Effect of Orlistat in Overweight and Obese Patients With Type 2 Diabetes Treated With Metformin

Patients with type 2 diabetes who were 40–65 years of age, had a BMI of 28–43 kg/m², had maintained a stable weight for ≥3 months, had HbA_1c between 7.5 and 12.0%, and had received metformin treatment at 1,000–2,550 mg/day for at least 6 weeks were eligible for the study. Sulfonylurea therapy in combination with metformin was permitted as long as the sulfonylurea dose was stable for 12 weeks before study entry. Patients receiving insulin, thiazolidinediones, or α-glucosidase inhibitors were excluded. Other exclusion criteria included any clinical condition that might affect study end points, including renal, hepatic, or endocrine disorders, poorly controlled hypertension (systolic blood pressure ≥160 mmHg or diastolic blood pressure ≥100 mmHg), active gastrointestinal disease, previous bariatric surgery, a history of bulimia, substance abuse, or the use of any weight loss medications. Women who were pregnant, lactating, or of childbearing potential were also excluded.

This multicenter, double-blind, randomized, placebo-controlled trial was conducted at 34 centers in the U.S. and 6 centers in Canada. The study consisted of a 2-week screening phase followed by a 52-week treatment phase. After screening, patients were randomized to 1 year of treatment with 120 mg orlistat (Xenical; Hoffman-La Roche, Nutley, NJ) or placebo three times daily with main meals. To ensure treatment group balance for initial glycemic control, patients were stratified into two groups based on HbA_1c levels (≤10 or >10%). A reduced-calorie diet (∼600 kcal daily deficit) based on American Diabetes Association recommendations and containing 30% of calories as fat, 50% as carbohydrate, and 20% as protein, with a maximum cholesterol content of 300 mg/day, was prescribed for all patients. The energy content of the prescribed diet was calculated from an estimate of patients’ initial energy requirements (using the Harris-Benedict equations) and was designed to promote a weight loss of 0.25–0.5 kg/ week. Daily calorie intake was reduced by an additional 200 kcal after 6 months to compensate for the reduction in energy requirements caused by weight loss, with a minimum intake of 1,200 kcal/day. Patients received dietary counseling at baseline and at regular intervals throughout the study and were encouraged to increase their level of physical activity. A multivitamin supplement was prescribed to be taken daily at least 2 h before or after the evening dose of study medication. Written informed consent was obtained from all patients. The study was conducted according to the principles of the Declaration of Helsinki and was approved by the Ethics Committee or Institutional Review Board at all study sites.

The initial screening visit included a medical history, physical examination, electrocardiogram, standard laboratory assessments, vital signs, and measurements of body weight, HbA_1c, and serum lipids. Body weight, blood pressure (recorded in a sitting position), and fasting serum glucose were measured at every clinic visit (weeks 0, 2, 4, 8, 12, 16, 20, 24, 30, 36, 42, 48, and 52). HbA_1c was measured at weeks 0, 12, 24, 36, and 52. Serum lipids (total, LDL, and HDL cholesterol, LDL/HDL ratio, and triglycerides) were obtained at weeks 0, 24, and 52.

Changes in body weight and HbA_1c were the primary efficacy variables. Fasting serum glucose and changes in diabetes medication were also used to assess efficacy. Other secondary efficacy variables were changes in serum lipids, blood pressure, and waist circumference (measured midway between the lateral lower rib margin and the iliac crest) at weeks 0, 24, and 52.

Diabetes medications were changed when hypoglycemia occurred or when deterioration in glycemic control was observed. The dosage of diabetes medication was increased if HbA_1c was >12.0% (or increased by >2.0% from baseline) at week 12, if HbA_1c was > 10.2% (or increased by >0.2%) at week 24, or if the serum glucose concentration was >19.4 mmol/l (350 mg/dl) at any clinic visit. The patients’ diabetes medications were also adjusted in response to documented hypoglycemia (two home blood glucose measurements of <3.3 mol/l [60 mg/dl] in 1 week or a single value <2.8 mmol/l [50 mg/dl] recorded by their physician). Additional measurements of HbA_1c and fasting serum glucose were made within 1 week of alterations in diabetes therapy.

All adverse events were recorded, regardless of their relationship to treatment. To ensure consistency in reporting of gastrointestinal events related to orlistat treatment, a glossary of standard terms was used, as defined in earlier studies (16,17). Standard laboratory measurements, including hematology, clinical chemistry, and urinalysis were assessed by a central laboratory.

