Pharmacotherapy for weight loss in adults with type 2 diabetes mellitus

Susan L Norris; Xuanping Zhang; Alison Avenell; Edward Gregg; Christopher H Schmid; Joseph Lau

doi:10.1002/14651858.CD004096.pub2

. 2005 Jan 24;2005(1):CD004096. doi: 10.1002/14651858.CD004096.pub2

Pharmacotherapy for weight loss in adults with type 2 diabetes mellitus

Susan L Norris ^1,^✉, Xuanping Zhang ², Alison Avenell ³, Edward Gregg ², Christopher H Schmid ⁴, Joseph Lau ⁵

Editor: Cochrane Metabolic and Endocrine Disorders Group

PMCID: PMC6718205 PMID: 15674929

Abstract

Background

Obesity is closely related to type 2 diabetes and long‐term weight reduction is an important part of the care delivered to obese persons with diabetes.

Objectives

To assess the efficacy of pharmacotherapy for weight loss in adults with type 2 diabetes.

Search methods

Computerized searches were performed of MEDLINE, EMBASE, Web of Science and other electronic bibliographic databases, supplemented with hand searches of reference lists and selected journals.

Selection criteria

Randomized, controlled trials were included where pharmacotherapy was used as the primary strategy for weight loss among adults with type 2 diabetes. Published and unpublished literature in any language and with any study design was included.

Data collection and analysis

Two reviewers abstracted data and the quality of included studies was evaluated by assessing potential attrition, as well as selection and measurement bias, and a Jadad score was obtained. Effects were combined using a random effects model.

Main results

A sufficient number of studies were available for a quantitative synthesis for fluoxetine, orlistat, and sibutramine. Twenty two randomized controlled trials were included in the review, with a total of 296 participants for fluoxitine, 2036 for orlistat, and 1047 for sibutramine. Pharmacotherapy produced modest reductions in weight for fluoxetine (5.1 kg (95% confidence interval [CI], 3.3 ‐ 6.9) at 24 to 26 weeks follow up; orlistat 2.0 kg (CI, 1.3 ‐ 2.8) at 12 to 57 weeks follow‐up, and sibutramine 5.1 kg (CI, 3.2 ‐ 7.0) at 12 to 52 weeks follow‐up. Glycated hemoglobin also modestly and significantly reduced for fluoxetine and orlistat. Gastrointestinal side effects were common with orlistat; tremor, somnolence and sweating with fluoxetine; and palpitations with sibutramine. Some studies, using a variety of study designs, were available on other drugs and a significant decrease in weight was noted in three studies of mazindol, one of phenmetrazine, two of phentermine. No studies were identified that fit inclusion criteria for pseudoephedrine, ephedra, sertraline, yohimbine, amphetamine or its derivatives, bupropion, topiramate, benzocaine, threachlorocitric acid, sertraline, and bromocriptine.

Authors' conclusions

Fluoxetine, orlistat, and sibutramine can achieve statistically significant weight loss over 12 to 57 weeks. The magnitude of weight loss is modest, however, and the long‐term health benefits remain unclear. The safety of sibutramine is uncertain. There is a paucity of data on other drugs for weight loss or control in persons with type 2 diabetes.

Plain language summary

Pharmacotherapy for weight loss in adults with type 2 diabetes mellitus

Obesity is closely related to type 2 diabetes and weight reduction is an important part of the care delivered to obese persons with diabetes. This review of drugs for weight loss among adults with type 2 diabetes revealed weight loss of between 2.0 and 5.1 kg for fluoxetine, orlistat and sibutramine at follow‐up of up to 57 weeks. The long‐term effects remain uncertain. Adverse events were common in all three drugs: gastrointestinal side effects with orlistat; tremor, somnolence, and sweating with fluoxetine; and palpitations with sibutramine. There were few studies examining other drugs used for weight loss in populations with diabetes.

Background

Description of the condition

Diabetes mellitus is a metabolic disorder resulting from a defect in insulin secretion, insulin action, or both. A consequence of this defect is chronic hyperglycaemia (i.e., elevated levels of plasma glucose) with disturbances of carbohydrate, fat, and protein metabolism. Long‐term complications of diabetes mellitus include retinopathy, nephropathy, and neuropathy, and the risk of cardiovascular disease increases over time. For a detailed overview of diabetes mellitus, please see under 'Additional information' in the information on the Metabolic and Endocrine Disorders Group on the Cochrane Library (see 'About the Cochrane Collaboration', 'Collaborative Review Groups‐CRGs'). For an explanation of methodological terms, see the main Glossary on The Cochrane Library.

