Bezafibrate and simvastatin combination therapy for diabetic dyslipidaemia: efficacy and safety


  • D. Gavish,

    1. From the 1Department of Medicine ‘A’, Wolfson Medical Center, Holon, and the 2Institute of Metabolic Diseases, Tel Aviv Sourasky Medical Center, Tel Aviv, Israel
    Search for more papers by this author
  • 1 E. Leibovitz,

    1. From the 1Department of Medicine ‘A’, Wolfson Medical Center, Holon, and the 2Institute of Metabolic Diseases, Tel Aviv Sourasky Medical Center, Tel Aviv, Israel
    Search for more papers by this author
  • 1 I. Shapira,

    1. From the 1Department of Medicine ‘A’, Wolfson Medical Center, Holon, and the 2Institute of Metabolic Diseases, Tel Aviv Sourasky Medical Center, Tel Aviv, Israel
    Search for more papers by this author
  • and 2 A. Rubinstein 2

    1. From the 1Department of Medicine ‘A’, Wolfson Medical Center, Holon, and the 2Institute of Metabolic Diseases, Tel Aviv Sourasky Medical Center, Tel Aviv, Israel
    Search for more papers by this author

Dov Gavish MD, Department of Medicine ‘A’, Wolfson Medical Center, Halohamim Street, Tel Giborim, Holon 58100, Israel (fax: 972-35028642; e-mail:


Abstract. Gavish D, Leibovitz E, Shapira I, Rubinstein A (Wolfson Medical Center, Holon, and Tel Aviv Sourasky Medical Center, Tel Aviv, Israel). Bezafibrate and simvastatin combination therapy for diabetic dyslipidaemia: efficacy and safety. J Intern Med 2000; 247: 563–569.

Objective. To determine the efficacy and safety of a statin–fibrate combination in diabetes patients.

Design.   An open 21-month trial in which each patient first received the single drug for 6 months and then a combination of the two for 1 year.

Setting. Three lipid clinics in university-based tertiary care hospitals.

Patients. One hundred and forty-eight patients with type 2 (non-insulin-dependent, NIDDM) diabetes mellitus under stable control for 3 months by means of diet and oral hypoglycaemic medication.

Intervention. Patients from one clinic (n = 48) received bezafibrate slow release (400 mg day–1), and patients from the other two clinics (n = 100) received simvastatin 20 mg day–1. Six months later, all patients were switched to a daily combination of 400 mg bezafibrate slow release and 20 mg simvastatin for 1 year.

Results. The combination of statin and fibrate led to a 23% reduction in total cholesterol, 42% reduction in triglycerides, 29% reduction in LDL-c, 25% increase in HDL-c, 10% decrease in fibrinogen and 19% reduction of Lp(a) levels, and a decrease in the cholesterol/HDL-c ratio (from 8.9 to 5.4) in all 148 patients. Cardiovascular (CV) event rate was significantly reduced from 9.5% during the first 6 months of the study to less than 2% during the last year of the study (whilst on combination Rx). Side-effects with all treatments included only two patients who developed myopathy when on the combined regimen and one on the single statin regimen. However, plasma creatinine phosphokinase (CPK) levels doubled (but remained within the normal range) in most of the patients on combination therapy, compared with only a mild increase in patients receiving a single medication.

Conclusions. The statin and fibrate combination was found to be more efficacious than a single medication for treatment of diabetic dyslipidaemia, as evidenced by improvement in the lipoprotein profile, reductions in Lp(a), fibrinogen and CV event rate, and almost no clinically significant side-effects.


Patients with type 2 (non-insulin-dependent, NIDDM) diabetes mellitus are at an increased risk for cardiovascular disease (CVD) complications and CVD-related death [1–3]. Diabetic dyslipidaemia is at least partially responsible for CVD complications [45]. It has been established that therapy designed to improve lipid levels can diminish the risk of CVD morbidity and mortality in hyperlipidaemic patients with or without prior coronary disease [6–8], and that the effect is at least pronounced in diabetic patients [9].