Statistical analyses were applied to data from the intent-to-treat population, defined as all patients who were randomized to treatment and had at least one follow-up assessment (18). The two primary efficacy variables were body weight and HbA_1c. Therefore an α-level of 0.025 was used in sample size estimations. The sample size was estimated to provide 80% power to detect treatment differences in body weight and HbA_1c separately at α = 0.025, assuming a dropout rate of 35%.

Changes in fasting serum glucose and in diabetes medication were also used to assess efficacy. Secondary efficacy variables were changes in serum lipids and blood pressure. The null hypothesis was that the expected mean changes in primary and secondary efficacy variables in the orlistat-treated patients were not different from those in the placebo-treated patients after 1 year of treatment. The significance of treatment differences for body weight and HbA_1c was tested using ANCOVA models that included fixed effects for treatment, center, treatment by center interaction, HbA_1c stratum, treatment by HbA_1c stratum interaction, and baseline covariate. As exploratory analyses, treatment differences in the change in HbA_1c were tested using an ANCOVA model that included fixed effects for treatment and change in diabetes medications. Treatment differences for changes in fasting serum glucose, serum lipids, blood pressure, dose of diabetes medication, and other secondary efficacy variables were tested using an ANCOVA model with fixed effects for treatment, center, treatment-by-center interaction, and baseline covariate. Placebo-adjusted 95% CIs of orlistat treatment effect, based on the least squares mean, were also determined for all efficacy variables. The frequency distributions of proportions of patients in each treatment group with decreases in body weight of ≥0.5 and ≥1.0% were compared using the Cochran Mantel-Haenszel test.

A total of 516 patients were enrolled and randomized to double-blind treatment with placebo (n = 261) or orlistat (n = 255). Of these, 254 patients in the placebo group and 250 patients in the orlistat group had at least one efficacy follow-up and were thus eligible for intent-to-treat analysis. A total of 311 patients completed 1 year of treatment. In the safety population, more patients withdrew prematurely in the placebo group (n = 115, 44%) than in the orlistat group (90, 35%) (P < 0.05). The most common reasons for premature withdrawal were treatment refusal (57 placebo vs. 36 orlistat), loss to follow-up (35 vs. 25), adverse events (12 vs. 25), protocol violations (5 vs. 0), and treatment failure (2 vs. 0).

Baseline characteristics were similar in the two treatment groups (Table 1). Similar proportions of patients in the placebo and orlistat groups were receiving metformin monotherapy (39 and 44%, respectively) or metformin in combination with sulfonylureas (60 and 55%, respectively) (P = NS between groups). Most individuals treated with sulfonylureas received either glyburide (n = 160) or glipizide (n = 99). Only 1% of patients in each group were receiving metformin in combination with a nonsulfonylurea diabetes medication. In general, glycemic control was poor at baseline, as evidenced by mean HbA_1c values of 8.8 and 8.9% in the placebo and orlistat groups, respectively.

After 1 year of treatment, patients in the orlistat plus diet group achieved greater weight loss than patients in the placebo plus diet group (4.7 ± 0.3 kg [4.6 ± 0.3%] vs. 1.8 ± 0.3 kg [1.7 ± 0.2%], P < 0.0001) (Fig. 1A). Differences in the rate of weight loss between groups were apparent within 4 weeks of starting double-blind treatment, with orlistat-treated patients losing weight more rapidly than placebo-treated patients over that period of time (mean rates of −0.48 and −0.33 kg/week, respectively, P < 0.05). More than twice as many orlistat-treated patients lost ≥5% of baseline body weight compared with the placebo group (39.0 vs. 15.7%, P = 0.008). In addition, more patients in the orlistat than in the placebo group lost ≥ 10% of baseline body weight (14.1 vs. 3.9%; P = 0.003).

Changes in treatment with diabetes medications over the 1-year study period were significantly different between the orlistat and placebo groups. Compared with placebo treatment, orlistat therapy was associated with reductions in both daily metformin dose (−16 ± 24 vs. 49 ± 24 mg/day, P = 0.013) and relative sulfonylurea dose, expressed as percent change, with doses standardized to a percentage of maximum daily dose (− 11.5 ± 3.6 vs. −0.9 ± 2.6%, P = 0.027). Moreover, twice as many patients in the orlistat than in the placebo group either reduced or discontinued one or more diabetes medications (17.1 vs. 8.2%). Conversely, more placebo- than orlistat-treated patients required additional or increased dosages of diabetes medication (21.7 vs. 12.2%). These changes in diabetes medication usage were significantly different between treatment groups (P = 0.0004).

The frequency distribution of changes in HbA_1c after 1 year of treatment were consistent with the weight loss data; more patients treated with orlistat had a mean reduction in HbA_1c of ≥0.5 or ≥1.0% compared with placebo (61.3 vs. 43.3%, P = 0.003 and 46.0 vs. 29.0%, P = 0.008, respectively).