The prevalence of both obesity and diabetes continues to increase. Obesity rates have risen threefold since 1980 in North America, parts of Europe, the Middle East, the Pacific Islands, Australasia, and China. More than one billion adults worldwide are overweight (body mass index (BMI (kg/m²) ≥ 25); at least 300 million of them are obese (BMI ≥ 30) (WHO 2002; WHO 2003). In the developed world, recent survey data (Flegal 2002) indicate that 65% of American adults are overweight and 31% obese (BM I≥ 30) (NHLBI 1998). The prevalence of diabetes is also rising, with worldwide prevalence estimated at 4.0% in 1995, but expected to rise to 5.4% by 2025 (King 1998). An estimated 135 million adults had diabetes in 1995, a number expected to rise to 300 million by 2025; this represents increases of 42% in developed countries and 170% in developing countries (King 1998). In the United States, the prevalence of diabetes increased 49% from 1990 to 2000 (U.S. DHHS 2002b). Of U.S. adults over age 20, 8.6% have diabetes, of whom one‐third are undiagnosed (U.S. DHHS 2002).

Both obesity and weight gain are major risk factors for diabetes (Maggio 1997; Pi‐Sunyer 1993) and every 1‐kg increase in weight (self‐reported) is associated with a 9% relative increase in the prevalence of diabetes (Mokdad 2000). Eighty to ninety percent of persons with type 2 diabetes are overweight (Wing 2000) and obesity worsens the metabolic and physiologic abnormalities associated with diabetes, particularly hyperglycemia, hyperlipidemia, and hypertension (Maggio 1997).

Description of the intervention

Weight loss is one cornerstone of diabetes care for overweight persons, as it improves insulin sensitivity and glycemic control (Pi‐Sunyer 2000), and moderate, intentional weight loss is associated with reduced mortality (Williamson 2000). Among persons with diabetes, weight loss improves lipid profiles by decreasing triglycerides and low‐density lipoprotein (LDL) cholesterol levels, and weight loss improves blood pressure (Maggio 1997), mental health, and quality of life (Wing 1987; Wing 1991). These benefits are clinically meaningful only if weight loss is sustained over time, however (Wing 1985). The findings of a reduced incidence of hypertension and diabetes in populations with impaired glucose tolerance or obesity that maintained weight loss over extended periods provide indirect evidence of this benefit (DPP 2002; HT Trials 1997; Tuomilehto 2001).

Dietary and behavioral treatment for weight loss can produce an average loss of 8% of initial body weight over 3 to 12 months (NHLBI 1998), but it is difficult to define effective weight control measures for the long term in general populations (NHLBI 1998; O'Meara 1998). The majority of obese patients regain most of the weight initially lost in successful interventions (Maggio 1997; Wing 1985; Wadden 1989).   Studies suggest that persons with diabetes lose less weight than persons without diabetes and regain their weight more rapidly, although the mechanisms responsible are unclear and the validity of this observation has not been systematically examined (Wing 2000). Obese or overweight persons with diabetes may face additional barriers than non‐diabetic persons trying to achieve weight loss. The use of insulin to achieve glycemic control may produce weight gain. The complex treatment regimens for diabetes, hypertension, and hyperlipidemia all complicate behavioral change aimed at weight reduction. In addition, Wing has noted that obese persons with diabetes who present for treatment are older and sicker than persons without diabetes (Wing 1985).

Obesity may be viewed as a chronic disease (NIH 1985); Greenway (Greenway 1999) suggests that obesity should therefore be treated as such and that optimal management may require long‐term pharmacotherapy. In patients who have failed behavioral therapy, adjunct treatment with drugs may help them reduce or maintain their weight while improving other parameters of health, including glycemic control and lipid profiles. Numerous anti‐obesity agents have been used for weight loss in general populations as well as in persons with diabetes (Yanovski 2002). These drugs act through a variety of mechanisms, including centrally acting appetite suppression (e.g., sibutramine and phentermine), increased energy expenditure (e.g., ephedrine and caffeine), and nutrient partitioning via decreased food absorption from the gastrointestinal tract (e.g., orlistat). Anti‐obesity drugs may be available over‐the‐counter or by prescription. Some drugs with other specific clinical indications are associated with weight loss (e.g., metformin), and many drugs are used for weight loss although they are not approved for that indication (i.e., off‐label usage, e.g., fluoxetine).