Statins comprise the most efficacious therapy for clinical event reduction [6–9] by reducing low-density lipoprotein cholesterol (LDL-c). Most diabetes patients with hyperlipidaemia tend to have higher triglyceride (TG) levels at baseline than non-diabetes patients who were studied in several large intervention investigations [6–9]. Patients with both type 2 diabetes mellitus and dyslipidaemia have lipid abnormalities characterized by low HDL-c, high TG, high very low-density lipoprotein cholesterol (VLDL-c), high intermediate-density lipoprotein cholesterol (IDL-C) and moderate LDL-c. However, patients with diabetes have a two- to threefold higher incidence of CVD complications than non-diabetes patients of the same age and gender, with the same level of LDL-c [45]. Two additional risk factors that are related to the high TG and low HDL-c in diabetes patients are higher fibrinogen and Lp(a) levels [1011]. Reassessment of the role of TG in CVD has indicated that the increased risk associated with TG was independent of HDL-c [1213]. Finally, triglycerides have emerged as an independent risk factor in women [14] and in type 2 diabetes patients [15].

The recent results from the UKPDS have demonstrated that coronary artery disease is a major cause of mortality in patients with type 2 diabetes mellitus and that the incidence of macrovascular complications is twice that of microvascular disease [1617]. Reduction of the macrovascular event rate is only moderately affected by control of hyperglycaemia, whilst LDL-c reduction was shown to have a significant capacity to reduce macrovascular complications in patients with type 2 diabetes [1617].

Statins are well-established treatments of hypercholesterolaemia [6–9], and fibrates are efficacious in the reduction of TG-rich lipoproteins and cause an increase in HDL-c [1819]. The recent Bezafibrate Coronary Angiographic Intervention Trial showed that the progression of congestive heart disease (CHD) and coronary events can be reduced with specific reduction in TG-rich lipoproteins, independent of any reduction in LDL-c [20].

The mode of action of fibrates was shown to be mediated by activation of specific transcription factors that belong to the nuclear receptor family termed peroxisome proliferator-activated receptors (PPARs) which cause induction of LPL expression, reduce formation of apoprotein CIII (APO CIII), increase synthesis of apoprotein AI and AII and reduce the amount of the pathogenic-dense LDL [2122]. Such effects are especially warranted in patients with diabetic dyslipidaemia who typically have low LPL activity, high APO CIII and high amounts of dense LDL [3–5] A statin–fibrate combination would appear, therefore, to be ideal for most diabetes patients with dyslipidaemia. The use of such combination medication has been reported by several groups, mostly in patients with combined dyslipidaemia [23] or when a synergistic effect in LDL-c lowering was required [24]. Whilst the use of the statin–fibrate combination, lovastatin and gemfibrozil, was associated with myopathy and rhabdomyolysis in 12 cases, several studies in patients with ischaemic heart disease (IHD) and hyperlipidaemia have shown that the combination medication is relatively safe [2324]. Information on the effect of combination medication in patients with diabetic dyslipidaemia together with high TG and total cholesterol levels is sparse. Our study was designed to compare the effect of single statin or fibrate medication given for 6 months with the effect of combination therapy given for 1 year in the same patients with diabetic dyslipidaemia.



Between 1994 and 1998, 148 patients aged 49–70 years with type 2 diabetes were enrolled in the study. These patients had fasting plasma glucose over 6 mmol L–1 on at least two occasions before being diagnosed with diabetes mellitus. There were 85 males and 63 females (all of whom were postmenopausal and not on any hormone replacement therapy), with a mean age of 59 years. All patients were recorded at our tertiary care lipid clinics after referral and follow-up due to dyslipidaemia.

Weight and body mass index (BMI) were stable for at least 3 months prior to inclusion. All patients received instructions for diet according to a standard protocol for patients with diabetes and dyslipidaemia (phase II AHA diet).

The study was approved by the ethics committee in the participating centres. All patients gave their informed consent for participating in this study.

Patients’ Hgb A1C , blood sugar levels and BMI were stabilized for two consecutive tests 3 months apart, prior to inclusion into the study ( Fig. 1).

Figure 1.

Design of study. All patients received combination Rx (n = 148); 48 patients received bezafibrate as a single drug; 100 patients received simvastatin as a single drug.

