Drug Therapy

The medical management of non-ST-elevation acute coronary syndromes focuses on blocking the coagulation cascade and inhibiting platelets. This—plus diagnostic angiography followed, if needed, by revascularization—has reduced the rates of death and recurrent ischemic events.¹ However, the combination of potent antithrombotic drugs and invasive procedures also increases the risk of bleeding.

This review discusses the incidence and complications associated with bleeding during the treatment of acute coronary syndromes and summarizes recommendations for preventing and managing bleeding in this setting.

THE TRUE INCIDENCE OF BLEEDING IS HARD TO DETERMINE

The optimal way to detect and analyze bleeding events in clinical trials and registries is highly debated. The reported incidences of bleeding during antithrombotic and antiplatelet therapy for non-ST-elevation acute coronary syndromes depend on how bleeding was defined, how the acute coronary syndromes were treated, and on other factors such as how the study was designed.

How was bleeding defined?

Since these classification schemes are based on different types of data, they yield different numbers when applied to the same study population. For instance, Rao et al⁴ pooled the data from the PURSUIT and PARAGON B trials (15,454 patients in all) and found that the incidence of severe bleeding (by the GUSTO criteria) was 1.2%, while the rate of major bleeding (by the TIMI criteria) was 8.2%.

What was the treatment strategy?

Another reason that the true incidence of bleeding is hard to determine is that different studies used treatment strategies that differed in the type, timing, and dose of antithrombotic agents and whether invasive procedures were used early. For example, if unfractionated heparin is used aggressively in regimens that are not adjusted for weight and with a higher target for the activated clotting time, the risk of bleeding is higher than with conservative dosing.^5–7

Subherwal et al⁸ evaluated the effect of treatment strategy on the incidence of bleeding in patients with non-ST-elevation acute coronary syndromes who received two or more antithrombotic drugs in the CRUSADE Quality Improvement Initiative. The risk of bleeding was higher with an invasive approach (catheterization) than with a conservative approach (no catheterization), regardless of baseline bleeding risk.

What type of study was it?

Another source of variation is the design of the study. Registries differ from clinical trials in patient characteristics and in the way data are gathered (prospectively vs retrospectively).

In registries, data are often collected retrospectively, whereas in clinical trials the data are prospectively collected. For this reason, the definition of bleeding in registries is often based on events that are easily identified through chart review, such as transfusion. This may lead to a lower reported rate of bleeding, since other, less serious bleeding events such as access-site hematomas and epistaxis may not be documented in the medical record.

On the other hand, registries often include older and sicker patients, who may be more prone to bleeding and who are often excluded from clinical trials. This may lead to a higher rate of reported bleeding.⁹

Where the study was conducted makes a difference as well, owing to regional practice differences. For example, Moscucci et al¹⁰ reported that the incidence of major bleeding in 24,045 patients with non-ST-elevation acute coronary syndromes in the GRACE registry (in 14 countries worldwide) was 3.9%. In contrast, Yang et al¹¹ reported that the rate of bleeding in the CRUSADE registry (in the United States) was 10.3%.

This difference was partly influenced by different definitions of bleeding. The GRACE registry defined major bleeding as life-threatening events requiring transfusion of two or more units of packed red blood cells, or resulting in an absolute decrease in the hematocrit of 10% or more or death, or hemorrhagic subdural hematoma. In contrast, the CRUSADE data reflect bleeding requiring transfusion. However, practice patterns such as greater use of invasive procedures in the United States may also be responsible.

Rao and colleagues¹² examined international variation in blood transfusion rates among patients with acute coronary syndromes. Patients outside the United States were significantly less likely to receive transfusions, even after adjusting for patient and practice differences.

Taking these confounders into account, it is reasonable to estimate that the frequency of bleeding in patients with non-ST-elevation acute coronary syndromes ranges from less than 1% to 10%.¹³

BLEEDING IS ASSOCIATED WITH POOR OUTCOMES

Regardless of the definition or the data source, hemorrhagic complications are associated with a higher risk of death and nonfatal adverse events, both in the short term and in the long term.

Short-term outcomes

A higher risk of death. In the GRACE registry study by Moscucci et al¹⁰ discussed above, patients who had major bleeding were significantly more likely to die during their hospitalization than those who did not (odds ratio [OR] 1.64, 95% confidence interval [CI] 1.18–2.28).

Rao et al¹⁴ evaluated pooled data from the multicenter international GUSTO IIb, PURSUIT, and PARAGON A and B trials and found that the effects of bleeding in non-ST-elevation acute coronary syndromes extended beyond the hospital stay. The more severe the bleeding (by the GUSTO criteria), the greater the adjusted hazard ratio (HR) for death within 30 days:

• With mild bleeding—HR 1.6, 95% CI 1.3–1.9
• With moderate bleeding—HR 2.7, 95% CI 2.3–3.4
• With severe bleeding—HR 10.6, 95% CI 8.3–13.6.

The pattern was the same for death within 6 months:

• With mild bleeding—HR 1.4, 95% CI 1.2–1.6
• With moderate bleeding—HR 2.1, 95% CI 1.8–2.4
• With severe bleeding, HR 7.5, 95% CI 6.1–9.3.

These findings were confirmed by Eikelboom et al¹⁵ in 34,146 patients with acute coronary syndromes in the OASIS registry, the OASIS-2 trial, and the CURE randomized trial. In the first 30 days, five times as many patients died (12.8% vs 2.5%; P < .0009) among those who developed major bleeding compared with those who did not. These investigators defined major bleeding as bleeding that was life-threatening or significantly disabling or that required transfusion of two or more units of packed red blood cells.

A higher risk of nonfatal adverse events. Bleeding after antithrombotic therapy for non-ST-elevation acute coronary syndromes has also been associated with nonfatal adverse events such as stroke and stent thrombosis.

For example, in the study by Eikelboom et al,¹⁵ major bleeding was associated with a higher risk of recurrent ischemic events. Approximately 1 in 5 patients in the OASIS trials who developed major bleeding during the first 30 days died or had a myocardial infarction or stroke by 30 days, compared with 1 in 20 of those who did not develop major bleeding during the first 30 days. However, after events that occurred during the first 30 days were excluded, the association between major bleeding and both myocardial infarction and stroke was no longer evident between 30 days and 6 months.

Manoukian et al¹⁶ evaluated the impact of major bleeding in 13,819 patients with highrisk acute coronary syndromes undergoing treatment with an early invasive strategy in the ACUITY trial. At 30 days, patients with major bleeding had higher rates of the composite end point of death, myocardial infarction, or unplanned revascularization for ischemia (23.1% vs 6.8%, P < .0001) and of stent thrombosis (3.4% vs 0.6%, P < .0001).

Long-term outcomes

The association between bleeding and adverse outcomes persists in the long term as well, although the mechanisms underlying this association are not well studied.

Kinnaird et al¹⁷ examined the data from 10,974 unselected patients who underwent percutaneous coronary intervention. At 1 year, the following percentages of patients had died:

• After TIMI major bleeding—17.2%
• After TIMI minor bleeding—9.1%
• After no bleeding—5.5%.

However, after adjustment for potential confounders, only transfusion remained a significant predictor of 1-year mortality.

Mehran et al¹⁸ evaluated 1-year mortality data from the ACUITY trial. Compared with the rate in patients who had no major bleeding and no myocardial infarction, the hazard ratios for death were:

• After major bleeding—HR 3.5, 95% CI 2.7–4.4
• After myocardial infarction—HR 3.1, 95% CI 2.4–3.9.

Interestingly, the risk of death associated with myocardial infarction abated after 7 days, while the risk associated with bleeding persisted beyond 30 days and remained constant throughout the first year following the bleeding event.

Similarly, Ndrepepa and colleagues¹⁹ examined pooled data from four ISAR trials using the TIMI bleeding scale and found that myocardial infarction, target vessel revascularization, and major bleeding all had similar discriminatory ability at predicting 1-year mortality.

In patients undergoing elective or urgent percutaneous coronary intervention in the REPLACE-2 trial,²⁰ independent predictors of death by 1 year were²¹:

• Major hemorrhage (OR 2.66, 95% CI 1.44–4.92)
• Periprocedural myocardial infarction (OR 2.46, 95% CI 1.44–4.20).

THEORIES OF HOW BLEEDING MAY CAUSE ADVERSE OUTCOMES

Several mechanisms have been proposed to explain the association between bleeding during treatment for acute coronary syndromes and adverse clinical outcomes.^13,22

The immediate effects of bleeding are thought to be hypotension and a reflex hyperadrenergic state to compensate for the loss of intravascular volume.²³ This physiologic response is believed to contribute to myocardial ischemia by further decreasing myocardial oxygen supply in obstructive coronary disease.

Trying to minimize blood loss, physicians may withhold anticoagulation and antiplatelet therapy, which in turn may lead to further ischemia.²⁴ To compensate for blood loss, physicians may also resort to blood transfusion. However, depletion of 2,3-diphosphoglycerate and nitric oxide in stored donor red blood cells is postulated to reduce oxygen delivery by increasing hemoglobin’s affinity for oxygen, leading to induced microvascular obstruction and adverse inflammatory reactions.^15,25

Recent data have also begun to elucidate the long-term effects of bleeding during acute coronary syndrome management. Patients with anemia during the acute phase of infarction have greater neurohormonal activation.²⁶ These adaptive responses to anemia may lead to eccentric left ventricular remodeling that may lead to higher oxygen consumption, increased diastolic wall stress, interstitial fibrosis, and accelerated myocyte loss.^27–30

Nevertheless, we must point out that although strong associations between bleeding and adverse outcomes have been established, direct causality has not.

TO PREVENT BLEEDING, START BY ASSESSING RISK

The CRUSADE bleeding risk score

The CRUSADE bleeding score (calculator available at http://www.crusadebleedingscore.org/) was developed and validated in more than 89,000 community-treated patients with non-ST-elevation acute coronary syndromes.⁸ It is based on eight variables:

• Sex (higher risk in women)
• History of diabetes (higher risk)
• Prior vascular disease (higher risk)
• Heart rate (the higher the rate, the higher the risk)
• Systolic blood pressure (higher risk with pressures above or below the 121–180 mm Hg range)
• Signs of congestive heart failure (higher risk)
• Baseline hematocrit (the lower the hematocrit, the higher the risk)
• Creatinine clearance (by the Cockcroft-Gault formula; the lower the creatinine clearance, the higher the risk).

Patients who are found to have bleeding scores suggesting a moderate or higher risk of bleeding should be considered for medications associated with a favorable bleeding profile, and for radial access at the time of coronary angiography. Scores are graded as follows⁸:

• < 21: Very low risk
• 21–30: Low risk
• 31–40: Moderate risk
• 41–50: High risk
• > 50: Very high risk.

The CRUSADE bleeding score is unique in that, unlike earlier risk stratification tools, it was developed in a “real world” population, not in subgroups or in a clinical trial. It can be calculated at baseline to help guide the selection of treatment.⁸

Adjusting the heparin regimen in patients at risk of bleeding

Both the joint American College of Cardiology/American Heart Association¹ and the European Society of Cardiology guidelines³¹ for the treatment of non-ST-elevation acute coronary syndromes recommend taking steps to prevent bleeding, such as adjusting the dosage of unfractionated heparin, using safer drugs, reducing the duration of antithrombotic treatment, and using combinations of antithrombotic and antiplatelet agents according to proven indications.³¹

In the CRUSADE registry, 42% of patients with non-ST-elevation acute coronary syndromes received at least one initial dose of antithrombotic drug outside the recommended range, resulting in an estimated 15% excess of bleeding events.³² Thus, proper dosing is a target for prevention.

Appropriate antithrombotic dosing takes into account the patient’s age, weight, and renal function. However, heparin dosage in the current guidelines¹ is based on weight only: a loading dose of 60 U/kg (maximum 4,000 U) by intravenous bolus, then 12 U/kg/hour (maximum 1,000 U/hour) to maintain an activated partial thromboplastin time of 50 to 70 seconds.¹

Renal dysfunction is particularly worrisome in patients with non-ST-elevation acute coronary syndromes because it is associated with higher rates of major bleeding and death. In the OASIS-5 trial,³³ the overall risk of death was approximately five times higher in patients in the lowest quartile of renal function (glomerular filtration rate < 58 mL/min/1.73 m²) than in the highest quartile (glomerular filtration rate ≥ 86 mL/min/1.73 m²).

Renal function must be evaluated not only on admission but also throughout the hospital stay. Patients presenting with acute coronary syndromes often experience fluctuations in renal function that would call for adjustment of heparin dosing, either increasing the dose to maximize the drug’s efficacy if renal function is recovering or decreasing the dose to prevent bleeding if renal function is deteriorating.

Clopidogrel vs prasugrel

Certain medications should be avoided when the risk of bleeding outweighs any potential benefit in terms of ischemia.

For example, in a randomized trial,³⁴ prasugrel (Effient), a potent thienopyridine, was associated with a significantly lower rate of the composite end point of stroke, myocardial infarction, or death than clopidogrel (Plavix) in patients with acute coronary syndromes undergoing percutaneous coronary interventions. However, it did not seem to offer any advantage in patients 75 years old and older, those with prior stroke or transient ischemic attack, or those weighing less than 60 kg, and it posed a substantially higher risk of bleeding.

With clopidogrel, the risk of acute bleeding is primarily in patients who undergo coronary artery bypass grafting within 5 days of receiving a dose.^35,36 Therefore, clopidogrel should be stopped 5 to 7 days before bypass surgery.¹ Importantly, there is no increased risk of recurrent ischemic events during this 5-day waiting period in patients who receive clopidogrel early. Therefore, the recommendation to stop clopidogrel before surgery does not negate the benefits of early treatment.³⁶

Lower-risk drugs: Fondaparinux and bivalirudin

At this time, only two agents have been studied in clinical trials that have specifically focused on reducing bleeding risk: fondaparinux (Arixtra) and bivalirudin (Angiomax).^20,37–39

Fondaparinux

OASIS-5 was a randomized, double-blind trial that compared fondaparinux and enoxaparin (Lovenox) in patients with acute coronary syndromes.³⁸ Fondaparinux was similar to enoxaparin in terms of the combined end point of death, myocardial infarction, or refractory ischemia at 9 days, and fewer patients on fondaparinux developed bleeding (2.2% vs 4.1%, HR 0.52; 95% CI 0.44–0.61). This difference persisted during long-term follow-up.

Importantly, fewer patients died in the fondaparinux group. At 180 days, 638 (6.5%) of the patients in the enoxaparin group had died, compared with 574 (5.8%) in the fondaparinux group, a difference of 64 deaths (P = .05). The authors found that 41 fewer patients in the fondaparinux group than in the enoxaparin group died after major bleeding, and 20 fewer patients in the fondaparinux group died after minor bleeding.³⁸ Thus, most of the difference in mortality rates between the two groups was attributed to a lower incidence of bleeding with fondaparinux.

