramide (augments levodopa’s effect)
MAOIs and narcotic analgesics (augments effects of both drugs)
Nitrates and phosphodiesterase type-5 inhibitors (augments effects of both drugs)
OCs and penicillins, phenobarbital, or tetracycline (reduces OCs’ effect)
Phenobarbital and simvastatin* (reduces simvastatin’s effect)
Quinolone and cation (reduces quinolone’s absorption and effect)
Repaglinide and diltiazem (augments repaglinide’s effect)
Simvastatin* or lovastatin and amiodarone or itraconazole (augments statin’s effect)
Theophylline and cimetidine or ciprofloxacin (augments theophylline’s effect)
MAOIs, monoamine oxidase inhibitors; OCs, oral contraceptives; PPIs, proton pump inhibitors; SSRIs, selective serotonin reuptake inhibitors.
*For a complete list of drugs that may interact with simvastatin, see US Food and Drug Administration.28

 

 

How is the drug metabolized?

DDIs may occur as a result of pharmaco-dynamic interaction (when 2 drugs act on the same receptor, site of action, or physiologic system) or pharmacokinetic changes (interference with absorption, albumin binding, distribution, metabolism, or elimination).²³ As already noted, age-related changes in pharmacokinetics and pharmacodynamics contribute to the high prevalence of DDIs in elderly patients.

In the liver, drug metabolism, particularly via the cytochrome P450 (CYP450) system, is the cornerstone of drug transformation.²³ Al-though the CYP system consists of “superfami-lies” with more than 100 types of enzymes, only a few are responsible for the majority of biotransformation.²³ The CYP system is also subject to genetic polymorphism, making some patients especially prone to DDIs.

P-glycoproteins (PGPs), which regulate drug absorption by transporting the drugs across cell membranes, also play a key role. PGP inhibitors or inducers help determine whether the accumulation of the molecule or the increased delivery of toxic metabolites leads to adverse effects.¹⁰

Reviewing the mechanism of action of any drug you prescribe for a patient taking other medications may alert you to a potential DDI—and the need to either switch the newly prescribed agent or alter the individual’s drug regimen in some other way.

CASE When John was readmitted to the hospital, he was taken off both the lovastatin and amiodarone and hydrated with forced alkaline diuresis. After a week, his symptoms resolved, and he was discharged soon after. His blood tests normalized 1 month later. The severe DDI he experienced occurred because lovastatin (which is metabolized primarily by CYP3A4) and amiodarone (a CYP3A4 inhibitor) were taken together. (Statins that are substrates of CYP3A4 have the greatest potential for interacting with drugs known to inhibit the CYP450 system [eg, cyclosporine, morphine derivatives, ketoconazole, and amiodarone].)

This adverse interaction could have been avoided if the physician who started John on amiodarone had been aware of the potential DDI—and switched him to an HMG-CoA inhibitor other than lovastatin. Pravastatin, which is not metabolized via CYP450, would have been an excellent choice.

Warfarin warrants special attention
Medications that have a particularly high potential for adverse interactions require special attention and patient monitoring, warfarin foremost among them. Warfarin metabolism and its anticoagulant effects can be dramatically changed if it is administered with a drug with a higher affinity for PGPs or an agent that competes with it within the CYP450 system.²⁴ Because of warfarin’s narrow therapeutic range, there are many drugs and drug classes that patients on warfarin should avoid (TABLE 3)—a fact that patients as well as their physicians need to be aware of.²⁴ Indeed, warfarin is often involved in drug-related hospital admissions for DDIs, especially in elderly patients and in those who are also taking nonsteroidal anti-inflamma-tory drugs (NSAIDs) or macrolides—2 of the many drug classes that patients taking warfarin should avoid.²⁴

TABLE 3
Patient on warfarin? Steer clear of these drugs^10,20,23

 

Drug class: agent(s)
Antiarrhythmics: amiodarone, propafenone
Antibiotics: ciprofloxacin, metronidazole, rifampin, trimethoprim/sulfamethoxazole
Anticonvulsants: carbamazepine, valproate
Antidepressants: fluoxetine, fluvoxamine, paroxetine, sertraline, trazodone
Antidiabetics: chlorpropamide
Antifungals: danazol, fluconazole, itraconazole, miconazole
Antimalarial agents: quinidine
Antineoplastics: azathioprine, fluorouracil, flutamide, ifosfamide, tamoxifen
Antiplatelet agents: ticlopidine
Antipsychotics: clozapine
Diuretics: spironolactone
GI drugs: cimetidine, esomeprazole, lansoprazole, omeprazole, pantoprazole, rabeprazole, ranitidine
Gout treatment: allopurinol
Hypolipidemics: atorvastatin, cholestyramine, ezetimibe, fenofibrate, fluvastatin, gemfibrozil, lovastatin, pravastatin, simvastatin
NSAIDs: aspirin, celecoxib, diclofenac, ibuprofen, indomethacin, ketoprofen, ketorolac, naproxen, piroxicam, sulindac
Thrombolytics: heparin, tissue plasminogen activator
Thyroid drugs: methimazole, propylthiouracil
Uricosuric agents: sulfinpyrazone
GI, gastrointestinal; NSAIDs, nonsteroidal anti-inflammatory drugs.

Keep an eye on these drug combinations, as well

Among the many combinations likely to result in clinically significant DDIs (TABLE 2), the following are worth mentioning:

 

Clopidogrel + certain proton pump inhibitors. The addition of a PPI to clopidogrel has been associated with a significant increase of recurrent infarction.²⁵ This may occur because clopidogrel is a prodrug and is converted in the liver to its active form by CYP2C19, an enzyme specifically inhibited by various PPIs—thereby altering the effectiveness of the antiplatelet agent. However, a recent analysis suggests that there is no need to avoid the concomitant use of a PPI and clopidogrel—and that the interference appears to be limited to omeprazole and esomeprazole.²⁶

Oral contraceptives (OCs) + penicillins, phenobarbital, or tetracycline. Each of these drugs reduces the effect of OCs, and women who are taking them concomitantly need to be advised to use another means of contraception.

Phenobarbital + simvastatin. Pheno-barbital (a CYP3A4 inducer) may reduce the efficacy of simvastatin.

Repaglinide + diltiazem. Diltiazem inhibits the metabolism of repaglinide (a CYP3A4 substrate), thus increasing the risk of hypoglycemia.

 

Simvastatin + amiodarone or itraconazole. Either of these antiarrhythmic agents decreases simvastatin metabolism, raising the risk of myopathy; with amiodarone, however, the likelihood of an adverse outcome is especially high. In 2008, the US Food and Drug Administration (FDA) issued a warning to healthcare professionals of the increased risk for rhabdomyolysis when simvastatin doses greater than 20 mg are administered together with amiodarone.²⁷ The agency issued a safety review of simvastatin, warning of its potential for DDIs with amiodarone and numerous other medications, earlier this year.²⁸

 

As John’s case illustrates, use of lovastatin with amiodarone should be avoided, as well.

Keep others safe: Report adverse events
When a DDI occurs despite your best efforts, you can help ensure that other patients do not experience the same adverse outcome by reporting it to MedWatch, the FDA’s voluntary safety information and adverse event reporting program. Go to https://www.accessdata.fda.gov/scripts/medwatch/medwatch-online.htm to file a report online.

CORRESPONDENCE Ignazio Grattagliano, MD, General Medicine, Department of Internal and Public Medicine, University of Bari, P.zza G. Cesare, 11 – 70124, Bari, Italy; [email protected]

 

CASE John L, a 63-year-old man taking lovastatin (40 mg/d) and ramipril (5 mg/d) for hypercholesterolemia and arterial hypertension was hospitalized with atrial fibrillation. Three days later, he was discharged, with a prescription for amiodarone (200 mg/d). After a month, he was readmitted to the hospital with dark urine and intensifying thigh weakness and achiness. Laboratory testing revealed aspartate aminotransferase and alanine aminotransferase levels 10 times the upper limit of normal, and elevated urine and serum myoglobin.

Drug-drug interactions (DDIs) like the one John experienced between lovastatin and amiodarone are a common cause of readmissions, as well as emergency department visits and hospitalizations, for everything from myopathy to electrolyte imbalance, gastrointestinal (GI) bleeding, hepatotoxicity, renal dysfunction, and changes in blood pressure and heart rate.^1-3

Yet many, if not most, DDIs can be avoided. With diligence and the right tools, you can do much to reduce the incidence of such interactions and adverse outcomes.

Polypharmacy and age pose the highest risks

The more medications a patient is taking, of course, the greater the likelihood of a clinically significant DDI. According to 1 study, 13% of patients taking 2 drugs develop a DDI; the incidence approaches 40% for patients taking 5 drugs, and exceeds 80% for patients taking 7 or more medications.⁴

In addition to polypharmacy, age alone is a key risk factor for DDIs.⁵ Pharmacokinetics and pharmacodynamics are frequently altered in older people, who may have slower intestinal transit time; diminished absorption capacity; decreased liver metabolism, mitochondrial function, and renal excretion; and alterations in volemia and body fat distribution.⁶ Although the speed at which these changes occur varies, aging is associated with a progressive deficiency in the regulation of most homeo-static mechanisms and an altered response to receptor stimulation.⁷

Whether age, multiple medications, or both are to blame, the impact on the elderly is striking. One recent retrospective study found 25% of elderly outpatients to be at risk for DDIs.⁸

Very young patients (<5 years) are also at risk for DDIs because of the immaturity of their enzymatic metabolic system.^5,9,10 Additional risk factors, detailed in TABLE 1, include the presence of an infection or other acute medical condition, a metabolic or endocrine disorder, and taking 1 or more drugs with a narrow therapeutic range. Women are also at higher risk for DDIs than their male counterparts, the result of a slower metabolic capacity and interference with sex hormones.^3-5,8,9 Pharmacogenetics may also play a big part in DDIs, and more and more studies are focusing on identifying patients at greatest risk.

TABLE 1
Risk factors for drug-drug interactions^3-5,8,9

 

Risk factor	Potential result
Acute medical condition (eg, dehydration, infection, alcoholism)	Augmented risk of elevated plasma drug concentration, increased catabolism, inhibition of hepatic drug metabolism
Age (very young [<5 years] and elderly)	Reduced metabolic capacity (greater accumulation of drugs)
Decreased renal and/or hepatic function	Decreased drug clearance/elimination; greater accumulation of drugs or their metabolites
Drug(s) with narrow therapeutic range	Increased risk for dose-related side effects
Female sex	Reduced metabolic capacity, interference with sex hormones
Metabolic or endocrine conditions (eg, fatty liver, obesity, hypothyroidism)	Altered hepatic metabolism, increased body distribution volumes, augmented risk of accumulation for hydrophobic molecules
Polypharmacy (≥3 medications)	Increased risk of metabolic and/or pharmacodynamic interference
Pharmacogenetics	Altered metabolic capacity (greater accumulation of drugs or their metabolites)

Lack of coordinated care also increases risk
Another risk factor involves the use of multiple providers.⁹ A woman may be treated by—and receive prescriptions from—an endocrinolo-gist, a gynecologist, and a family physician (FP), for instance, and get medications from a local pharmacy, a nationwide discount chain, and a mail order pharmacy. As with the number of medications being taken, the greater the number of health care professionals a patient sees, the higher the risk.

To mitigate the risk, encourage your patients to fill all their prescriptions at the same pharmacy—and for your part, take a complete medication history before writing a new prescription.

Medication history in doubt? Schedule a “brown bag review”
Ask patients to provide the name and dose of every medication they’re taking. Inquire specifically about over-the-counter (OTC) cough and cold remedies and complementary and alternative medicines, including herbal remedies, vitamins, and supplements. Patients often neglect to mention nonprescription remedies and may not even think of them as medicine, but OTC products with the potential to interact adversely with prescription drugs may otherwise remain undetected.^11-13

 

Consider a “brown bag review” for patients who don’t know what dosage they’re taking or have difficulty identifying the drugs other physicians have prescribed. Ask them to put all their medications in a brown bag and bring them in on their next visit.^14,15

 

Steps to take to reduce risk

Software systems. A number of free and low-cost software systems identify potential DDIs (See “Check for drug interactions: Software programs to consider”). While such electronic programs can indeed lower the risk,^16-18 they cannot be counted on to detect or avert every possible adverse interaction.

The downside. One problem is that some software programs fail to distinguish between clinically significant and nonsignificant interactions, causing some prescribers to override system alerts—and possibly miss an important warning.¹⁹ Another problem: While most systems do an excellent job of checking to see whether 2 drugs can be safely taken together, few are capable of checking for all potential interactions among multiple medications. What’s more, many drugs have not been evaluated for their potential to interact with other agents, so the absence of reported interactions is no guarantee of a lack of DDIs.

 

Other strategies to consider:

Minimize the number of prescriptions. While it may not be possible to avoid prescribing a new agent for a patient who is already taking multiple medications, limiting the number of new drugs to those that are absolutely essential will help to minimize DDIs. Whenever possible, select a compound with the desired effects. Prescribe a single agent with antihypertensive as well as uricosuric effects for a patient with elevated blood pressure and uric acid levels rather than 2 different drugs (eg, losartan instead of an anti-hypertensive agent plus allopurinol).

Alter the dosing regimen. Several active molecules may cause DDIs by interfering with intestinal absorption or GI transit time if they’re taken closely together. For example, a quinolone should not be administered at the same time as a cation because of possible chelation in the GI tract. If a patient needs both, however, you may be able to avert a DDI by advising the patient to take them at least 2 hours apart.²⁰ Another possibility is to temporarily discontinue a maintenance medication if it has the potential to interact adversely with a drug that is needed for only a short duration.

Choose a different drug (or drug class). Some drug classes should never be mixed—nitrates and phosphodiesterase type-5 inhibitors, taken together, greatly increase the risk of vasodilation and may result in severe hypotension, for example. There are also drug classes with a low potential for DDIs, including cholinesterase inhibitors and anti-hypertensives (angiotensin-converting enzyme inhibitors, angiotensin receptor blockers, beta-blockers, and thiazides).²¹

 

Frequently, though, medications within the same drug class do not share the same potential for DDIs. In such cases, an adverse outcome can often be averted by being aware of combinations likely to result in clinically significant DDIs (TABLE 2) and, whenever possible, prescribing another agent. If a patient taking carbamazepine needs a macro-lide antibiotic, for instance, azithromycin is a better choice than erythromycin. That’s because erythromycin inhibits the hepatic metabolism of the anticonvulsant, increasing the serum level of carbamazepine, while azithromycin does not interfere with carbamazepine metabolism.²²

TABLE 2
Clinically significant drug-drug interactions^21,22

 

Combination (effect)
Allopurinol and captopril (augments allopurinol’s effect)
Antidepressants (SSRIs, MAOIs) and antiepileptics (augments antidepressant effect)
Clopidogrel and omeprazole or esomeprazole (reduces clopidogrel’s effect)
Erythromycin and carbamazepine (augments carbamazepine’s effect)
Erythromycin and terfenadine (augments terfenadine’s effect)
Ketoconazole and PPIs (reduces ketoconazole’s absorption)
Levodopa and metoclopramide (augments levodopa’s effect)
MAOIs and narcotic analgesics (augments effects of both drugs)
Nitrates and phosphodiesterase type-5 inhibitors (augments effects of both drugs)
OCs and penicillins, phenobarbital, or tetracycline (reduces OCs’ effect)
Phenobarbital and simvastatin* (reduces simvastatin’s effect)
Quinolone and cation (reduces quinolone’s absorption and effect)
Repaglinide and diltiazem (augments repaglinide’s effect)
Simvastatin* or lovastatin and amiodarone or itraconazole (augments statin’s effect)
Theophylline and cimetidine or ciprofloxacin (augments theophylline’s effect)
MAOIs, monoamine oxidase inhibitors; OCs, oral contraceptives; PPIs, proton pump inhibitors; SSRIs, selective serotonin reuptake inhibitors.
*For a complete list of drugs that may interact with simvastatin, see US Food and Drug Administration.28

 

 

How is the drug metabolized?

