In this study, we showed that treatment with bezafibrate significantly reduced HbA1c levels in patients with diabetes complicated by dyslipidemia. This effect was observed in a baseline HbA1c-dependent manner, regardless of concurrent use of diabetes drugs. Patients with a shorter duration of diabetes showed greater decreases in HbA1c levels. Further, the rate of change in triglyceride levels was related to changes in HbA1c levels. Adverse drug reactions occurred in 306 patients (5.1%), but these were not severe in 289 (94.4%). Taken together, these results suggest that bezafibrate, which improves both lipid and blood glucose metabolism, should be considered suitable for the treatment of dyslipidemia in patients with diabetes.
The importance of the rigorous control of glycemia in diabetes is confirmed by the striking findings that a 1% reduction in HbA1c level is associated with a 21% decrease in risk of death related to diabetes , and that a 1-mmol/l increase in TG level is associated with a 32% increased risk of a cardiovascular event . HbA1c correlates significantly with both TG and HDL-C in patients with type 2 diabetes [21, 22]. Further, patients with type 2 diabetes tend to have higher TG and lower HDL-C values than dyslipidemic patients without type 2 diabetes . TG levels are significantly higher and HDL-C levels significantly lower in patients with relatively poor glycemic control than in those with adequate control . These findings confirm the importance of controlling both TG and HDL-C, in addition to blood glucose, in the treatment of type 2 diabetes.
Statins have been widely used to treat dyslipidemia, and evidence supports their role in preventing cardiovascular events [23–26]. However, fibrates, including bezafibrate, generally have a stronger effect on TG and HDL-C than statins. Atorvastatin 20 mg/day and rosuvastatin 10 mg/day reduced TG levels by 4.6% and 36.4%, and increased or decreased HDL-C levels by 4.5% and 2.0%, respectively . Although rosuvastatin (5-20 mg/day) decreased TG levels by 29%-31% and increased HDL-C levels by 12.4%-16.7%, bezafibrate (400 mg/day) decreased TG levels by 45% and increased HDL-C levels by 43% . The present study also revealed that bezafibrate decreased the rate of change in TG level by 45.4% and the rate of increase in HDL-C level by 14.0% (Table 3). These results suggest that the effect of fibrates on TG and HDL-C levels is stronger than that of statins.
Although controversial, evidence suggests that fibrates have a lower association with cardiovascular events than statins. Indeed, a recent meta-analysis of 18 fibrate trials  revealed that fibrates significantly reduce the risk of cardiovascular events, but not uniformly [30–34]. In the ACCORD study, cardiovascular disease risk in 5,581 patients  receiving combination therapy with a statin plus a fenofibrate was no different than that in patients with statin monotherapy. Further, in the BIP study of 3,090 patients , bezafibrate did not significantly reduce the primary endpoint (fatal or nonfatal myocardial infarction or sudden death).
Common findings between these studies [30–33] are as follows: first, baseline TG values were not high, averaging 1.8 mmol/l in ACCORD and 1.6 mmol/l in BIP. Second, subgroup analysis of patients with baseline TG values ≥ 2.2 mmol/l exhibited potent (31%) risk reduction in ACCORD and 39.5% (P = 0.02) in BIP. Results of the FIELD study were similar . We therefore conclude that fibrates can play a prominent role in reducing the risk of cardiovascular events in patients with high TG levels. Further, recent reports indicate that statin therapy is associated with a slightly higher risk of developing diabetes [35, 36]. These findings suggest that if a patient is at low risk for cardiovascular disease, the potential for statins to increase diabetes risk should be taken into account.
Bezafibrate reduces blood glucose levels in patients with type 2 diabetes [14, 15]. Elkeles et al. reported that bezafibrate significantly decreased the combined incidence of ischemic change in ECG and documented myocardial infarction in patients with type 2 diabetes . Further, Tenenbaum et al. reported that bezafibrate decreased the development and delayed the onset of type 2 diabetes in patients with impaired fasting glucose . Therefore, bezafibrate may be more suitable than statins for treating dyslipidemic patients with type 2 diabetes if their cardiovascular risk is low. To our knowledge, our present study is the first to demonstrate the beneficial effects of bezafibrate on lipid and glucose metabolism in a large number of patients with diabetes.