Unfortunately, despite its safe bleeding profile, fondaparinux has fallen out of favor for use in acute coronary syndromes, owing to a higher risk of catheter thrombosis in the fondaparinux group (0.9%) than in those undergoing percutaneous coronary interventions with enoxaparin alone (0.4%) in the OASIS-5 trial.⁴⁰

The direct thrombin inhibitor bivalirudin has been studied in three large randomized trials in patients undergoing percutaneous coronary interventions.^20,37,41

The ACUITY trial³⁷ was a prospective, open-label, randomized, multicenter trial that compared three regimens in patients with moderate or high-risk non-ST-elevation acute coronary syndromes:

• Heparin plus a glycoprotein IIb/IIIa inhibitor
• Bivalirudin plus a glycoprotein IIb/IIIa inhibitor
• Bivalirudin alone.

Bivalirudin alone was as effective as heparin plus a glycoprotein IIb/IIIa inhibitor with respect to the composite ischemia end point, which at 30 days had occurred in 7.8% vs 7.3% of the patients in these treatment groups (P = .32, RR 1.08; 95% CI 0.93–1.24), and it was superior with respect to major bleeding (3.0% vs 5.7%, P < .001, RR 0.53; 95% CI 0.43–0.65).

The HORIZONS-AMI study⁴¹ was a prospective, open-label, randomized, multicenter trial that compared bivalirudin alone vs heparin plus a glycoprotein IIb/IIIa inhibitor in patients with ST-elevation acute coronary syndromes who were undergoing primary percutaneous coronary interventions. The two primary end points were major bleeding and net adverse events.

At 1 year, patients assigned to bivalirudin had a lower rate of major bleeding than did controls (5.8% vs 9.2%, HR 0.61, 95% CI 0.48–0.78, P < .0001), with similar rates of major adverse cardiac events in both groups (11.9% vs 11.9%, HR 1.00, 95% CI 0.82– 1.21, P = .98).⁴¹

Both OASIS 5 and HORIZONS-AMI are examples of clinical trials in which strategies that reduced bleeding risk were also associated with improved survival.

For cardiac catheterization, inserting the catheter in the wrist poses less risk

Bleeding is currently the most common noncardiac complication in patients undergoing percutaneous coronary interventions, and it most often occurs at the vascular access site.¹⁷

Rao et al¹² evaluated data from 593,094 procedures in the National Cardiovascular Data Registry and found that, compared with the femoral approach, the use of transradial percutaneous coronary intervention was associated with a similar rate of procedural success (OR 1.02, 95% CI 0.93–1.12) but a significantly lower risk of bleeding complications (OR 0.42, 95% CI 0.31–0.56) after multivariable adjustment.

The use of smaller sheath sizes (4F–6F) and preferential use of bivalirudin over unfractionated heparin and glycoprotein IIb/IIIa inhibitor therapy are other methods described to decrease the risk of bleeding after percutaneous coronary interventions.^20,41–49

IF BLEEDING OCCURS

Once a bleeding complication occurs, cessation of therapy is a potential option. Stopping or reversing antithrombotic and antiplatelet therapy is warranted in the event of major bleeding (eg, gastrointestinal, retroperitoneal, intracranial).³¹

Stopping antithrombotic and antiplatelet therapy

Whether bleeding is minor or major, the risk of a recurrent thrombotic event must be considered, especially in patients who have undergone revascularization, stent implantation, or both. The risk of acute thrombotic events after interrupting antithrombotic or antiplatelet agents is considered greatest 4 to 5 days following revascularization or percutaneous coronary intervention.¹⁵ If bleeding can be controlled with local treatment such as pressure, packing, or dressing, antithrombotic and antiplatelet therapy need not be interrupted.⁵⁰

Current guidelines recommend strict control of hemorrhage for at least 24 hours before reintroducing antiplatelet or antithrombotic agents.

It is also important to remember that in the setting of gastrointestinal bleeding due to peptic ulcer disease, adjunctive proton pump inhibitors are recommended after restarting antiplatelet or antithrombotic therapy or both.

Importantly, evidence-based antithrombotic medications (especially dual antiplatelet therapy) should be restarted once the acute bleeding event has resolved.³¹

Reversal of anticoagulant and antiplatelet therapies

Unfractionated heparin is reversed with infusion of protamine sulfate at a dose of 1 mg per 100 U of unfractionated heparin given over the previous 4 hours.^51,52 The rate of protamine sulfate infusion should be less than 100 mg over 2 hours, with 50% of the dose given initially and subsequent doses titrated according to bleeding response.^52,53 Protamine sulfate is associated with a risk of hypotension and bradycardia, and for this reason it should be given no faster than 5 mg/min.

Low-molecular-weight heparin (LMWH) can be inhibited by 1 mg of protamine sulfate for each 1 mg of LMWH given over the previous 4 hours.^51,52

However, protamine sulfate only partially neutralizes the anticoagulant effect of LMWH. In cases in which protamine sulfate is unsuccessful in abating bleeding associated with LMWH use, guidelines allow for the use of recombinant factor VIIa (NovoSeven).³¹ In healthy volunteers given fondaparinux, recombinant factor VIIa normalized coagulation times and thrombin generation within 1.5 hours, with a sustained effect for 6 hours.⁵²

It is important to note that the use of this agent has not been fully studied, it is very costly (a single dose of 40 μg/kg costs from $3,000 to $4,000), and it is linked to reports of increased risk of thrombotic complications.^54,55

Antiplatelet agents are more complex to reverse. The antiplatelet actions of aspirin and clopidogrel wear off as new platelets are produced. Approximately 10% of a patient’s platelet count is produced daily; thus, the antiplatelet effects of aspirin and clopidogrel can persist for 5 to 10 days.^31,56

If these agents need to be reversed quickly to stop bleeding, according to expert consensus the aspirin effect can be reversed by transfusion of one unit of platelets. The antiplatelet effect of clopidogrel is more significant than that of aspirin; thus, two units of platelets are recommended.⁵⁶

Glycoprotein IIb/IIIa inhibitors. If a major bleeding event requires the reversal of glycoprotein IIb/IIIa inhibitor therapy, the treatment must take into consideration the pharmacodynamics of the target drug. Both eptifibatide (Integrilin) and tirofiban (Aggrastat) competitively inhibit glycoprotein IIb/IIIa receptors; thus, their effects depend on dosing, elimination, and time. Due to the stoichiometry of both drugs, transfusion of platelets is ineffective. Both eptifibatide and tirofiban are eliminated by the kidney; thus, normal renal function is key to the amount of time it takes for platelet function to return to baseline.⁵⁷ Evidence suggests that fibrinogen-rich plasma can be administered to restore platelet function.^31,58,59

Abciximab (ReoPro). Whereas reversal of eptifibatide and tirofiban focuses on overcoming competitive inhibition, neutralization of abciximab involves overcoming its high receptor affinity. At 24 hours after abciximab infusion is stopped, platelet aggregation may still be inhibited by up to 50%. Fortunately, owing to abciximab’s short plasma half-life and its dilution in serum, platelet transfusion is effective in reversing its antiplatelet effects.^31,57

Long considered beneficial to critically ill patients, blood transfusion to maintain hematocrit levels during acute coronary syndromes has come under intense scrutiny. Randomized trials have shown that transfusion should not be given aggressively to critically ill patients.⁶⁰ In acute coronary syndromes, there are only observational data.

Rao et al⁶¹ used detailed clinical data from 24,112 patients with acute coronary syndromes in the GUSTO IIb, PURSUIT, and PARAGON B trials to determine the association between blood transfusion and outcomes in patients who developed moderate to severe bleeding, anemia, or both during their hospitalization. The rates of death in the hospital and at 30 days were significantly higher in patients who received a transfusion (30-day mortality HR 3.94; 95% CI 3.36–4.75). However, there was no significant association between transfusion and the 30-day mortality rate if the nadir hematocrit was 25% or less.

Of note: no randomized clinical trial has evaluated transfusion strategies in acute coronary syndromes at this time. Until such data are available, it is reasonable to follow published guidelines and to avoid transfusion in stable patients with ischemic heart disease unless the hematocrit is 25% or less.³¹ The risks and benefits of blood transfusion should be carefully weighed. Routine use of transfusion to maintain predefined hemoglobin levels is not recommended in stable patients.

The medical management of non-ST-elevation acute coronary syndromes focuses on blocking the coagulation cascade and inhibiting platelets. This—plus diagnostic angiography followed, if needed, by revascularization—has reduced the rates of death and recurrent ischemic events.¹ However, the combination of potent antithrombotic drugs and invasive procedures also increases the risk of bleeding.

This review discusses the incidence and complications associated with bleeding during the treatment of acute coronary syndromes and summarizes recommendations for preventing and managing bleeding in this setting.

THE TRUE INCIDENCE OF BLEEDING IS HARD TO DETERMINE

The optimal way to detect and analyze bleeding events in clinical trials and registries is highly debated. The reported incidences of bleeding during antithrombotic and antiplatelet therapy for non-ST-elevation acute coronary syndromes depend on how bleeding was defined, how the acute coronary syndromes were treated, and on other factors such as how the study was designed.

How was bleeding defined?

Since these classification schemes are based on different types of data, they yield different numbers when applied to the same study population. For instance, Rao et al⁴ pooled the data from the PURSUIT and PARAGON B trials (15,454 patients in all) and found that the incidence of severe bleeding (by the GUSTO criteria) was 1.2%, while the rate of major bleeding (by the TIMI criteria) was 8.2%.

What was the treatment strategy?

Another reason that the true incidence of bleeding is hard to determine is that different studies used treatment strategies that differed in the type, timing, and dose of antithrombotic agents and whether invasive procedures were used early. For example, if unfractionated heparin is used aggressively in regimens that are not adjusted for weight and with a higher target for the activated clotting time, the risk of bleeding is higher than with conservative dosing.^5–7

Subherwal et al⁸ evaluated the effect of treatment strategy on the incidence of bleeding in patients with non-ST-elevation acute coronary syndromes who received two or more antithrombotic drugs in the CRUSADE Quality Improvement Initiative. The risk of bleeding was higher with an invasive approach (catheterization) than with a conservative approach (no catheterization), regardless of baseline bleeding risk.

What type of study was it?

Another source of variation is the design of the study. Registries differ from clinical trials in patient characteristics and in the way data are gathered (prospectively vs retrospectively).

In registries, data are often collected retrospectively, whereas in clinical trials the data are prospectively collected. For this reason, the definition of bleeding in registries is often based on events that are easily identified through chart review, such as transfusion. This may lead to a lower reported rate of bleeding, since other, less serious bleeding events such as access-site hematomas and epistaxis may not be documented in the medical record.

On the other hand, registries often include older and sicker patients, who may be more prone to bleeding and who are often excluded from clinical trials. This may lead to a higher rate of reported bleeding.⁹

Where the study was conducted makes a difference as well, owing to regional practice differences. For example, Moscucci et al¹⁰ reported that the incidence of major bleeding in 24,045 patients with non-ST-elevation acute coronary syndromes in the GRACE registry (in 14 countries worldwide) was 3.9%. In contrast, Yang et al¹¹ reported that the rate of bleeding in the CRUSADE registry (in the United States) was 10.3%.

This difference was partly influenced by different definitions of bleeding. The GRACE registry defined major bleeding as life-threatening events requiring transfusion of two or more units of packed red blood cells, or resulting in an absolute decrease in the hematocrit of 10% or more or death, or hemorrhagic subdural hematoma. In contrast, the CRUSADE data reflect bleeding requiring transfusion. However, practice patterns such as greater use of invasive procedures in the United States may also be responsible.

Rao and colleagues¹² examined international variation in blood transfusion rates among patients with acute coronary syndromes. Patients outside the United States were significantly less likely to receive transfusions, even after adjusting for patient and practice differences.

Taking these confounders into account, it is reasonable to estimate that the frequency of bleeding in patients with non-ST-elevation acute coronary syndromes ranges from less than 1% to 10%.¹³

BLEEDING IS ASSOCIATED WITH POOR OUTCOMES

Regardless of the definition or the data source, hemorrhagic complications are associated with a higher risk of death and nonfatal adverse events, both in the short term and in the long term.

Short-term outcomes

A higher risk of death. In the GRACE registry study by Moscucci et al¹⁰ discussed above, patients who had major bleeding were significantly more likely to die during their hospitalization than those who did not (odds ratio [OR] 1.64, 95% confidence interval [CI] 1.18–2.28).

Rao et al¹⁴ evaluated pooled data from the multicenter international GUSTO IIb, PURSUIT, and PARAGON A and B trials and found that the effects of bleeding in non-ST-elevation acute coronary syndromes extended beyond the hospital stay. The more severe the bleeding (by the GUSTO criteria), the greater the adjusted hazard ratio (HR) for death within 30 days:

• With mild bleeding—HR 1.6, 95% CI 1.3–1.9
• With moderate bleeding—HR 2.7, 95% CI 2.3–3.4
• With severe bleeding—HR 10.6, 95% CI 8.3–13.6.

The pattern was the same for death within 6 months:

• With mild bleeding—HR 1.4, 95% CI 1.2–1.6
• With moderate bleeding—HR 2.1, 95% CI 1.8–2.4
• With severe bleeding, HR 7.5, 95% CI 6.1–9.3.

These findings were confirmed by Eikelboom et al¹⁵ in 34,146 patients with acute coronary syndromes in the OASIS registry, the OASIS-2 trial, and the CURE randomized trial. In the first 30 days, five times as many patients died (12.8% vs 2.5%; P < .0009) among those who developed major bleeding compared with those who did not. These investigators defined major bleeding as bleeding that was life-threatening or significantly disabling or that required transfusion of two or more units of packed red blood cells.

A higher risk of nonfatal adverse events. Bleeding after antithrombotic therapy for non-ST-elevation acute coronary syndromes has also been associated with nonfatal adverse events such as stroke and stent thrombosis.

For example, in the study by Eikelboom et al,¹⁵ major bleeding was associated with a higher risk of recurrent ischemic events. Approximately 1 in 5 patients in the OASIS trials who developed major bleeding during the first 30 days died or had a myocardial infarction or stroke by 30 days, compared with 1 in 20 of those who did not develop major bleeding during the first 30 days. However, after events that occurred during the first 30 days were excluded, the association between major bleeding and both myocardial infarction and stroke was no longer evident between 30 days and 6 months.

Manoukian et al¹⁶ evaluated the impact of major bleeding in 13,819 patients with highrisk acute coronary syndromes undergoing treatment with an early invasive strategy in the ACUITY trial. At 30 days, patients with major bleeding had higher rates of the composite end point of death, myocardial infarction, or unplanned revascularization for ischemia (23.1% vs 6.8%, P < .0001) and of stent thrombosis (3.4% vs 0.6%, P < .0001).

Long-term outcomes

The association between bleeding and adverse outcomes persists in the long term as well, although the mechanisms underlying this association are not well studied.

Kinnaird et al¹⁷ examined the data from 10,974 unselected patients who underwent percutaneous coronary intervention. At 1 year, the following percentages of patients had died:

• After TIMI major bleeding—17.2%
• After TIMI minor bleeding—9.1%
• After no bleeding—5.5%.

However, after adjustment for potential confounders, only transfusion remained a significant predictor of 1-year mortality.

Mehran et al¹⁸ evaluated 1-year mortality data from the ACUITY trial. Compared with the rate in patients who had no major bleeding and no myocardial infarction, the hazard ratios for death were:

• After major bleeding—HR 3.5, 95% CI 2.7–4.4
• After myocardial infarction—HR 3.1, 95% CI 2.4–3.9.