DDIs may occur as a result of pharmaco-dynamic interaction (when 2 drugs act on the same receptor, site of action, or physiologic system) or pharmacokinetic changes (interference with absorption, albumin binding, distribution, metabolism, or elimination).²³ As already noted, age-related changes in pharmacokinetics and pharmacodynamics contribute to the high prevalence of DDIs in elderly patients.

In the liver, drug metabolism, particularly via the cytochrome P450 (CYP450) system, is the cornerstone of drug transformation.²³ Al-though the CYP system consists of “superfami-lies” with more than 100 types of enzymes, only a few are responsible for the majority of biotransformation.²³ The CYP system is also subject to genetic polymorphism, making some patients especially prone to DDIs.

P-glycoproteins (PGPs), which regulate drug absorption by transporting the drugs across cell membranes, also play a key role. PGP inhibitors or inducers help determine whether the accumulation of the molecule or the increased delivery of toxic metabolites leads to adverse effects.¹⁰

Reviewing the mechanism of action of any drug you prescribe for a patient taking other medications may alert you to a potential DDI—and the need to either switch the newly prescribed agent or alter the individual’s drug regimen in some other way.

CASE When John was readmitted to the hospital, he was taken off both the lovastatin and amiodarone and hydrated with forced alkaline diuresis. After a week, his symptoms resolved, and he was discharged soon after. His blood tests normalized 1 month later. The severe DDI he experienced occurred because lovastatin (which is metabolized primarily by CYP3A4) and amiodarone (a CYP3A4 inhibitor) were taken together. (Statins that are substrates of CYP3A4 have the greatest potential for interacting with drugs known to inhibit the CYP450 system [eg, cyclosporine, morphine derivatives, ketoconazole, and amiodarone].)

This adverse interaction could have been avoided if the physician who started John on amiodarone had been aware of the potential DDI—and switched him to an HMG-CoA inhibitor other than lovastatin. Pravastatin, which is not metabolized via CYP450, would have been an excellent choice.

Warfarin warrants special attention
Medications that have a particularly high potential for adverse interactions require special attention and patient monitoring, warfarin foremost among them. Warfarin metabolism and its anticoagulant effects can be dramatically changed if it is administered with a drug with a higher affinity for PGPs or an agent that competes with it within the CYP450 system.²⁴ Because of warfarin’s narrow therapeutic range, there are many drugs and drug classes that patients on warfarin should avoid (TABLE 3)—a fact that patients as well as their physicians need to be aware of.²⁴ Indeed, warfarin is often involved in drug-related hospital admissions for DDIs, especially in elderly patients and in those who are also taking nonsteroidal anti-inflamma-tory drugs (NSAIDs) or macrolides—2 of the many drug classes that patients taking warfarin should avoid.²⁴

TABLE 3
Patient on warfarin? Steer clear of these drugs^10,20,23

 

Drug class: agent(s)
Antiarrhythmics: amiodarone, propafenone
Antibiotics: ciprofloxacin, metronidazole, rifampin, trimethoprim/sulfamethoxazole
Anticonvulsants: carbamazepine, valproate
Antidepressants: fluoxetine, fluvoxamine, paroxetine, sertraline, trazodone
Antidiabetics: chlorpropamide
Antifungals: danazol, fluconazole, itraconazole, miconazole
Antimalarial agents: quinidine
Antineoplastics: azathioprine, fluorouracil, flutamide, ifosfamide, tamoxifen
Antiplatelet agents: ticlopidine
Antipsychotics: clozapine
Diuretics: spironolactone
GI drugs: cimetidine, esomeprazole, lansoprazole, omeprazole, pantoprazole, rabeprazole, ranitidine
Gout treatment: allopurinol
Hypolipidemics: atorvastatin, cholestyramine, ezetimibe, fenofibrate, fluvastatin, gemfibrozil, lovastatin, pravastatin, simvastatin
NSAIDs: aspirin, celecoxib, diclofenac, ibuprofen, indomethacin, ketoprofen, ketorolac, naproxen, piroxicam, sulindac
Thrombolytics: heparin, tissue plasminogen activator
Thyroid drugs: methimazole, propylthiouracil
Uricosuric agents: sulfinpyrazone
GI, gastrointestinal; NSAIDs, nonsteroidal anti-inflammatory drugs.

Keep an eye on these drug combinations, as well

Among the many combinations likely to result in clinically significant DDIs (TABLE 2), the following are worth mentioning:

 

Clopidogrel + certain proton pump inhibitors. The addition of a PPI to clopidogrel has been associated with a significant increase of recurrent infarction.²⁵ This may occur because clopidogrel is a prodrug and is converted in the liver to its active form by CYP2C19, an enzyme specifically inhibited by various PPIs—thereby altering the effectiveness of the antiplatelet agent. However, a recent analysis suggests that there is no need to avoid the concomitant use of a PPI and clopidogrel—and that the interference appears to be limited to omeprazole and esomeprazole.²⁶

Oral contraceptives (OCs) + penicillins, phenobarbital, or tetracycline. Each of these drugs reduces the effect of OCs, and women who are taking them concomitantly need to be advised to use another means of contraception.

Phenobarbital + simvastatin. Pheno-barbital (a CYP3A4 inducer) may reduce the efficacy of simvastatin.

Repaglinide + diltiazem. Diltiazem inhibits the metabolism of repaglinide (a CYP3A4 substrate), thus increasing the risk of hypoglycemia.

 

Simvastatin + amiodarone or itraconazole. Either of these antiarrhythmic agents decreases simvastatin metabolism, raising the risk of myopathy; with amiodarone, however, the likelihood of an adverse outcome is especially high. In 2008, the US Food and Drug Administration (FDA) issued a warning to healthcare professionals of the increased risk for rhabdomyolysis when simvastatin doses greater than 20 mg are administered together with amiodarone.²⁷ The agency issued a safety review of simvastatin, warning of its potential for DDIs with amiodarone and numerous other medications, earlier this year.²⁸

 

As John’s case illustrates, use of lovastatin with amiodarone should be avoided, as well.

Keep others safe: Report adverse events
When a DDI occurs despite your best efforts, you can help ensure that other patients do not experience the same adverse outcome by reporting it to MedWatch, the FDA’s voluntary safety information and adverse event reporting program. Go to https://www.accessdata.fda.gov/scripts/medwatch/medwatch-online.htm to file a report online.

CORRESPONDENCE Ignazio Grattagliano, MD, General Medicine, Department of Internal and Public Medicine, University of Bari, P.zza G. Cesare, 11 – 70124, Bari, Italy; [email protected]

Each year, venous thrombosis develops in approximately 1 in 1000 people.¹ The cause: An alteration in blood composition, venous stasis, or vascular damage—commonly known as Virchow’s triad. Changes in blood composition are associated with hereditary thrombophilias, such as factor V Leiden mutation or a deficiency in protein C or S or antithrombin III (AT). Changes in blood flow (stasis) and vessel damage stem from acquired conditions that commonly lead to hypercoagulability—pregnancy, malignancy, and estrogen use among them.

Regardless of the reason a patient is at elevated risk, however, the goal is the same: to prevent the development of thrombi, thereby reducing the increased morbidity and mortality associated with thromboembolism. Achieving that goal requires an understanding of both inherited and acquired risk factors, familiarity with diagnostic tools, and knowledge of appropriate treatment. This review, which begins with hereditary hypercoagulable states before turning to acquired conditions associated with hypercoagulability, will help toward that end.

Keep these thrombophilias on your radar screen

Factor V Leiden mutation is the most common inherited hypercoagulable disorder, and the most common form of activated protein C resistance. The mutation is found in an estimated 5% to 10% of the general population.² Among patients with thromboembolic disorders, however, the incidence is considerably higher, with estimates ranging from 21% to 60%.^3,4 Factor V Leiden mutation is more prevalent among Caucasian populations, and rarely found in people of Asian or African descent.²

Prothrombin G20210A mutation, an autosomal-dominant disorder, is the second most common inherited hypercoagulable state.⁴ This mutation is associated with an increase in prothrombin levels, causing an elevation in thrombin and, in turn, a heightened risk of thrombosis.

Although prothrombin G20210A mutation is found in only 1% to 2% of the general population, its prevalence among those with a history of thromboembolic events is estimated at 5% to 19%.⁴ This disorder, too, varies significantly by ethnicity: People from southern Europe are twice as likely to be affected as northern Europeans, and the mutation is rarely found in people of Asian or African descent.^2,5

Protein C deficiency. Protein C, a vitamin K-dependent anticoagulant produced in the liver, is activated when thrombomodulin binds with thrombin in the presence of protein S, which serves as the cofactor. Protein C is required to inactivate clotting factors V and VIII.² The deficiency is an autosomal-dominant disorder, and is more likely to result in venous than arterial thrombosis.²

Protein C deficiency affects 1 in every 200 to 500 people in the general population; among patients with a history of venous thrombosis, its prevalence is 2% to 9%.^2,3 Generally, people with a protein C deficiency begin developing thrombi in their late teens, and about 75% suffer from 1 or more thrombotic events during the course of their lives.²

There are 2 types of protein C deficiency: Patients with type I have a decreased production of protein C, while those with type II have normal levels of the protein, but in a form that is dysfunctional.⁶

AT deficiency. AT, a natural anticoagulant synthesized by the liver and endothelial cells, is responsible for inactivating several clotting factors, including thrombin and factors IXa, Xa, XIa, and XIIa. Like protein C deficiency, AT deficiency is an autosomal-dominant disorder with 2 subtypes. Individuals with type I deficiency have normal plasma levels of AT, but the anticoagulant has reduced biological activity or is dysfunctional; those with type II deficiency have decreased plasma levels of fully functional AT.² Both types are more likely to lead to venous than arterial thrombosis.

AT deficiency affects approximately 1 in 2000 to 5000 people, including 2.8% of patients who develop venous thrombosis.⁷ Nearly two-thirds of those with AT deficiency develop thrombi before the age of 35.²

Protein S deficiency. Endothelial cells are responsible for the synthesis of protein S, which, like protein C (for which it serves as a cofactor), is vitamin K-dependent. Protein S deficiency, also an autosomal-dominant disorder, has 3 subtypes: Type I, also known as classical deficiency, is characterized by reduced free and total levels of functional protein S; type II patients have a normal total level of protein S, but a decreased amount of free protein; and type III patients have normal levels of both free and total protein S, but the available proteins are dysfunctional.^2,6

 

The prevalence of this hypercoagulable state in the general population is unknown, and protein S deficiencies have been found in only 1% of patients with a history of deep vein thrombosis (DVT).^3,8 Although this hypercoagulable state is less common than other hereditary thrombophilias, 74% of people with this disorder develop DVT—half of them before the age of 25.²

Hyperhomocysteinemia. Elevations in homocysteine may occur as a result of a hereditary disorder (deficiencies in cystathionine beta-synthase or methylene-tetrahydrofolate reductase). Hyperhomocysteinemia may also be an acquired condition, associated with deficiencies in vitamins B6, B12, or folic acid; chronic kidney disease; hypothyroidism; and certain malignancies.

The prevalence of hyperhomocysteinemia varies, based on the underlying disorder. Only about 0.3% of the general population has a cystathionine beta-synthase deficiency. Methylene-tetrahydrofolate reductase deficiency, however, is common among Italian and Hispanic populations (occurring in about 20%), but rare (<1%) among African American people.³

When to test for thrombophilias
Idiopathic venous thrombosis is probably the most common reason for ordering testing for inherited hypercoagulable states, and an underlying thrombophilia is found in about 50% of cases.⁹ Other indications for testing include thrombus development in an unusual site (eg, splanchic, renal, retinal, or ovarian veins; cerebral venous sinuses; or upper limbs), recurrent venous thromboembolism (VTE), venous thrombosis at an early age (<45 years) or in a patient with a strong family history of VTE, and unexplained recurrent pregnancy loss.

Testing may also be considered for relatives of patients with known inherited hypercoagulability disorders, but this should be done only if the results could affect a treatment decision: Can a man safely undergo surgery with a high postoperative risk for thrombosis, for example? Should a woman take oral contraceptives (OCs), start hormone replacement therapy (HRT), or attempt another pregnancy?

Overall, testing for inherited hypercoagulable states should focus on the identification of individuals most likely to benefit from it; only tests yielding useful data should be performed. Testing of asymptomatic individuals for the sole purpose of initiating long-term prophylactic therapy is not recommended.

Which tests for which patients?
In some cases, selective assays may be more useful than test panels, for reasons associated with patient presentation as well as cost. Proper selection of specific tests should be individualized based on the patient’s age, thrombotic presentation, and family history, and on potential effects on patient management.

For patients with heparin resistance, cerebral vein thrombosis, intra-abdominal vein thrombosis, or recurrent superficial thrombophlebitis, AT testing may be in order. Patients with recurrent superficial thrombophlebitis may also benefit from testing for factor V Leiden mutation, and protein C and protein S testing may be beneficial for patients with warfarin skin necrosis, recurrent superficial thrombophlebitis, or neonatal purpura fulminans. Cerebral vein thrombosis in the general population and in women using OCs, in particular, is suggestive of a G20210A mutation,⁹ and is a possible indication for testing. Hyperhomocysteinemia testing may be considered for patients with premature arterial and venous thrombosis, as well as mental retardation, skeletal abnormalities, and vitamin B6 or B12 deficiency.

Many physicians prefer to order testing in stages, starting with tests for the most common thrombophilias. When ordering tests for the conditions detailed above and in the TABLE, be aware that test panels vary among facilities, so it may be necessary to check with the testing laboratory to ensure that it offers the tests that are indicated for a particular patient.