Because obese and overweight adults with type 2 diabetes benefit from weight loss but may have more difficulty losing weight than persons without diabetes, we need to define the scope of our knowledge about the efficacy of pharmacologic interventions for losing weight or preventing weight gain in these populations. We must determine which, if any, drugs are effective in obese and overweight persons, particularly in the long term, and we must define the nature and incidence of side effects. In addition, we must define areas of uncertainty where further research is needed.

Why it is important to do this review

We have identified four recent reviews of anti‐obesity pharmacotherapy for type 2 diabetes. Scheen and Lefebvre (Scheen 2000) and Scheen and Ernest (Scheen 2002) discussed the effects of anti‐obesity drugs on weight loss, glycemic control, and cardiovascular risk profile for obese persons with type 2 diabetes. Greenway 1999 reviewed the use of a broad range of anti‐obesity agents among persons with diabetes. Hauner 1999 discussed both the impact of antidiabetic agents on weight and the effect of weight management drugs on glycemic control in obese diabetic patients.

None of these articles was a systematic review, involved quantitative syntheses, or assessed the quality of individual studies. In addition, none examined a broad range of outcomes, such as morbidity, mortality, and quality of life. Thus, to date we have not located any quality systematic reviews on the efficacy of drugs for weight loss or weight maintenance in overweight and obese adults with type 2 diabetes.

Objectives

To assess the efficacy of pharmacotherapy for weight loss and the maintenance of weight loss in adults with type 2 diabetes.

Primary research question

What drugs are effective in achieving or maintaining weight loss in overweight and obese adults with type 2 diabetes?

Secondary research questions

What additional interventions are delivered with drug therapy and how do they affect outcomes?
What side effects/complications of the drugs are reported?
How does the follow‐up interval relate to outcomes?
What are the effects of the weight loss interventions on glycemic control, blood pressure, and lipid profiles?

Methods

Criteria for considering studies for this review

Types of studies

Randomized controlled trials only were included in the review of efficacy as these minimize the potential effects of bias on our results. All study designs, however, were included in the review of adverse events: those with a contemporaneous comparison group (randomized controlled trial, non‐randomized trial, or observational study with a concurrent comparison group), or a pre‐versus‐post design, a cross‐sectional design, case‐control studies, or a case series. We recognize the potential for bias from confounding and secular trends in studies without randomization, but because observational studies yield important information on adverse events related to treatment, particularly on rare, long‐term side effects (Elphick 2002), we searched for, and synthesized in narrative form, the available observational data on side effects. In our protocol we indicated that we would include additional comparative study designs if we had found an insufficient number of randomized, controlled trials. Since we identified a sufficient number of randomized trials, we only included this study design in the review for efficacy.

Length of follow‐up and timing of outcomes measurement

We included studies of any duration and length of follow‐up. We defined follow‐up from the time of randomization (or for studies without randomization, from the time of entrance into the study) until the last outcomes measurement. We recognize that long‐term outcomes are of paramount importance, but examination of the efficacy of pharmacotherapy in the short term also has value. For example, if weight loss can be demonstrated with drugs in the short term, pharmacotherapy may be combined with behavioral interventions for long‐term weight control. In addition, an exploration of the relationships between the population and intervention characteristics and the efficacy of these drugs in the short‐term may provide insights into how to achieve longer‐term success.

Full text and abstracts

Both full‐text publications and abstracts are included in this review. Because it is more difficult to assess the quality of abstracts, data from these publications was analyzed both separately and combined with full‐text publications.

Publication status

We examined published data only, as we had no success in obtaining unpublished data from private or public sector sources.

Types of participants

Age

Studies included adults aged 18 years or older with type 2 diabetes. If the type of diabetes was not specified, studies were included if they involved adults with diabetes, with or without insulin treatment. Persons labelled with "NIDDM" were assumed to have type 2 diabetes. Studies involving only "IDDM" participants were excluded unless there was information to indicate that they have type 2 disease (e.g., concurrent use of oral hypoglycemic agents and insulin). Studies that include participants without diabetes were also included if there were outcome data on the subpopulation with diabetes.