After 1 year, mean reduction in HbA_1c was 0.75 ± 0.08% (P = 0.0001 vs. baseline) in the orlistat group compared with 0.41 ± 0.08% (P = 0.012 vs. baseline) in the placebo group (data not shown). Although the decrease in HbA_1c in patients treated with orlistat was not statistically significantly different from the decrease observed in the placebo group, the comparison was confounded by changes in diabetes medications. The greater reductions in diabetes medication in the orlistat group would be expected to dampen the mean reductions in HbA_1c, thereby underestimating the true effect of orlistat on glycemic control. Conversely, increasing medication doses in the placebo group could exaggerate the treatment effect on HbA_1c in those patients. Two approaches were taken to control for the potential confounding effect of differences in diabetes treatment: 1) HbA_1c values that were obtained after changes in diabetes medications occurred were excluded from the analysis, carrying the last observation before such changes forward; and 2) an ANCOVA model, including fixed effects for treatment and change in diabetes medication, was applied to the change in HbA_1c from baseline to the end of treatment. The results of these two analyses were consistent: in the first analysis, the mean reduction in HbA_1c was significantly greater in the orlistat group than in the placebo group (0.73 ± 0.08 vs. 0.36 ± 0.09%, P = 0.0024); in the latter analysis, the change in HbA_1c, adjusted for changes in diabetes medication, was also significantly different between the two treatment groups (0.90 ± 0.08 vs. 0.61 ± 0.08%, P = 0.014) (not shown).

After 1 year, fasting serum glucose decreased significantly more in the orlistat-treated patients compared with the placebo group (−2.0 ± 0.2 vs. −0.7 ± 0.2 mmol/l, P = 0.001). This significant difference between the study groups was apparent within 2 weeks of starting treatment and was then maintained for the duration of the study (Fig. 1B).

Cardiovascular risk factors are shown in Table 2. Treatment with orlistat was associated with significantly greater improvements in total and LDL cholesterol and LDL/HDL ratio compared with placebo. Changes in HDL cholesterol and triglyceride levels did not differ significantly between the two treatment groups. At 1 year, the decline in systolic blood pressure was greater in orlistat- than in placebo-treated patients. The change in diastolic blood pressure was not significantly different between the orlistat and placebo groups.

Overall, more patients in the placebo group withdrew prematurely from the study than in the orlistat group (44 vs. 35%, P < 0.05). Of these patients, more orlistat- than placebo-treated patients (10 vs. 5%, P < 0.05) discontinued treatment because of an adverse event. Adverse event profiles were similar in the two treatment groups, with the exception of a higher incidence of gastrointestinal effects associated with orlistat therapy. A total of 83% of patients in the orlistat group experienced at least one gastrointestinal event compared with 62% of patients in the placebo group. Common gastrointestinal events included oily or liquid stools, oily spotting, and fecal urgency. More than 75% of patients who had gastrointestinal events reported only a single episode. The majority of gastrointestinal events were mild, transient, and occurred during the early phase of treatment.

Hypoglycemic episodes occurred in 10% of orlistat-treated patients compared with 4% of placebo-treated patients. No patient required medical intervention for hypoglycemia, with symptoms reported as mild to moderate in intensity in both groups.

Weight loss is an important element in the management of overweight and obese patients with type 2 diabetes because of the beneficial effects of weight loss on insulin sensitivity and glycemic control. Unfortunately, weight loss is particularly difficult to achieve in patients with type 2 diabetes using conventional diet therapy, and pharmacological treatment of type 2 diabetes is usually associated with weight gain (5,8, 9). Metformin is considered by many to be the drug of choice in the treatment of overweight and obese type 2 diabetic patients and has been shown to decrease food consumption in diabetic patients (19,20). The favorable effects of this agent on glycemic control may therefore be due in part to reduced energy intake. Orlistat has been shown to produce weight loss and improvement in glycemic control in patients with type 2 diabetes treated with sulfonylureas (14). The present study is the first randomized-controlled trial to evaluate the efficacy of orlistat therapy in conjunction with a reduced-calorie diet in overweight and obese metformin-treated type 2 diabetic patients. The results of the study indicate that, compared with placebo, orlistat therapy enhances weight loss, improves glycemic control, and improves other important cardiovascular risk factors (LDL cholesterol and blood pressure).