TABLE
Suspect a hereditary hypercoagulable disorder? Testing considerations to keep in mind^{2,3,8,9,11,12}

Disorder	Potential indication(s) for testing	Timing of test	Interaction with warfarin	Interaction with LMWH and UFH
Factor V Leiden mutation	Recurrent superficial thrombophlebitis	Not during an acute event	Modified aPTT test does not interact	Modified aPTT test does not interact
Prothrombin G20210A mutation	Cerebral vein thrombosis	Not during an acute event	No	No
Protein C deficiency	Warfarin skin necrosis Recurrent superficial thrombophlebitis Neonatal purpura fulminans	7-10 days after cessation of warfarin therapy	Yes	No
Antithrombin III deficiency	Heparin resistance Cerebral vein thrombosis Intra-abdominal vein thrombosis Recurrent superficial thrombophlebitis	7-10 days after cessation of warfarin therapy	Yes	Yes
Protein S deficiency	Warfarin skin necrosis Recurrent superficial thrombophlebitis Neonatal purpura fulminans	7-10 days after cessation of warfarin therapy	Yes	No
Hyperhomocysteinemia	Premature arterial/venous thrombosis Skeletal abnormalities Vitamin B6 or B12 deficiency	Not during an acute event	No	No
aPTT, activated partial thromboplastin time; LMWH, low-molecular-weight heparin; UFH, unfractionated heparin.

Timing of tests, and other specifics
Acute thrombosis, anticoagulation therapy, and some disease states—eg, liver disease, nephritic syndrome, disseminated intravascular coagulation, and acute illness—can affect levels of AT, protein C, and protein S. Within the first 48 hours of warfarin therapy, for example, a patient’s protein C and protein S levels decline about 50%; after 2 weeks, the levels rise to about 70% of their normal range. Because of warfarin’s effect on these proteins, evaluation for these deficiencies should be performed at least 1 week to 10 days after cessation of a 3- to 6-month course of anticoagulation therapy. Abnormal findings should be confirmed with a second test approximately 3 weeks later.^2,3,9

 

Several tests are used to detect factor V Leiden mutation, including the polymerase chain reaction-based test and the modified activated partial thromboplastin time (aPTT)-based test. The latter can be given to patients who are receiving anticoagulants. The results indicate the ratio of activated protein C (APC), and a finding of <2.0 is considered abnormal.⁸ The presence of lupus anticoagulant—autoantibodies that bind to phospholipids and proteins associated with the cell membrane—may yield a false-positive result on the modified aPTT test.

While the prothrombin G20210A mutation is linked to elevations in prothrombin, measuring prothrombin levels is not an accurate way to test for this hypercoagulable state. The prothrombin G20210A mutation generally is detected through DNA analysis instead.^2,8

Hyperhomocysteinemia is diagnosed by measuring fasting plasma levels of homocysteine. Although the results are not standardized, they generally correlate with the severity of disease (mild, 15-30 μmol/L; moderate, >30-100 μmol/L; severe, >100 μmol/L). However, plasma levels may be falsely elevated during an acute thrombotic episode and decreased by supplementation with folic acid, B6, and B12. The decrease, however, does not indicate a reduction in the risk of thrombosis.

Treatment, yes, but for how long?
Most thrombotic events are initially treated with a combination of a heparin product (low-molecular-weight heparin [LMWH] or unfractionated heparin [UFH]) and warfarin—commonly known as bridging therapy. The heparin is discontinued once the patient’s international normalized ratio (INR) has been maintained at a therapeutic level for more than 24 hours, which usually takes about 5 days. Warfarin therapy, however, should continue for at least 3 to 6 months, depending on the severity and cause of the thrombosis.

Individuals with an AT, protein C, or protein S deficiency may be at increased risk of recurrence, and long-term anticoagulant therapy may be warranted.⁵ Under current guidelines from the American College of Chest Physicians (ACCP), however, hereditary thrombophilias are not considered to be major determinants of recurrence or a major factor guiding the duration of anticoagulation therapy.¹⁰ Practitioners must use their judgment to determine the need for treatment; if anticoagulation therapy is initiated, the duration should be based on an individual assessment of benefit and risk.

Recognizing and responding to acquired risks

Several conditions associated with vessel wall changes and venous stasis—the hallmarks of acquired hypercoagulable states—put patients at increased risk of venous thrombosis. Following is a review of the most likely risk factors, including antiphospholipid antibody syndrome (APS), previously known as lupus anticoagulant syndrome; heparin-induced thrombocytopenia (HIT); pregnancy; trauma; estrogen; and malignancy.

When to test for—and treat—APS
APS, a systemic autoimmune disorder that can result in arterial or venous thrombosis or pregnancy loss and morbidity, is characterized by the presence of autoantibodies. Patients of all ages may be affected by APS, one of the most common acquired hypercoagulability disorders. APS affects an estimated 28% of the general population.² About 15% of recurrent pregnancy losses and 20% of recurrent thromboses in young adults are attributed to this autoimmune disorder.¹¹

A definitive diagnosis of APS requires a history of either vascular thrombosis or pregnancy morbidity—defined as miscarriage after the 10th gestational week, consecutive fetal losses before the 10th gestational week, or placental insufficiency and premature birth before 34 weeks.^11,12 APS testing may be useful in patients with cerebral vein thrombosis, intra-abdominal vein thrombosis, or unexplained recurrent fetal loss.⁹

In addition to clinical criteria, a diagnosis of APS is based on the presence of plasma antibodies on 2 or more occasions at least 12 weeks apart. APS encompasses 3 types of antiphospholipid antibodies—lupus anticoagulant antibodies, anticardiolipin antibodies, and anti-beta 2-glycoprotein I antibodies—which can be detected with 2 different tests. Coagulation assays are used to identify lupus anticoagulant antibodies because they prolong clotting time; however, immunoassays are used to measure immunologic reactivity to phospholipids to determine the presence of anticardiolipin antibodies and anti-beta 2-glycoprotein I antibodies.^11,12

Treatment for APS generally involves anticoagulant therapy for the prevention and treatment of acute thrombotic events, or as prophylaxis during pregnancy. ACCP guidelines call for initiating warfarin therapy with a target INR of 2.5 (range 2.0-3.0) in patients with no other risk factors. In patients who have had recurrent thromboembolic events or have additional risk factors, a target INR of 3.0 (range 2.5-3.5) is suggested.¹⁰ LMWH and UFH are also options for use in the event of recurrence or for prophylaxis during pregnancy.^11,13,14

 

HIT can be benign, or life threatening
HIT—defined by a decrease in platelet count to less than 150,000 (or a 50% drop from baseline) after initiation of heparin therapy—may or may not be benign. HIT type I (previously called heparin-associated thrombocytopenia), which affects approximately 10% of patients treated with heparin, is transient, asymptomatic, and not associated with an increased risk of thrombosis. Type I typically occurs within the first 2 days of heparin therapy.¹⁵

HIT type II is an immune-mediated response that does increase the risk of thrombosis. Patients usually develop type II 5 to 12 days after initiation of heparin, but in rare instances onset is delayed, occurring up to 40 days after heparin therapy. Approximately 5% of patients on heparin develop HIT type II, and the risk increases with frequent heparin use. Unlike other states of thrombocytopenia, HIT rarely causes bleeding. However, patients with HIT type II are at risk for a paradoxical thrombotic syndrome that may become life threatening.^13,15

To diagnose HIT, an enzyme-linked immunosorbent assay or other specific blood tests must be used to confirm the presence of circulating antibodies.¹³ The diagnosis is based on the following criteria: (1) thrombocytopenia, (2) exclusion of other possible causes of thrombocytopenia, and (3) resolution of thrombocytopenia after discontinuation of heparin.^13,15

When HIT is suspected, all heparin-containing products must be discontinued immediately and alternative anticoagulant therapy (typically, with danaparoid, lepirudin, or argatroban) should be initiated to reduce the risk of thrombosis. Warfarin alone should not be used for the treatment of HIT because of its association with worsening thrombosis and venous limb gangrene.¹³ However, warfarin should be initiated while the patient is receiving danaparoid or a thrombin-specific inhibitor—with at least 5 days of overlapping therapy recommended. Duration of therapy has not been well defined, but an overall course of at least 2 to 3 months is recommended to reduce the risk of recurrent thrombosis.^13,15

Pregnancy raises risk, but limits Tx options
By altering the body’s normal physiologic state in a way that may lead to hypercoagulability, pregnancy increases the risk of VTE 6-fold.¹⁶ The risk continues throughout pregnancy and peaks during puerperium, the 6-week period after delivery. Cesarean delivery, prolonged immobility, and obesity elevate the risk.^13,16

Treatment options for acute thrombotic events during pregnancy are limited because warfarin is contraindicated. Current ACCP guidelines recommend substituting UFH or LMWH for oral anticoagulant therapy when treatment for an acute thrombotic event is required.¹⁰ While no pharmacologic prophylaxis is currently recommended for pregnant patients with thrombophilias but no history of thrombotic events,^10,17 there are cases when it may be necessary. Patients with an AT deficiency, for example, may require prenatal and postpartum prophylaxis,¹⁰ and patients who deliver by cesarean and have 1 or more additional risk factors should receive prophylaxis for the duration of their hospitalization. Women with multiple risk factors, in addition to pregnancy and cesarean section, should receive pharmacologic prophylaxis for up to 4 to 6 weeks postpartum.

Choices for postpartum anticoagulation for the treatment of an acute thrombotic event should be based on guidelines developed for nonpregnant patients, with this caveat: For women who are breastfeeding, safety during lactation must be considered. Neither UFH nor warfarin enters breast milk, so both are safe for such patients.¹⁸ LMWH, although smaller in molecular weight than UFH, is considered moderately safe during breastfeeding, as well.¹⁸

Prophylaxis is vital for trauma patients
Trauma and major injuries increase the risk of thrombosis by approximately 50%. Patients who are hospitalized after a major trauma are at high risk for the development of a VTE. At the greatest risk are those with spinal cord injuries (62%), pelvic fractures (61%), and leg fractures (80%).¹⁶ Current ACCP guidelines recommend the use of LMWH as soon as it is safe for trauma patients, and continuing it until discharge in patients with no apparent contraindication. If a patient has an active bleed or other contraindication, mechanical thromboprophylaxis is indicated until the bleeding risk decreases.¹⁵

Estrogens increase platelet aggregation
Estrogens are considered a risk factor because of their effect on both natural anticoagulants and clotting factors. A reduction in AT activity and increasing concentrations of clotting factors VII, X, and XII result from the use of estrogens. Estrogen is also thought to be responsible for the increase in platelet count and aggregation associated with the use of combination OCs.¹⁹ In fact, OC use is associated with about a 3-fold overall risk of thrombosis, a risk reported to be highest during the first year of use.¹⁶ Among hormonal contraceptives, the transdermal formulation has the highest risk.^20-22 Hormone replacement therapy taken during menopause confers approximately a 2- to 4-fold increase in risk for VTE, and selective estrogen receptor modulators are associated with a 2-fold risk.¹⁶ Physicians should educate patients about the risks associated with these agents and signs and symptoms of thrombosis.

 

Cancer and hypercoagulability: Which patients need treatment?
Although the pathophysiologic process is not fully understood, a link between cancer and hypercoagulability has long been recognized. In fact, malignancy—the second most common cause of acquired hypercoagulability—is associated with 10% to 20% of spontaneous DVTs.²

One possible mechanism is the interaction of tumor cells with thrombin and plasmin-generating systems, directly influencing thrombus formation.⁷

Cancer patients also have an elevated risk for thrombosis related to immobilization, infection, treatment with antineoplastic agents, surgery, and the insertion of central venous catheters. Approximately 30% of patients with central venous catheters develop a DVT of the arm.¹⁶

Anticoagulant therapy in cancer patients varies, depending on the severity and circumstances of the patient. According to American Society of Clinical Oncology, National Comprehensive Cancer Center Network, and ACCP guidelines, LMWH is the preferred initial treatment for thromboses in patients with cancer— that is, in the first 3 to 6 months of therapy after a thrombotic event.^17,23,24 The guidelines also mention warfarin as an alternative for long-term (>6 months) anticoagulant therapy, if no contraindications exist.

Because cancer is usually a long-term illness, anticoagulant therapy should be continued indefinitely, or until the cancer has resolved.^10,23,24 Prophylaxis is recommended for cancer patients who are bedridden with an acute medical illness, but should not be routinely used in patients with indwelling venous catheters or those receiving chemotherapy or hormonal therapy.^13,15

CORRESPONDENCE Haley M. Phillippe, PharmD, BCPS, University of Alabama Birmingham School of Medicine-Huntsville Campus, Department of Family Medicine, 301 Governors Drive, Huntsville, AL 35801; [email protected]

Each year, venous thrombosis develops in approximately 1 in 1000 people.¹ The cause: An alteration in blood composition, venous stasis, or vascular damage—commonly known as Virchow’s triad. Changes in blood composition are associated with hereditary thrombophilias, such as factor V Leiden mutation or a deficiency in protein C or S or antithrombin III (AT). Changes in blood flow (stasis) and vessel damage stem from acquired conditions that commonly lead to hypercoagulability—pregnancy, malignancy, and estrogen use among them.

Regardless of the reason a patient is at elevated risk, however, the goal is the same: to prevent the development of thrombi, thereby reducing the increased morbidity and mortality associated with thromboembolism. Achieving that goal requires an understanding of both inherited and acquired risk factors, familiarity with diagnostic tools, and knowledge of appropriate treatment. This review, which begins with hereditary hypercoagulable states before turning to acquired conditions associated with hypercoagulability, will help toward that end.

Keep these thrombophilias on your radar screen

Factor V Leiden mutation is the most common inherited hypercoagulable disorder, and the most common form of activated protein C resistance. The mutation is found in an estimated 5% to 10% of the general population.² Among patients with thromboembolic disorders, however, the incidence is considerably higher, with estimates ranging from 21% to 60%.^3,4 Factor V Leiden mutation is more prevalent among Caucasian populations, and rarely found in people of Asian or African descent.²

Prothrombin G20210A mutation, an autosomal-dominant disorder, is the second most common inherited hypercoagulable state.⁴ This mutation is associated with an increase in prothrombin levels, causing an elevation in thrombin and, in turn, a heightened risk of thrombosis.

Although prothrombin G20210A mutation is found in only 1% to 2% of the general population, its prevalence among those with a history of thromboembolic events is estimated at 5% to 19%.⁴ This disorder, too, varies significantly by ethnicity: People from southern Europe are twice as likely to be affected as northern Europeans, and the mutation is rarely found in people of Asian or African descent.^2,5

Protein C deficiency. Protein C, a vitamin K-dependent anticoagulant produced in the liver, is activated when thrombomodulin binds with thrombin in the presence of protein S, which serves as the cofactor. Protein C is required to inactivate clotting factors V and VIII.² The deficiency is an autosomal-dominant disorder, and is more likely to result in venous than arterial thrombosis.²

Protein C deficiency affects 1 in every 200 to 500 people in the general population; among patients with a history of venous thrombosis, its prevalence is 2% to 9%.^2,3 Generally, people with a protein C deficiency begin developing thrombi in their late teens, and about 75% suffer from 1 or more thrombotic events during the course of their lives.²

There are 2 types of protein C deficiency: Patients with type I have a decreased production of protein C, while those with type II have normal levels of the protein, but in a form that is dysfunctional.⁶

AT deficiency. AT, a natural anticoagulant synthesized by the liver and endothelial cells, is responsible for inactivating several clotting factors, including thrombin and factors IXa, Xa, XIa, and XIIa. Like protein C deficiency, AT deficiency is an autosomal-dominant disorder with 2 subtypes. Individuals with type I deficiency have normal plasma levels of AT, but the anticoagulant has reduced biological activity or is dysfunctional; those with type II deficiency have decreased plasma levels of fully functional AT.² Both types are more likely to lead to venous than arterial thrombosis.