Interestingly, the risk of death associated with myocardial infarction abated after 7 days, while the risk associated with bleeding persisted beyond 30 days and remained constant throughout the first year following the bleeding event.

Similarly, Ndrepepa and colleagues¹⁹ examined pooled data from four ISAR trials using the TIMI bleeding scale and found that myocardial infarction, target vessel revascularization, and major bleeding all had similar discriminatory ability at predicting 1-year mortality.

In patients undergoing elective or urgent percutaneous coronary intervention in the REPLACE-2 trial,²⁰ independent predictors of death by 1 year were²¹:

• Major hemorrhage (OR 2.66, 95% CI 1.44–4.92)
• Periprocedural myocardial infarction (OR 2.46, 95% CI 1.44–4.20).

THEORIES OF HOW BLEEDING MAY CAUSE ADVERSE OUTCOMES

Several mechanisms have been proposed to explain the association between bleeding during treatment for acute coronary syndromes and adverse clinical outcomes.^13,22

The immediate effects of bleeding are thought to be hypotension and a reflex hyperadrenergic state to compensate for the loss of intravascular volume.²³ This physiologic response is believed to contribute to myocardial ischemia by further decreasing myocardial oxygen supply in obstructive coronary disease.

Trying to minimize blood loss, physicians may withhold anticoagulation and antiplatelet therapy, which in turn may lead to further ischemia.²⁴ To compensate for blood loss, physicians may also resort to blood transfusion. However, depletion of 2,3-diphosphoglycerate and nitric oxide in stored donor red blood cells is postulated to reduce oxygen delivery by increasing hemoglobin’s affinity for oxygen, leading to induced microvascular obstruction and adverse inflammatory reactions.^15,25

Recent data have also begun to elucidate the long-term effects of bleeding during acute coronary syndrome management. Patients with anemia during the acute phase of infarction have greater neurohormonal activation.²⁶ These adaptive responses to anemia may lead to eccentric left ventricular remodeling that may lead to higher oxygen consumption, increased diastolic wall stress, interstitial fibrosis, and accelerated myocyte loss.^27–30

Nevertheless, we must point out that although strong associations between bleeding and adverse outcomes have been established, direct causality has not.

TO PREVENT BLEEDING, START BY ASSESSING RISK

The CRUSADE bleeding risk score

The CRUSADE bleeding score (calculator available at http://www.crusadebleedingscore.org/) was developed and validated in more than 89,000 community-treated patients with non-ST-elevation acute coronary syndromes.⁸ It is based on eight variables:

• Sex (higher risk in women)
• History of diabetes (higher risk)
• Prior vascular disease (higher risk)
• Heart rate (the higher the rate, the higher the risk)
• Systolic blood pressure (higher risk with pressures above or below the 121–180 mm Hg range)
• Signs of congestive heart failure (higher risk)
• Baseline hematocrit (the lower the hematocrit, the higher the risk)
• Creatinine clearance (by the Cockcroft-Gault formula; the lower the creatinine clearance, the higher the risk).

Patients who are found to have bleeding scores suggesting a moderate or higher risk of bleeding should be considered for medications associated with a favorable bleeding profile, and for radial access at the time of coronary angiography. Scores are graded as follows⁸:

• < 21: Very low risk
• 21–30: Low risk
• 31–40: Moderate risk
• 41–50: High risk
• > 50: Very high risk.

The CRUSADE bleeding score is unique in that, unlike earlier risk stratification tools, it was developed in a “real world” population, not in subgroups or in a clinical trial. It can be calculated at baseline to help guide the selection of treatment.⁸

Adjusting the heparin regimen in patients at risk of bleeding

Both the joint American College of Cardiology/American Heart Association¹ and the European Society of Cardiology guidelines³¹ for the treatment of non-ST-elevation acute coronary syndromes recommend taking steps to prevent bleeding, such as adjusting the dosage of unfractionated heparin, using safer drugs, reducing the duration of antithrombotic treatment, and using combinations of antithrombotic and antiplatelet agents according to proven indications.³¹

In the CRUSADE registry, 42% of patients with non-ST-elevation acute coronary syndromes received at least one initial dose of antithrombotic drug outside the recommended range, resulting in an estimated 15% excess of bleeding events.³² Thus, proper dosing is a target for prevention.

Appropriate antithrombotic dosing takes into account the patient’s age, weight, and renal function. However, heparin dosage in the current guidelines¹ is based on weight only: a loading dose of 60 U/kg (maximum 4,000 U) by intravenous bolus, then 12 U/kg/hour (maximum 1,000 U/hour) to maintain an activated partial thromboplastin time of 50 to 70 seconds.¹

Renal dysfunction is particularly worrisome in patients with non-ST-elevation acute coronary syndromes because it is associated with higher rates of major bleeding and death. In the OASIS-5 trial,³³ the overall risk of death was approximately five times higher in patients in the lowest quartile of renal function (glomerular filtration rate < 58 mL/min/1.73 m²) than in the highest quartile (glomerular filtration rate ≥ 86 mL/min/1.73 m²).

Renal function must be evaluated not only on admission but also throughout the hospital stay. Patients presenting with acute coronary syndromes often experience fluctuations in renal function that would call for adjustment of heparin dosing, either increasing the dose to maximize the drug’s efficacy if renal function is recovering or decreasing the dose to prevent bleeding if renal function is deteriorating.

Clopidogrel vs prasugrel

Certain medications should be avoided when the risk of bleeding outweighs any potential benefit in terms of ischemia.

For example, in a randomized trial,³⁴ prasugrel (Effient), a potent thienopyridine, was associated with a significantly lower rate of the composite end point of stroke, myocardial infarction, or death than clopidogrel (Plavix) in patients with acute coronary syndromes undergoing percutaneous coronary interventions. However, it did not seem to offer any advantage in patients 75 years old and older, those with prior stroke or transient ischemic attack, or those weighing less than 60 kg, and it posed a substantially higher risk of bleeding.

With clopidogrel, the risk of acute bleeding is primarily in patients who undergo coronary artery bypass grafting within 5 days of receiving a dose.^35,36 Therefore, clopidogrel should be stopped 5 to 7 days before bypass surgery.¹ Importantly, there is no increased risk of recurrent ischemic events during this 5-day waiting period in patients who receive clopidogrel early. Therefore, the recommendation to stop clopidogrel before surgery does not negate the benefits of early treatment.³⁶

Lower-risk drugs: Fondaparinux and bivalirudin

At this time, only two agents have been studied in clinical trials that have specifically focused on reducing bleeding risk: fondaparinux (Arixtra) and bivalirudin (Angiomax).^20,37–39

Fondaparinux

OASIS-5 was a randomized, double-blind trial that compared fondaparinux and enoxaparin (Lovenox) in patients with acute coronary syndromes.³⁸ Fondaparinux was similar to enoxaparin in terms of the combined end point of death, myocardial infarction, or refractory ischemia at 9 days, and fewer patients on fondaparinux developed bleeding (2.2% vs 4.1%, HR 0.52; 95% CI 0.44–0.61). This difference persisted during long-term follow-up.

Importantly, fewer patients died in the fondaparinux group. At 180 days, 638 (6.5%) of the patients in the enoxaparin group had died, compared with 574 (5.8%) in the fondaparinux group, a difference of 64 deaths (P = .05). The authors found that 41 fewer patients in the fondaparinux group than in the enoxaparin group died after major bleeding, and 20 fewer patients in the fondaparinux group died after minor bleeding.³⁸ Thus, most of the difference in mortality rates between the two groups was attributed to a lower incidence of bleeding with fondaparinux.

Unfortunately, despite its safe bleeding profile, fondaparinux has fallen out of favor for use in acute coronary syndromes, owing to a higher risk of catheter thrombosis in the fondaparinux group (0.9%) than in those undergoing percutaneous coronary interventions with enoxaparin alone (0.4%) in the OASIS-5 trial.⁴⁰

The direct thrombin inhibitor bivalirudin has been studied in three large randomized trials in patients undergoing percutaneous coronary interventions.^20,37,41

The ACUITY trial³⁷ was a prospective, open-label, randomized, multicenter trial that compared three regimens in patients with moderate or high-risk non-ST-elevation acute coronary syndromes:

• Heparin plus a glycoprotein IIb/IIIa inhibitor
• Bivalirudin plus a glycoprotein IIb/IIIa inhibitor
• Bivalirudin alone.

Bivalirudin alone was as effective as heparin plus a glycoprotein IIb/IIIa inhibitor with respect to the composite ischemia end point, which at 30 days had occurred in 7.8% vs 7.3% of the patients in these treatment groups (P = .32, RR 1.08; 95% CI 0.93–1.24), and it was superior with respect to major bleeding (3.0% vs 5.7%, P < .001, RR 0.53; 95% CI 0.43–0.65).

The HORIZONS-AMI study⁴¹ was a prospective, open-label, randomized, multicenter trial that compared bivalirudin alone vs heparin plus a glycoprotein IIb/IIIa inhibitor in patients with ST-elevation acute coronary syndromes who were undergoing primary percutaneous coronary interventions. The two primary end points were major bleeding and net adverse events.

At 1 year, patients assigned to bivalirudin had a lower rate of major bleeding than did controls (5.8% vs 9.2%, HR 0.61, 95% CI 0.48–0.78, P < .0001), with similar rates of major adverse cardiac events in both groups (11.9% vs 11.9%, HR 1.00, 95% CI 0.82– 1.21, P = .98).⁴¹

Both OASIS 5 and HORIZONS-AMI are examples of clinical trials in which strategies that reduced bleeding risk were also associated with improved survival.

For cardiac catheterization, inserting the catheter in the wrist poses less risk

Bleeding is currently the most common noncardiac complication in patients undergoing percutaneous coronary interventions, and it most often occurs at the vascular access site.¹⁷

Rao et al¹² evaluated data from 593,094 procedures in the National Cardiovascular Data Registry and found that, compared with the femoral approach, the use of transradial percutaneous coronary intervention was associated with a similar rate of procedural success (OR 1.02, 95% CI 0.93–1.12) but a significantly lower risk of bleeding complications (OR 0.42, 95% CI 0.31–0.56) after multivariable adjustment.

The use of smaller sheath sizes (4F–6F) and preferential use of bivalirudin over unfractionated heparin and glycoprotein IIb/IIIa inhibitor therapy are other methods described to decrease the risk of bleeding after percutaneous coronary interventions.^20,41–49

IF BLEEDING OCCURS

Once a bleeding complication occurs, cessation of therapy is a potential option. Stopping or reversing antithrombotic and antiplatelet therapy is warranted in the event of major bleeding (eg, gastrointestinal, retroperitoneal, intracranial).³¹

Stopping antithrombotic and antiplatelet therapy

Whether bleeding is minor or major, the risk of a recurrent thrombotic event must be considered, especially in patients who have undergone revascularization, stent implantation, or both. The risk of acute thrombotic events after interrupting antithrombotic or antiplatelet agents is considered greatest 4 to 5 days following revascularization or percutaneous coronary intervention.¹⁵ If bleeding can be controlled with local treatment such as pressure, packing, or dressing, antithrombotic and antiplatelet therapy need not be interrupted.⁵⁰

Current guidelines recommend strict control of hemorrhage for at least 24 hours before reintroducing antiplatelet or antithrombotic agents.

It is also important to remember that in the setting of gastrointestinal bleeding due to peptic ulcer disease, adjunctive proton pump inhibitors are recommended after restarting antiplatelet or antithrombotic therapy or both.

Importantly, evidence-based antithrombotic medications (especially dual antiplatelet therapy) should be restarted once the acute bleeding event has resolved.³¹

Reversal of anticoagulant and antiplatelet therapies

Unfractionated heparin is reversed with infusion of protamine sulfate at a dose of 1 mg per 100 U of unfractionated heparin given over the previous 4 hours.^51,52 The rate of protamine sulfate infusion should be less than 100 mg over 2 hours, with 50% of the dose given initially and subsequent doses titrated according to bleeding response.^52,53 Protamine sulfate is associated with a risk of hypotension and bradycardia, and for this reason it should be given no faster than 5 mg/min.

Low-molecular-weight heparin (LMWH) can be inhibited by 1 mg of protamine sulfate for each 1 mg of LMWH given over the previous 4 hours.^51,52

However, protamine sulfate only partially neutralizes the anticoagulant effect of LMWH. In cases in which protamine sulfate is unsuccessful in abating bleeding associated with LMWH use, guidelines allow for the use of recombinant factor VIIa (NovoSeven).³¹ In healthy volunteers given fondaparinux, recombinant factor VIIa normalized coagulation times and thrombin generation within 1.5 hours, with a sustained effect for 6 hours.⁵²

It is important to note that the use of this agent has not been fully studied, it is very costly (a single dose of 40 μg/kg costs from $3,000 to $4,000), and it is linked to reports of increased risk of thrombotic complications.^54,55

Antiplatelet agents are more complex to reverse. The antiplatelet actions of aspirin and clopidogrel wear off as new platelets are produced. Approximately 10% of a patient’s platelet count is produced daily; thus, the antiplatelet effects of aspirin and clopidogrel can persist for 5 to 10 days.^31,56

If these agents need to be reversed quickly to stop bleeding, according to expert consensus the aspirin effect can be reversed by transfusion of one unit of platelets. The antiplatelet effect of clopidogrel is more significant than that of aspirin; thus, two units of platelets are recommended.⁵⁶

Glycoprotein IIb/IIIa inhibitors. If a major bleeding event requires the reversal of glycoprotein IIb/IIIa inhibitor therapy, the treatment must take into consideration the pharmacodynamics of the target drug. Both eptifibatide (Integrilin) and tirofiban (Aggrastat) competitively inhibit glycoprotein IIb/IIIa receptors; thus, their effects depend on dosing, elimination, and time. Due to the stoichiometry of both drugs, transfusion of platelets is ineffective. Both eptifibatide and tirofiban are eliminated by the kidney; thus, normal renal function is key to the amount of time it takes for platelet function to return to baseline.⁵⁷ Evidence suggests that fibrinogen-rich plasma can be administered to restore platelet function.^31,58,59

Abciximab (ReoPro). Whereas reversal of eptifibatide and tirofiban focuses on overcoming competitive inhibition, neutralization of abciximab involves overcoming its high receptor affinity. At 24 hours after abciximab infusion is stopped, platelet aggregation may still be inhibited by up to 50%. Fortunately, owing to abciximab’s short plasma half-life and its dilution in serum, platelet transfusion is effective in reversing its antiplatelet effects.^31,57

Long considered beneficial to critically ill patients, blood transfusion to maintain hematocrit levels during acute coronary syndromes has come under intense scrutiny. Randomized trials have shown that transfusion should not be given aggressively to critically ill patients.⁶⁰ In acute coronary syndromes, there are only observational data.

Rao et al⁶¹ used detailed clinical data from 24,112 patients with acute coronary syndromes in the GUSTO IIb, PURSUIT, and PARAGON B trials to determine the association between blood transfusion and outcomes in patients who developed moderate to severe bleeding, anemia, or both during their hospitalization. The rates of death in the hospital and at 30 days were significantly higher in patients who received a transfusion (30-day mortality HR 3.94; 95% CI 3.36–4.75). However, there was no significant association between transfusion and the 30-day mortality rate if the nadir hematocrit was 25% or less.