Patient characteristics are shown in Table 1. Excluded from the study were patients with a history of stroke or MI in the 6 months prior to inclusion, patients with malignancies or patients receiving steroid treatment for any reason. Three tertiary care lipid clinics received patients to the study. Two clinics initiated treatment with simvastatin as the first single medication (100 patients – group I), and the third clinic initiated slow-release (SR) bezafibrate (Bezalip retard) as the initial treatment (48 patients – group II). The mean triglyceride levels were slightly higher in group II than in group I. No other differences existed between the groups ( Table 1). Lp(a) was tested only in group I. All patients were on single medication for a period of 6 months (simvastatin 20 mg day–1 or bezafibrate SR 400 mg day–1), followed by a period of 12 months on combination therapy. Monitoring for effect and safety was done every 3 months.

Table 1.  Patient characteristics for patients with type 2 diabetes mellitus according to treatment group and sex
Group IGroup IIAll
  1. ACE, angiotensin-converting enzyme; CVD, cardiovascular disease; HTN, hypertension; BMI, body mass index; LDL-c, low-density lipoprotein cholesterol; Ca blockers, calcium channel blockers; TC, total cholesterol; TG, triglycerides; HDL-c, high-density lipoprotein cholesterol.

Age (years)60 ± 954 ± 558 ± 962 ± 859 ± 660 ± 861 ± 958 ± 659 ± 8
Prior CVD [n (%)]32 (55)22 (52)54 (54)14 (52)9 (43)23 (48)46 (54)31 (49)77 (52)
Smoking [n (%)]11 (19)12 (29)23 (23)10 (37)9 (42)19 (40)21 (25)21 (33)42 (28)
HTN [n (%)]28 (48)13 (31)41 (41)11 (41)10 (48)21 (44)39 (46)23 (37)62 (42)
HbA1c8.0 ± 0.88.0 ± 0.68.0 ± 0.88.2 ± 18.4 ± 18.3 ± 1.38.0 ± 1.08.1 ± 2.08.0 ± 1.0
BMI25 ± 224 ± 325 ± 325 ± 326 ± 226 ± 325 ± 324.5 ± 325.5 ± 3
Sulphonylurea [n (%)]20 (34)20 (48)40 (40)12 (44)10 (48)22 (46)32 (38)30 (48)62 (43)
Metformin [n (%)]52 (90)38 (90)90 (90)26 (96)20 (96)46 (96)78 (92)58 (92)136 (92)
Acarbose [n (%)]10 (17)8 (19)18 (18)6 (22)4 (19)10 (21)16 (19)12 (19)28 (19)
Insulin [n (%)]10 (17)4 (10)14 (14)4 (15)3 (14)7 (15)14 (17)7 (11)21 (14)
ACE inhibitors [n (%)]33 (57)36 (86)69 (69)20 (74)16 (76)36 (75)53 (62)52 (83)105 (71)
Aspirin [n (%)]55 (95)40 (95)95 (95)26 (96)20 (95)46 (96)81 (95)60 (95)141 (95)
β-blockers [n (%)]6 (10)4 (9)10 (10)3 (11)2 (9)5 (10)9 (11)6 (10)15 (10)
Ca blockers [n (%)]22 (38)17 (40)39 (39)6 (22)3 (14)9 (19)28 (33)20 (32)48 (32)
TC (mmol L–1) 6.8 ± 0.77.0 ± 0.56.9 ± 0.76.8 ± 1.07.2 ± 0.87.0 ± 0.66.8 ± 0.97.1 ± 0.97.0 ± 0.8
TG (mmol L–1) 3.8 ± 0.83.9 ± 0.73.8 ± 0.84.6 ± 0.95.3 ± 0.65.2 ± 0.94.2 ± 2.54.3 ± 2.04.3 ± 2.0
LDL-c (mmol L–1) 4.5 ± 0.64.6 ± 0.74.5 ± 0.74.4 ± 2.04.8 ± 2.04.6 ± 1.44.4 ± 2.04.7 ± 1.24.5 ± 1.6
HDL-c (mmol L–1) 0.9 ± 0.21.0 ± 0.20.9 ± 0.20.9 ± 0.20.9 ± 0.20.9 ± 0.20.9 ± 0.20.9 ± 0.30.9 ± 0.2
TC/HDL9.0 ± 2.08.9 ± 2.09.0 ± 0.28.8 ± 2.28.4 ± 2.08.6 ± 2.08.9 ± 2.08.8 ± 2.08.9 ± 2.0
Fibrinogen (mg dL–1) 366 ± 80355 ± 81364 ± 82376 ± 46382 ± 15381 ± 17374 ± 80360 ± 82369 ± 80
Lp(a) (mg dL–1) 30 ± 1438 ± 1632 ± 19N.D.30 ± 1438 ± 1632 ± 19  