AT deficiency affects approximately 1 in 2000 to 5000 people, including 2.8% of patients who develop venous thrombosis.⁷ Nearly two-thirds of those with AT deficiency develop thrombi before the age of 35.²

Protein S deficiency. Endothelial cells are responsible for the synthesis of protein S, which, like protein C (for which it serves as a cofactor), is vitamin K-dependent. Protein S deficiency, also an autosomal-dominant disorder, has 3 subtypes: Type I, also known as classical deficiency, is characterized by reduced free and total levels of functional protein S; type II patients have a normal total level of protein S, but a decreased amount of free protein; and type III patients have normal levels of both free and total protein S, but the available proteins are dysfunctional.^2,6

 

The prevalence of this hypercoagulable state in the general population is unknown, and protein S deficiencies have been found in only 1% of patients with a history of deep vein thrombosis (DVT).^3,8 Although this hypercoagulable state is less common than other hereditary thrombophilias, 74% of people with this disorder develop DVT—half of them before the age of 25.²

Hyperhomocysteinemia. Elevations in homocysteine may occur as a result of a hereditary disorder (deficiencies in cystathionine beta-synthase or methylene-tetrahydrofolate reductase). Hyperhomocysteinemia may also be an acquired condition, associated with deficiencies in vitamins B6, B12, or folic acid; chronic kidney disease; hypothyroidism; and certain malignancies.

The prevalence of hyperhomocysteinemia varies, based on the underlying disorder. Only about 0.3% of the general population has a cystathionine beta-synthase deficiency. Methylene-tetrahydrofolate reductase deficiency, however, is common among Italian and Hispanic populations (occurring in about 20%), but rare (<1%) among African American people.³

When to test for thrombophilias
Idiopathic venous thrombosis is probably the most common reason for ordering testing for inherited hypercoagulable states, and an underlying thrombophilia is found in about 50% of cases.⁹ Other indications for testing include thrombus development in an unusual site (eg, splanchic, renal, retinal, or ovarian veins; cerebral venous sinuses; or upper limbs), recurrent venous thromboembolism (VTE), venous thrombosis at an early age (<45 years) or in a patient with a strong family history of VTE, and unexplained recurrent pregnancy loss.

Testing may also be considered for relatives of patients with known inherited hypercoagulability disorders, but this should be done only if the results could affect a treatment decision: Can a man safely undergo surgery with a high postoperative risk for thrombosis, for example? Should a woman take oral contraceptives (OCs), start hormone replacement therapy (HRT), or attempt another pregnancy?

Overall, testing for inherited hypercoagulable states should focus on the identification of individuals most likely to benefit from it; only tests yielding useful data should be performed. Testing of asymptomatic individuals for the sole purpose of initiating long-term prophylactic therapy is not recommended.

Which tests for which patients?
In some cases, selective assays may be more useful than test panels, for reasons associated with patient presentation as well as cost. Proper selection of specific tests should be individualized based on the patient’s age, thrombotic presentation, and family history, and on potential effects on patient management.

For patients with heparin resistance, cerebral vein thrombosis, intra-abdominal vein thrombosis, or recurrent superficial thrombophlebitis, AT testing may be in order. Patients with recurrent superficial thrombophlebitis may also benefit from testing for factor V Leiden mutation, and protein C and protein S testing may be beneficial for patients with warfarin skin necrosis, recurrent superficial thrombophlebitis, or neonatal purpura fulminans. Cerebral vein thrombosis in the general population and in women using OCs, in particular, is suggestive of a G20210A mutation,⁹ and is a possible indication for testing. Hyperhomocysteinemia testing may be considered for patients with premature arterial and venous thrombosis, as well as mental retardation, skeletal abnormalities, and vitamin B6 or B12 deficiency.

Many physicians prefer to order testing in stages, starting with tests for the most common thrombophilias. When ordering tests for the conditions detailed above and in the TABLE, be aware that test panels vary among facilities, so it may be necessary to check with the testing laboratory to ensure that it offers the tests that are indicated for a particular patient.

TABLE
Suspect a hereditary hypercoagulable disorder? Testing considerations to keep in mind^{2,3,8,9,11,12}

Disorder	Potential indication(s) for testing	Timing of test	Interaction with warfarin	Interaction with LMWH and UFH
Factor V Leiden mutation	Recurrent superficial thrombophlebitis	Not during an acute event	Modified aPTT test does not interact	Modified aPTT test does not interact
Prothrombin G20210A mutation	Cerebral vein thrombosis	Not during an acute event	No	No
Protein C deficiency	Warfarin skin necrosis Recurrent superficial thrombophlebitis Neonatal purpura fulminans	7-10 days after cessation of warfarin therapy	Yes	No
Antithrombin III deficiency	Heparin resistance Cerebral vein thrombosis Intra-abdominal vein thrombosis Recurrent superficial thrombophlebitis	7-10 days after cessation of warfarin therapy	Yes	Yes
Protein S deficiency	Warfarin skin necrosis Recurrent superficial thrombophlebitis Neonatal purpura fulminans	7-10 days after cessation of warfarin therapy	Yes	No
Hyperhomocysteinemia	Premature arterial/venous thrombosis Skeletal abnormalities Vitamin B6 or B12 deficiency	Not during an acute event	No	No
aPTT, activated partial thromboplastin time; LMWH, low-molecular-weight heparin; UFH, unfractionated heparin.

Timing of tests, and other specifics
Acute thrombosis, anticoagulation therapy, and some disease states—eg, liver disease, nephritic syndrome, disseminated intravascular coagulation, and acute illness—can affect levels of AT, protein C, and protein S. Within the first 48 hours of warfarin therapy, for example, a patient’s protein C and protein S levels decline about 50%; after 2 weeks, the levels rise to about 70% of their normal range. Because of warfarin’s effect on these proteins, evaluation for these deficiencies should be performed at least 1 week to 10 days after cessation of a 3- to 6-month course of anticoagulation therapy. Abnormal findings should be confirmed with a second test approximately 3 weeks later.^2,3,9

 

Several tests are used to detect factor V Leiden mutation, including the polymerase chain reaction-based test and the modified activated partial thromboplastin time (aPTT)-based test. The latter can be given to patients who are receiving anticoagulants. The results indicate the ratio of activated protein C (APC), and a finding of <2.0 is considered abnormal.⁸ The presence of lupus anticoagulant—autoantibodies that bind to phospholipids and proteins associated with the cell membrane—may yield a false-positive result on the modified aPTT test.

While the prothrombin G20210A mutation is linked to elevations in prothrombin, measuring prothrombin levels is not an accurate way to test for this hypercoagulable state. The prothrombin G20210A mutation generally is detected through DNA analysis instead.^2,8

Hyperhomocysteinemia is diagnosed by measuring fasting plasma levels of homocysteine. Although the results are not standardized, they generally correlate with the severity of disease (mild, 15-30 μmol/L; moderate, >30-100 μmol/L; severe, >100 μmol/L). However, plasma levels may be falsely elevated during an acute thrombotic episode and decreased by supplementation with folic acid, B6, and B12. The decrease, however, does not indicate a reduction in the risk of thrombosis.

Treatment, yes, but for how long?
Most thrombotic events are initially treated with a combination of a heparin product (low-molecular-weight heparin [LMWH] or unfractionated heparin [UFH]) and warfarin—commonly known as bridging therapy. The heparin is discontinued once the patient’s international normalized ratio (INR) has been maintained at a therapeutic level for more than 24 hours, which usually takes about 5 days. Warfarin therapy, however, should continue for at least 3 to 6 months, depending on the severity and cause of the thrombosis.

Individuals with an AT, protein C, or protein S deficiency may be at increased risk of recurrence, and long-term anticoagulant therapy may be warranted.⁵ Under current guidelines from the American College of Chest Physicians (ACCP), however, hereditary thrombophilias are not considered to be major determinants of recurrence or a major factor guiding the duration of anticoagulation therapy.¹⁰ Practitioners must use their judgment to determine the need for treatment; if anticoagulation therapy is initiated, the duration should be based on an individual assessment of benefit and risk.

Recognizing and responding to acquired risks

Several conditions associated with vessel wall changes and venous stasis—the hallmarks of acquired hypercoagulable states—put patients at increased risk of venous thrombosis. Following is a review of the most likely risk factors, including antiphospholipid antibody syndrome (APS), previously known as lupus anticoagulant syndrome; heparin-induced thrombocytopenia (HIT); pregnancy; trauma; estrogen; and malignancy.

When to test for—and treat—APS
APS, a systemic autoimmune disorder that can result in arterial or venous thrombosis or pregnancy loss and morbidity, is characterized by the presence of autoantibodies. Patients of all ages may be affected by APS, one of the most common acquired hypercoagulability disorders. APS affects an estimated 28% of the general population.² About 15% of recurrent pregnancy losses and 20% of recurrent thromboses in young adults are attributed to this autoimmune disorder.¹¹

A definitive diagnosis of APS requires a history of either vascular thrombosis or pregnancy morbidity—defined as miscarriage after the 10th gestational week, consecutive fetal losses before the 10th gestational week, or placental insufficiency and premature birth before 34 weeks.^11,12 APS testing may be useful in patients with cerebral vein thrombosis, intra-abdominal vein thrombosis, or unexplained recurrent fetal loss.⁹

In addition to clinical criteria, a diagnosis of APS is based on the presence of plasma antibodies on 2 or more occasions at least 12 weeks apart. APS encompasses 3 types of antiphospholipid antibodies—lupus anticoagulant antibodies, anticardiolipin antibodies, and anti-beta 2-glycoprotein I antibodies—which can be detected with 2 different tests. Coagulation assays are used to identify lupus anticoagulant antibodies because they prolong clotting time; however, immunoassays are used to measure immunologic reactivity to phospholipids to determine the presence of anticardiolipin antibodies and anti-beta 2-glycoprotein I antibodies.^11,12

Treatment for APS generally involves anticoagulant therapy for the prevention and treatment of acute thrombotic events, or as prophylaxis during pregnancy. ACCP guidelines call for initiating warfarin therapy with a target INR of 2.5 (range 2.0-3.0) in patients with no other risk factors. In patients who have had recurrent thromboembolic events or have additional risk factors, a target INR of 3.0 (range 2.5-3.5) is suggested.¹⁰ LMWH and UFH are also options for use in the event of recurrence or for prophylaxis during pregnancy.^11,13,14

 

HIT can be benign, or life threatening
HIT—defined by a decrease in platelet count to less than 150,000 (or a 50% drop from baseline) after initiation of heparin therapy—may or may not be benign. HIT type I (previously called heparin-associated thrombocytopenia), which affects approximately 10% of patients treated with heparin, is transient, asymptomatic, and not associated with an increased risk of thrombosis. Type I typically occurs within the first 2 days of heparin therapy.¹⁵

HIT type II is an immune-mediated response that does increase the risk of thrombosis. Patients usually develop type II 5 to 12 days after initiation of heparin, but in rare instances onset is delayed, occurring up to 40 days after heparin therapy. Approximately 5% of patients on heparin develop HIT type II, and the risk increases with frequent heparin use. Unlike other states of thrombocytopenia, HIT rarely causes bleeding. However, patients with HIT type II are at risk for a paradoxical thrombotic syndrome that may become life threatening.^13,15

To diagnose HIT, an enzyme-linked immunosorbent assay or other specific blood tests must be used to confirm the presence of circulating antibodies.¹³ The diagnosis is based on the following criteria: (1) thrombocytopenia, (2) exclusion of other possible causes of thrombocytopenia, and (3) resolution of thrombocytopenia after discontinuation of heparin.^13,15

When HIT is suspected, all heparin-containing products must be discontinued immediately and alternative anticoagulant therapy (typically, with danaparoid, lepirudin, or argatroban) should be initiated to reduce the risk of thrombosis. Warfarin alone should not be used for the treatment of HIT because of its association with worsening thrombosis and venous limb gangrene.¹³ However, warfarin should be initiated while the patient is receiving danaparoid or a thrombin-specific inhibitor—with at least 5 days of overlapping therapy recommended. Duration of therapy has not been well defined, but an overall course of at least 2 to 3 months is recommended to reduce the risk of recurrent thrombosis.^13,15

Pregnancy raises risk, but limits Tx options
By altering the body’s normal physiologic state in a way that may lead to hypercoagulability, pregnancy increases the risk of VTE 6-fold.¹⁶ The risk continues throughout pregnancy and peaks during puerperium, the 6-week period after delivery. Cesarean delivery, prolonged immobility, and obesity elevate the risk.^13,16

Treatment options for acute thrombotic events during pregnancy are limited because warfarin is contraindicated. Current ACCP guidelines recommend substituting UFH or LMWH for oral anticoagulant therapy when treatment for an acute thrombotic event is required.¹⁰ While no pharmacologic prophylaxis is currently recommended for pregnant patients with thrombophilias but no history of thrombotic events,^10,17 there are cases when it may be necessary. Patients with an AT deficiency, for example, may require prenatal and postpartum prophylaxis,¹⁰ and patients who deliver by cesarean and have 1 or more additional risk factors should receive prophylaxis for the duration of their hospitalization. Women with multiple risk factors, in addition to pregnancy and cesarean section, should receive pharmacologic prophylaxis for up to 4 to 6 weeks postpartum.

Choices for postpartum anticoagulation for the treatment of an acute thrombotic event should be based on guidelines developed for nonpregnant patients, with this caveat: For women who are breastfeeding, safety during lactation must be considered. Neither UFH nor warfarin enters breast milk, so both are safe for such patients.¹⁸ LMWH, although smaller in molecular weight than UFH, is considered moderately safe during breastfeeding, as well.¹⁸

Prophylaxis is vital for trauma patients
Trauma and major injuries increase the risk of thrombosis by approximately 50%. Patients who are hospitalized after a major trauma are at high risk for the development of a VTE. At the greatest risk are those with spinal cord injuries (62%), pelvic fractures (61%), and leg fractures (80%).¹⁶ Current ACCP guidelines recommend the use of LMWH as soon as it is safe for trauma patients, and continuing it until discharge in patients with no apparent contraindication. If a patient has an active bleed or other contraindication, mechanical thromboprophylaxis is indicated until the bleeding risk decreases.¹⁵

Estrogens increase platelet aggregation
Estrogens are considered a risk factor because of their effect on both natural anticoagulants and clotting factors. A reduction in AT activity and increasing concentrations of clotting factors VII, X, and XII result from the use of estrogens. Estrogen is also thought to be responsible for the increase in platelet count and aggregation associated with the use of combination OCs.¹⁹ In fact, OC use is associated with about a 3-fold overall risk of thrombosis, a risk reported to be highest during the first year of use.¹⁶ Among hormonal contraceptives, the transdermal formulation has the highest risk.^20-22 Hormone replacement therapy taken during menopause confers approximately a 2- to 4-fold increase in risk for VTE, and selective estrogen receptor modulators are associated with a 2-fold risk.¹⁶ Physicians should educate patients about the risks associated with these agents and signs and symptoms of thrombosis.