Of note: no randomized clinical trial has evaluated transfusion strategies in acute coronary syndromes at this time. Until such data are available, it is reasonable to follow published guidelines and to avoid transfusion in stable patients with ischemic heart disease unless the hematocrit is 25% or less.³¹ The risks and benefits of blood transfusion should be carefully weighed. Routine use of transfusion to maintain predefined hemoglobin levels is not recommended in stable patients.

The medical management of non-ST-elevation acute coronary syndromes focuses on blocking the coagulation cascade and inhibiting platelets. This—plus diagnostic angiography followed, if needed, by revascularization—has reduced the rates of death and recurrent ischemic events.¹ However, the combination of potent antithrombotic drugs and invasive procedures also increases the risk of bleeding.

This review discusses the incidence and complications associated with bleeding during the treatment of acute coronary syndromes and summarizes recommendations for preventing and managing bleeding in this setting.

THE TRUE INCIDENCE OF BLEEDING IS HARD TO DETERMINE

The optimal way to detect and analyze bleeding events in clinical trials and registries is highly debated. The reported incidences of bleeding during antithrombotic and antiplatelet therapy for non-ST-elevation acute coronary syndromes depend on how bleeding was defined, how the acute coronary syndromes were treated, and on other factors such as how the study was designed.

How was bleeding defined?

Since these classification schemes are based on different types of data, they yield different numbers when applied to the same study population. For instance, Rao et al⁴ pooled the data from the PURSUIT and PARAGON B trials (15,454 patients in all) and found that the incidence of severe bleeding (by the GUSTO criteria) was 1.2%, while the rate of major bleeding (by the TIMI criteria) was 8.2%.

What was the treatment strategy?

Another reason that the true incidence of bleeding is hard to determine is that different studies used treatment strategies that differed in the type, timing, and dose of antithrombotic agents and whether invasive procedures were used early. For example, if unfractionated heparin is used aggressively in regimens that are not adjusted for weight and with a higher target for the activated clotting time, the risk of bleeding is higher than with conservative dosing.^5–7

Subherwal et al⁸ evaluated the effect of treatment strategy on the incidence of bleeding in patients with non-ST-elevation acute coronary syndromes who received two or more antithrombotic drugs in the CRUSADE Quality Improvement Initiative. The risk of bleeding was higher with an invasive approach (catheterization) than with a conservative approach (no catheterization), regardless of baseline bleeding risk.

What type of study was it?

Another source of variation is the design of the study. Registries differ from clinical trials in patient characteristics and in the way data are gathered (prospectively vs retrospectively).

In registries, data are often collected retrospectively, whereas in clinical trials the data are prospectively collected. For this reason, the definition of bleeding in registries is often based on events that are easily identified through chart review, such as transfusion. This may lead to a lower reported rate of bleeding, since other, less serious bleeding events such as access-site hematomas and epistaxis may not be documented in the medical record.

On the other hand, registries often include older and sicker patients, who may be more prone to bleeding and who are often excluded from clinical trials. This may lead to a higher rate of reported bleeding.⁹

Where the study was conducted makes a difference as well, owing to regional practice differences. For example, Moscucci et al¹⁰ reported that the incidence of major bleeding in 24,045 patients with non-ST-elevation acute coronary syndromes in the GRACE registry (in 14 countries worldwide) was 3.9%. In contrast, Yang et al¹¹ reported that the rate of bleeding in the CRUSADE registry (in the United States) was 10.3%.

This difference was partly influenced by different definitions of bleeding. The GRACE registry defined major bleeding as life-threatening events requiring transfusion of two or more units of packed red blood cells, or resulting in an absolute decrease in the hematocrit of 10% or more or death, or hemorrhagic subdural hematoma. In contrast, the CRUSADE data reflect bleeding requiring transfusion. However, practice patterns such as greater use of invasive procedures in the United States may also be responsible.

Rao and colleagues¹² examined international variation in blood transfusion rates among patients with acute coronary syndromes. Patients outside the United States were significantly less likely to receive transfusions, even after adjusting for patient and practice differences.

Taking these confounders into account, it is reasonable to estimate that the frequency of bleeding in patients with non-ST-elevation acute coronary syndromes ranges from less than 1% to 10%.¹³

BLEEDING IS ASSOCIATED WITH POOR OUTCOMES

Regardless of the definition or the data source, hemorrhagic complications are associated with a higher risk of death and nonfatal adverse events, both in the short term and in the long term.

Short-term outcomes

A higher risk of death. In the GRACE registry study by Moscucci et al¹⁰ discussed above, patients who had major bleeding were significantly more likely to die during their hospitalization than those who did not (odds ratio [OR] 1.64, 95% confidence interval [CI] 1.18–2.28).

Rao et al¹⁴ evaluated pooled data from the multicenter international GUSTO IIb, PURSUIT, and PARAGON A and B trials and found that the effects of bleeding in non-ST-elevation acute coronary syndromes extended beyond the hospital stay. The more severe the bleeding (by the GUSTO criteria), the greater the adjusted hazard ratio (HR) for death within 30 days:

• With mild bleeding—HR 1.6, 95% CI 1.3–1.9
• With moderate bleeding—HR 2.7, 95% CI 2.3–3.4
• With severe bleeding—HR 10.6, 95% CI 8.3–13.6.

The pattern was the same for death within 6 months:

• With mild bleeding—HR 1.4, 95% CI 1.2–1.6
• With moderate bleeding—HR 2.1, 95% CI 1.8–2.4
• With severe bleeding, HR 7.5, 95% CI 6.1–9.3.

These findings were confirmed by Eikelboom et al¹⁵ in 34,146 patients with acute coronary syndromes in the OASIS registry, the OASIS-2 trial, and the CURE randomized trial. In the first 30 days, five times as many patients died (12.8% vs 2.5%; P < .0009) among those who developed major bleeding compared with those who did not. These investigators defined major bleeding as bleeding that was life-threatening or significantly disabling or that required transfusion of two or more units of packed red blood cells.

A higher risk of nonfatal adverse events. Bleeding after antithrombotic therapy for non-ST-elevation acute coronary syndromes has also been associated with nonfatal adverse events such as stroke and stent thrombosis.

For example, in the study by Eikelboom et al,¹⁵ major bleeding was associated with a higher risk of recurrent ischemic events. Approximately 1 in 5 patients in the OASIS trials who developed major bleeding during the first 30 days died or had a myocardial infarction or stroke by 30 days, compared with 1 in 20 of those who did not develop major bleeding during the first 30 days. However, after events that occurred during the first 30 days were excluded, the association between major bleeding and both myocardial infarction and stroke was no longer evident between 30 days and 6 months.

Manoukian et al¹⁶ evaluated the impact of major bleeding in 13,819 patients with highrisk acute coronary syndromes undergoing treatment with an early invasive strategy in the ACUITY trial. At 30 days, patients with major bleeding had higher rates of the composite end point of death, myocardial infarction, or unplanned revascularization for ischemia (23.1% vs 6.8%, P < .0001) and of stent thrombosis (3.4% vs 0.6%, P < .0001).

Long-term outcomes

The association between bleeding and adverse outcomes persists in the long term as well, although the mechanisms underlying this association are not well studied.

Kinnaird et al¹⁷ examined the data from 10,974 unselected patients who underwent percutaneous coronary intervention. At 1 year, the following percentages of patients had died:

• After TIMI major bleeding—17.2%
• After TIMI minor bleeding—9.1%
• After no bleeding—5.5%.

However, after adjustment for potential confounders, only transfusion remained a significant predictor of 1-year mortality.

Mehran et al¹⁸ evaluated 1-year mortality data from the ACUITY trial. Compared with the rate in patients who had no major bleeding and no myocardial infarction, the hazard ratios for death were:

• After major bleeding—HR 3.5, 95% CI 2.7–4.4
• After myocardial infarction—HR 3.1, 95% CI 2.4–3.9.

Interestingly, the risk of death associated with myocardial infarction abated after 7 days, while the risk associated with bleeding persisted beyond 30 days and remained constant throughout the first year following the bleeding event.

Similarly, Ndrepepa and colleagues¹⁹ examined pooled data from four ISAR trials using the TIMI bleeding scale and found that myocardial infarction, target vessel revascularization, and major bleeding all had similar discriminatory ability at predicting 1-year mortality.

In patients undergoing elective or urgent percutaneous coronary intervention in the REPLACE-2 trial,²⁰ independent predictors of death by 1 year were²¹:

• Major hemorrhage (OR 2.66, 95% CI 1.44–4.92)
• Periprocedural myocardial infarction (OR 2.46, 95% CI 1.44–4.20).

THEORIES OF HOW BLEEDING MAY CAUSE ADVERSE OUTCOMES

Several mechanisms have been proposed to explain the association between bleeding during treatment for acute coronary syndromes and adverse clinical outcomes.^13,22

The immediate effects of bleeding are thought to be hypotension and a reflex hyperadrenergic state to compensate for the loss of intravascular volume.²³ This physiologic response is believed to contribute to myocardial ischemia by further decreasing myocardial oxygen supply in obstructive coronary disease.

Trying to minimize blood loss, physicians may withhold anticoagulation and antiplatelet therapy, which in turn may lead to further ischemia.²⁴ To compensate for blood loss, physicians may also resort to blood transfusion. However, depletion of 2,3-diphosphoglycerate and nitric oxide in stored donor red blood cells is postulated to reduce oxygen delivery by increasing hemoglobin’s affinity for oxygen, leading to induced microvascular obstruction and adverse inflammatory reactions.^15,25

Recent data have also begun to elucidate the long-term effects of bleeding during acute coronary syndrome management. Patients with anemia during the acute phase of infarction have greater neurohormonal activation.²⁶ These adaptive responses to anemia may lead to eccentric left ventricular remodeling that may lead to higher oxygen consumption, increased diastolic wall stress, interstitial fibrosis, and accelerated myocyte loss.^27–30

Nevertheless, we must point out that although strong associations between bleeding and adverse outcomes have been established, direct causality has not.

TO PREVENT BLEEDING, START BY ASSESSING RISK

The CRUSADE bleeding risk score

The CRUSADE bleeding score (calculator available at http://www.crusadebleedingscore.org/) was developed and validated in more than 89,000 community-treated patients with non-ST-elevation acute coronary syndromes.⁸ It is based on eight variables:

• Sex (higher risk in women)
• History of diabetes (higher risk)
• Prior vascular disease (higher risk)
• Heart rate (the higher the rate, the higher the risk)
• Systolic blood pressure (higher risk with pressures above or below the 121–180 mm Hg range)
• Signs of congestive heart failure (higher risk)
• Baseline hematocrit (the lower the hematocrit, the higher the risk)
• Creatinine clearance (by the Cockcroft-Gault formula; the lower the creatinine clearance, the higher the risk).

Patients who are found to have bleeding scores suggesting a moderate or higher risk of bleeding should be considered for medications associated with a favorable bleeding profile, and for radial access at the time of coronary angiography. Scores are graded as follows⁸:

• < 21: Very low risk
• 21–30: Low risk
• 31–40: Moderate risk
• 41–50: High risk
• > 50: Very high risk.

The CRUSADE bleeding score is unique in that, unlike earlier risk stratification tools, it was developed in a “real world” population, not in subgroups or in a clinical trial. It can be calculated at baseline to help guide the selection of treatment.⁸

Adjusting the heparin regimen in patients at risk of bleeding

Both the joint American College of Cardiology/American Heart Association¹ and the European Society of Cardiology guidelines³¹ for the treatment of non-ST-elevation acute coronary syndromes recommend taking steps to prevent bleeding, such as adjusting the dosage of unfractionated heparin, using safer drugs, reducing the duration of antithrombotic treatment, and using combinations of antithrombotic and antiplatelet agents according to proven indications.³¹

In the CRUSADE registry, 42% of patients with non-ST-elevation acute coronary syndromes received at least one initial dose of antithrombotic drug outside the recommended range, resulting in an estimated 15% excess of bleeding events.³² Thus, proper dosing is a target for prevention.

Appropriate antithrombotic dosing takes into account the patient’s age, weight, and renal function. However, heparin dosage in the current guidelines¹ is based on weight only: a loading dose of 60 U/kg (maximum 4,000 U) by intravenous bolus, then 12 U/kg/hour (maximum 1,000 U/hour) to maintain an activated partial thromboplastin time of 50 to 70 seconds.¹

Renal dysfunction is particularly worrisome in patients with non-ST-elevation acute coronary syndromes because it is associated with higher rates of major bleeding and death. In the OASIS-5 trial,³³ the overall risk of death was approximately five times higher in patients in the lowest quartile of renal function (glomerular filtration rate < 58 mL/min/1.73 m²) than in the highest quartile (glomerular filtration rate ≥ 86 mL/min/1.73 m²).

Renal function must be evaluated not only on admission but also throughout the hospital stay. Patients presenting with acute coronary syndromes often experience fluctuations in renal function that would call for adjustment of heparin dosing, either increasing the dose to maximize the drug’s efficacy if renal function is recovering or decreasing the dose to prevent bleeding if renal function is deteriorating.

Clopidogrel vs prasugrel

Certain medications should be avoided when the risk of bleeding outweighs any potential benefit in terms of ischemia.

For example, in a randomized trial,³⁴ prasugrel (Effient), a potent thienopyridine, was associated with a significantly lower rate of the composite end point of stroke, myocardial infarction, or death than clopidogrel (Plavix) in patients with acute coronary syndromes undergoing percutaneous coronary interventions. However, it did not seem to offer any advantage in patients 75 years old and older, those with prior stroke or transient ischemic attack, or those weighing less than 60 kg, and it posed a substantially higher risk of bleeding.

With clopidogrel, the risk of acute bleeding is primarily in patients who undergo coronary artery bypass grafting within 5 days of receiving a dose.^35,36 Therefore, clopidogrel should be stopped 5 to 7 days before bypass surgery.¹ Importantly, there is no increased risk of recurrent ischemic events during this 5-day waiting period in patients who receive clopidogrel early. Therefore, the recommendation to stop clopidogrel before surgery does not negate the benefits of early treatment.³⁶

Lower-risk drugs: Fondaparinux and bivalirudin

At this time, only two agents have been studied in clinical trials that have specifically focused on reducing bleeding risk: fondaparinux (Arixtra) and bivalirudin (Angiomax).^20,37–39

Fondaparinux

OASIS-5 was a randomized, double-blind trial that compared fondaparinux and enoxaparin (Lovenox) in patients with acute coronary syndromes.³⁸ Fondaparinux was similar to enoxaparin in terms of the combined end point of death, myocardial infarction, or refractory ischemia at 9 days, and fewer patients on fondaparinux developed bleeding (2.2% vs 4.1%, HR 0.52; 95% CI 0.44–0.61). This difference persisted during long-term follow-up.

Importantly, fewer patients died in the fondaparinux group. At 180 days, 638 (6.5%) of the patients in the enoxaparin group had died, compared with 574 (5.8%) in the fondaparinux group, a difference of 64 deaths (P = .05). The authors found that 41 fewer patients in the fondaparinux group than in the enoxaparin group died after major bleeding, and 20 fewer patients in the fondaparinux group died after minor bleeding.³⁸ Thus, most of the difference in mortality rates between the two groups was attributed to a lower incidence of bleeding with fondaparinux.