Laboratory testing

Analytical methods. Fasting blood samples were collected into chilled tubes after a 12-h fast prior to inclusion and every 12 weeks after inclusion. Total cholesterol and TG levels were measured by an enzymatic method [25]. HDL-c was measured following precipitation of VLDL and LDL-c by sodium phosphotungstate and magnesium chloride. Day-to-day analysis and interday analysis variations did not exceed a standard deviation of 0.05 mmol L–1, and all coefficients of variations were < 3.5%. Only four patients (2.7%) had a triglyceride level > 4.5 mmol L–1 (but below 5 mmol L–1), a level at which the indirect LDL-c calculation becomes unreliable. However, exclusion of those patients did not change the final results, so we continued to use Friedwald’s formula in the patients with TG level above 4.5 mmol L–1 as well, as did the investigators in the Copenhagen male study [13]. In patients with acute events occurring during the follow-up monitoring period (only five episodes), blood tests were performed 8 weeks after the event to avoid interference with the study results. Fibrinogen and Lp(a) were measured using ELISA methods as described previously [2627]. Routine haematological studies, serum glucose and creatinine determinations, serum aspartate aminotransferase (AST), alanine aminotransferase (ALT) and creatinine phosphokinase (CPK) were measured at every visit and assays were performed in the local laboratories.

Data analysis

Data analysis was performed with the SPSS for Windows and the SPSI PL programs. Data are given as means ± SD. Two-tailed Student’s t-tests were used to compare groups.


The effects of therapy on plasma lipids and lipoproteins (mean ± SD) are shown in Table 2. In both groups, the effects of the combination therapy were significantly more pronounced, with a 42% reduction in plasma TG, 23% reduction of total cholesterol, 29% reduction of LDL cholesterol, 19% reduction of Lp(a) levels, and a 10% reduction in fibrinogen, whilst HDL-c increased by 25%.

Table 2.  Risk factor reduction by treatment groups (mean ± SD)
Group I (n = 100) Group II (n = 48) All (n = 148)
  • *

    P < 0.01 (compared to baseline);

  • , P < 0.01 (comparison of single drug to combination);

  • , P < 0.05 (compared to baseline).

Triglycerides (mmol L–1) 3.8 ± 0.83.5 ± 0.62.4 ± 0.5 *5.2 ± 1.53.2 ± 0.9 *2.8 ± 0.7 *4.3 ± 1.32.5 ± 0.6 *
Cholesterol (mmol L–1) 6.9 ± 0.75.4 ± 0.6 *5.2 ± 0.6 *7.0 ± 0.66.7 ± 0.5 *5.6 ± 0.4 *7.0 ± 0.75.4 ± 0.5 *
LDL-c (mmol L–1) 4.5 ± 0.73.1 ± 0.6 *3.2 ± 0.5 *4.6 ± 1.64.4 ± 0.4 *3.4 ± 0.4 *4.5 ± 1.03.2 ± 0.5 *
HDL-c (mmol L–1) 0.9 ± 0.20.9 ± 0.21.0 ± 0.1 *0.8 ± 0.21.0 ± 0.2 *1.1 ± 0.2 *0.8 ± 0.21.0 ± 0.2 *
Cholesterol/HDL ratio9.0 ± 2.26.6 ± 1.3 *5.6 ± 0.9 *8.5 ± 2.06.7 ± 1.35.4 ± 1.0 *8.9 ± 2.05.5 ± 1.0 *
LDL to HDL ratio6.2 ± 1.54.0 ± 1.0 *3.6 ± 0.7 *5.4 ± 2.54.2 ± 1.2 *3.2 ± 0.56.0 ± 2.03.4 ± 0.6 *
Fibrinogen (mg dL–1) 364 ± 82365 ± 74323 ± 51 *380 ± 16360 ± 19360 ± 17 *369 ± 80334 ± 54 *
Lp(a) (mg dL–1) 32 ± 2832 ± 2926 ± 24 *Not doneNot doneNot done32 ± 2826 ± 24

There was no difference between males and females in both groups, and no effect of additional medications between groups ( Table 1). Figure 2 shows the results presented as a percentage change from baseline.