 

Cancer and hypercoagulability: Which patients need treatment?
Although the pathophysiologic process is not fully understood, a link between cancer and hypercoagulability has long been recognized. In fact, malignancy—the second most common cause of acquired hypercoagulability—is associated with 10% to 20% of spontaneous DVTs.²

One possible mechanism is the interaction of tumor cells with thrombin and plasmin-generating systems, directly influencing thrombus formation.⁷

Cancer patients also have an elevated risk for thrombosis related to immobilization, infection, treatment with antineoplastic agents, surgery, and the insertion of central venous catheters. Approximately 30% of patients with central venous catheters develop a DVT of the arm.¹⁶

Anticoagulant therapy in cancer patients varies, depending on the severity and circumstances of the patient. According to American Society of Clinical Oncology, National Comprehensive Cancer Center Network, and ACCP guidelines, LMWH is the preferred initial treatment for thromboses in patients with cancer— that is, in the first 3 to 6 months of therapy after a thrombotic event.^17,23,24 The guidelines also mention warfarin as an alternative for long-term (>6 months) anticoagulant therapy, if no contraindications exist.

Because cancer is usually a long-term illness, anticoagulant therapy should be continued indefinitely, or until the cancer has resolved.^10,23,24 Prophylaxis is recommended for cancer patients who are bedridden with an acute medical illness, but should not be routinely used in patients with indwelling venous catheters or those receiving chemotherapy or hormonal therapy.^13,15

CORRESPONDENCE Haley M. Phillippe, PharmD, BCPS, University of Alabama Birmingham School of Medicine-Huntsville Campus, Department of Family Medicine, 301 Governors Drive, Huntsville, AL 35801; [email protected]

Each year, venous thrombosis develops in approximately 1 in 1000 people.¹ The cause: An alteration in blood composition, venous stasis, or vascular damage—commonly known as Virchow’s triad. Changes in blood composition are associated with hereditary thrombophilias, such as factor V Leiden mutation or a deficiency in protein C or S or antithrombin III (AT). Changes in blood flow (stasis) and vessel damage stem from acquired conditions that commonly lead to hypercoagulability—pregnancy, malignancy, and estrogen use among them.

Regardless of the reason a patient is at elevated risk, however, the goal is the same: to prevent the development of thrombi, thereby reducing the increased morbidity and mortality associated with thromboembolism. Achieving that goal requires an understanding of both inherited and acquired risk factors, familiarity with diagnostic tools, and knowledge of appropriate treatment. This review, which begins with hereditary hypercoagulable states before turning to acquired conditions associated with hypercoagulability, will help toward that end.

Keep these thrombophilias on your radar screen

Factor V Leiden mutation is the most common inherited hypercoagulable disorder, and the most common form of activated protein C resistance. The mutation is found in an estimated 5% to 10% of the general population.² Among patients with thromboembolic disorders, however, the incidence is considerably higher, with estimates ranging from 21% to 60%.^3,4 Factor V Leiden mutation is more prevalent among Caucasian populations, and rarely found in people of Asian or African descent.²

Prothrombin G20210A mutation, an autosomal-dominant disorder, is the second most common inherited hypercoagulable state.⁴ This mutation is associated with an increase in prothrombin levels, causing an elevation in thrombin and, in turn, a heightened risk of thrombosis.

Although prothrombin G20210A mutation is found in only 1% to 2% of the general population, its prevalence among those with a history of thromboembolic events is estimated at 5% to 19%.⁴ This disorder, too, varies significantly by ethnicity: People from southern Europe are twice as likely to be affected as northern Europeans, and the mutation is rarely found in people of Asian or African descent.^2,5

Protein C deficiency. Protein C, a vitamin K-dependent anticoagulant produced in the liver, is activated when thrombomodulin binds with thrombin in the presence of protein S, which serves as the cofactor. Protein C is required to inactivate clotting factors V and VIII.² The deficiency is an autosomal-dominant disorder, and is more likely to result in venous than arterial thrombosis.²

Protein C deficiency affects 1 in every 200 to 500 people in the general population; among patients with a history of venous thrombosis, its prevalence is 2% to 9%.^2,3 Generally, people with a protein C deficiency begin developing thrombi in their late teens, and about 75% suffer from 1 or more thrombotic events during the course of their lives.²

There are 2 types of protein C deficiency: Patients with type I have a decreased production of protein C, while those with type II have normal levels of the protein, but in a form that is dysfunctional.⁶

AT deficiency. AT, a natural anticoagulant synthesized by the liver and endothelial cells, is responsible for inactivating several clotting factors, including thrombin and factors IXa, Xa, XIa, and XIIa. Like protein C deficiency, AT deficiency is an autosomal-dominant disorder with 2 subtypes. Individuals with type I deficiency have normal plasma levels of AT, but the anticoagulant has reduced biological activity or is dysfunctional; those with type II deficiency have decreased plasma levels of fully functional AT.² Both types are more likely to lead to venous than arterial thrombosis.

AT deficiency affects approximately 1 in 2000 to 5000 people, including 2.8% of patients who develop venous thrombosis.⁷ Nearly two-thirds of those with AT deficiency develop thrombi before the age of 35.²

Protein S deficiency. Endothelial cells are responsible for the synthesis of protein S, which, like protein C (for which it serves as a cofactor), is vitamin K-dependent. Protein S deficiency, also an autosomal-dominant disorder, has 3 subtypes: Type I, also known as classical deficiency, is characterized by reduced free and total levels of functional protein S; type II patients have a normal total level of protein S, but a decreased amount of free protein; and type III patients have normal levels of both free and total protein S, but the available proteins are dysfunctional.^2,6

 

The prevalence of this hypercoagulable state in the general population is unknown, and protein S deficiencies have been found in only 1% of patients with a history of deep vein thrombosis (DVT).^3,8 Although this hypercoagulable state is less common than other hereditary thrombophilias, 74% of people with this disorder develop DVT—half of them before the age of 25.²

Hyperhomocysteinemia. Elevations in homocysteine may occur as a result of a hereditary disorder (deficiencies in cystathionine beta-synthase or methylene-tetrahydrofolate reductase). Hyperhomocysteinemia may also be an acquired condition, associated with deficiencies in vitamins B6, B12, or folic acid; chronic kidney disease; hypothyroidism; and certain malignancies.

The prevalence of hyperhomocysteinemia varies, based on the underlying disorder. Only about 0.3% of the general population has a cystathionine beta-synthase deficiency. Methylene-tetrahydrofolate reductase deficiency, however, is common among Italian and Hispanic populations (occurring in about 20%), but rare (<1%) among African American people.³

When to test for thrombophilias
Idiopathic venous thrombosis is probably the most common reason for ordering testing for inherited hypercoagulable states, and an underlying thrombophilia is found in about 50% of cases.⁹ Other indications for testing include thrombus development in an unusual site (eg, splanchic, renal, retinal, or ovarian veins; cerebral venous sinuses; or upper limbs), recurrent venous thromboembolism (VTE), venous thrombosis at an early age (<45 years) or in a patient with a strong family history of VTE, and unexplained recurrent pregnancy loss.

Testing may also be considered for relatives of patients with known inherited hypercoagulability disorders, but this should be done only if the results could affect a treatment decision: Can a man safely undergo surgery with a high postoperative risk for thrombosis, for example? Should a woman take oral contraceptives (OCs), start hormone replacement therapy (HRT), or attempt another pregnancy?

Overall, testing for inherited hypercoagulable states should focus on the identification of individuals most likely to benefit from it; only tests yielding useful data should be performed. Testing of asymptomatic individuals for the sole purpose of initiating long-term prophylactic therapy is not recommended.

Which tests for which patients?
In some cases, selective assays may be more useful than test panels, for reasons associated with patient presentation as well as cost. Proper selection of specific tests should be individualized based on the patient’s age, thrombotic presentation, and family history, and on potential effects on patient management.

For patients with heparin resistance, cerebral vein thrombosis, intra-abdominal vein thrombosis, or recurrent superficial thrombophlebitis, AT testing may be in order. Patients with recurrent superficial thrombophlebitis may also benefit from testing for factor V Leiden mutation, and protein C and protein S testing may be beneficial for patients with warfarin skin necrosis, recurrent superficial thrombophlebitis, or neonatal purpura fulminans. Cerebral vein thrombosis in the general population and in women using OCs, in particular, is suggestive of a G20210A mutation,⁹ and is a possible indication for testing. Hyperhomocysteinemia testing may be considered for patients with premature arterial and venous thrombosis, as well as mental retardation, skeletal abnormalities, and vitamin B6 or B12 deficiency.

Many physicians prefer to order testing in stages, starting with tests for the most common thrombophilias. When ordering tests for the conditions detailed above and in the TABLE, be aware that test panels vary among facilities, so it may be necessary to check with the testing laboratory to ensure that it offers the tests that are indicated for a particular patient.

TABLE
Suspect a hereditary hypercoagulable disorder? Testing considerations to keep in mind^{2,3,8,9,11,12}

Disorder	Potential indication(s) for testing	Timing of test	Interaction with warfarin	Interaction with LMWH and UFH
Factor V Leiden mutation	Recurrent superficial thrombophlebitis	Not during an acute event	Modified aPTT test does not interact	Modified aPTT test does not interact
Prothrombin G20210A mutation	Cerebral vein thrombosis	Not during an acute event	No	No
Protein C deficiency	Warfarin skin necrosis Recurrent superficial thrombophlebitis Neonatal purpura fulminans	7-10 days after cessation of warfarin therapy	Yes	No
Antithrombin III deficiency	Heparin resistance Cerebral vein thrombosis Intra-abdominal vein thrombosis Recurrent superficial thrombophlebitis	7-10 days after cessation of warfarin therapy	Yes	Yes
Protein S deficiency	Warfarin skin necrosis Recurrent superficial thrombophlebitis Neonatal purpura fulminans	7-10 days after cessation of warfarin therapy	Yes	No
Hyperhomocysteinemia	Premature arterial/venous thrombosis Skeletal abnormalities Vitamin B6 or B12 deficiency	Not during an acute event	No	No
aPTT, activated partial thromboplastin time; LMWH, low-molecular-weight heparin; UFH, unfractionated heparin.

Timing of tests, and other specifics
Acute thrombosis, anticoagulation therapy, and some disease states—eg, liver disease, nephritic syndrome, disseminated intravascular coagulation, and acute illness—can affect levels of AT, protein C, and protein S. Within the first 48 hours of warfarin therapy, for example, a patient’s protein C and protein S levels decline about 50%; after 2 weeks, the levels rise to about 70% of their normal range. Because of warfarin’s effect on these proteins, evaluation for these deficiencies should be performed at least 1 week to 10 days after cessation of a 3- to 6-month course of anticoagulation therapy. Abnormal findings should be confirmed with a second test approximately 3 weeks later.^2,3,9

 

Several tests are used to detect factor V Leiden mutation, including the polymerase chain reaction-based test and the modified activated partial thromboplastin time (aPTT)-based test. The latter can be given to patients who are receiving anticoagulants. The results indicate the ratio of activated protein C (APC), and a finding of <2.0 is considered abnormal.⁸ The presence of lupus anticoagulant—autoantibodies that bind to phospholipids and proteins associated with the cell membrane—may yield a false-positive result on the modified aPTT test.

While the prothrombin G20210A mutation is linked to elevations in prothrombin, measuring prothrombin levels is not an accurate way to test for this hypercoagulable state. The prothrombin G20210A mutation generally is detected through DNA analysis instead.^2,8

Hyperhomocysteinemia is diagnosed by measuring fasting plasma levels of homocysteine. Although the results are not standardized, they generally correlate with the severity of disease (mild, 15-30 μmol/L; moderate, >30-100 μmol/L; severe, >100 μmol/L). However, plasma levels may be falsely elevated during an acute thrombotic episode and decreased by supplementation with folic acid, B6, and B12. The decrease, however, does not indicate a reduction in the risk of thrombosis.

Treatment, yes, but for how long?
Most thrombotic events are initially treated with a combination of a heparin product (low-molecular-weight heparin [LMWH] or unfractionated heparin [UFH]) and warfarin—commonly known as bridging therapy. The heparin is discontinued once the patient’s international normalized ratio (INR) has been maintained at a therapeutic level for more than 24 hours, which usually takes about 5 days. Warfarin therapy, however, should continue for at least 3 to 6 months, depending on the severity and cause of the thrombosis.

Individuals with an AT, protein C, or protein S deficiency may be at increased risk of recurrence, and long-term anticoagulant therapy may be warranted.⁵ Under current guidelines from the American College of Chest Physicians (ACCP), however, hereditary thrombophilias are not considered to be major determinants of recurrence or a major factor guiding the duration of anticoagulation therapy.¹⁰ Practitioners must use their judgment to determine the need for treatment; if anticoagulation therapy is initiated, the duration should be based on an individual assessment of benefit and risk.

Recognizing and responding to acquired risks

Several conditions associated with vessel wall changes and venous stasis—the hallmarks of acquired hypercoagulable states—put patients at increased risk of venous thrombosis. Following is a review of the most likely risk factors, including antiphospholipid antibody syndrome (APS), previously known as lupus anticoagulant syndrome; heparin-induced thrombocytopenia (HIT); pregnancy; trauma; estrogen; and malignancy.