Unfortunately, despite its safe bleeding profile, fondaparinux has fallen out of favor for use in acute coronary syndromes, owing to a higher risk of catheter thrombosis in the fondaparinux group (0.9%) than in those undergoing percutaneous coronary interventions with enoxaparin alone (0.4%) in the OASIS-5 trial.⁴⁰

The direct thrombin inhibitor bivalirudin has been studied in three large randomized trials in patients undergoing percutaneous coronary interventions.^20,37,41

The ACUITY trial³⁷ was a prospective, open-label, randomized, multicenter trial that compared three regimens in patients with moderate or high-risk non-ST-elevation acute coronary syndromes:

• Heparin plus a glycoprotein IIb/IIIa inhibitor
• Bivalirudin plus a glycoprotein IIb/IIIa inhibitor
• Bivalirudin alone.

Bivalirudin alone was as effective as heparin plus a glycoprotein IIb/IIIa inhibitor with respect to the composite ischemia end point, which at 30 days had occurred in 7.8% vs 7.3% of the patients in these treatment groups (P = .32, RR 1.08; 95% CI 0.93–1.24), and it was superior with respect to major bleeding (3.0% vs 5.7%, P < .001, RR 0.53; 95% CI 0.43–0.65).

The HORIZONS-AMI study⁴¹ was a prospective, open-label, randomized, multicenter trial that compared bivalirudin alone vs heparin plus a glycoprotein IIb/IIIa inhibitor in patients with ST-elevation acute coronary syndromes who were undergoing primary percutaneous coronary interventions. The two primary end points were major bleeding and net adverse events.

At 1 year, patients assigned to bivalirudin had a lower rate of major bleeding than did controls (5.8% vs 9.2%, HR 0.61, 95% CI 0.48–0.78, P < .0001), with similar rates of major adverse cardiac events in both groups (11.9% vs 11.9%, HR 1.00, 95% CI 0.82– 1.21, P = .98).⁴¹

Both OASIS 5 and HORIZONS-AMI are examples of clinical trials in which strategies that reduced bleeding risk were also associated with improved survival.

For cardiac catheterization, inserting the catheter in the wrist poses less risk

Bleeding is currently the most common noncardiac complication in patients undergoing percutaneous coronary interventions, and it most often occurs at the vascular access site.¹⁷

Rao et al¹² evaluated data from 593,094 procedures in the National Cardiovascular Data Registry and found that, compared with the femoral approach, the use of transradial percutaneous coronary intervention was associated with a similar rate of procedural success (OR 1.02, 95% CI 0.93–1.12) but a significantly lower risk of bleeding complications (OR 0.42, 95% CI 0.31–0.56) after multivariable adjustment.

The use of smaller sheath sizes (4F–6F) and preferential use of bivalirudin over unfractionated heparin and glycoprotein IIb/IIIa inhibitor therapy are other methods described to decrease the risk of bleeding after percutaneous coronary interventions.^20,41–49

IF BLEEDING OCCURS

Once a bleeding complication occurs, cessation of therapy is a potential option. Stopping or reversing antithrombotic and antiplatelet therapy is warranted in the event of major bleeding (eg, gastrointestinal, retroperitoneal, intracranial).³¹

Stopping antithrombotic and antiplatelet therapy

Whether bleeding is minor or major, the risk of a recurrent thrombotic event must be considered, especially in patients who have undergone revascularization, stent implantation, or both. The risk of acute thrombotic events after interrupting antithrombotic or antiplatelet agents is considered greatest 4 to 5 days following revascularization or percutaneous coronary intervention.¹⁵ If bleeding can be controlled with local treatment such as pressure, packing, or dressing, antithrombotic and antiplatelet therapy need not be interrupted.⁵⁰

Current guidelines recommend strict control of hemorrhage for at least 24 hours before reintroducing antiplatelet or antithrombotic agents.

It is also important to remember that in the setting of gastrointestinal bleeding due to peptic ulcer disease, adjunctive proton pump inhibitors are recommended after restarting antiplatelet or antithrombotic therapy or both.

Importantly, evidence-based antithrombotic medications (especially dual antiplatelet therapy) should be restarted once the acute bleeding event has resolved.³¹

Reversal of anticoagulant and antiplatelet therapies

Unfractionated heparin is reversed with infusion of protamine sulfate at a dose of 1 mg per 100 U of unfractionated heparin given over the previous 4 hours.^51,52 The rate of protamine sulfate infusion should be less than 100 mg over 2 hours, with 50% of the dose given initially and subsequent doses titrated according to bleeding response.^52,53 Protamine sulfate is associated with a risk of hypotension and bradycardia, and for this reason it should be given no faster than 5 mg/min.

Low-molecular-weight heparin (LMWH) can be inhibited by 1 mg of protamine sulfate for each 1 mg of LMWH given over the previous 4 hours.^51,52

However, protamine sulfate only partially neutralizes the anticoagulant effect of LMWH. In cases in which protamine sulfate is unsuccessful in abating bleeding associated with LMWH use, guidelines allow for the use of recombinant factor VIIa (NovoSeven).³¹ In healthy volunteers given fondaparinux, recombinant factor VIIa normalized coagulation times and thrombin generation within 1.5 hours, with a sustained effect for 6 hours.⁵²

It is important to note that the use of this agent has not been fully studied, it is very costly (a single dose of 40 μg/kg costs from $3,000 to $4,000), and it is linked to reports of increased risk of thrombotic complications.^54,55

Antiplatelet agents are more complex to reverse. The antiplatelet actions of aspirin and clopidogrel wear off as new platelets are produced. Approximately 10% of a patient’s platelet count is produced daily; thus, the antiplatelet effects of aspirin and clopidogrel can persist for 5 to 10 days.^31,56

If these agents need to be reversed quickly to stop bleeding, according to expert consensus the aspirin effect can be reversed by transfusion of one unit of platelets. The antiplatelet effect of clopidogrel is more significant than that of aspirin; thus, two units of platelets are recommended.⁵⁶

Glycoprotein IIb/IIIa inhibitors. If a major bleeding event requires the reversal of glycoprotein IIb/IIIa inhibitor therapy, the treatment must take into consideration the pharmacodynamics of the target drug. Both eptifibatide (Integrilin) and tirofiban (Aggrastat) competitively inhibit glycoprotein IIb/IIIa receptors; thus, their effects depend on dosing, elimination, and time. Due to the stoichiometry of both drugs, transfusion of platelets is ineffective. Both eptifibatide and tirofiban are eliminated by the kidney; thus, normal renal function is key to the amount of time it takes for platelet function to return to baseline.⁵⁷ Evidence suggests that fibrinogen-rich plasma can be administered to restore platelet function.^31,58,59

Abciximab (ReoPro). Whereas reversal of eptifibatide and tirofiban focuses on overcoming competitive inhibition, neutralization of abciximab involves overcoming its high receptor affinity. At 24 hours after abciximab infusion is stopped, platelet aggregation may still be inhibited by up to 50%. Fortunately, owing to abciximab’s short plasma half-life and its dilution in serum, platelet transfusion is effective in reversing its antiplatelet effects.^31,57

Long considered beneficial to critically ill patients, blood transfusion to maintain hematocrit levels during acute coronary syndromes has come under intense scrutiny. Randomized trials have shown that transfusion should not be given aggressively to critically ill patients.⁶⁰ In acute coronary syndromes, there are only observational data.

Rao et al⁶¹ used detailed clinical data from 24,112 patients with acute coronary syndromes in the GUSTO IIb, PURSUIT, and PARAGON B trials to determine the association between blood transfusion and outcomes in patients who developed moderate to severe bleeding, anemia, or both during their hospitalization. The rates of death in the hospital and at 30 days were significantly higher in patients who received a transfusion (30-day mortality HR 3.94; 95% CI 3.36–4.75). However, there was no significant association between transfusion and the 30-day mortality rate if the nadir hematocrit was 25% or less.

Of note: no randomized clinical trial has evaluated transfusion strategies in acute coronary syndromes at this time. Until such data are available, it is reasonable to follow published guidelines and to avoid transfusion in stable patients with ischemic heart disease unless the hematocrit is 25% or less.³¹ The risks and benefits of blood transfusion should be carefully weighed. Routine use of transfusion to maintain predefined hemoglobin levels is not recommended in stable patients.

Yes. A number of large randomized controlled trials have shown inhaled corticosteroids to be beneficial in low doses for patients who have mild persistent asthma, and therefore these drugs are strongly recommended in this situation.¹

Asthma care providers should, however, consider this “yes” in the context of asthma severity, the goals of therapy, and the benefits and risks associated with inhaled corticosteroids.

CLASSIFICATION OF ASTHMA SEVERITY

Although the studies of asthma prevalence had methodologic limitations and therefore the true prevalence of mild persistent asthma cannot be determined, it is common. Fuhlbrigge et al² reported that most asthma patients have some form of persistent asthma. In contrast, Dusser et al³ reviewed available studies and concluded that most patients with asthma have either intermittent or mild persistent asthma.

GOALS: REDUCE IMPAIRMENT AND RISK

The goals of asthma management are to:

Reduce impairment by controlling symptoms so that normal activity levels can be maintained, by minimizing the need for short-acting bronchodilator use, and by maintaining normal pulmonary function; and to

Reduce risk by preventing progressive loss of lung function and recurrent exacerbations, and by optimizing pharmacotherapy while minimizing potential adverse effects.¹

EVIDENCE OF BENEFIT

The OPTIMA trial⁴ (Low Dose Inhaled Budesonide and Formoterol in Mild Persistent Asthma) was a double-blind, randomized trial carried out in 198 centers in 17 countries. Compared with those randomized to receive placebo, patients who were randomized to receive an inhaled corticosteroid, ie, budesonide (Pulmicort) 100 μg twice daily, had 60% fewer severe exacerbations (relative risk [RR] 0.4, 95% confidence interval [CI] 0.27–0.59) and 48% fewer days when their asthma was poorly controlled (RR 0.52, 95% CI 0.4–0.67). Adding a long-acting beta-agonist did not change this outcome.

The START study⁵ (Inhaled Steroid Treatment as Regular Therapy in Early Asthma) showed that, compared with placebo, starting inhaled budesonide within the first 2 years of asthma symptoms in patients with mild persistent asthma was associated with better asthma control and less need for additional asthma medication.

The IMPACT study⁶ (Improving Asthma Control Trial) showed that inhaled steroids need to be taken daily, on a regular schedule, rather than intermittently as needed. Patients received either inhaled budesonide as needed, budesonide 200 μg twice daily every day, or zafirlukast (Accolate) 20 mg twice daily. Daily budesonide therapy resulted in better asthma control, less bronchial hyperresponsiveness, and less airway inflammation compared with intermittent use, zafirlukast therapy, or placebo. Daily zafirlukast and intermittent steroid treatment produced similar results for all outcomes measured.

Despite this strong evidence supporting regular use of inhaled corticosteroids in patients with mild persistent asthma, many patients choose to take them intermittently.

Suissa et al⁷ found, in a large observational cohort study, that fewer patients died of asthma if they were receiving low-dose inhaled corticosteroids than if they were not. The rate of death due to asthma was lower in patients who had used more inhaled corticosteroids over the previous year, and the death rate was higher in those who had discontinued inhaled corticosteroids in the previous 3 months than in those who continued using them.

STEROIDS DO NOT SLOW THE LOSS OF LUNG FUNCTION

Compared with people without asthma, asthma patients have substantially lower values of forced expiratory volume in the first second of expiration (FEV₁). They also have a faster rate of functional decline: the average decrease in FEV₁ in asthma patients is 38 mL per year, compared with 22 mL per year in nonasthmatic people.⁹

Although inhaled corticosteroids have been shown to increase lung function in asthma patients in the short term, there is little convincing evidence to suggest that they affect the rate of decline in the long term.¹⁰ In fact, airway inflammation and bronchial hyperresponsiveness return to baseline within 2 weeks after inhaled corticosteroids are discontinued.¹⁰

DO INHALED CORTICOSTEROIDS STUNT CHILDREN’S GROWTH?

The safety of long-term low-dose inhaled corticosteroids is well established in adults. However, two large randomized controlled trials found that children treated with low-dose inhaled steroids (budesonide 200–400 μg per day) grew 1 to 1.5 cm less over 3 to 5 years of treatment than children receiving placebo.¹¹ However, this effect was primarily evident within the first year of therapy, and growth velocity was similar to that with placebo at the end of the treatment period (4 to 6 years).¹²

Agertoft and Pedersen¹³ found that taking inhaled corticosteroids long-term is unlikely to have an effect on final height. Children who took inhaled budesonide (up to an average daily dose of 500 μg) into adulthood ended up no shorter than those who did not.

Based on these and other data, inhaled corticosteroids are generally considered safe at recommended doses. However, the decision to prescribe them for long-term therapy should be based on the risks and benefits to the individual patient.¹

ALTERNATIVE DRUGS FOR MILD PERSISTENT ASTHMA

Leukotriene-modifying drugs include the leukotriene receptor antagonists montelukast (Singulair) and zafirlukast and the 5-lipoxygenase inhibitor zileuton (Zyflo CR). These drugs have been associated with statistically significant improvement in FEV₁ compared with placebo in patients with mild to moderate asthma, reductions in both blood and sputum eosinophils,¹⁴ and attenuation of bronchoconstriction with exercise.¹¹

Large randomized trials comparing leukotriene modifier therapy with low-dose inhaled steroids in adults and children with mild persistent asthma have found that although outcomes improve with either therapy, the improvement is statistically superior with inhaled steroids for most asthma-control measures. ^6,8 Low-dose inhaled steroid therapy in patients with mild persistent and moderate persistent asthma has been associated with superior clinical outcomes as well as greater improvement in pulmonary function than treatment with antileukotriene drugs (Table 2).⁸

Asthma is heterogeneous, and properly selected patients with mild persistent asthma may achieve good control with leukotrienemodifier monotherapy.¹⁵ Alternatives for patients with mild persistent asthma include the methylxanthine theophylline, but this drug is less desirable due to its narrow therapeutic index. ¹ The inhaled cromones nedocromil (Tilade) and cromolyn (Intal) were other options in this patient population, but their short half-lives made them less practical, and US production has been discontinued.

THE BOTTOM LINE

Though leukotriene receptor antagonists can be effective, the daily use of inhaled corticosteroids results in higher asthma control test scores, more symptom-free days, greater pre-bronchodilator FEV₁, and decreased percentage of sputum eosinophils⁶ in patients with mild persistent asthma, and the addition of a long-acting beta agonist does not provide additional benefit.⁴ Furthermore, daily use of inhaled corticosteroids in these patients has also been associated with a lower rate of asthma-related deaths and with less need for systemic corticosteroid therapy,^7,8 even though inhaled corticosteroids have not yet been shown to alter the progressive loss of lung function.¹⁰

Yes. A number of large randomized controlled trials have shown inhaled corticosteroids to be beneficial in low doses for patients who have mild persistent asthma, and therefore these drugs are strongly recommended in this situation.¹

Asthma care providers should, however, consider this “yes” in the context of asthma severity, the goals of therapy, and the benefits and risks associated with inhaled corticosteroids.