Consistent with studies cited above, we showed here that bezafibrate improved blood glucose and lipid levels. All lipid values except LDL-C improved significantly from baseline, with TG, HDL-C levels and the TG/HDL ratio in particular showing significant changes. Further, HbA1c, FBG, and HOMA-R improved significantly from baseline. HbA1c levels decreased by -0.47% in all patients and by -0.76% in patients with baseline HbA1c levels ≥7.0%. Notably, bezafibrate showed a potent hypoglycemic effect regardless of concurrent diabetes drug use. On analysis stratified by diabetes duration, bezafibrate alone reduced HbA1c in all subgroups, most notably in those with diabetes for < 1 year. Therefore, bezafibrate monotherapy may be appropriate for treating early-onset type 2 diabetes coexisting with hypertriglyceridemia. Alternatively, hyperglycemic patients with inadequate blood glucose control might benefit from concomitant administration of bezafibrate and a diabetes drug.
We also demonstrated here that as TG levels decreased, those of HbA1c decreased by 0.47% in all patients (Figure 1 and Table 3). Ogawa reported that bezafibrate decreased TG levels by 50% and decreased HbA1c levels from 7.2% to 6.9% in patients with type 2 diabetes . Taniguchi reported similar results . In contrast, the three-year SENDCAP study of 164 patients with type 2 diabetes reported that while bezafibrate decreased TG significantly, the change in HbA1c levels over the course of the study was not significant between bezafibrate and the control groups . Although long-term studies will be required to confirm this observation, average baseline TG levels were lower than those cited above, leading us to conclude that this may have caused the change in HbA1c levels.
Bays et al. reported the following two relationships between fatty acid levels and type 2 diabetes : (1) chronically increased plasma free-fatty acid induces hepatic and muscle insulin resistance and impairs insulin secretion; and (2) enlarged fat cells become insulin-resistant, with diminished capacity to store fat. When the storage capacity in adipocyte tissue is exceeded, lipids flow over into muscle and liver, causing muscle and hepatic insulin resistance. Bezafibrate has been reported to increase fatty acid degradation via beta-oxidation in the peroxisomes and mitochondria [40–42]. Further, Van der Ziji et al. reported that pancreatic fatty acid accumulation is related to β cell dysfunction , and Fernandes-Santos et al. reported that bezafibrate prevented pancreatic fat accumulation and hypertrophy in mice while attenuating glucose intolerance and insulin resistance . These studies suggest that plasma free fatty acid and ectopic fat accumulation involve glucose tolerance, and that bezafibrate improves glucose tolerance through fatty acid degradation via activating beta-oxidation.
ADRs were observed in 306 (5.1%) of 5,978 patients in the safety analysis group. The most frequent were increases in blood levels of creatine phosphokinase (0.8%), creatinine (0.8%) and urea (0.5%). No specific differences were observed related to the use or non-use of diabetes drugs. Betteridge and O' Bryan-Tear  and Beggs et al.  reported, respectively, that 7.7% (7/91) and 5.4% (7/130) of ADRs were caused by bezafibrate administration. The rate of ADRs in the present study is comparable to their result.
In the present study, concurrent use of the biguanide drug metformin was 16.6%. The guidelines of the American Diabetes Association and the European Association for the Study of Diabetes recommend initial treatment of type 2 diabetes mellitus with metformin . Since 2009, metformin treatment has been restricted to cases that do not respond to sulfonylureas. The maximum dose was 750 mg/day, which, until 2010 in Japan at least, is much less than the dose administered in Western countries. Therefore, the results of clinical trials in Western countries cannot be directly applied to Japanese patients. Further, the majority of Japanese patients with type 2 diabetes mellitus are less obese and less insulin-resistant than European and American patients . Currently in Japan, the maximum dose of metformin is 2,250 mg/day.
Our study has several important limitations. First, it was conducted under a prospective observational cohort design with no control arm, and it was not possible to eliminate all confounding factors. Interpretation of our findings, therefore, requires caution until further studies with controls can be conducted. Second, changes in body weight, modification of lifestyle, and use of diabetes drugs were not recorded, and the effects of these variables on glucose metabolism cannot be ruled out. It is, however, important to note that other studies [15, 16] reported result similar to ours presented here. Moreover, of the 254 patients not treated with diabetes drugs in the present study, average HbA1c levels decreased by 0.56% after administration of bezafibrate (Table 4), strongly suggesting that bezafibrate alone lowers blood glucose levels. Third, for logistic, economic and other reasons, laboratory measurements could only be performed at each of the 1,066 institutions and not in a central laboratory. Although the JDS, in collaboration with the Japan Society of Clinical Chemistry, developed a national standardization scheme for determining HbA1c levels in 1995 in Japan , and Nihei et al. reported that commutability among the most frequently used analytical techniques in Japan was secured at a specific level , differences in HbA1c levels determined by different methods and laboratories cannot be ruled out. Nevertheless, we believe that the present results demonstrate convincingly the beneficial effects of bezafibrate in patients with type 2 diabetes in clinical settings.