CURRENT DRUG THERAPY W.H. WILSON TANG, MD* CME Section of Heart Failure and Cardiac CREDIT Transplantation, Department of Cardiovascular Medicine, Cleveland Clinic Do thiazolidinediones cause heart failure? A critical review ■ ABSTRACT HE POSSIBILITY that the glucose-lowering T thiazolidinedione drugs (TZDs, also Concern has been raised about whether the fluid retention called “glitazones”) may cause or exacerbate caused by thiazolidinediones (TZDs, ie, rosiglitazone and congestive heart failure has led some physi- pioglitazone) can cause or exacerbate congestive heart cians to avoid using these drugs. Although failure. Although fluid retention is a worrisome side effect many patients do retain fluid while taking of TZDs, current evidence does not link fluid retention TZDs, the mechanism does not seem to be caused by TZDs with worsening heart function. TZDs have cardiogenic, the pattern is mostly that of slow- many benefits for patients with diabetes and can even be ly developing peripheral edema rather than used cautiously in patients with mild heart failure, with “flash” pulmonary edema, and the fluid reten- careful monitoring of volume status. tion often resolves if the drug is stopped. Furthermore, TZDs may provide benefits ■ KEY POINTS beyond glycemic control, including reducing the rate of heart attack, stroke, and death. In randomized trials, the incidence of fluid retention and This article reviews the evidence on peripheral edema ranged from about 5% when current TZDs with regard to their association with TZDs were used as monotherapy to about 15% when fluid retention and heart failure, their possible they were added to insulin therapy. benefits in patients with heart failure, and how to monitor for their side effects and man- age them. The incidence of congestive heart failure reported in clinical trials is less than 1% and appears to be related to ■ DIABETES AND HEART FAILURE underlying dysfunction, with decompensation caused by ARE RELATED sodium retention and fluid accumulation rather than a direct cardiac suppressive effect. Diabetes mellitus and heart failure are increas- ing in prevalence, and many patients have both. Fluid retention can often be reversed by stopping the TZD. More than 40% of patients hospitalized for Other strategies have been tried, including lowering its decompensated heart failure have a known his- dose or adding a loop or thiazide diuretic, spironolactone, tory of diabetes mellitus, and many more have or angiotensin-converting enzyme inhibitor. unrecognized abnormal glucose metabolism. The two conditions are likely linked. The Prospective Pioglitazone Clinical Trial in Diabetes can lead to heart failure either by promoting atherosclerosis and coronary artery Macrovascular Events (PROactive; Lancet 2005; disease with resulting ischemic heart failure, 366:1279–1289) found that the use of a TZD was associated with a 16% reduction in the combined end point of heart attack, stroke, and death, despite increased *Dr. Tang has indicated that he previously served as a consultant for GlaxoSmithKline Pharmaceuticals, Amylin Pharmaceuticals, and rates of edema and hospitalization for heart failure. F. Hoffmann-La Roche Ltd. He is a member of the speakers’ bureau for Takeda Pharmaceuticals. 390 CLEVELAND CLINIC JOURNAL OF MEDICINE VOLUME 73 • NUMBER 4 APRIL 2006 Downloaded from www.ccjm.org on October 2, 2021. For personal use only. All other uses require permission. or via direct effects on the cardiac muscles (so- TABLE 1 called diabetic cardiomyopathy). Bell1 described heart failure as “the frequent, forgot- Risks and benefits of thiazolidinediones ten, and often fatal complication of diabetes,” in patients with heart disease and called for its early detection and timely Benefits treatment in patients with diabetes. Increased: Heart failure may be more likely to devel- Coronary vasodilation op in diabetic patients with poor glycemic Glucose uptake control.2 The United Kingdom Prospective 3 Decreased: Diabetes Study, involving more than 3,600 Angiotensin II levels patients with diabetes, found that the inci- Blood pressure dence of heart failure declined by 16% for Endothelin-1 levels each 1% reduction in hemoglobin A1c. Ischemic-reperfusion injury Many patients with diabetes and heart Left ventricular end-diastolic pressure failure need to take multiple drugs for each Left ventricular hypertrophy condition, and the possible side effects and Myocardial infarct size interactions can be challenging to manage. Peripheral vascular resistance Compounding the problem, care is compart- Tumor necrosis factor alpha levels mentalized: cardiologists deal with the heart Improved: failure, endocrinologists deal with the dia- Endothelial function betes, and one specialist may be unaware of Lipid profile the impact—good or bad—of the treatment Risks prescribed by the other specialist. Increased: Moreover, little attention has been paid to Adiposity and weight the side effects of antidiabetes drugs in Atrial natriuretic peptide and B-type natriuretic peptide levels patients with chronic heart failure. Insulin has Circulating