Fibrates: What Have We Learned in the Past 40 Years?

James M. Backes, Pharm.D.; Cheryl A. Gibson, Ph.D.; Janelle F. Ruisinger, Pharm.D.; Patrick M. Moriarty, M.D.


Pharmacotherapy. 2007;27(3):412-424. 

In This Article

Possible Reasons for the Mixed Results

The question that remained after these studies was why did the fibric acid derivatives produce such lackluster results in many of the trials. Although fibrates as a class do not provide marked LDL reductions, they do substantially increase HDL levels and profoundly reduce plasma triglyceride concentrations.[57] In addition, the fibrates generate significant improvements in many of the emerging risk factors, including CRP level, fibrinogen level, and small dense LDL particles.[8,14,15,58] The overall metabolic effects of the fibrates would suggest that they are ideal agents for individuals with type 2 diabetes mellitus or metabolic syndrome. We may never know the complete answer to the question; however, the following key points address the primary reasons why the fibrate trials have provided such varied results.

Study Populations

A major reason for mixed results may lie in the diverse study populations of the fibrate trials. Although data from large fibrate trials are not entirely consistent, two major subpopulations that appear to receive the greatest clinical benefit from fibrates are those with mixed dyslipidemia (low HDL levels and elevated plasma triglyceride concentrations) and/or impaired glucose homeostasis (i.e., type 2 diabetes mellitus and prediabetes or metabolic syndrome).[8,53,59,60,61,62,63] When reviewing baseline lipid levels of patients participating in major trials, it is evident that a fibrate was not always the best choice of drug therapy ( Table 4 ).[2,8,50–53] The primary lipid abnormalities for most of the study subjects were elevated LDL and total cholesterol levels. At the time of the early trials, however, pharmacologic therapy for lowering LDL levels was limited primarily to resins and niacin. Further, most major fibrate trials had very limited numbers of patients with diabetes mellitus (i.e., the CDP, WHO, HHS, BIP studies). Enrolling more patients with mixed dyslipidemia and glucose impairment would have better matched the favorable effects of the fibrates and possibly provided more positive findings. Finally, fibrates have demonstrated reductions in some of the emerging cardiovascular risk factors (i.e., CRP level, fibrinogen concentration, small dense LDL particles)[4,8,10,12] that are strongly associated with mixed dyslipidemia and metabolic syndrome.[64] This may further explain why fibrates have generally provided greater efficacy at reducing cardiovascular events among these populations in post hoc analyses.

After assessment of baseline lipid levels, subjects in the VA-HIT study were the best suited for fibrate therapy of all major studies evaluating fibric acid derivatives. The primary lipid abnormality among this secondary prevention group was a low HDL concentration, whereas LDL concentrations were nearly at goal (≤ 100 mg/dl according to the National Cholesterol Education Program Adult Treatment Panel [NCEPATP]-II[65]). Compared with those in VA-HIT, most other study populations in fibrate trials had a markedly higher LDL concentration along with moderately higher HDL concentrations, which are not as conducive for fibrate therapy. Clinical outcomes of fibrate trials also reflect this view. Although significant reductions in cardiovascular events were noted with treatment in other fibrate studies (i.e., the HHS, WHO, FIELD studies), overall mortality remained unchanged or increased compared with the control group. In VA-HIT, a significant 22% reduction in the composite end point of CHD death or nonfatal myocardial infarction was observed with those receiving gemfibrozil, along with a strong trend toward a reduction in overall mortality, although the study was not powered to measure this outcome. Moreover, significant reductions were also seen for stroke, transient ischemic attack, and carotid endarterectomy in the treatment group.