CLASSIFICATION OF ASTHMA SEVERITY

Although the studies of asthma prevalence had methodologic limitations and therefore the true prevalence of mild persistent asthma cannot be determined, it is common. Fuhlbrigge et al² reported that most asthma patients have some form of persistent asthma. In contrast, Dusser et al³ reviewed available studies and concluded that most patients with asthma have either intermittent or mild persistent asthma.

GOALS: REDUCE IMPAIRMENT AND RISK

The goals of asthma management are to:

Reduce impairment by controlling symptoms so that normal activity levels can be maintained, by minimizing the need for short-acting bronchodilator use, and by maintaining normal pulmonary function; and to

Reduce risk by preventing progressive loss of lung function and recurrent exacerbations, and by optimizing pharmacotherapy while minimizing potential adverse effects.¹

EVIDENCE OF BENEFIT

The OPTIMA trial⁴ (Low Dose Inhaled Budesonide and Formoterol in Mild Persistent Asthma) was a double-blind, randomized trial carried out in 198 centers in 17 countries. Compared with those randomized to receive placebo, patients who were randomized to receive an inhaled corticosteroid, ie, budesonide (Pulmicort) 100 μg twice daily, had 60% fewer severe exacerbations (relative risk [RR] 0.4, 95% confidence interval [CI] 0.27–0.59) and 48% fewer days when their asthma was poorly controlled (RR 0.52, 95% CI 0.4–0.67). Adding a long-acting beta-agonist did not change this outcome.

The START study⁵ (Inhaled Steroid Treatment as Regular Therapy in Early Asthma) showed that, compared with placebo, starting inhaled budesonide within the first 2 years of asthma symptoms in patients with mild persistent asthma was associated with better asthma control and less need for additional asthma medication.

The IMPACT study⁶ (Improving Asthma Control Trial) showed that inhaled steroids need to be taken daily, on a regular schedule, rather than intermittently as needed. Patients received either inhaled budesonide as needed, budesonide 200 μg twice daily every day, or zafirlukast (Accolate) 20 mg twice daily. Daily budesonide therapy resulted in better asthma control, less bronchial hyperresponsiveness, and less airway inflammation compared with intermittent use, zafirlukast therapy, or placebo. Daily zafirlukast and intermittent steroid treatment produced similar results for all outcomes measured.

Despite this strong evidence supporting regular use of inhaled corticosteroids in patients with mild persistent asthma, many patients choose to take them intermittently.

Suissa et al⁷ found, in a large observational cohort study, that fewer patients died of asthma if they were receiving low-dose inhaled corticosteroids than if they were not. The rate of death due to asthma was lower in patients who had used more inhaled corticosteroids over the previous year, and the death rate was higher in those who had discontinued inhaled corticosteroids in the previous 3 months than in those who continued using them.

STEROIDS DO NOT SLOW THE LOSS OF LUNG FUNCTION

Compared with people without asthma, asthma patients have substantially lower values of forced expiratory volume in the first second of expiration (FEV₁). They also have a faster rate of functional decline: the average decrease in FEV₁ in asthma patients is 38 mL per year, compared with 22 mL per year in nonasthmatic people.⁹

Although inhaled corticosteroids have been shown to increase lung function in asthma patients in the short term, there is little convincing evidence to suggest that they affect the rate of decline in the long term.¹⁰ In fact, airway inflammation and bronchial hyperresponsiveness return to baseline within 2 weeks after inhaled corticosteroids are discontinued.¹⁰

DO INHALED CORTICOSTEROIDS STUNT CHILDREN’S GROWTH?

The safety of long-term low-dose inhaled corticosteroids is well established in adults. However, two large randomized controlled trials found that children treated with low-dose inhaled steroids (budesonide 200–400 μg per day) grew 1 to 1.5 cm less over 3 to 5 years of treatment than children receiving placebo.¹¹ However, this effect was primarily evident within the first year of therapy, and growth velocity was similar to that with placebo at the end of the treatment period (4 to 6 years).¹²

Agertoft and Pedersen¹³ found that taking inhaled corticosteroids long-term is unlikely to have an effect on final height. Children who took inhaled budesonide (up to an average daily dose of 500 μg) into adulthood ended up no shorter than those who did not.

Based on these and other data, inhaled corticosteroids are generally considered safe at recommended doses. However, the decision to prescribe them for long-term therapy should be based on the risks and benefits to the individual patient.¹

ALTERNATIVE DRUGS FOR MILD PERSISTENT ASTHMA

Leukotriene-modifying drugs include the leukotriene receptor antagonists montelukast (Singulair) and zafirlukast and the 5-lipoxygenase inhibitor zileuton (Zyflo CR). These drugs have been associated with statistically significant improvement in FEV₁ compared with placebo in patients with mild to moderate asthma, reductions in both blood and sputum eosinophils,¹⁴ and attenuation of bronchoconstriction with exercise.¹¹

Large randomized trials comparing leukotriene modifier therapy with low-dose inhaled steroids in adults and children with mild persistent asthma have found that although outcomes improve with either therapy, the improvement is statistically superior with inhaled steroids for most asthma-control measures. ^6,8 Low-dose inhaled steroid therapy in patients with mild persistent and moderate persistent asthma has been associated with superior clinical outcomes as well as greater improvement in pulmonary function than treatment with antileukotriene drugs (Table 2).⁸

Asthma is heterogeneous, and properly selected patients with mild persistent asthma may achieve good control with leukotrienemodifier monotherapy.¹⁵ Alternatives for patients with mild persistent asthma include the methylxanthine theophylline, but this drug is less desirable due to its narrow therapeutic index. ¹ The inhaled cromones nedocromil (Tilade) and cromolyn (Intal) were other options in this patient population, but their short half-lives made them less practical, and US production has been discontinued.

THE BOTTOM LINE

Though leukotriene receptor antagonists can be effective, the daily use of inhaled corticosteroids results in higher asthma control test scores, more symptom-free days, greater pre-bronchodilator FEV₁, and decreased percentage of sputum eosinophils⁶ in patients with mild persistent asthma, and the addition of a long-acting beta agonist does not provide additional benefit.⁴ Furthermore, daily use of inhaled corticosteroids in these patients has also been associated with a lower rate of asthma-related deaths and with less need for systemic corticosteroid therapy,^7,8 even though inhaled corticosteroids have not yet been shown to alter the progressive loss of lung function.¹⁰

Yes. A number of large randomized controlled trials have shown inhaled corticosteroids to be beneficial in low doses for patients who have mild persistent asthma, and therefore these drugs are strongly recommended in this situation.¹

Asthma care providers should, however, consider this “yes” in the context of asthma severity, the goals of therapy, and the benefits and risks associated with inhaled corticosteroids.

CLASSIFICATION OF ASTHMA SEVERITY

Although the studies of asthma prevalence had methodologic limitations and therefore the true prevalence of mild persistent asthma cannot be determined, it is common. Fuhlbrigge et al² reported that most asthma patients have some form of persistent asthma. In contrast, Dusser et al³ reviewed available studies and concluded that most patients with asthma have either intermittent or mild persistent asthma.

GOALS: REDUCE IMPAIRMENT AND RISK

The goals of asthma management are to:

Reduce impairment by controlling symptoms so that normal activity levels can be maintained, by minimizing the need for short-acting bronchodilator use, and by maintaining normal pulmonary function; and to

Reduce risk by preventing progressive loss of lung function and recurrent exacerbations, and by optimizing pharmacotherapy while minimizing potential adverse effects.¹

EVIDENCE OF BENEFIT

The OPTIMA trial⁴ (Low Dose Inhaled Budesonide and Formoterol in Mild Persistent Asthma) was a double-blind, randomized trial carried out in 198 centers in 17 countries. Compared with those randomized to receive placebo, patients who were randomized to receive an inhaled corticosteroid, ie, budesonide (Pulmicort) 100 μg twice daily, had 60% fewer severe exacerbations (relative risk [RR] 0.4, 95% confidence interval [CI] 0.27–0.59) and 48% fewer days when their asthma was poorly controlled (RR 0.52, 95% CI 0.4–0.67). Adding a long-acting beta-agonist did not change this outcome.

The START study⁵ (Inhaled Steroid Treatment as Regular Therapy in Early Asthma) showed that, compared with placebo, starting inhaled budesonide within the first 2 years of asthma symptoms in patients with mild persistent asthma was associated with better asthma control and less need for additional asthma medication.

The IMPACT study⁶ (Improving Asthma Control Trial) showed that inhaled steroids need to be taken daily, on a regular schedule, rather than intermittently as needed. Patients received either inhaled budesonide as needed, budesonide 200 μg twice daily every day, or zafirlukast (Accolate) 20 mg twice daily. Daily budesonide therapy resulted in better asthma control, less bronchial hyperresponsiveness, and less airway inflammation compared with intermittent use, zafirlukast therapy, or placebo. Daily zafirlukast and intermittent steroid treatment produced similar results for all outcomes measured.

Despite this strong evidence supporting regular use of inhaled corticosteroids in patients with mild persistent asthma, many patients choose to take them intermittently.

Suissa et al⁷ found, in a large observational cohort study, that fewer patients died of asthma if they were receiving low-dose inhaled corticosteroids than if they were not. The rate of death due to asthma was lower in patients who had used more inhaled corticosteroids over the previous year, and the death rate was higher in those who had discontinued inhaled corticosteroids in the previous 3 months than in those who continued using them.

STEROIDS DO NOT SLOW THE LOSS OF LUNG FUNCTION

Compared with people without asthma, asthma patients have substantially lower values of forced expiratory volume in the first second of expiration (FEV₁). They also have a faster rate of functional decline: the average decrease in FEV₁ in asthma patients is 38 mL per year, compared with 22 mL per year in nonasthmatic people.⁹

Although inhaled corticosteroids have been shown to increase lung function in asthma patients in the short term, there is little convincing evidence to suggest that they affect the rate of decline in the long term.¹⁰ In fact, airway inflammation and bronchial hyperresponsiveness return to baseline within 2 weeks after inhaled corticosteroids are discontinued.¹⁰

DO INHALED CORTICOSTEROIDS STUNT CHILDREN’S GROWTH?

The safety of long-term low-dose inhaled corticosteroids is well established in adults. However, two large randomized controlled trials found that children treated with low-dose inhaled steroids (budesonide 200–400 μg per day) grew 1 to 1.5 cm less over 3 to 5 years of treatment than children receiving placebo.¹¹ However, this effect was primarily evident within the first year of therapy, and growth velocity was similar to that with placebo at the end of the treatment period (4 to 6 years).¹²

Agertoft and Pedersen¹³ found that taking inhaled corticosteroids long-term is unlikely to have an effect on final height. Children who took inhaled budesonide (up to an average daily dose of 500 μg) into adulthood ended up no shorter than those who did not.

Based on these and other data, inhaled corticosteroids are generally considered safe at recommended doses. However, the decision to prescribe them for long-term therapy should be based on the risks and benefits to the individual patient.¹

ALTERNATIVE DRUGS FOR MILD PERSISTENT ASTHMA

Leukotriene-modifying drugs include the leukotriene receptor antagonists montelukast (Singulair) and zafirlukast and the 5-lipoxygenase inhibitor zileuton (Zyflo CR). These drugs have been associated with statistically significant improvement in FEV₁ compared with placebo in patients with mild to moderate asthma, reductions in both blood and sputum eosinophils,¹⁴ and attenuation of bronchoconstriction with exercise.¹¹

Large randomized trials comparing leukotriene modifier therapy with low-dose inhaled steroids in adults and children with mild persistent asthma have found that although outcomes improve with either therapy, the improvement is statistically superior with inhaled steroids for most asthma-control measures. ^6,8 Low-dose inhaled steroid therapy in patients with mild persistent and moderate persistent asthma has been associated with superior clinical outcomes as well as greater improvement in pulmonary function than treatment with antileukotriene drugs (Table 2).⁸

Asthma is heterogeneous, and properly selected patients with mild persistent asthma may achieve good control with leukotrienemodifier monotherapy.¹⁵ Alternatives for patients with mild persistent asthma include the methylxanthine theophylline, but this drug is less desirable due to its narrow therapeutic index. ¹ The inhaled cromones nedocromil (Tilade) and cromolyn (Intal) were other options in this patient population, but their short half-lives made them less practical, and US production has been discontinued.

THE BOTTOM LINE

Though leukotriene receptor antagonists can be effective, the daily use of inhaled corticosteroids results in higher asthma control test scores, more symptom-free days, greater pre-bronchodilator FEV₁, and decreased percentage of sputum eosinophils⁶ in patients with mild persistent asthma, and the addition of a long-acting beta agonist does not provide additional benefit.⁴ Furthermore, daily use of inhaled corticosteroids in these patients has also been associated with a lower rate of asthma-related deaths and with less need for systemic corticosteroid therapy,^7,8 even though inhaled corticosteroids have not yet been shown to alter the progressive loss of lung function.¹⁰

In patients with “borderline” hypertension or those in whom the duration of blood pressure elevation is hard to ascertain, the finding of end-organ damage has traditionally been used as an argument to institute aggressive antihypertensive therapy. In this setting, retinal hypertensive disease, an S4 gallop, and left ventricular hypertrophy (LVH) are often specifically sought.

LVH and otherwise unexplained chronic kidney disease in patients with hypertension have generally been believed to be products of the elevated arterial pressure, and primary treatment has targeted pressure control. Bauml and Underwood, in this issue of the Journal, emphasize some published clinical trial data indicating that LVH may be an independent risk factor for poorer cardiovascular outcome. Even more provocative is the suggestion that LVH can be reversed, as can the associated increased risk of cardiovascular morbidity, independently of the hypertension.

Given our current understanding that LVH, under some conditions, can be induced by products of the renin-angiotensin system, this would suggest that pharmacologic blockade of this enzyme system should have extra benefit, above that seen from other antihypertensive agents. Conceivably, this may be true only in patients with LVH, and the time course of benefit may not directly parallel that seen with the control of hypertension. That theoretically may explain the lack of uniform advantage of angiotensin blockade over other effective antihypertensive approaches.

Since electrocardiography is a specific but not very sensitive test for LVH, the authors suggest that patients with hypertension be routinely screened for LVH using echocardiography. I am not sure the weight of the evidence supports this approach at present, particularly in the current frenzy of cost containment. Nonetheless, this concept warrants consideration, and at the least, large patient databases might be screened retrospectively to further validate or refute the concept that hypertension-associated LVH is an independent, reversible risk factor for cardiovascular morbidity.

In patients with “borderline” hypertension or those in whom the duration of blood pressure elevation is hard to ascertain, the finding of end-organ damage has traditionally been used as an argument to institute aggressive antihypertensive therapy. In this setting, retinal hypertensive disease, an S4 gallop, and left ventricular hypertrophy (LVH) are often specifically sought.

LVH and otherwise unexplained chronic kidney disease in patients with hypertension have generally been believed to be products of the elevated arterial pressure, and primary treatment has targeted pressure control. Bauml and Underwood, in this issue of the Journal, emphasize some published clinical trial data indicating that LVH may be an independent risk factor for poorer cardiovascular outcome. Even more provocative is the suggestion that LVH can be reversed, as can the associated increased risk of cardiovascular morbidity, independently of the hypertension.