Figure 2.

Percentage change from baseline in risk factors: triglycerides (TG), total cholesterol (Chol), low-density lipoprotein cholesterol (LDL-c), high-density lipoprotein cholesterol (HDL-c) and creatinine phosphokinase (CPK).

The effects on CPK were as follows: CPK was measured as 74 ± 33 IU L–1 at baseline and changed to 88 ± 48 IU L–1 with statin alone and to 76 ± 24 IU L–1 with fibrate alone. The combination medication raised plasma CPK to 113 ± 87 IU L–1 (a 52% change from baseline). However, all mean changes were within the accepted normal range for CPK levels at our laboratory. There were no significant changes in liver function tests (LFTs) or complete blood count (CBC) as a result of the various drug protocols.

All elevations of enzymes and cases of clinical myopathy are shown in Table 3. There was a very low incidence of side-effects with combination medication after 1 year. Table 4 summarizes the total number of clinical events observed during the study period, noteworthy amongst them being the relatively high incidence of cardiovascular events (9.5%). Most of the events occurred within the first 6 months of the study, whilst the patients were on a single drug, with a slightly higher event rate in the bezafibrate group (12.5% – six events) and only 6% in the statin group (P-value – NS). There was a significant reduction in the event rate during the year whilst the patients were on the combination therapy (only two events – under 2%). More events occurred in males than in females (10/85 [12%] vs. 6/63 [6.3%]; P < 0.05).

Table 3.  Number and percentage of patients suffering from side-effects according to treatment group
GroupMild CPK elevation> × 2 CPK elevationLiver function abnormalityClinical myopathy
All16 (11%)5 (3.4%)4 (3%)3 (2%)
Statins only3 (3%)1 (1%)1 (1%)1 (1%)
Fibrates only1 (2%)1 (2%)1 (2%)
Combination12 (8%)3 (2%)2 (1.4%)2 (1.4%)
Table 4.  All events (cardiovascular and non-cardiovascular) occurring during the 18 months of follow-up
(n = 100)
(n = 48)
(n = 148)
  • MI, myocardial infarction; AP, angina pectoris ; CABG, coronary artery bypass graft; TIA, transient ischaemic attack.

  • a

    The four patients who needed angiography during follow-up include two patients who had undergone angioplasty and one patient who underwent CABG.

Cardiovascular events
All6 (6%)6 (12%)2 (1.4%)14 (9.5%)
Death1 (1%)1 (2.1%)02 (1.4%)
Acute MI01 (2.1%)1 (0.7%)2 (1.4%)
Unstable AP2 (2%)1 (2.1%)1 (0.7%)4 (2.8%)
Angiography2 (2%)2 (4.2%)04 (2.8%) a
Angioplasty2 (2%)2 (1.4%) a
CABG1 0.7%) a
TIA1 (1%)1 (2.1%)02 (1.4%)
Non-cardiovascular events
All1 (1%)1 (2.1%)2 (1.4%)4 (2.8%)
Pneumonia1 (1%)1 (0.7%)2 (1.4%)
Urosepsis1 (0.7%)1 (0.7%)
Acute cholecystitis1 (2.1%)1 (0.7%)


Patients with diabetic dyslipidaemia represent a population at very high risk for cardiovascular complications. In this group of patients, cholesterol and triglyceride level control, as well as glycaemic control, is essential in reducing the rate of cardiovascular events.