Post hoc analyses were performed on the data from some of the trials to further address whether fibrates are more effective at reducing clinical events among individuals in special populations or with mixed dyslipidemia. The frequency of CHD among persons with non-insulin-dependent diabetes mellitus (NIDDM) in HHS was investigated.[60] Although the overall number of patients with NIDDM was small (135 patients), the rate of a myocardial infarction or CHD death was less with gemfibrozil (3.4%) than with placebo (10.5%, p=0.19). In another subgroup analysis of HHS, investigators found the LDL:HDL ratio and plasma triglyceride concentrations to be effective predictors of CHD events.[61] Patients with an LDL:HDL ratio greater than 5 and plasma triglyceride concentrations greater than 200 mg/dl had a relative risk of 3.8 (95% confidence interval 2.2-6.6) compared with those with an LDL:HDL ratio of 5 or less and plasma triglyceride concentrations of 200 mg/dl or less. When treated with gemfibrozil, the higher-risk subgroup (154 patients) experienced a 71% reduction (p<0.005) in CHD events compared with the corresponding subgroup receiving placebo (1262 patients).

Similar findings in event reductions were seen among patients with both a low baseline HDL concentration (< 42 mg/dl) and elevated plasma triglyceride concentrations (> 200 mg/dl). An 18-year mortality follow-up was recently published of participants in the HHS.[62] Investigators found those in the original gemfibrozil group with a body mass index and triglyceride concentration in the highest tertile had significantly lower rates of CHD mortality (71% lower relative risk) and all-cause mortality (33% lower relative risk) compared with those in the original placebo group. The authors concluded that the long-term mortality data suggest individuals originally randomly assigned to receive gemfibrozil benefited from treatment especially in the presence of characteristics related to the metabolic syndrome.

Specific populations were targeted for additional analysis in VA-HIT. One group of authors evaluated data among persons with diabetes mellitus and those without diabetes who had hyperinsulinemia.[59] The composite end point consisted of CHD death, stroke, and nonfatal myocardial infarction. This outcome was significantly reduced (24%, p<0.001) with gemfibrozil therapy when evaluating the entire study population in the initial VA-HIT analysis. Among patients with diabetes mellitus, the composite end point was reduced 32%, whereas those without diabetes experienced an 18% reduction (p=0.26 between groups). For the composite end point, the majority of benefit derived from treatment was due to marked reductions in CHD death (41%, p=0.02) and stroke (40%, p=0.046). Gemfibrozil also showed a substantial reduction in clinical events compared with placebo (35%, p=0.04) for patients without diabetes mellitus but with the highest fasting plasma insulin level. The VA-HIT subgroup analysis suggests that fibrate therapy may be more efficacious at reducing cardio-vascular events among persons with diabetes and those without diabetes who have elevated plasma insulin concentrations.

A subgroup analysis for subjects with elevated plasma triglyceride concentrations was also performed on the BIP study and, like previous post hoc fibrate data, demonstrated similar results.[8] Among 234 patients with plasma triglyceride concentrations 200 mg/dl or greater, event rate (fatal and nonfatal myocardial infarction, sudden death) was reduced by 39.5% (p=0.02) compared with that in 225 patients in the placebo group. Significant reductions in clinical events (42%, p=0.02) were also noted for those with plasma triglyceride concentrations of 200 mg/dl or greater and HDL concentration of less than 35 mg/dl (184 patients). A very recent subgroup analysis reaffirmed these findings.[63] The authors further evaluated the BIP data by investigating the effects of bezafibrate among 1470 patients with metabolic syndrome and, as in the previous subanalysis, found that those receiving active treatment experienced a 29% reduction in myocardial infarction (p=0.02) and 26% fewer CHD deaths (p=0.056). The investigators also noted that more marked reductions in cardiac mortality were seen in patients with a higher number of metabolic syndrome features.