Type 2 diabetes

The acceptable diagnostic criteria for diabetes includes those described by the standards of the National Diabetes Data Group (Data Group 1979), the World Health Organization (WHO Committee 1980; WHO Committee 1985; Alberti 1998), or the American Diabetes Association (Expert Committee). If the diagnostic criteria were not given in the study, the authors' statement of the diagnosis of diabetes among participants was accepted.

Overweight or obese

Participants were overweight as defined in the study; there was no minimum weight or BMI at baseline.

Types of interventions

Any drug therapy delivered for the primary purpose of losing and/or controlling weight was included. Studies that combined pharmacotherapy with other weight loss strategies, including behavioral, educational, lifestyle (diet and exercise), or surgical interventions, were included. Both prescription and over‐the‐counter medications were included. Drugs that were not approved for weight loss, but which were used for the primary purpose of weight loss were included (i.e., off‐label usage of the drug, e.g., fluoxetine).

The drugs examined included:

Centrally acting appetite suppressants:

Amphetamine/dextroamphetamine;
Bupropion;
Diethylpropion;
Fluoxetine;
Mazindol;
Methamphetamine/benzphetamine;
Phenmetrazine/Phendimetrazine;
Phentermine;
Sibutramine;
Topiramate;
Yohimbine.

Peripheral effect on appetite:

Benzocaine.

Nutrient partitioning:

Orlistat/tetrahydrolipstatin;
Treacholorocitric acid.

Increase thermogenesis:

Ephedra alkaloids;
Caffeine.

Combined drug therapy:

Ephedrine and caffeine.

Comparison groups:

Studies that involved a comparison group with a different intervention were included regardless of the nature of the comparison intervention. We included studies with a range of comparison groups as we wanted to determine which interventions were more effective than others.

The comparison group could receive:

placebo;
no intervention;
usual care;

Any other weight loss intervention: behavioral strategy, dietary program, physical activity program, surgery, other.

Types of outcome measures

Primary outcomes

Weight or BMI must be measured at both baseline and follow‐up in order for the study to be included in this review.

weight and body fat distribution: weight (kg), BMI (kg/m²);
drug‐related morbidity: severe (necessitating withdrawal) or minor;
quality of life.

Secondary outcomes

glycemic control: glycated hemoglobin, fasting blood sugar;
serum lipids;
blood pressure;
non drug‐related morbidity;
mortality.

Exclusion criteria

study populations with binge eating or other eating disorders were excluded from this review.
drugs withdrawn from market in the U.S. were excluded, including fenfluramine, dexfenfluramine, and phenylpropanolamine.
investigational drugs, defined as those drugs not yet approved for use in the U.S., were excluded (e.g., leptin, beta‐2 agonists such as BRL‐26830A).
herbal supplements, including ginseng and a number of other herbal supplements that are not regulated by the United States Food and Drug Administration, were excluded.
drugs that may produce weight loss but whose primary purpose is another clinical indication were excluded. These include metformin, acarbose, and benfluorex, all of which may produce weight loss but are used primarily for glycemic control. We recognize that the clinical indications for these drugs may change and that in the future they may be regarded as drugs whose primary purpose includes weight loss.

Search methods for identification of studies

Electronic searches

A number of electronic databases were screened for potentially relevant titles and abstracts. There were no language restrictions on our searches. Conference proceedings and abstracts were included in the review but not in the primary pooled analysis, because they had insufficient detail to evaluate the intervention and the quality of the study. These are summarized in narrative form and presented as potentially important studies that may appear in future in the literature. Dissertations were excluded, as they were difficult to locate in full text.

The following electronic databases were searched between the date in parentheses and June 30, 2004.

The Cochrane Library (Issue 3, 2003), including Cochrane Controlled Trials Register, DARE, CRG specialized registers;
MEDLINE (1966) (includes Healthstar);
EMBASE (1974);
CINAHL (1982);
Web of Science (1981);
Biosis (1980);
International Pharmaceutical Abstracts (1970).

For the detailed MEDLINE search strategy see under Appendix 1. The search strategy was improved with minor modifications, from the protocol.

Other searches are available upon request.

Searching other resources

The following journals, believed to be of high topic relevance, were hand searched from 1980 to February 2003: Diabetes Care; International Journal of Obesity and Related Metabolic Disorders (prior to 1992 this journal was the International Journal of Obesity);Obesity Research (journal commenced in 1993);American Journal of Clinical Nutrition;Journal of the American Dietetic Association.