When to test for—and treat—APS
APS, a systemic autoimmune disorder that can result in arterial or venous thrombosis or pregnancy loss and morbidity, is characterized by the presence of autoantibodies. Patients of all ages may be affected by APS, one of the most common acquired hypercoagulability disorders. APS affects an estimated 28% of the general population.² About 15% of recurrent pregnancy losses and 20% of recurrent thromboses in young adults are attributed to this autoimmune disorder.¹¹

A definitive diagnosis of APS requires a history of either vascular thrombosis or pregnancy morbidity—defined as miscarriage after the 10th gestational week, consecutive fetal losses before the 10th gestational week, or placental insufficiency and premature birth before 34 weeks.^11,12 APS testing may be useful in patients with cerebral vein thrombosis, intra-abdominal vein thrombosis, or unexplained recurrent fetal loss.⁹

In addition to clinical criteria, a diagnosis of APS is based on the presence of plasma antibodies on 2 or more occasions at least 12 weeks apart. APS encompasses 3 types of antiphospholipid antibodies—lupus anticoagulant antibodies, anticardiolipin antibodies, and anti-beta 2-glycoprotein I antibodies—which can be detected with 2 different tests. Coagulation assays are used to identify lupus anticoagulant antibodies because they prolong clotting time; however, immunoassays are used to measure immunologic reactivity to phospholipids to determine the presence of anticardiolipin antibodies and anti-beta 2-glycoprotein I antibodies.^11,12

Treatment for APS generally involves anticoagulant therapy for the prevention and treatment of acute thrombotic events, or as prophylaxis during pregnancy. ACCP guidelines call for initiating warfarin therapy with a target INR of 2.5 (range 2.0-3.0) in patients with no other risk factors. In patients who have had recurrent thromboembolic events or have additional risk factors, a target INR of 3.0 (range 2.5-3.5) is suggested.¹⁰ LMWH and UFH are also options for use in the event of recurrence or for prophylaxis during pregnancy.^11,13,14

 

HIT can be benign, or life threatening
HIT—defined by a decrease in platelet count to less than 150,000 (or a 50% drop from baseline) after initiation of heparin therapy—may or may not be benign. HIT type I (previously called heparin-associated thrombocytopenia), which affects approximately 10% of patients treated with heparin, is transient, asymptomatic, and not associated with an increased risk of thrombosis. Type I typically occurs within the first 2 days of heparin therapy.¹⁵

HIT type II is an immune-mediated response that does increase the risk of thrombosis. Patients usually develop type II 5 to 12 days after initiation of heparin, but in rare instances onset is delayed, occurring up to 40 days after heparin therapy. Approximately 5% of patients on heparin develop HIT type II, and the risk increases with frequent heparin use. Unlike other states of thrombocytopenia, HIT rarely causes bleeding. However, patients with HIT type II are at risk for a paradoxical thrombotic syndrome that may become life threatening.^13,15

To diagnose HIT, an enzyme-linked immunosorbent assay or other specific blood tests must be used to confirm the presence of circulating antibodies.¹³ The diagnosis is based on the following criteria: (1) thrombocytopenia, (2) exclusion of other possible causes of thrombocytopenia, and (3) resolution of thrombocytopenia after discontinuation of heparin.^13,15

When HIT is suspected, all heparin-containing products must be discontinued immediately and alternative anticoagulant therapy (typically, with danaparoid, lepirudin, or argatroban) should be initiated to reduce the risk of thrombosis. Warfarin alone should not be used for the treatment of HIT because of its association with worsening thrombosis and venous limb gangrene.¹³ However, warfarin should be initiated while the patient is receiving danaparoid or a thrombin-specific inhibitor—with at least 5 days of overlapping therapy recommended. Duration of therapy has not been well defined, but an overall course of at least 2 to 3 months is recommended to reduce the risk of recurrent thrombosis.^13,15

Pregnancy raises risk, but limits Tx options
By altering the body’s normal physiologic state in a way that may lead to hypercoagulability, pregnancy increases the risk of VTE 6-fold.¹⁶ The risk continues throughout pregnancy and peaks during puerperium, the 6-week period after delivery. Cesarean delivery, prolonged immobility, and obesity elevate the risk.^13,16

Treatment options for acute thrombotic events during pregnancy are limited because warfarin is contraindicated. Current ACCP guidelines recommend substituting UFH or LMWH for oral anticoagulant therapy when treatment for an acute thrombotic event is required.¹⁰ While no pharmacologic prophylaxis is currently recommended for pregnant patients with thrombophilias but no history of thrombotic events,^10,17 there are cases when it may be necessary. Patients with an AT deficiency, for example, may require prenatal and postpartum prophylaxis,¹⁰ and patients who deliver by cesarean and have 1 or more additional risk factors should receive prophylaxis for the duration of their hospitalization. Women with multiple risk factors, in addition to pregnancy and cesarean section, should receive pharmacologic prophylaxis for up to 4 to 6 weeks postpartum.

Choices for postpartum anticoagulation for the treatment of an acute thrombotic event should be based on guidelines developed for nonpregnant patients, with this caveat: For women who are breastfeeding, safety during lactation must be considered. Neither UFH nor warfarin enters breast milk, so both are safe for such patients.¹⁸ LMWH, although smaller in molecular weight than UFH, is considered moderately safe during breastfeeding, as well.¹⁸

Prophylaxis is vital for trauma patients
Trauma and major injuries increase the risk of thrombosis by approximately 50%. Patients who are hospitalized after a major trauma are at high risk for the development of a VTE. At the greatest risk are those with spinal cord injuries (62%), pelvic fractures (61%), and leg fractures (80%).¹⁶ Current ACCP guidelines recommend the use of LMWH as soon as it is safe for trauma patients, and continuing it until discharge in patients with no apparent contraindication. If a patient has an active bleed or other contraindication, mechanical thromboprophylaxis is indicated until the bleeding risk decreases.¹⁵

Estrogens increase platelet aggregation
Estrogens are considered a risk factor because of their effect on both natural anticoagulants and clotting factors. A reduction in AT activity and increasing concentrations of clotting factors VII, X, and XII result from the use of estrogens. Estrogen is also thought to be responsible for the increase in platelet count and aggregation associated with the use of combination OCs.¹⁹ In fact, OC use is associated with about a 3-fold overall risk of thrombosis, a risk reported to be highest during the first year of use.¹⁶ Among hormonal contraceptives, the transdermal formulation has the highest risk.^20-22 Hormone replacement therapy taken during menopause confers approximately a 2- to 4-fold increase in risk for VTE, and selective estrogen receptor modulators are associated with a 2-fold risk.¹⁶ Physicians should educate patients about the risks associated with these agents and signs and symptoms of thrombosis.

 

Cancer and hypercoagulability: Which patients need treatment?
Although the pathophysiologic process is not fully understood, a link between cancer and hypercoagulability has long been recognized. In fact, malignancy—the second most common cause of acquired hypercoagulability—is associated with 10% to 20% of spontaneous DVTs.²

One possible mechanism is the interaction of tumor cells with thrombin and plasmin-generating systems, directly influencing thrombus formation.⁷

Cancer patients also have an elevated risk for thrombosis related to immobilization, infection, treatment with antineoplastic agents, surgery, and the insertion of central venous catheters. Approximately 30% of patients with central venous catheters develop a DVT of the arm.¹⁶

Anticoagulant therapy in cancer patients varies, depending on the severity and circumstances of the patient. According to American Society of Clinical Oncology, National Comprehensive Cancer Center Network, and ACCP guidelines, LMWH is the preferred initial treatment for thromboses in patients with cancer— that is, in the first 3 to 6 months of therapy after a thrombotic event.^17,23,24 The guidelines also mention warfarin as an alternative for long-term (>6 months) anticoagulant therapy, if no contraindications exist.

Because cancer is usually a long-term illness, anticoagulant therapy should be continued indefinitely, or until the cancer has resolved.^10,23,24 Prophylaxis is recommended for cancer patients who are bedridden with an acute medical illness, but should not be routinely used in patients with indwelling venous catheters or those receiving chemotherapy or hormonal therapy.^13,15

CORRESPONDENCE Haley M. Phillippe, PharmD, BCPS, University of Alabama Birmingham School of Medicine-Huntsville Campus, Department of Family Medicine, 301 Governors Drive, Huntsville, AL 35801; [email protected]

A patient of yours has nonspecific back pain that fails to improve with the usual self-care measures. He asks you whether osteopathic manipulation might help. Would you be prepared to discuss the relevant clinical evidence?

For a patient such as this, expert guidelines do recommend referral for osteopathic spinal manipulation, which, if performed by a qualified physician, may be efficacious and cost effective. Limited data show that osteopathic manipulation may also be effective for nonspinal disorders.

We conducted a systematic review of the evidence for osteopathic manipulative treatment (OMT) as applied to several conditions. Specifically, we searched PubMed for English language articles published between 1970 and December 2007, using the keywords osteopathy, osteopathic medicine, osteopathic manipulation, spinal manipulation, and somatic dysfunction. Our findings follow.

How OMT contributes to wellness

Osteopathic manipulative procedures are based on the premise that the neuromuscular system is vital to maintaining homeostasis. Changes in the musculoskeletal system can affect other organs (somatovisceral reflex), and visceral pathology can manifest as abnormalities in musculoskeletal tissue texture and articular motion (viscerosomatic reflex).¹ These musculoskeletal changes are diagnosed as somatic dysfunction and are assigned International Statistical Classification of Diseases and Related Health Problems (ICD-9) codes corresponding to the area of the body in which these changes are palpated.² Similarly, OMT therapeutic procedures are assigned Evaluation and Management (E&M) codes corresponding to the number of body areas treated.

OMT comprises more than 100 different techniques used to treat somatic dysfunction. Some techniques are similar to those used by chiropractors and physical or massage therapists; others are unique to osteopathically trained physicians.

OMT has multiple physiologic effects. Mechanically, OMT causes articular release, freeing joint motion. Neuromuscularly, OMT generates afferent input into the dorsal root ganglion, diminishing motor neuron discharge and relaxing muscle fibers.³ Vascularly, OMT may increase nitric oxide concentration in the blood, promoting vasodilatation and increasing blood flow to peripheral vascular tissue.⁴ Neurochemically, OMT can transiently increase serum levels of anandamide, stimulating cannabinoid receptors in the brain.⁵

What the evidence says about OMT for back pain

Joint clinical practice guidelines issued in 2007 by the American College of Physicians and the American Pain Society give a weak recommendation based on moderate-quality evidence that manipulation is an appropriate nonpharmacologic modality for treating nonspecific acute and chronic low back pain that fails to improve with self-care.⁶

The Institute for Clinical Systems Improvement guidelines for back pain, updated in 2008, recommend referral to a spine therapy professional for manipulative treatment of nonspecific low back pain that has failed to improve with self-care after 2 weeks, or for a patient experiencing incapacitating pain. The guidelines suggest that referred patients usually demonstrate improvement within 3 to 4 visits and typically require no more than 6 visits.⁷

DO family practitioners appear to use OMT more often for pain in the back than for pain in other areas of the body.⁸ Although a large number of randomized controlled trials (RCTs) have examined the role of spinal manipulation for adults with back pain, regardless of the type of practitioner, fewer trials have focused on manipulation specifically performed by osteopathically trained physicians.

Pain reduction is significant. A meta-analysis was conducted on 8 RCTs involving patients with back pain of at least 3 weeks’ duration, with 318 patients assigned to receive OMT vs 231 controls. Subjects in the OMT group received a variable number of OMT sessions over a given time frame per study protocol, while subjects in the control group were allowed to pursue standard care for back pain, including nonsteroidal anti-inflammatory agents (NSAIDs), muscle relaxants, narcotics, physical therapy, and home exercises. The authors found a significant (30%) overall reduction in pain rating in the OMT group compared with various control therapies at 4 and 12 weeks’ follow-up (95% confidence interval [CI], -0.47 to -0.13; P=.001).⁹

Another study randomized 155 patients with subacute low back pain to receive standard care or standard care plus 8 sessions of OMT over 2 months. At follow-up, both groups had similar pain ratings on a visual analog scale, but participants in the OMT group required significantly less NSAIDs, muscle relaxants, and physical therapy.¹⁰

 

A few RCTs have investigated the role of OMT in adults with chronic low back pain. One study randomized 91 patients with non-specific back pain of more than 3 months’ duration to receive 7 sessions of OMT, 7 sessions of sham manipulative therapy, or usual care. (Sham manipulation consisted of range of motion and light touch without therapeutic intention.) Both OMT and sham therapy significantly decreased back pain at 1 month (P=.01 and P=.003, respectively), 3 months (P=.001, P=.01), and 6 months (P=.02, P=.02) compared with usual care.¹¹

A study conducted in the United Kingdom randomized 201 patients with spinal pain of 2 to 12 weeks’ duration to receive usual care or usual care plus 3 OMT sessions. At 2 months’ follow-up, the OMT group, compared with the usual care group, exhibited a significant reduction in spinal pain levels (95% CI, 0.7-9.8; P=.02) and in psychological distress secondary to spinal pain (95% CI, 2.7-10.7; P=.001). Both measures were rated on a scale of 0 to 100.¹²

A follow-up cost analysis between the usual care and usual care/OMT group found a nonsignificant difference in mean health care costs due to spinal pain for the duration of the study, estimated to be 58 £ ($88.13 US) in the usual care group and 47 £ ($71.42 US) in the OMT group.¹³ Authors of other studies have inferred potential health care cost savings associated with OMT for back pain based on workers’ compensation claims, lost work time, provider services, medication use, or length of hospital stay.¹⁴

Evidence basis for OMT in other disorders

Headache
One study randomized 22 subjects with tension-type headaches lasting longer than 6 months to 10-minute sessions of OMT, sham therapy, or supine rest. Participants rated their discomfort on a scale of 0 (absence of headache) to 7 (debilitating headache) before and after study intervention. Only the OMT group showed a significant immediate post-treatment reduction in patient-rated headache severity (P<.003).¹⁵

A more recent study examined 26 patients with tension headaches of similar severity and frequency at baseline. All 26 subjects received training in progressive muscular relaxation home exercises, while 14 subjects also received 3 OMT sessions over 3 weeks. At 6 weeks’ follow-up, the OMT group noted 1.79 headache-free days per week, compared with 0.21 headache-free days per week in the control group (P=.016).¹⁶

Neck pain
Fifty-eight patients with neck pain lasting longer than 3 weeks who sought care at an emergency department were asked to rate their pain intensity on an 11-point numerical scale before and after randomization to receive either 30 mg intramuscular ketorolac or a 5-minute OMT session. Both groups experienced a reduction in pain, 1.7±1.6 (95% CI, 1.1-2.3; P<.001) and 2.8 ±1.7 (95% CI, 2.1-3.4; P<.002), respectively. However, patients receiving OMT showed a significantly greater reduction in pain intensity compared with those receiving ketorolac (95% CI, 0.2-1.9; P=.02).¹⁷

Otitis media
One study examined the role of OMT in children who had experienced 3 episodes of acute otitis media (AOM) in the 6 months before study enrollment, or 4 episodes in the prior 12 months. Fifty-seven children ages 6 months to 6 years were randomized to receive usual care or usual care plus 7 OMT sessions over 6 months. The OMT group showed a significantly reduced number of AOM episodes and reduced referral for myringotomy/ ventilation tube placement compared with the control group. Additionally, final tympanograms showed an increased frequency of more normal tympanogram types in the OMT group (95% CI, 0.08-1.02; P=.02).¹⁸

Pediatric asthma
Using the registry of an asthma clinic, 1 study selected 140 subjects ages 5 to 17 years and randomized them to receive 1 OMT or 1 sham session, with peak expiratory flows (PEF) measured before and after treatment. The OMT group showed a significant mean increase in PEF from 364 to 377 L/min (95% CI, 7.3-18.7) compared with no change in the sham group.¹⁹

In the same year (2005), a Cochrane review analyzed 3 previous trials including 156 children and adults and found no significant difference in lung function measures with OMT or other manipulative or sham treatments.²⁰

Infantile colic
One study randomized 28 infants ages 1 to 12 weeks diagnosed with colic to receive 4 weekly OMT sessions or no treatment. At 4 weeks’ follow-up, the OMT group showed a significant reduction in parent-reported daily number of hours their infants spent crying, from 2.39 to 0.89 hours (P<.001), and a significant increase in the daily number of hours infants spent sleeping, from 11.55 to 12.9 hours (P<.002). The control group showed a nonsignificant reduction in daily number of hours infants spent crying, from 2.06 to 1.56 hours, and a nonsignificant increase in daily number of hours spent sleeping, from 11.86 to 12.04 hours.²¹

 

IBS, fibromyalgia …
A number of very small RCTs with equivocal results, pilot studies, and retrospective reviews have investigated the use of OMT in postsurgical functionality, irritable bowel syndrome, fibromyalgia, infantile torticollis, muscle spasticity, joint pain, labor pain, back pain during pregnancy, adult asthma, chronic obstructive pulmonary disease, and other medical conditions. Results to date have not been meaningful enough to recommend a place for OMT in the management of these disorders.