In our subjects, there was improvement in a number of metabolic and clinical variables related to diabetes management: 1) a greater reduction in fasting serum glucose; 2) a greater reduction in HbA_1c when changes in diabetes medication were taken into account; 3) a greater proportion of patients achieving decreases in HbA_1c of ≥0.5 and ≥1.0%; and 4) a greater percentage of patients who decreased or discontinued at least one diabetes medication and a lower percentage of patients who required an increase in diabetes medication dose or an additional agent. The benefit of orlistat was observed within 2 weeks of initiation of treatment, before there was a significant difference in weight loss between study groups. This finding is consistent with previous studies demonstrating that acute negative energy balance itself lowers fasting blood glucose and improves insulin-mediated glucose uptake in patients with type 2 diabetes (21,22). Therefore, our results suggest that greater negative energy balance occurs with orlistat than with placebo in the early phase of treatment, before there are detectable differences in weight loss.

The improvement in glycemic control associated with orlistat therapy in this study (HbA_1c lowering of 0.3–0.4% compared with placebo) is of a smaller magnitude than the reported improvements in HbA_1c and fasting serum glucose produced by sulfonylureas or metformin, but similar to the improvements observed in some studies of the effects of thiazolidinediones and α-glucosidase inhibitors when these agents were used as monotherapy or in combination with insulin or sulfonylureas (23–26). The overall lowering of HbA_1c in the orlistat patients in the present study (0.75%) was significantly greater than that observed in the placebo group when changes in diabetes medications were taken into account. The UKPDS has demonstrated that relatively small reductions (<1%) in HbA_1c were associated with reduced microvascular complications (13). Moreover, no threshold of risk was observed for any end point in the UKPDS study (27). Oral monotherapy of type 2 diabetes fails to produce satisfactory glycemic control in at least 75% of patients with type 2 diabetes using <8.0% as a target and is presumably even less successful in achieving HbA_1c values < 7.0% (28). Thus, the vast majority of patients will require combination therapy to achieve optimal glycemic control. In fact, it is commonplace for three drugs to be required for satisfactory glycemic control (29).

Orlistat therapy also had beneficial effects on other risk factors for cardiovascular disease. Compared with placebo, orlistat therapy was associated with a 6–7% improvement in serum total and LDL cholesterol concentrations. This is consistent with the results of several previous studies in both diabetic and nondiabetic patients and is presumably caused by both weight loss and orlistat’s independent effect on cholesterol absorption (14,16,30). The lowering of LDL cholesterol is of clinical importance because serum LDL cholesterol concentration is an important predictor of cardiovascular events in patients with diabetes (13) and is particularly noteworthy in view of a substantial (13%) increase in HDL cholesterol. Orlistat therapy was also associated with a greater decrease in systolic blood pressure than placebo therapy. Therefore, orlistat improves glycemic control and other risk factors for coronary heart disease, the major cause of mortality in patients with diabetes.

Metformin is associated with gastrointestinal events that tend to subside with time (31). Orlistat therapy, due to its effect on dietary fat absorption, is also associated with transient gastrointestinal events. Because of this, we were concerned that the additive effects of the two agents could cause excessive gastrointestinal symptoms and thereby lead to premature study withdrawal. However, although gastrointestinal events were more frequent in the metformin plus orlistat than in the metformin plus placebo groups, combined therapy did not lead to a greater drop-out rate. In fact, the overall rate of withdrawal was actually higher in the placebo group than in the orlistat group (43 vs. 35%, P < 0.05), a finding consistent with earlier reports (16). This indicates that gastrointestinal events are generally not a cause for discontinuation of orlistat therapy. In fact, the lower withdrawal rate in the orlistat group suggests that patients receiving orlistat perceived a benefit that was not apparent to patients taking placebo. Whether this benefit was related to greater weight loss or to better glycemic control is not clear. The overall difference in the incidence of gastrointestinal side effects between the orlistat and placebo groups in the present study was similar to that previously observed between orlistat and placebo groups in type 2 diabetic patients receiving sulfonylurea monotherapy (14). Therefore, a decision to use orlistat in the management of overweight and obese patients with type 2 diabetes should not be based on the type of diabetes medications the patients are receiving.

Modest weight loss is a reasonable goal for many patients with type 2 diabetes. However, diabetic subjects have greater difficulty losing weight than nondiabetic individuals. The weight loss achieved in the current study was associated with significant improvements in glycemic control and other cardiovascular risk factors. Although lifestyle changes to decrease energy intake and increase physical activity should remain the cornerstone of weight loss therapy, data from the present study suggest that antiobesity medications can enhance weight loss and provide benefits in the management of patients with type 2 diabetes.

This study was supported by a grant from F. Hoffmann-La Roche and has previously been published in abstract form [Miles JM, Aronne LJ, Hollander P, Klein S: Effect of overweight and obese type 2 diabetes patients treated with metformin. Diabetes 50 (Suppl. 2):A442–A443, 2001].