In our study, 9.5% had a cardiovascular event during the 18 months of follow-up. Most of the events occurred within the first 6 months, whilst patients were on a single hypolipidaemic drug, and it seems that the group on bezafibrate SR had a higher rate of events (12%) than did the simvastatin group (6%). The relatively high rate of cardiovascular events is similar to that described for these individuals in the literature [1–6]. We treated the study patients with a combination of simvastatin and bezafibrate, each of which was shown to improve the lipoprotein profile and reduce the occurrence of cardiovascular events [192324]. Our results demonstrated that combination therapy carries a significant reduction (of 29%) in plasma LDL-c, a 25% increase in HDL-c, a 42% reduction in TG and a 10% reduction in fibrinogen, with an up to 19% reduction in Lp(a). This improved lipid profile was not reached with either simvastatin 20 mg day–1 alone or with bezafibrate 400 mg day–1 alone. The ratio of TC/HDL-c was significantly reduced (by 38%) with combination medication, as compared with only a 22% reduction using a single medication. Simvastatin as a single medication was shown to cause a significant reduction in cardiovascular events [6]. Bezafibrate as a single medication was shown to have additional possible protective effects unrelated to the effect on LDL-c [20]. Therefore it can be assumed that a combination of statin and fibrate could better improve the lipid profile control, as well as the outcome, in this high-risk population of diabetes patients. It is noteworthy that the CV rate was reduced to 2% in the year of combination treatment.

An interesting observation in our study is that the plasma fibrinogen levels, which were practically unaffected by a single drug (simvastatin or bezafibrate), were significantly reduced by a combination of the two. Whilst the results of previous reports on the effects of statins on fibrinogen were conflicting, it was to be expected that fibrates as a single medication would reduce the fibrinogen level [27], and our patients did have lower fibrinogen levels after they were given fibrates. The combination medication, however, enhanced the effect even further, possibly by enhancing the reduction of TG-rich lipoproteins via a combined effect on synthesis and on the clearance of lipoproteins [18–21].

Plasma Lp(a) is also considered as a risk factor for cardiovascular disease [11]. Although statins fail to reduce Lp(a) [26], several reports have suggested that fibrates may cause some reduction in these levels [18]. Our findings confirmed that statins alone had no effect on Lp(a) and fibrinogen levels, but a lasting effect (for more than 12 months) was observed when combination treatment was used, with a reduction of Lp(a) by 19% and fibrinogen by 10% (P = 0.05). Such effects of combination medication on both lipoproteins and the pro-atherogenic Lp(a) and fibrinogen support the contention that this treatment approach could be used for high-risk diabetes patients with dyslipidaemia with the aim of reducing the rate of clinical cardiovascular events more effectively than a single medication.

In testing drug safety, we showed that the rate of side-effects attributed to medications was similar for single and combination medication (less than 2%). Minor LFT abnormalities were seen only in two (1.4%) patients on combination medication, and this was similar to the incidence for single medications. No patient needed to stop treatment due to LFT abnormalities. Elevated CPK and clinical myopathy were previously reported in individuals on statins medication [6] as well as those on combination medication [28]. In our study, only three patients had myopathy (two were on combination medication and one on simvastatin alone), which was reversible after cessation of drugs. However, a 40% increase in CPK (albeit within the normal range) was observed in 70% of the patients on combination medication as compared with a 20% increase amongst those on a single medication. This increase in CPK was clinically insignificant and caused neither myopathy nor a need to stop treatment at up to 12 months of follow-up. However, due to the persistent increase in CPK seen in patients on the combination medication, a follow-up of CPK levels and myopathy should be continued at least every 3 months in all patients on combination medication of statins and fibrates so that myopathy can be detected.

In conclusion, we have shown that a combination of statin and fibrate is more effective than a single medication in reducing lipids (TG, total cholesterol), atherogenic lipoproteins (LDL-c, VLDL-c, IDL-C), atherogenic fibrinogen and Lp(a) and in raising HDL-c in high-risk diabetes patients with dyslipidaemia. Such a drug combination is relatively safe and may reduce cardiovascular events amongst these patients. The relatively high incidence of cardiovascular events (9.5%) during the first phase of the study (whilst the patients were on single drug therapy) emphasizes the importance of appropriate treatment specific for the prevention of cardiovascular complications in diabetic patients with dyslipidaemia. Our data suggest that with the combination treatment of statin and fibrate, a significant reduction of events (less than 2% year–1) can be achieved.

Received 6 September 1999; accepted 19 October 1999.