Although post hoc data from fibrate trials suggest much greater reductions in clinical events among those with metabolic syndrome or type 2 diabetes, results of the FIELD study did not fully support these findings. All subjects enrolled in the FIELD study had a previous diagnosis of type 2 diabetes, and although fenofibrate produced moderate reductions in cardiovascular events, these lower rates were modest compared with post hoc results of previous fibrate studies evaluating persons with metabolic syndrome or type 2 diabetes. A number of reasons may explain why this population did not experience greater benefit with fenofibrate. First, a significantly greater number of subjects in the placebo arm began other lipid-altering agents (primarily statins) during the trial compared with those receiving fenofibrate (17% vs 8%, p<0.0001), potentially masking the effects of fibrate therapy. Second, although all subjects had type 2 diabetes, baseline lipid levels ( Table 4 ) were not entirely conducive to fibrate therapy. As noted from post hoc data from earlier fibrate trials, individuals appear to have a more favorable clinical response when baseline plasma triglyceride concentrations exceed 200 mg/dl and HDL concentration is less than 40 mg/dl. Subgroup analysis of the FIELD study indicated patients with low HDL levels (< 40 mg/dl for men and < 50 mg/dl for women) or plasma triglyceride concentrations of 150 mg/dl or greater experienced a lower rate of cardiovascular events; however, subjects meeting the criteria for metabolic syndrome experienced more clinical events compared with those without metabolic syndrome. Further analyses of these subgroups are required to fully explain these conflicting findings.

Targeting Low-Density Lipoprotein Cholesterol

Fibrates provide substantial improvements in HDL and plasma triglyceride concentrations but have mixed effects on LDL concentrations. Despite growing evidence indicating the importance of modifying HDL and plasma triglyceride concentrations,[66,67,68] LDL remains the primary target of therapy.[6] The NCEP ATP-III emphasizes the importance of achieving the LDL goal before addressing other lipoprotein abnormalities.[6] The results of the Collaborative Atorvastatin Diabetes Study (CARDS) illustrate this point well.[69] Men and women with type 2 diabetes and one additional CHD risk factor, but free of cardiovascular disease, were randomly assigned to receive atorvastatin 10 mg/day (1428 patients) or placebo (1410 patients). Characteristics of this study population were similar to those of participants in the FIELD study, although one major difference was 22% of FIELD subjects had type 2 diabetes mellitus and CVD. Baseline lipid levels between the two studies were nearly comparable; however, those in CARDS had substantially higher HDL concentrations (54 vs 43 mg/dl). After an approximate trial duration of 4 years, CARDS was stopped prematurely because prespecified rules for efficacy were met. Statin therapy resulted in significant reductions of major cardiovascular events (37%, p=0.001), lower rates of acute CHD events (36%), coronary revascularizations (31%), and fatal stroke (48%), and a nearly significant reduction in overall mortality (27%, p=0.059). Although numerous confounding factors prevent a direct comparison between studies, the results of the FIELD study and CARDS suggest statin therapy is more effective in this population and reinforces the LDL recommendations from NCEP ATP-III. Further, similar findings demonstrating the benefits of statins in those with diabetes were also observed in the Heart Protection Study and the Anglo-Scandinavian Cardiac Outcomes Trial-Lipid Lowering Arm (ASCOT-LLA).[70,71]

Differences Among Agents

Moderate differences among individual fibrates may have also contributed to inconsistency in safety and efficacy reported in long-term trials. Gemfibrozil has provided the most impressive data at reducing clinical events (HHS, VA-HIT studies) for the fibrates. However, concerns regarding drug interactions, especially with concomitant statin use, and potential for higher rates of muscle toxicity may limit use of this agent in clinical practice. Fenofibrate appears to be more effective than gemfibrozil at favorably altering lipoproteins and does not significantly affect statin metabolism[16,17,18,44,72]; however, the results of the FIELD study did not provide the robust evidence required to justify exclusive use of this agent. In addition, safety concerns regarding slight but significant increases in pancreatitis and pulmonary embolism in the FIELD study indicate a need for practitioners to be cognizant of these possible adverse effects.

Similar increases in the frequency of pulmonary embolism were noted with clofibrate in the CDP. However, a class effect cannot be ruled out since the rates of pulmonary embolism and pancreatitis were not reported in most other major fibrate trials. In the BIP, HHS, and VA-HIT studies, the rates of adverse effects in the active treatment group were generally similar to those of the placebo groups. Other differences among fibrates, which may affect clinical outcomes in long-term trials, include altering concentrations of nontraditional cardiovascular risk factors such as homocysteine, lipoprotein(a), and uric acid concentrations.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: