Effect of Sibutramine on Weight Management and Metabolic Control in Type 2 Diabetes: A meta-analysis of clinical studies

Randomized controlled clinical trials were searched in the current list of medical literature from the Cochrane Library, Medline, and EMBASE, and reference lists of all relevant articles were also checked. Abbott Laboratories were also contacted in order to identify unpublished data.

The key indexing terms used were “sibutramine, obesity, diabetes, clinical trials.” Inclusion criteria were a full publication in English, a randomized controlled trial (blind, parallel, or crossover), a diagnosis of type 2 diabetes, the presence of overweight or obesity, and minimum duration of treatment with sibutramine of at least 3 months. Trials lasted until October 2004. One investigator extracted data and a second checked data extraction. Studies published in abstract form only were not included because it was impossible to judge study quality from abstract alone. The randomized controlled trial quality was assessed following Verhagen Delphi criteria (8). The outcome measures were body weight; waist circumference; fasting blood glucose and serum insulin; HbA_1c; fasting serum triglycerides; total, LDL, and HDL cholesterol; systolic (SBP) and diastolic (DBP) blood pressure; and heart rate. Results are presented as means ± SD.

The meta-analysis was done according to the methods described by Hedges and Olkin (9) to determine the effect size for each study. The mean of a control group (M_c) was subtracted from the mean of the experimental group (M_e) and divided by the pooled SD of both groups: d = (M_e − M_c)/SD. In this case, SD is the square root of the weighted average of the two variances: s² = [(n_e − 1)(s_e)² + (n_c − 1)(s_c)²]/(n_e + n_c − 2). The degree of homogeneity or heterogeneity of the dataset and the χ² test was also calculated.

We found 35 articles in databases matching the search string and selected eight clinical studies that met inclusion criteria. Table 1 shows the study protocols of the trials considered in the meta-analysis. One 6-month randomized controlled trial comparing sibutramine (15 mg/day) with placebo in subjects receiving sulfonylureas (10) and a 12-month randomized prospective placebo-controlled double-blind study (sibutramine 15 and 20 mg/day) in metformin-treated obese subjects (11) with type 2 diabetes were identified. In another three studies, either metformin or sulfonylurea or insulin or diet only was the treatment of the patients (12–14). Another trial had been performed with 10 mg sibutramine for 12 months in diabetic patients treated with glibenclamide (15). In the last two studies examined, sibutramine (15 and 10 mg) had been compared with only a hypocaloric diet (16,17) for 1 year. In all the studies, diet was reduced in calories and specific dietary recommendations were given.

A total of 1,093 obese patients with type 2 diabetes (552 treated with sibutramine M163 F201 and 541 treated with placebo M123 F225) were analyzed. The mean age of the sibutramine-treated patients in all the studies was slightly lower than in the placebo group, and therefore the overall effect resulted in a significant difference between the two groups (P = 0.0078). In the study of Gokcel et al. (14), the enrolled patients were only women. The two groups were similar in terms of race; sex; body weight; waist circumference; fasting blood glucose and serum insulin; HbA_1c; fasting serum triglycerides; total, LDL, and HDL cholesterol; SBP and DBP; and heart rate.

In the sibutramine-placebo comparison, body weight decreased on average 5.533 ± 0.225 kg after sibutramine treatment and −0.900 ± 0.169 kg in the placebo group, and waist circumference (5.320 ± 0.310 vs. −1.130 ± 0.160 cm) significantly decreased after sibutramine. The overall effect size (standard mean difference) was 0.87 (95% CI 1.00–0.74; P = 0.0000) on weight change (Fig. 1A) and 0.67 (0.83–0.51; P = 0.0000) for waist circumference (Fig. 1B).

Mean changes in basal blood glucose show a small but significant variation (−0.17 [95% CI 0.03–0.32]; P = 0.0187) (Fig. 2A), whereas HbA_1c significantly decreased after sibutramine treatment. The overall effect size on HbA_1c was −0.28% (−0.13 to −0.42; P = 0.0002), with some heterogeneity (P = 0.0104) among the studies (Fig. 2B). Treatment benefits were seen in plasma triglycerides and HDL, without significant variations in serum total and LDL cholesterol. The overall effect size on serum triglycerides was −0.24 (−0.09 to −0.39; P = 0.0024) and 0.20 (0.05–0.35; P = 0.0087) for HDL (Fig. 3).

The eight trials did not report differences in SBP between the sibutramine and the placebo but showed a weak increase in DBP in the sibutramine group (0.22 [95% CI 0.07–0.38; P = 0.0050]) (Fig. 4A). Moreover, recording of heart rate showed that sibutramine produced a small increase relative to placebo, either between groups or within groups over time. The overall effect size on heart rate was 0.53 (0.39–0.67; P = 0.0000) (Fig. 4B).

Obesity is a risk factor for type 2 diabetes, dyslipidemia, hypertension, and cardiovascular and coronary artery disease (18,19). In type 2 diabetic patients with obesity, loss of quite modest amounts of weight has been shown to bring about regression of coronary atherosclerosis (20), to increase HDL cholesterol, and to reduce triglycerides. The results of our meta-analysis clearly confirm that treatment with sibutramine is able to determine a considerable decrease in body weight that differs significantly from that reported in the placebo group. This observation further supports the therapeutic efficacy of sibutramine in reducing body weight (21–24), even in a population of obese type 2 diabetic subjects who are considered less compliant and with a limited adherence to the lifestyle advertisements. Weight loss remains the cornerstone for the great majority of overweight diabetic patients, and glucose control and fasting glucose levels improve before major weight loss has occurred. According to Fujioka et al. (12), one can expect a 0.5% reduction in HbA_1c and a 1.1-mmol/l fall in fasting plasma glucose for every 4.5-kg loss in weight.