Given our current understanding that LVH, under some conditions, can be induced by products of the renin-angiotensin system, this would suggest that pharmacologic blockade of this enzyme system should have extra benefit, above that seen from other antihypertensive agents. Conceivably, this may be true only in patients with LVH, and the time course of benefit may not directly parallel that seen with the control of hypertension. That theoretically may explain the lack of uniform advantage of angiotensin blockade over other effective antihypertensive approaches.

Since electrocardiography is a specific but not very sensitive test for LVH, the authors suggest that patients with hypertension be routinely screened for LVH using echocardiography. I am not sure the weight of the evidence supports this approach at present, particularly in the current frenzy of cost containment. Nonetheless, this concept warrants consideration, and at the least, large patient databases might be screened retrospectively to further validate or refute the concept that hypertension-associated LVH is an independent, reversible risk factor for cardiovascular morbidity.

In patients with “borderline” hypertension or those in whom the duration of blood pressure elevation is hard to ascertain, the finding of end-organ damage has traditionally been used as an argument to institute aggressive antihypertensive therapy. In this setting, retinal hypertensive disease, an S4 gallop, and left ventricular hypertrophy (LVH) are often specifically sought.

LVH and otherwise unexplained chronic kidney disease in patients with hypertension have generally been believed to be products of the elevated arterial pressure, and primary treatment has targeted pressure control. Bauml and Underwood, in this issue of the Journal, emphasize some published clinical trial data indicating that LVH may be an independent risk factor for poorer cardiovascular outcome. Even more provocative is the suggestion that LVH can be reversed, as can the associated increased risk of cardiovascular morbidity, independently of the hypertension.

Given our current understanding that LVH, under some conditions, can be induced by products of the renin-angiotensin system, this would suggest that pharmacologic blockade of this enzyme system should have extra benefit, above that seen from other antihypertensive agents. Conceivably, this may be true only in patients with LVH, and the time course of benefit may not directly parallel that seen with the control of hypertension. That theoretically may explain the lack of uniform advantage of angiotensin blockade over other effective antihypertensive approaches.

Since electrocardiography is a specific but not very sensitive test for LVH, the authors suggest that patients with hypertension be routinely screened for LVH using echocardiography. I am not sure the weight of the evidence supports this approach at present, particularly in the current frenzy of cost containment. Nonetheless, this concept warrants consideration, and at the least, large patient databases might be screened retrospectively to further validate or refute the concept that hypertension-associated LVH is an independent, reversible risk factor for cardiovascular morbidity.

Left ventricular hypertrophy (LVH) strongly predicts cardiovascular morbidity and overall mortality in hypertensive patients. ^1–7 Antihypertensive treatment that causes LVH to regress decreases the rates of adverse cardiovascular events and improves survival, independent of how much the blood pressure is lowered.^8–11 It is clinically important to recognize that LVH is a modifiable risk factor and that management is more complex than just blood pressure control.

This paper reviews the definition of LVH, compares the diagnostic tests for it, and discusses the current evidence-based approach to managing this dangerous risk factor.

A CHRONICALLY ELEVATED CARDIAC WORKLOAD CAUSES LVH

LVH is an abnormal increase in the mass of the left ventricular myocardium caused by a chronically increased workload on the heart.¹² This most commonly results from the heart pumping against an elevated afterload, as in hypertension and aortic stenosis. Another notable cause is increased filling of the left ventricle (ie, diastolic overload), which is the underlying mechanism for LVH in patients with aortic or mitral regurgitation and dilated cardiomyopathy. Coronary artery disease can also play a role in the pathogenesis of LVH, as the normal myocardium attempts to compensate for the ischemic or infarcted tissue.¹³

The development of myocardial fibrosis appears to be pathophysiologically linked to the renin-angiotensin-aldosterone system. Specifically, there is evidence that angiotensin II has a profibrotic effect on the myocardium of hypertensive patients.¹⁵ This may explain why angiotensin-converting enzyme (ACE) inhibitors and angiotensin II receptor blockers (ARBs) are among the most potent agents for treating LVH, as we will discuss later in this review.

DIAGNOSIS BY ELECTROCARDIOGRAPHY, ECHOCARDIOGRAPHY, OR MRI

Many different criteria for electrocardiographic LVH have been proposed over the years. Most use the voltage in one or more leads, with or without additional factors such as QRS duration, secondary ST-T wave abnormalities, or left atrial abnormalities. The most well known electrocardiographic criteria are the Cornell voltage,²¹ the Cornell product,²² the Sokolow-Lyon index,²³ and the Romhilt-Estes point score system (Table 1).²⁴

• Cornell voltage—median sensitivity 15%, median specificity 96%
• Cornell product—median sensitivity 19.5%, median specificity 91%
• Sokolow-Lyon voltage—median sensitivity 21%, median specificity 89%
• Romhilt-Estes point score—median sensitivity 17%, median specificity 95%.

Of note, the ranges of the published values were extremely broad. For example, the ranges in sensitivity were:

• Cornell voltage—2% to 41%
• Cornell product—8% to 32%
• Sokolow-Lyon voltage—4% to 51%
• Romhilt-Estes point score—0% to 41%.

While the studies with the extreme values may have had issues of small sample size or poor study quality, the wide range in values may primarily be the result of diverse study populations as well as different validation methods and cutoff values to define LVH. Regardless, the overall message of high specificity and low sensitivity is indisputable.

Electrocardiography is insensitive for diagnosing LVH because it relies on measuring the electrical activity of the heart by electrodes on the surface of the skin to predict the left ventricular mass. The intracardiac electrical activity is problematic to measure externally because the measurements are affected by everything between the myocardium and the electrodes, most notably fat, fluid, and air. Because of this effect, electrocardiography underdiagnoses LVH in patients with obesity, pleural effusions, pericardial effusions, anasarca, or chronic obstructive pulmonary disease. In addition, the diagnosis of LVH by electrocardiography is strongly influenced by age and ethnicity.^25–26

While electrocardiography is not sensitive and cannot be used to rule out LVH, it still has a role in its diagnosis and management. In the landmark Losartan Intervention for Endpoint Reduction in Hypertension (LIFE) study, regression of LVH (diagnosed electrocardiographically by the Sokolow-Lyon index or the Cornell product criteria) in response to losartan (Cozaar) improved cardiovascular outcomes independent of blood pressure.¹⁰ Based on this, it is reasonable that all hypertensive patients and other patients at risk of LVH who undergo electrocardiography be screened with these two criteria.

Echocardiography, if available, should be the test of choice to assess for LVH. It is much more sensitive than electrocardiography and can also detect other abnormalities such as left ventricular dysfunction and valvular disease.

This test uses transthoracic or transesophageal ultrasonography to measure the left ventricular end-diastolic diameter, posterior wall thickness, and interventricular septum thickness. From these measurements and the patient’s height and weight, the left ventricular mass index can be calculated.²⁷

Several different cutoff values for the left ventricular mass index have been proposed; the LIFE study used values of > 104 g/m² in women and > 116 g/m² in men to define LVH.

When using echocardiography to assess for LVH, it is imperative that the left ventricular mass index be used and not just the left ventricular wall thickness, as often happens in clinical practice. This is necessary because diagnosis by wall thickness alone is not a good indicator of LVH, with a concordance between wall thickness and a left ventricular mass index of only 60%.²⁸ In addition, wall thickness tends to underestimate LVH in women and overestimate it in men.

Is echocardiography cost-effective?

Despite its clear advantages, an important consideration about echocardiography as a screening test for all hypertensive patients is its cost.

A suggested way to reduce cost is to measure the left ventricular mass index only.²⁹ A limited echocardiographic examination is much less expensive than a complete two-dimensional echocardiogram ($255 vs $431 per the 2009 Medicare Ambulatory Payment Classification³⁰) and should be the examination performed if the patient has no other clinical indication for echocardiography.

Another way to control cost is to stratify patients by risk and to do echocardiography only in those who would benefit most from it. Based on the prevalence of LVH, one study concluded that echocardiography is most cost-effective in men 50 years or older.³¹

Further study is necessary to more precisely define the cost-effectiveness of echocardiographic screening for LVH in terms of potentially preventable cardiovascular morbidity and death.

Cardiac MRI: The costly gold standard

Cardiac MRI is the gold standard test for LVH, as it is even more accurate and reproducible than echocardiography.³² It can precisely estimate a patient's left ventricular mass and assess for other structural cardiac abnormalities.

MRI’s use, however, is severely restricted in clinical practice due to its high cost and limited availability. While it may never be used for general screening for LVH, it certainly has a role in clinical research and for assessing cardiac anatomy in special clinical situations.

TREATMENT SHOULD INCLUDE AN ACE INHIBITOR OR ARB

Once LVH has been diagnosed, the next step is to decide on an appropriate treatment plan.

While the choice of therapy will always depend on other comorbidities, a 2003 metaanalysis of antihypertensive medications in the treatment of LVH (controlling for the degree of blood pressure lowering) showed that ARBs were the most efficacious class of agents for reducing the left ventricular mass.³³ Specifically, ARBs decreased the mass by 13%, followed by calcium-channel blockers at 11%, ACE inhibitors at 10%, diuretics at 8%, and beta-blockers at 6%. In pairwise comparison, ARBs, calcium-channel blockers, and ACE inhibitors were all significantly more effective in reducing the left ventricular mass than beta-blockers.

As previously discussed, LVH appears to be pathophysiologically linked to myocardial fibrosis and the renin-angiotensin-aldosterone system. For this reason and based on the data presented above regarding the degree of LVH regression, ACE inhibitors or ARBs should be used as the first-line agents for LVH unless they are contraindicated in the individual patient.

The LIFE study

The LIFE study offers the strongest evidence that treating LVH is beneficial. It showed that in hypertensive patients with electrocardiographic LVH by the Cornell product or Sokolow-Lyon criteria, treatment with antihypertensive drugs that resulted in less-severe LVH on electrocardiography was associated with lower rates of cardiovascular morbidity and death, independent of the blood pressure achieved or the drug used.¹⁰

The end point in this study was a composite of stroke, myocardial infarction, and cardiovascular death. Regression of electrocardiographic LVH in hypertensive patients has also been shown to decrease the incidence of diabetes mellitus,³⁴ atrial fibrillation,³⁵ and hospitalizations for heart failure.³⁶

The LIFE study also examined the prognostic implications of treating LVH detected by echocardiography. In this prospective cohort substudy, patients who had a lower left ventricular mass index during treatment with antihypertensive drugs had lower rates of cardiovascular morbidity and all-cause mortality, independent of the effects of blood pressure and treatment used.¹¹

These results suggest that there may be a role not only for treating LVH, but also for monitoring for a reduction in the left ventricular mass index as a goal of therapy (similar to the way hemoglobin A_1c is used in diabetic patients). If the index is used in this way, one could potentially adjust the dose of current drugs, switch classes, or add an additional drug based on a persistently elevated left ventricular mass index in order to optimize the patient's overall cardiovascular risk. A randomized controlled trial of therapy directed by the mass index vs conventional therapy of LVH would be necessary to assess the clinical utility of this approach.

RECOMMENDATIONS

LVH is a common and potentially modifiable cardiovascular risk factor often overlooked in clinical practice. Ideally, all hypertensive patients should be screened with echocardiography to look for LVH, using the calculated left ventricular mass index rather than wall thickness alone to make the diagnosis. While electrocardiography is specific and also has prognostic implications, it is not sensitive enough to be used alone to screen for LVH.

Once the diagnosis of LVH is made, the initial therapy should be an ARB or an ACE inhibitor. Response to therapy can be assessed by monitoring for a reduction in left ventricular mass index or regression of electrocardiographic LVH.

Treatment-induced regression of LVH decreases adverse cardiovascular events and improves overall survival. When modifying medications in hypertensive patients, it is important to remember that the treatment of LVH is not synonymous with blood pressure control.

Left ventricular hypertrophy (LVH) strongly predicts cardiovascular morbidity and overall mortality in hypertensive patients. ^1–7 Antihypertensive treatment that causes LVH to regress decreases the rates of adverse cardiovascular events and improves survival, independent of how much the blood pressure is lowered.^8–11 It is clinically important to recognize that LVH is a modifiable risk factor and that management is more complex than just blood pressure control.

This paper reviews the definition of LVH, compares the diagnostic tests for it, and discusses the current evidence-based approach to managing this dangerous risk factor.

A CHRONICALLY ELEVATED CARDIAC WORKLOAD CAUSES LVH

LVH is an abnormal increase in the mass of the left ventricular myocardium caused by a chronically increased workload on the heart.¹² This most commonly results from the heart pumping against an elevated afterload, as in hypertension and aortic stenosis. Another notable cause is increased filling of the left ventricle (ie, diastolic overload), which is the underlying mechanism for LVH in patients with aortic or mitral regurgitation and dilated cardiomyopathy. Coronary artery disease can also play a role in the pathogenesis of LVH, as the normal myocardium attempts to compensate for the ischemic or infarcted tissue.¹³

The development of myocardial fibrosis appears to be pathophysiologically linked to the renin-angiotensin-aldosterone system. Specifically, there is evidence that angiotensin II has a profibrotic effect on the myocardium of hypertensive patients.¹⁵ This may explain why angiotensin-converting enzyme (ACE) inhibitors and angiotensin II receptor blockers (ARBs) are among the most potent agents for treating LVH, as we will discuss later in this review.

DIAGNOSIS BY ELECTROCARDIOGRAPHY, ECHOCARDIOGRAPHY, OR MRI

Many different criteria for electrocardiographic LVH have been proposed over the years. Most use the voltage in one or more leads, with or without additional factors such as QRS duration, secondary ST-T wave abnormalities, or left atrial abnormalities. The most well known electrocardiographic criteria are the Cornell voltage,²¹ the Cornell product,²² the Sokolow-Lyon index,²³ and the Romhilt-Estes point score system (Table 1).²⁴

• Cornell voltage—median sensitivity 15%, median specificity 96%
• Cornell product—median sensitivity 19.5%, median specificity 91%
• Sokolow-Lyon voltage—median sensitivity 21%, median specificity 89%
• Romhilt-Estes point score—median sensitivity 17%, median specificity 95%.

Of note, the ranges of the published values were extremely broad. For example, the ranges in sensitivity were:

• Cornell voltage—2% to 41%
• Cornell product—8% to 32%
• Sokolow-Lyon voltage—4% to 51%
• Romhilt-Estes point score—0% to 41%.

While the studies with the extreme values may have had issues of small sample size or poor study quality, the wide range in values may primarily be the result of diverse study populations as well as different validation methods and cutoff values to define LVH. Regardless, the overall message of high specificity and low sensitivity is indisputable.

Electrocardiography is insensitive for diagnosing LVH because it relies on measuring the electrical activity of the heart by electrodes on the surface of the skin to predict the left ventricular mass. The intracardiac electrical activity is problematic to measure externally because the measurements are affected by everything between the myocardium and the electrodes, most notably fat, fluid, and air. Because of this effect, electrocardiography underdiagnoses LVH in patients with obesity, pleural effusions, pericardial effusions, anasarca, or chronic obstructive pulmonary disease. In addition, the diagnosis of LVH by electrocardiography is strongly influenced by age and ethnicity.^25–26

While electrocardiography is not sensitive and cannot be used to rule out LVH, it still has a role in its diagnosis and management. In the landmark Losartan Intervention for Endpoint Reduction in Hypertension (LIFE) study, regression of LVH (diagnosed electrocardiographically by the Sokolow-Lyon index or the Cornell product criteria) in response to losartan (Cozaar) improved cardiovascular outcomes independent of blood pressure.¹⁰ Based on this, it is reasonable that all hypertensive patients and other patients at risk of LVH who undergo electrocardiography be screened with these two criteria.