Limitations of evidence for OMT

Studies of OMT and other forms of spinal manipulation and manual modalities have been criticized for inconsistent quality.²² Sample sizes of published studies tend to be small, rendering statistical analysis problematic.

Pretrial bias of participants may also influence outcome measures. Patients tend to have preformed opinions regarding the efficacy of manual modalities.²²

The lack of validation of a placebo control has historically been problematic, and the use of sham treatment is an attempt to overcome this.²³ Some studies lack objective parameters for outcomes, relying on subjective patient ratings. Finally, severity of illness in chronic conditions such as back pain varies over time, affecting study results in follow-up.²⁴

CORRESPONDENCE Sarah Cole, DO, 12680 olive Boulevard, Suite 300, St. Louis, MO 63141; [email protected]

A patient of yours has nonspecific back pain that fails to improve with the usual self-care measures. He asks you whether osteopathic manipulation might help. Would you be prepared to discuss the relevant clinical evidence?

For a patient such as this, expert guidelines do recommend referral for osteopathic spinal manipulation, which, if performed by a qualified physician, may be efficacious and cost effective. Limited data show that osteopathic manipulation may also be effective for nonspinal disorders.

We conducted a systematic review of the evidence for osteopathic manipulative treatment (OMT) as applied to several conditions. Specifically, we searched PubMed for English language articles published between 1970 and December 2007, using the keywords osteopathy, osteopathic medicine, osteopathic manipulation, spinal manipulation, and somatic dysfunction. Our findings follow.

How OMT contributes to wellness

Osteopathic manipulative procedures are based on the premise that the neuromuscular system is vital to maintaining homeostasis. Changes in the musculoskeletal system can affect other organs (somatovisceral reflex), and visceral pathology can manifest as abnormalities in musculoskeletal tissue texture and articular motion (viscerosomatic reflex).¹ These musculoskeletal changes are diagnosed as somatic dysfunction and are assigned International Statistical Classification of Diseases and Related Health Problems (ICD-9) codes corresponding to the area of the body in which these changes are palpated.² Similarly, OMT therapeutic procedures are assigned Evaluation and Management (E&M) codes corresponding to the number of body areas treated.

OMT comprises more than 100 different techniques used to treat somatic dysfunction. Some techniques are similar to those used by chiropractors and physical or massage therapists; others are unique to osteopathically trained physicians.

OMT has multiple physiologic effects. Mechanically, OMT causes articular release, freeing joint motion. Neuromuscularly, OMT generates afferent input into the dorsal root ganglion, diminishing motor neuron discharge and relaxing muscle fibers.³ Vascularly, OMT may increase nitric oxide concentration in the blood, promoting vasodilatation and increasing blood flow to peripheral vascular tissue.⁴ Neurochemically, OMT can transiently increase serum levels of anandamide, stimulating cannabinoid receptors in the brain.⁵

What the evidence says about OMT for back pain

Joint clinical practice guidelines issued in 2007 by the American College of Physicians and the American Pain Society give a weak recommendation based on moderate-quality evidence that manipulation is an appropriate nonpharmacologic modality for treating nonspecific acute and chronic low back pain that fails to improve with self-care.⁶

The Institute for Clinical Systems Improvement guidelines for back pain, updated in 2008, recommend referral to a spine therapy professional for manipulative treatment of nonspecific low back pain that has failed to improve with self-care after 2 weeks, or for a patient experiencing incapacitating pain. The guidelines suggest that referred patients usually demonstrate improvement within 3 to 4 visits and typically require no more than 6 visits.⁷

DO family practitioners appear to use OMT more often for pain in the back than for pain in other areas of the body.⁸ Although a large number of randomized controlled trials (RCTs) have examined the role of spinal manipulation for adults with back pain, regardless of the type of practitioner, fewer trials have focused on manipulation specifically performed by osteopathically trained physicians.

Pain reduction is significant. A meta-analysis was conducted on 8 RCTs involving patients with back pain of at least 3 weeks’ duration, with 318 patients assigned to receive OMT vs 231 controls. Subjects in the OMT group received a variable number of OMT sessions over a given time frame per study protocol, while subjects in the control group were allowed to pursue standard care for back pain, including nonsteroidal anti-inflammatory agents (NSAIDs), muscle relaxants, narcotics, physical therapy, and home exercises. The authors found a significant (30%) overall reduction in pain rating in the OMT group compared with various control therapies at 4 and 12 weeks’ follow-up (95% confidence interval [CI], -0.47 to -0.13; P=.001).⁹

Another study randomized 155 patients with subacute low back pain to receive standard care or standard care plus 8 sessions of OMT over 2 months. At follow-up, both groups had similar pain ratings on a visual analog scale, but participants in the OMT group required significantly less NSAIDs, muscle relaxants, and physical therapy.¹⁰

 

A few RCTs have investigated the role of OMT in adults with chronic low back pain. One study randomized 91 patients with non-specific back pain of more than 3 months’ duration to receive 7 sessions of OMT, 7 sessions of sham manipulative therapy, or usual care. (Sham manipulation consisted of range of motion and light touch without therapeutic intention.) Both OMT and sham therapy significantly decreased back pain at 1 month (P=.01 and P=.003, respectively), 3 months (P=.001, P=.01), and 6 months (P=.02, P=.02) compared with usual care.¹¹

A study conducted in the United Kingdom randomized 201 patients with spinal pain of 2 to 12 weeks’ duration to receive usual care or usual care plus 3 OMT sessions. At 2 months’ follow-up, the OMT group, compared with the usual care group, exhibited a significant reduction in spinal pain levels (95% CI, 0.7-9.8; P=.02) and in psychological distress secondary to spinal pain (95% CI, 2.7-10.7; P=.001). Both measures were rated on a scale of 0 to 100.¹²

A follow-up cost analysis between the usual care and usual care/OMT group found a nonsignificant difference in mean health care costs due to spinal pain for the duration of the study, estimated to be 58 £ ($88.13 US) in the usual care group and 47 £ ($71.42 US) in the OMT group.¹³ Authors of other studies have inferred potential health care cost savings associated with OMT for back pain based on workers’ compensation claims, lost work time, provider services, medication use, or length of hospital stay.¹⁴

Evidence basis for OMT in other disorders

Headache
One study randomized 22 subjects with tension-type headaches lasting longer than 6 months to 10-minute sessions of OMT, sham therapy, or supine rest. Participants rated their discomfort on a scale of 0 (absence of headache) to 7 (debilitating headache) before and after study intervention. Only the OMT group showed a significant immediate post-treatment reduction in patient-rated headache severity (P<.003).¹⁵

A more recent study examined 26 patients with tension headaches of similar severity and frequency at baseline. All 26 subjects received training in progressive muscular relaxation home exercises, while 14 subjects also received 3 OMT sessions over 3 weeks. At 6 weeks’ follow-up, the OMT group noted 1.79 headache-free days per week, compared with 0.21 headache-free days per week in the control group (P=.016).¹⁶

Neck pain
Fifty-eight patients with neck pain lasting longer than 3 weeks who sought care at an emergency department were asked to rate their pain intensity on an 11-point numerical scale before and after randomization to receive either 30 mg intramuscular ketorolac or a 5-minute OMT session. Both groups experienced a reduction in pain, 1.7±1.6 (95% CI, 1.1-2.3; P<.001) and 2.8 ±1.7 (95% CI, 2.1-3.4; P<.002), respectively. However, patients receiving OMT showed a significantly greater reduction in pain intensity compared with those receiving ketorolac (95% CI, 0.2-1.9; P=.02).¹⁷

Otitis media
One study examined the role of OMT in children who had experienced 3 episodes of acute otitis media (AOM) in the 6 months before study enrollment, or 4 episodes in the prior 12 months. Fifty-seven children ages 6 months to 6 years were randomized to receive usual care or usual care plus 7 OMT sessions over 6 months. The OMT group showed a significantly reduced number of AOM episodes and reduced referral for myringotomy/ ventilation tube placement compared with the control group. Additionally, final tympanograms showed an increased frequency of more normal tympanogram types in the OMT group (95% CI, 0.08-1.02; P=.02).¹⁸

Pediatric asthma
Using the registry of an asthma clinic, 1 study selected 140 subjects ages 5 to 17 years and randomized them to receive 1 OMT or 1 sham session, with peak expiratory flows (PEF) measured before and after treatment. The OMT group showed a significant mean increase in PEF from 364 to 377 L/min (95% CI, 7.3-18.7) compared with no change in the sham group.¹⁹

In the same year (2005), a Cochrane review analyzed 3 previous trials including 156 children and adults and found no significant difference in lung function measures with OMT or other manipulative or sham treatments.²⁰

Infantile colic
One study randomized 28 infants ages 1 to 12 weeks diagnosed with colic to receive 4 weekly OMT sessions or no treatment. At 4 weeks’ follow-up, the OMT group showed a significant reduction in parent-reported daily number of hours their infants spent crying, from 2.39 to 0.89 hours (P<.001), and a significant increase in the daily number of hours infants spent sleeping, from 11.55 to 12.9 hours (P<.002). The control group showed a nonsignificant reduction in daily number of hours infants spent crying, from 2.06 to 1.56 hours, and a nonsignificant increase in daily number of hours spent sleeping, from 11.86 to 12.04 hours.²¹

 

IBS, fibromyalgia …
A number of very small RCTs with equivocal results, pilot studies, and retrospective reviews have investigated the use of OMT in postsurgical functionality, irritable bowel syndrome, fibromyalgia, infantile torticollis, muscle spasticity, joint pain, labor pain, back pain during pregnancy, adult asthma, chronic obstructive pulmonary disease, and other medical conditions. Results to date have not been meaningful enough to recommend a place for OMT in the management of these disorders.

Limitations of evidence for OMT

Studies of OMT and other forms of spinal manipulation and manual modalities have been criticized for inconsistent quality.²² Sample sizes of published studies tend to be small, rendering statistical analysis problematic.

Pretrial bias of participants may also influence outcome measures. Patients tend to have preformed opinions regarding the efficacy of manual modalities.²²

The lack of validation of a placebo control has historically been problematic, and the use of sham treatment is an attempt to overcome this.²³ Some studies lack objective parameters for outcomes, relying on subjective patient ratings. Finally, severity of illness in chronic conditions such as back pain varies over time, affecting study results in follow-up.²⁴

CORRESPONDENCE Sarah Cole, DO, 12680 olive Boulevard, Suite 300, St. Louis, MO 63141; [email protected]

A patient of yours has nonspecific back pain that fails to improve with the usual self-care measures. He asks you whether osteopathic manipulation might help. Would you be prepared to discuss the relevant clinical evidence?

For a patient such as this, expert guidelines do recommend referral for osteopathic spinal manipulation, which, if performed by a qualified physician, may be efficacious and cost effective. Limited data show that osteopathic manipulation may also be effective for nonspinal disorders.

We conducted a systematic review of the evidence for osteopathic manipulative treatment (OMT) as applied to several conditions. Specifically, we searched PubMed for English language articles published between 1970 and December 2007, using the keywords osteopathy, osteopathic medicine, osteopathic manipulation, spinal manipulation, and somatic dysfunction. Our findings follow.

How OMT contributes to wellness

Osteopathic manipulative procedures are based on the premise that the neuromuscular system is vital to maintaining homeostasis. Changes in the musculoskeletal system can affect other organs (somatovisceral reflex), and visceral pathology can manifest as abnormalities in musculoskeletal tissue texture and articular motion (viscerosomatic reflex).¹ These musculoskeletal changes are diagnosed as somatic dysfunction and are assigned International Statistical Classification of Diseases and Related Health Problems (ICD-9) codes corresponding to the area of the body in which these changes are palpated.² Similarly, OMT therapeutic procedures are assigned Evaluation and Management (E&M) codes corresponding to the number of body areas treated.

OMT comprises more than 100 different techniques used to treat somatic dysfunction. Some techniques are similar to those used by chiropractors and physical or massage therapists; others are unique to osteopathically trained physicians.

OMT has multiple physiologic effects. Mechanically, OMT causes articular release, freeing joint motion. Neuromuscularly, OMT generates afferent input into the dorsal root ganglion, diminishing motor neuron discharge and relaxing muscle fibers.³ Vascularly, OMT may increase nitric oxide concentration in the blood, promoting vasodilatation and increasing blood flow to peripheral vascular tissue.⁴ Neurochemically, OMT can transiently increase serum levels of anandamide, stimulating cannabinoid receptors in the brain.⁵

What the evidence says about OMT for back pain

Joint clinical practice guidelines issued in 2007 by the American College of Physicians and the American Pain Society give a weak recommendation based on moderate-quality evidence that manipulation is an appropriate nonpharmacologic modality for treating nonspecific acute and chronic low back pain that fails to improve with self-care.⁶

The Institute for Clinical Systems Improvement guidelines for back pain, updated in 2008, recommend referral to a spine therapy professional for manipulative treatment of nonspecific low back pain that has failed to improve with self-care after 2 weeks, or for a patient experiencing incapacitating pain. The guidelines suggest that referred patients usually demonstrate improvement within 3 to 4 visits and typically require no more than 6 visits.⁷

DO family practitioners appear to use OMT more often for pain in the back than for pain in other areas of the body.⁸ Although a large number of randomized controlled trials (RCTs) have examined the role of spinal manipulation for adults with back pain, regardless of the type of practitioner, fewer trials have focused on manipulation specifically performed by osteopathically trained physicians.