These results demonstrate that the addition of sibutramine to the treatment with diet, oral antidiabetic agents, or insulin may improve glycemic control. This is a significant finding that could have implications on how these patients are treated in the future. Given that glycemic control reduces HbA_1c levels, one would postulate that sibutramine and the resulting weight loss observed significantly ameliorated the metabolic control. In the Norfolk study, Khaw et al. (25) showed that there is a continuous relationship between HbA_1c levels and cardiovascular disease and total mortality. This relationship was apparent even among persons without diabetes. However, the U.K. Prospective Diabetes Study only found a weak association between glucose control and the risk of myocardial infarction (16% reduction; P = 0.052) (26). Each 1% reduction in HbA_1c levels resulted in a 14% decrease in myocardial infarctions compared with a 37% decrease in microvascular events. Since glycemic control lowers the risk of microvascular events to a greater extent than macrovascular events (27), pharmacologic interventions targeting hypertension and dyslipidemia, preferentially affecting macrovascular disease, have markedly reduced cardiovascular events in patients with type 2 diabetes. In fact, lipid-lowering and hypertension trials in type 2 diabetes have shown significant cardiovascular benefits in persons with long-standing diabetes. Strict glycemic control must therefore be coupled with proven therapies to reduce cardiovascular events by addressing the atherogenic dyslipidemia and hypertension associated with type 2 diabetes.

The “lipid triad” (small dense LDL, hypertriglyceridemia, and low HDL levels) found in the metabolic syndrome and type 2 diabetes is associated with nearly twice the incidence of cardiovascular events compared with an isolated LDL alone (28). Patients with type 2 diabetes have average LDL levels but have an increased number of small dense LDL particles, which are highly susceptible to oxidative modification and are associated with a threefold increase in cardiovascular disease. In addition, low HDL levels deprive patients with type 2 diabetes of maximum protection from reverse cholesterol transport and exposure to potent antioxidants (paroxanase). In this context, the body weight–lowering effect of sibutramine was associated with a marked decrease in plasma triglycerides and a significant increase in HDL cholesterol in obese type 2 diabetic subjects. These results, far from indicating sibutramine as a therapeutic mean to prevent the macrovascular complications in obese type 2 diabetic patients, clearly suggest that a pharmacological approach in a weight management program may be useful in metabolic control and cardiovascular complications.

At present, a major emphasis should first be on reaching appropriate blood pressure targets in patients with type 2 diabetes. Recent data from the National Health and Nutrition Examination Survey registry suggest that <15% of diabetic patients meet Joint National Committee 6 blood pressure targets (29). Finally, the new guidelines established by the Joint National Committee 7 (30) on prevention, detection, evaluation, and treatment of high blood pressure stress the use of aggressive combination drug therapy to achieve blood pressure goals in high-risk patients. Sibutramine causes weight loss, acting at the central nervous system by promoting energy expenditure and inhibiting food intake. It also increases the sympathetic drive, heart rate, and blood pressure through its norepinephrine reuptake inhibitory effect (12,31–33), and this fact limits its use in hypertensive patients. Despite the existence of contradictory results reporting that sibutramine does not influence blood pressure (34,35), a recent meta-analysis (36) confirmed that in sibutramine-treated obese subjects, there is a slight but significant increase in both SBP and DBP. In our meta-analysis of controlled trials, no clear evidence for an effect of sibutramine treatment on SBP and only a small enhancement of DBP were observed, probably because these effects are masked or abolished by the decrease in body weight, which often implies a reduction in blood pressure. However, the decrease in body weight is not sufficient to produce a reduction in the activation of the sympathetic drive by sibutramine. In fact, the heart rate, which may represent a surrogate index of the sympathetic activity, is significantly higher in sibutramine-treated patients. Moreover, we have to take into account that the cardiovascular effect of sibutramine results from a complex interaction of peripheral and central effects (37). In fact, it has been described that in young healthy subject with low sympathetic activity, the peripheral effect dominates, leading to an increased resting blood pressure. On the other hand, when sympathetic activity is increased, such as was found in a larger subgroup of obese hypertensive patients (38), the inhibitory clonidine-like effects of sibutramine seem to prevail. This hypothesis could explain the observation that sibutramine only slightly increases or leaves unmodified the resting blood pressure in a significant proportion of patients.

In summary, although only eight studies were included in the meta-analysis, the results are remarkable in revealing a substantial effect of sibutramine on weight loss as well as a considerable benefit on glycemic control and lipid profile, therefore reducing some of the cardiovascular risk factors in type 2 diabetic patients. The reviewed data further enforce the recommendations that weight management may be the most important therapeutic task for most obese type 2 diabetic individuals.

We are grateful to Dr. Benito Chinea for his contribution on statistical analysis of the IBIS Informatica Srl, Milan. We are also grateful for valuable assistance from Dr. Francesca Gigli.