plasma volume, total body water long been associated with sympathetic overac- Fluid retention, peripheral edema tivation and sodium retention. Some sulfonyl- Left ventricular end-diastolic volume urea drugs are thought to abolish ischemic Decreased: preconditioning, leaving the myocardium Hematocrit (transiently) more susceptible to injury. Metformin is con- traindicated in patients with heart failure because it is thought to cause lactic acidosis if the patient has renal dysfunction, which is utilization.4 common in patients with heart failure. In placebo-controlled trials, TZDs in max- imal doses reduced hemoglobin A1c levels as ■ BENEFITS OF TZDs effectively as sulfonylureas and metformin. Furthermore, TZDs may provide clinical The two TZDs available in the United States— benefits beyond glycemic control (TABLE 1): rosiglitazone (Avandia) and pioglitazone they modestly reduce blood pressure, and (Actos)—are increasingly used as first-line and pioglitazone improves the overall lipid profile, second-line agents for treating type 2 diabetes possibly reducing the risk of atherosclerosis. mellitus, making up a substantial proportion of In patients with heart failure, other theo- the oral antidiabetic drug market share. retical advantages of TZDs are that they TZDs directly improve insulin sensitivity, reduce afterload, improve neurohormonal pancreatic beta-cell function, and endothelial function (reducing levels of angiotensin II and function. They work in the cell nuclei by bind- endothelin), and reduce levels of tumor necro- ing and activating peroxisome proliferator- sis factor alpha.5 They may improve myocar- activated receptor gamma (PPAR gamma), dial insulin resistance by raising myocardial which presumably regulates the transcription glucose uptake and reducing reliance on fatty of insulin-responsive genes involved in the acid oxidation in the failing heart. In animal control of glucose production, transport, and models, TZDs promote regression of left ven- CLEVELAND CLINIC JOURNAL OF MEDICINE VOLUME 73 • NUMBER 4 APRIL 2006 391 Downloaded from www.ccjm.org on October 2, 2021. For personal use only. All other uses require permission. THIAZOLIDINEDIONES TANG TABLE 2 fluid retention, which has led to concern that they may induce or worsen heart failure. Case Frequency of edema and congestive reports of marked edema and pulmonary heart failure in double-blind studies edema began to appear as TZDs gained wider in patients with diabetes use. (Troglitazone, the first TZD to become available, was withdrawn because of cases of TREATMENT N EDEMA WITHDRAWAL HEART (%) DUE TO EDEMA FAILURE hepatic dysfunction, another issue entirely.) (NO.) (%) In the clinical trials that led to the approval of rosiglitazone and pioglitazone,7,8 the inci- Rosiglitazone dence of fluid retention was about 5% when Alone 2,526 4.8 1 0.2 these drugs were used as monotherapy, 4% to + metformin 338 4.4 1 0.3 + sulfonylurea 405 4.0 0 0.5 7% when they were used in combination with + insulin 408 14.7 1 2.5 metformin or a sulfonylurea, and 15% when Placebo 601 1.3 1 0.2 used in combination with insulin (TABLE 2). Metformin alone 225 2.2 0 0 However, Niemeyer and Janney9 studied Sulfonylurea alone 626 1.0 0 0.5 166 patients started on TZDs at a Veterans Insulin alone 203 5.4 0 1.0 Administration medical center and found Pioglitazone that 30 (18.1%) developed fluid retention, of Alone 606 4.8 0 0 whom 16 had to discontinue TZD therapy. + metformin 168 6.0 0 0.6 The two available TZDs seem to be + sulfonylurea 373 7.2 1 0 approximately equal in causing fluid reten- + insulin 379 15.3 1 1.1 tion. Goldberg et al10 compared rosiglitazone Placebo 259 1.2 0 0 and pioglitazone as monotherapy in a study of Metformin alone 160 2.5 0 0 802 patients with type 2 diabetes without Sulfonylurea alone 187 2.1 0 1.1 heart failure. At the end of 6 months, both Insulin alone 187 7.0 0 0 groups had gained weight (2.7 kg ± 0.2 with DATA FROM AVANDIA [PACKAGE INSERT]. RESEARCH TRIANGLE PARK, NC: rosiglitazone and 3.0 kg ± 0.2 with pioglita- GLAXOSMITHKLINE; 2000, AND ACTOS [PACKAGE INSERT]. LINCOLNSHIRE, IL: TAKEDA PHARMACEUTICAL AMERICA, INC.; 2002. zone) and had increased categorical edema (0.29 ± .03 with rosiglitazone and 0.34 ± .04 with pioglitazone). Overall, 13% of patients reported worsening edema. The differences in tricular hypertrophy and reduce left ventricu- weight gain and edema were not significant lar remodeling following acute myocardial between the two groups. One episode of con- infarction. gestive heart failure was reported. The Prospective Pioglitazone Clinical Fluid retention also occurred in phase II Trial in Macrovascular Events (PROactive)6 clinical trials of the two dual PPAR- included more than 5,000 patients with type 2 alpha/gamma agonists, muraglitazar (9.2% vs diabetes and evidence of macrovascular dis- 7.2% with pioglitazone11) and tesaglitazar ease (meaning they were at high risk of car- (4.2%–6.8% vs 4.2% with pioglitazone12). diovascular events). They were randomized to receive either oral pioglitazone (titrated from ■ RISK OF HEART FAILURE 15 to 45 mg) or placebo.