Pain reduction is significant. A meta-analysis was conducted on 8 RCTs involving patients with back pain of at least 3 weeks’ duration, with 318 patients assigned to receive OMT vs 231 controls. Subjects in the OMT group received a variable number of OMT sessions over a given time frame per study protocol, while subjects in the control group were allowed to pursue standard care for back pain, including nonsteroidal anti-inflammatory agents (NSAIDs), muscle relaxants, narcotics, physical therapy, and home exercises. The authors found a significant (30%) overall reduction in pain rating in the OMT group compared with various control therapies at 4 and 12 weeks’ follow-up (95% confidence interval [CI], -0.47 to -0.13; P=.001).⁹

Another study randomized 155 patients with subacute low back pain to receive standard care or standard care plus 8 sessions of OMT over 2 months. At follow-up, both groups had similar pain ratings on a visual analog scale, but participants in the OMT group required significantly less NSAIDs, muscle relaxants, and physical therapy.¹⁰

 

A few RCTs have investigated the role of OMT in adults with chronic low back pain. One study randomized 91 patients with non-specific back pain of more than 3 months’ duration to receive 7 sessions of OMT, 7 sessions of sham manipulative therapy, or usual care. (Sham manipulation consisted of range of motion and light touch without therapeutic intention.) Both OMT and sham therapy significantly decreased back pain at 1 month (P=.01 and P=.003, respectively), 3 months (P=.001, P=.01), and 6 months (P=.02, P=.02) compared with usual care.¹¹

A study conducted in the United Kingdom randomized 201 patients with spinal pain of 2 to 12 weeks’ duration to receive usual care or usual care plus 3 OMT sessions. At 2 months’ follow-up, the OMT group, compared with the usual care group, exhibited a significant reduction in spinal pain levels (95% CI, 0.7-9.8; P=.02) and in psychological distress secondary to spinal pain (95% CI, 2.7-10.7; P=.001). Both measures were rated on a scale of 0 to 100.¹²

A follow-up cost analysis between the usual care and usual care/OMT group found a nonsignificant difference in mean health care costs due to spinal pain for the duration of the study, estimated to be 58 £ ($88.13 US) in the usual care group and 47 £ ($71.42 US) in the OMT group.¹³ Authors of other studies have inferred potential health care cost savings associated with OMT for back pain based on workers’ compensation claims, lost work time, provider services, medication use, or length of hospital stay.¹⁴

Evidence basis for OMT in other disorders

Headache
One study randomized 22 subjects with tension-type headaches lasting longer than 6 months to 10-minute sessions of OMT, sham therapy, or supine rest. Participants rated their discomfort on a scale of 0 (absence of headache) to 7 (debilitating headache) before and after study intervention. Only the OMT group showed a significant immediate post-treatment reduction in patient-rated headache severity (P<.003).¹⁵

A more recent study examined 26 patients with tension headaches of similar severity and frequency at baseline. All 26 subjects received training in progressive muscular relaxation home exercises, while 14 subjects also received 3 OMT sessions over 3 weeks. At 6 weeks’ follow-up, the OMT group noted 1.79 headache-free days per week, compared with 0.21 headache-free days per week in the control group (P=.016).¹⁶

Neck pain
Fifty-eight patients with neck pain lasting longer than 3 weeks who sought care at an emergency department were asked to rate their pain intensity on an 11-point numerical scale before and after randomization to receive either 30 mg intramuscular ketorolac or a 5-minute OMT session. Both groups experienced a reduction in pain, 1.7±1.6 (95% CI, 1.1-2.3; P<.001) and 2.8 ±1.7 (95% CI, 2.1-3.4; P<.002), respectively. However, patients receiving OMT showed a significantly greater reduction in pain intensity compared with those receiving ketorolac (95% CI, 0.2-1.9; P=.02).¹⁷

Otitis media
One study examined the role of OMT in children who had experienced 3 episodes of acute otitis media (AOM) in the 6 months before study enrollment, or 4 episodes in the prior 12 months. Fifty-seven children ages 6 months to 6 years were randomized to receive usual care or usual care plus 7 OMT sessions over 6 months. The OMT group showed a significantly reduced number of AOM episodes and reduced referral for myringotomy/ ventilation tube placement compared with the control group. Additionally, final tympanograms showed an increased frequency of more normal tympanogram types in the OMT group (95% CI, 0.08-1.02; P=.02).¹⁸

Pediatric asthma
Using the registry of an asthma clinic, 1 study selected 140 subjects ages 5 to 17 years and randomized them to receive 1 OMT or 1 sham session, with peak expiratory flows (PEF) measured before and after treatment. The OMT group showed a significant mean increase in PEF from 364 to 377 L/min (95% CI, 7.3-18.7) compared with no change in the sham group.¹⁹

In the same year (2005), a Cochrane review analyzed 3 previous trials including 156 children and adults and found no significant difference in lung function measures with OMT or other manipulative or sham treatments.²⁰

Infantile colic
One study randomized 28 infants ages 1 to 12 weeks diagnosed with colic to receive 4 weekly OMT sessions or no treatment. At 4 weeks’ follow-up, the OMT group showed a significant reduction in parent-reported daily number of hours their infants spent crying, from 2.39 to 0.89 hours (P<.001), and a significant increase in the daily number of hours infants spent sleeping, from 11.55 to 12.9 hours (P<.002). The control group showed a nonsignificant reduction in daily number of hours infants spent crying, from 2.06 to 1.56 hours, and a nonsignificant increase in daily number of hours spent sleeping, from 11.86 to 12.04 hours.²¹

 

IBS, fibromyalgia …
A number of very small RCTs with equivocal results, pilot studies, and retrospective reviews have investigated the use of OMT in postsurgical functionality, irritable bowel syndrome, fibromyalgia, infantile torticollis, muscle spasticity, joint pain, labor pain, back pain during pregnancy, adult asthma, chronic obstructive pulmonary disease, and other medical conditions. Results to date have not been meaningful enough to recommend a place for OMT in the management of these disorders.

Limitations of evidence for OMT

Studies of OMT and other forms of spinal manipulation and manual modalities have been criticized for inconsistent quality.²² Sample sizes of published studies tend to be small, rendering statistical analysis problematic.

Pretrial bias of participants may also influence outcome measures. Patients tend to have preformed opinions regarding the efficacy of manual modalities.²²

The lack of validation of a placebo control has historically been problematic, and the use of sham treatment is an attempt to overcome this.²³ Some studies lack objective parameters for outcomes, relying on subjective patient ratings. Finally, severity of illness in chronic conditions such as back pain varies over time, affecting study results in follow-up.²⁴

CORRESPONDENCE Sarah Cole, DO, 12680 olive Boulevard, Suite 300, St. Louis, MO 63141; [email protected]

A mother brings her 5-year-old boy in to your office because she is concerned about a rash on his legs that seems to be worsening. She tells you that he had a runny nose and a mild cough a week earlier, but that those symptoms resolved before the rash developed. He has also complained of “belly pain.”

The boy’s mother says he’s been less active and more irritable since the onset of the rash, and that he is hardly eating. She also tells you that earlier in the day, her son told her that it hurts to walk.

You dig deeper…
A complete review of systems is otherwise negative. The 5-year-old was born at term without complication. He has met all developmental milestones and his immunizations are up to date. He takes no medications.

The boy’s vital signs are normal. He has an erythematous maculopapular rash distributed on his legs symmetrically; it is palpable, nontender, and nonblanching. You detect no abnormalities in abdominal, neurologic, or musculoskeletal examinations.

A complete blood count (CBC) and basic metabolic panel (BMP) reveal mild leukocytosis with a normal differential. Urinalysis shows moderate blood and trace protein. Laboratory results are otherwise normal.

WHAT IS YOUR DIAGNOSIS?

A classic presentation revealed: Henoch-Schönlein purpura

The combination of rash (FIGURE), abdominal pain, arthralgia, and evidence of renal involvement are the classic symptoms for Henoch-Schönlein purpura (HSP),¹ the most common systemic vasculitis of childhood.² Fever may also be present. Boys are affected nearly twice as often as girls. HSP occurs only rarely among adults, and men and women are affected equally. The estimated annual incidence in children is 20 cases per 100,000, with most cases occurring in those between the ages of 4 and 6 years.³

Although the exact cause of HSP is unknown, 75% of cases follow an upper respiratory infection. Streptococcus, Mycoplasma, adenovirus, parvovirus, Epstein-Barr virus, and varicella have all been implicated as offending pathogens. A number of case reports have also described the disease after the use of certain drugs, including penicillin, ampicillin, and quinine; and after the administration of vaccines, including those for typhoid, measles, yellow fever, and cholera. In all cases, the underlying mechanism is thought to be a systemic rise in immunoglobulin A (IgA), which forms immune complexes deposited in arterioles, capillaries, and venules.⁴ The precise interactions in the mechanism of disease are yet to be determined.

FIGURE
Rash and leg pain

Like the patient described in the text, this 11-year-old girl had a similar rash on her legs, as well as leg and abdominal pain.

HSP is usually benign, self-limited
HSP signs and symptoms may occur in any order over a period of days to weeks. Symptoms tend to be less severe in younger children than in older children and adults. Besides abdominal and renal systems, the lungs and central nervous system (CNS) may be affected, although rarely in children. Pulmonary involvement, if it does occur, usually manifests as diffuse alveolar hemorrhage or interstitial pneumonia or fibrosis.⁵ CNS vasculitis may lead to cerebral hemorrhage.⁶

HSP usually is a benign, self-limited disorder. The average disease course is 4 weeks, although it may be as brief as 3 days or as long as 2 years. Up to 94% of children (and 89% of adults) recover fully.⁷ However, the potential for chronic renal disease does exist.⁸

The recurrence rate of HSP is generally reported as 40%;⁸ patients who develop nephritis have higher recurrence rates.⁹

Rash is almost always present
Nearly all of those affected with HSP exhibit a rash. In children, it may begin as urticarial or maculopapular skin lesions. The rash typically is a palpable purpura, with lesions measuring 2 to 10 mm in diameter. Classically, the rash appears symmetrically on the extensor surfaces of the arms, legs, and buttocks. Lesions may also occur on the face and ears; however, the rash usually spares the trunk.² The rash generally fades over several days and gives way to darkened pigment. The purpuric lesions resolve more quickly with bed rest and tend to reappear when the patient resumes activity.

Bullous or necrotic lesions—although reported to occur in up to 60% of affected adults—are uncommon in children. If you had noted such lesions in this case, you would have had to rule out toxic vasculitides and meningococcal septicemia or other septic emboli.¹⁰

Arthralgia affects lower extremities
Arthralgia is the second most common clinical manifestation of HSP, occurring in approximately 82% of patients. Joint pain associated with HSP is likely due to periarticular soft tissue edema, and most commonly occurs in the hips, knees, or ankles.^1,11 The arthralgia of HSP is transient and self-limited and does not cause permanent damage to the joints. Like the rash, joint pain tends to decrease with bed rest and increase with activity.¹¹

 

Gastrointestinal involvement is rarely serious
Gastrointestinal involvement has been reported in 63% of patients with HSP. Symptoms typically include a colicky abdominal pain, associated with nausea and vomiting. These symptoms probably arise secondary to edema of the bowel wall and vasculitis of the gastrointestinal tract. The pain most often develops within 8 days of the appearance of the rash, but it has been reported to appear within weeks to months of cutaneous changes.¹²

Intussusception is the most common gastrointestinal complication in patients with HSP, with a reported overall incidence of 3.5% in hospitalized patients with severe abdominal pain. Gastrointestinal bleeding, presenting as either melena or hematochezia, has been reported in 25% of patients with HSP. Occult bleeding may occur in up to 50% of patients.¹²

More serious complications are rare, and they include bowel infarction and perforation, usually in the jejunum or ileum, and pancreatitis, cholecystitis, or protein-losing enteropathy secondary to HSP.

Renal involvement may require close follow-up
Renal disease has been reported to affect 30% to 70% of all HSP patients.¹³ Onset occurs within weeks to months after other symptoms of HSP. Most patients with renal involvement have only mild disease, such as asymptomatic hematuria and proteinuria.^1,7,13 HSP is thought to account for approximately 15% of all glomerulopathies in childhood.¹³

Although renal disease in HSP is generally benign, such complications as nephrotic syndrome, hypertension, and acute and chronic renal failure may occur. Adults are much more susceptible to the latter complication.

When the diagnosis is unclear, renal biopsy may help confirm the presence of disease. Evidence of HSP is identical to that seen with IgA nephropathy.^11,13 The percentage of glomeruli showing crescents on renal biopsy seems to be the most important prognostic finding, with crescent formation involving more than 50% of the glomeruli carrying a poor prognosis.¹⁴

Diagnosis is mainly clinical

Clinical diagnosis is not difficult with the classic 4 symptoms present (rash, abdominal pain, arthralgia, and evidence of renal involvement). However, when the presentation is less straightforward, confirmation of the diagnosis may depend on biopsy of the affected organ (eg, skin, kidney) demonstrating leukocytoclastic vasculitis with IgA deposition.

The differential diagnosis of HSP is large, and includes acute abdomen, meningococcal meningitis or septicemia, rheumatoid arthritis, idiopathic thrombocytopenic purpura, and systemic lupus erythematosus.¹⁵

Lab tests are minimally helpful. No lab test or imaging study is sensitive or specific for HSP. An elevated serum IgA level suggests the disease. A CBC may show leukocytosis and thrombocytosis, but test results may also be normal. The erythrocyte sedimentation rate may be elevated. On a BMP, blood urea nitrogen and creatinine levels may be elevated secondary to renal involvement or dehydration associated with HSP. Finally, hematuria or proteinuria show up on urinalysis in 30% to 70% of patients with HSP.^2,12,13

What treatment is indicated?

Most patients with HSP recover completely without any specific intervention other than reassurance, bed rest, and supportive care. Arthralgia usually responds to nonsteroidal anti-inflammatory drugs and corticosteroids. Hospitalization is warranted when patients have a depleted volume status or inadequately controlled pain. With an otherwise un-complicated illness, watchful management suffices. Gastrointestinal and renal complications may require more aggressive therapy.

Gastrointestinal complications. As noted earlier, the most common gastrointestinal complication is intussusception. No definite measures for preventing intussusception appear in the literature. Some evidence supports the use of corticosteroids for severe abdominal symptoms,^16,17 but in general, corticosteroids are not indicated for extrarenal manifestations. More serious complications such as bowel infarction or perforation, pancreatitis, or cholecystitis are rare.

Renal disease. Much research has focused on treatment options for patients with renal complications, due to the possibility of long-term debilitating effects. Although no evidence supports corticosteroid use for patients with mild renal involvement,¹⁶ for patients with severe renal disease—defined as crescenteric nephritis on biopsy, usually complicated by oliguria and hypertension—corticosteroids may help prevent irreversible glomerular injury.

Other agents that have been used for severe renal disease include azathioprine, cyclophosphamide, and dipyridamole. These drugs have shown some success in resolving symptoms, but their use remains controversial.^16,18 Finally, plasmapheresis has led to significant clinical improvement for a small number of patients with severe, rapidly progressing HSP.

Caveat. The effectiveness of all treatments for HSP remains in question, given that patients’ symptoms may simply resolve spontaneously. A clear answer will depend on further research.

Rest, fluids, and ibuprofen for your patient
You send your patient home with his mother, and advise her to provide him with supportive care, including oral hydration. You tell her to encourage rest and to give symptomatic pain relief with an over-the-counter medication such as ibuprofen.

 

You see the child again in 3 days, by which time his arthralgias are improving. Repeat urinalysis shows continuing moderate blood and trace protein. You follow him biweekly for 2 weeks, and then weekly for 4 more weeks. Six weeks after his initial presentation, his symptoms completely resolve, and urinalysis shows no evidence of hematuria or proteinuria. His blood pressure is normal at all visits.

You see the patient monthly for the next 5 months. Symptoms do not recur, and urinalysis results are normal. You counsel the boy’s mother regarding the risk of recurrence, and advise her to contact the office immediately if symptoms return.

CORRESPONDENCE Shailendra K. Saxena, MD, PhD, Department of Family Medicine, 10828 John Galt Boulevard, Omaha, NE 68137; [email protected]