Echocardiography, if available, should be the test of choice to assess for LVH. It is much more sensitive than electrocardiography and can also detect other abnormalities such as left ventricular dysfunction and valvular disease.

This test uses transthoracic or transesophageal ultrasonography to measure the left ventricular end-diastolic diameter, posterior wall thickness, and interventricular septum thickness. From these measurements and the patient’s height and weight, the left ventricular mass index can be calculated.²⁷

Several different cutoff values for the left ventricular mass index have been proposed; the LIFE study used values of > 104 g/m² in women and > 116 g/m² in men to define LVH.

When using echocardiography to assess for LVH, it is imperative that the left ventricular mass index be used and not just the left ventricular wall thickness, as often happens in clinical practice. This is necessary because diagnosis by wall thickness alone is not a good indicator of LVH, with a concordance between wall thickness and a left ventricular mass index of only 60%.²⁸ In addition, wall thickness tends to underestimate LVH in women and overestimate it in men.

Is echocardiography cost-effective?

Despite its clear advantages, an important consideration about echocardiography as a screening test for all hypertensive patients is its cost.

A suggested way to reduce cost is to measure the left ventricular mass index only.²⁹ A limited echocardiographic examination is much less expensive than a complete two-dimensional echocardiogram ($255 vs $431 per the 2009 Medicare Ambulatory Payment Classification³⁰) and should be the examination performed if the patient has no other clinical indication for echocardiography.

Another way to control cost is to stratify patients by risk and to do echocardiography only in those who would benefit most from it. Based on the prevalence of LVH, one study concluded that echocardiography is most cost-effective in men 50 years or older.³¹

Further study is necessary to more precisely define the cost-effectiveness of echocardiographic screening for LVH in terms of potentially preventable cardiovascular morbidity and death.

Cardiac MRI: The costly gold standard

Cardiac MRI is the gold standard test for LVH, as it is even more accurate and reproducible than echocardiography.³² It can precisely estimate a patient's left ventricular mass and assess for other structural cardiac abnormalities.

MRI’s use, however, is severely restricted in clinical practice due to its high cost and limited availability. While it may never be used for general screening for LVH, it certainly has a role in clinical research and for assessing cardiac anatomy in special clinical situations.

TREATMENT SHOULD INCLUDE AN ACE INHIBITOR OR ARB

Once LVH has been diagnosed, the next step is to decide on an appropriate treatment plan.

While the choice of therapy will always depend on other comorbidities, a 2003 metaanalysis of antihypertensive medications in the treatment of LVH (controlling for the degree of blood pressure lowering) showed that ARBs were the most efficacious class of agents for reducing the left ventricular mass.³³ Specifically, ARBs decreased the mass by 13%, followed by calcium-channel blockers at 11%, ACE inhibitors at 10%, diuretics at 8%, and beta-blockers at 6%. In pairwise comparison, ARBs, calcium-channel blockers, and ACE inhibitors were all significantly more effective in reducing the left ventricular mass than beta-blockers.

As previously discussed, LVH appears to be pathophysiologically linked to myocardial fibrosis and the renin-angiotensin-aldosterone system. For this reason and based on the data presented above regarding the degree of LVH regression, ACE inhibitors or ARBs should be used as the first-line agents for LVH unless they are contraindicated in the individual patient.

The LIFE study

The LIFE study offers the strongest evidence that treating LVH is beneficial. It showed that in hypertensive patients with electrocardiographic LVH by the Cornell product or Sokolow-Lyon criteria, treatment with antihypertensive drugs that resulted in less-severe LVH on electrocardiography was associated with lower rates of cardiovascular morbidity and death, independent of the blood pressure achieved or the drug used.¹⁰

The end point in this study was a composite of stroke, myocardial infarction, and cardiovascular death. Regression of electrocardiographic LVH in hypertensive patients has also been shown to decrease the incidence of diabetes mellitus,³⁴ atrial fibrillation,³⁵ and hospitalizations for heart failure.³⁶

The LIFE study also examined the prognostic implications of treating LVH detected by echocardiography. In this prospective cohort substudy, patients who had a lower left ventricular mass index during treatment with antihypertensive drugs had lower rates of cardiovascular morbidity and all-cause mortality, independent of the effects of blood pressure and treatment used.¹¹

These results suggest that there may be a role not only for treating LVH, but also for monitoring for a reduction in the left ventricular mass index as a goal of therapy (similar to the way hemoglobin A_1c is used in diabetic patients). If the index is used in this way, one could potentially adjust the dose of current drugs, switch classes, or add an additional drug based on a persistently elevated left ventricular mass index in order to optimize the patient's overall cardiovascular risk. A randomized controlled trial of therapy directed by the mass index vs conventional therapy of LVH would be necessary to assess the clinical utility of this approach.

RECOMMENDATIONS

LVH is a common and potentially modifiable cardiovascular risk factor often overlooked in clinical practice. Ideally, all hypertensive patients should be screened with echocardiography to look for LVH, using the calculated left ventricular mass index rather than wall thickness alone to make the diagnosis. While electrocardiography is specific and also has prognostic implications, it is not sensitive enough to be used alone to screen for LVH.

Once the diagnosis of LVH is made, the initial therapy should be an ARB or an ACE inhibitor. Response to therapy can be assessed by monitoring for a reduction in left ventricular mass index or regression of electrocardiographic LVH.

Treatment-induced regression of LVH decreases adverse cardiovascular events and improves overall survival. When modifying medications in hypertensive patients, it is important to remember that the treatment of LVH is not synonymous with blood pressure control.

Left ventricular hypertrophy (LVH) strongly predicts cardiovascular morbidity and overall mortality in hypertensive patients. ^1–7 Antihypertensive treatment that causes LVH to regress decreases the rates of adverse cardiovascular events and improves survival, independent of how much the blood pressure is lowered.^8–11 It is clinically important to recognize that LVH is a modifiable risk factor and that management is more complex than just blood pressure control.

This paper reviews the definition of LVH, compares the diagnostic tests for it, and discusses the current evidence-based approach to managing this dangerous risk factor.

A CHRONICALLY ELEVATED CARDIAC WORKLOAD CAUSES LVH

LVH is an abnormal increase in the mass of the left ventricular myocardium caused by a chronically increased workload on the heart.¹² This most commonly results from the heart pumping against an elevated afterload, as in hypertension and aortic stenosis. Another notable cause is increased filling of the left ventricle (ie, diastolic overload), which is the underlying mechanism for LVH in patients with aortic or mitral regurgitation and dilated cardiomyopathy. Coronary artery disease can also play a role in the pathogenesis of LVH, as the normal myocardium attempts to compensate for the ischemic or infarcted tissue.¹³

The development of myocardial fibrosis appears to be pathophysiologically linked to the renin-angiotensin-aldosterone system. Specifically, there is evidence that angiotensin II has a profibrotic effect on the myocardium of hypertensive patients.¹⁵ This may explain why angiotensin-converting enzyme (ACE) inhibitors and angiotensin II receptor blockers (ARBs) are among the most potent agents for treating LVH, as we will discuss later in this review.

DIAGNOSIS BY ELECTROCARDIOGRAPHY, ECHOCARDIOGRAPHY, OR MRI

Many different criteria for electrocardiographic LVH have been proposed over the years. Most use the voltage in one or more leads, with or without additional factors such as QRS duration, secondary ST-T wave abnormalities, or left atrial abnormalities. The most well known electrocardiographic criteria are the Cornell voltage,²¹ the Cornell product,²² the Sokolow-Lyon index,²³ and the Romhilt-Estes point score system (Table 1).²⁴

• Cornell voltage—median sensitivity 15%, median specificity 96%
• Cornell product—median sensitivity 19.5%, median specificity 91%
• Sokolow-Lyon voltage—median sensitivity 21%, median specificity 89%
• Romhilt-Estes point score—median sensitivity 17%, median specificity 95%.

Of note, the ranges of the published values were extremely broad. For example, the ranges in sensitivity were:

• Cornell voltage—2% to 41%
• Cornell product—8% to 32%
• Sokolow-Lyon voltage—4% to 51%
• Romhilt-Estes point score—0% to 41%.

While the studies with the extreme values may have had issues of small sample size or poor study quality, the wide range in values may primarily be the result of diverse study populations as well as different validation methods and cutoff values to define LVH. Regardless, the overall message of high specificity and low sensitivity is indisputable.

Electrocardiography is insensitive for diagnosing LVH because it relies on measuring the electrical activity of the heart by electrodes on the surface of the skin to predict the left ventricular mass. The intracardiac electrical activity is problematic to measure externally because the measurements are affected by everything between the myocardium and the electrodes, most notably fat, fluid, and air. Because of this effect, electrocardiography underdiagnoses LVH in patients with obesity, pleural effusions, pericardial effusions, anasarca, or chronic obstructive pulmonary disease. In addition, the diagnosis of LVH by electrocardiography is strongly influenced by age and ethnicity.^25–26

While electrocardiography is not sensitive and cannot be used to rule out LVH, it still has a role in its diagnosis and management. In the landmark Losartan Intervention for Endpoint Reduction in Hypertension (LIFE) study, regression of LVH (diagnosed electrocardiographically by the Sokolow-Lyon index or the Cornell product criteria) in response to losartan (Cozaar) improved cardiovascular outcomes independent of blood pressure.¹⁰ Based on this, it is reasonable that all hypertensive patients and other patients at risk of LVH who undergo electrocardiography be screened with these two criteria.

Echocardiography, if available, should be the test of choice to assess for LVH. It is much more sensitive than electrocardiography and can also detect other abnormalities such as left ventricular dysfunction and valvular disease.

This test uses transthoracic or transesophageal ultrasonography to measure the left ventricular end-diastolic diameter, posterior wall thickness, and interventricular septum thickness. From these measurements and the patient’s height and weight, the left ventricular mass index can be calculated.²⁷

Several different cutoff values for the left ventricular mass index have been proposed; the LIFE study used values of > 104 g/m² in women and > 116 g/m² in men to define LVH.

When using echocardiography to assess for LVH, it is imperative that the left ventricular mass index be used and not just the left ventricular wall thickness, as often happens in clinical practice. This is necessary because diagnosis by wall thickness alone is not a good indicator of LVH, with a concordance between wall thickness and a left ventricular mass index of only 60%.²⁸ In addition, wall thickness tends to underestimate LVH in women and overestimate it in men.

Is echocardiography cost-effective?

Despite its clear advantages, an important consideration about echocardiography as a screening test for all hypertensive patients is its cost.

A suggested way to reduce cost is to measure the left ventricular mass index only.²⁹ A limited echocardiographic examination is much less expensive than a complete two-dimensional echocardiogram ($255 vs $431 per the 2009 Medicare Ambulatory Payment Classification³⁰) and should be the examination performed if the patient has no other clinical indication for echocardiography.

Another way to control cost is to stratify patients by risk and to do echocardiography only in those who would benefit most from it. Based on the prevalence of LVH, one study concluded that echocardiography is most cost-effective in men 50 years or older.³¹

Further study is necessary to more precisely define the cost-effectiveness of echocardiographic screening for LVH in terms of potentially preventable cardiovascular morbidity and death.

Cardiac MRI: The costly gold standard

Cardiac MRI is the gold standard test for LVH, as it is even more accurate and reproducible than echocardiography.³² It can precisely estimate a patient's left ventricular mass and assess for other structural cardiac abnormalities.

MRI’s use, however, is severely restricted in clinical practice due to its high cost and limited availability. While it may never be used for general screening for LVH, it certainly has a role in clinical research and for assessing cardiac anatomy in special clinical situations.

TREATMENT SHOULD INCLUDE AN ACE INHIBITOR OR ARB

Once LVH has been diagnosed, the next step is to decide on an appropriate treatment plan.

While the choice of therapy will always depend on other comorbidities, a 2003 metaanalysis of antihypertensive medications in the treatment of LVH (controlling for the degree of blood pressure lowering) showed that ARBs were the most efficacious class of agents for reducing the left ventricular mass.³³ Specifically, ARBs decreased the mass by 13%, followed by calcium-channel blockers at 11%, ACE inhibitors at 10%, diuretics at 8%, and beta-blockers at 6%. In pairwise comparison, ARBs, calcium-channel blockers, and ACE inhibitors were all significantly more effective in reducing the left ventricular mass than beta-blockers.

As previously discussed, LVH appears to be pathophysiologically linked to myocardial fibrosis and the renin-angiotensin-aldosterone system. For this reason and based on the data presented above regarding the degree of LVH regression, ACE inhibitors or ARBs should be used as the first-line agents for LVH unless they are contraindicated in the individual patient.

The LIFE study

The LIFE study offers the strongest evidence that treating LVH is beneficial. It showed that in hypertensive patients with electrocardiographic LVH by the Cornell product or Sokolow-Lyon criteria, treatment with antihypertensive drugs that resulted in less-severe LVH on electrocardiography was associated with lower rates of cardiovascular morbidity and death, independent of the blood pressure achieved or the drug used.¹⁰

The end point in this study was a composite of stroke, myocardial infarction, and cardiovascular death. Regression of electrocardiographic LVH in hypertensive patients has also been shown to decrease the incidence of diabetes mellitus,³⁴ atrial fibrillation,³⁵ and hospitalizations for heart failure.³⁶

The LIFE study also examined the prognostic implications of treating LVH detected by echocardiography. In this prospective cohort substudy, patients who had a lower left ventricular mass index during treatment with antihypertensive drugs had lower rates of cardiovascular morbidity and all-cause mortality, independent of the effects of blood pressure and treatment used.¹¹

These results suggest that there may be a role not only for treating LVH, but also for monitoring for a reduction in the left ventricular mass index as a goal of therapy (similar to the way hemoglobin A_1c is used in diabetic patients). If the index is used in this way, one could potentially adjust the dose of current drugs, switch classes, or add an additional drug based on a persistently elevated left ventricular mass index in order to optimize the patient's overall cardiovascular risk. A randomized controlled trial of therapy directed by the mass index vs conventional therapy of LVH would be necessary to assess the clinical utility of this approach.

RECOMMENDATIONS

LVH is a common and potentially modifiable cardiovascular risk factor often overlooked in clinical practice. Ideally, all hypertensive patients should be screened with echocardiography to look for LVH, using the calculated left ventricular mass index rather than wall thickness alone to make the diagnosis. While electrocardiography is specific and also has prognostic implications, it is not sensitive enough to be used alone to screen for LVH.

Once the diagnosis of LVH is made, the initial therapy should be an ARB or an ACE inhibitor. Response to therapy can be assessed by monitoring for a reduction in left ventricular mass index or regression of electrocardiographic LVH.

Treatment-induced regression of LVH decreases adverse cardiovascular events and improves overall survival. When modifying medications in hypertensive patients, it is important to remember that the treatment of LVH is not synonymous with blood pressure control.