Potential missing and unpublished studies were sought by contacting experts in the field and authors of relevant identified studies as well as drug manufacturers. The reference lists of all relevant review articles and of the studies included in the review were reviewed. The National Heart, Lung, and Blood Institute 1998 review (NHLBI 1998) and a review by the University of York, National Health Centre for Reviews and Dissemination (York CRD 1997) were examined for relevant citations.

Data collection and analysis

Selection of studies

Search results for MEDLINE and CINAHL were examined by two authors (SLN and XZ) and the remaining databases by one author (SLN). Potentially relevant full‐text articles were then reviewed for inclusion (SLN); if there was uncertainty about inclusion, a second author (AA) reviewed the paper and consensus was achieved. Due to resource constraints and the need for efficiency, only SLN reviewed the full text for potential inclusion (this is a change from the protocol). AA provided secondary review and consensus was achieved for studies when SLN had any uncertainty. After consensus was reached between AA and SN, XZ screened all included papers to confirm inclusion.

Data extraction and management

For studies that fulfilled inclusion criteria, two reviewers abstracted the relevant data using a standardized template. Extraction was not blinded, as there is no evidence that such blinding decreases bias in conducting systematic reviews and meta‐analyses (Berlin 1997; Irwig 1994). We attempted to contact study authors for missing data or when we needed clarification of the data presented.     For continuous outcomes we extracted for each study group the baseline sample size, pre‐ and post‐intervention mean, and a measure of dispersion (SD [standard deviation], standard error of the mean (SEM), or 95% confidence interval) for the intervention and comparison groups. If the post‐intervention measures of dispersion were not available, they were assumed to be the same as the pre‐intervention measures. When necessary, mean and SD were approximated from figures using an image scanner to optimize resolution. For dichotomous variables (e.g., mortality) the number of participants, person‐years, and the number of events were extracted for each study group.

Assessment of risk of bias in included studies

Internal validity

Internal validity was assessed by two reviewers for each study. For randomized controlled trials (RCTs), the component assessment method of Cochrane was used (Clarke 2003) as well as the Jadad score (Jaded 1998). For the former method, the following quality criteria were assessed as "met" or "unmet":   1. Minimisation of selection bias: a) Was the randomization procedure adequate? b) Was the allocation concealment adequate?   2. Minimisation of performance bias: Were the participants and those administering the treatment blind to the intervention?   3. Minimisation of attrition bias: a) Were withdrawals and dropouts completely described? b) Was analysis by intention‐to‐treat?   4. Minimisation of detection bias: Were outcome assessors blind to the intervention?

The risk of bias was assessed as low (A) (all criteria were met), moderate (B) (one or more criteria were only partly met), or high (C) (one or more criteria were not met).

For studies that were not RCTs, the comparability of groups at baseline and attrition were noted. Studies were not excluded because of poor quality; where data were sufficient the effect of potentially biasing factors on outcomes was examined.

Other design issues

In addition to the above‐mentioned components of the assessment of internal validity, we noted whether the study used an intention‐to‐treat analysis and whether the last‐outcome measurement was carried forward (LOCF) to subsequent follow‐up measurements.

Assessment of heterogeneity

Data were pooled using the random effects model and using the DerSimonian and Laird formula for calculating between‐study variance (DerSimonian 1954). Each study was weighted by the inverse of the study variance. Heterogeneity between trial results was tested using a standard chi‐square test (Cochran 1954) with a significance level of alpha = 0.1 in view of the low power of such tests. When we found heterogeneity we tried to explain it by examining individual study characteristics and those of subgroups of the main body of evidence. When heterogeneity was thought to be too great to meaningfully pool the results quantitatively, the results are presented in a narrative fashion.

Assessment of reporting biases

Funnel plots were used in exploratory data analysis to assess for the potential existence of small sample bias. An asymmetrical funnel plot, however, has several explanations, including true heterogeneity of effect with respect to study size, poor methodological design of small studies (Sterne 2001; Tang 2000; Thornton 2000), and publication bias. Thus, we did not place undue emphasis on this tool.

Data synthesis

Statistical pooling

Where data from RCTs were thought to be sufficiently homogeneous with respect to interventions and outcomes, we calculated pooled effect sizes. For continuous variables reported on the same scale we calculated weighted mean differences. The absolute differences in outcome between each follow‐up and the baseline measure for the intervention and comparison study group (ΔI and ΔC) were calculated and inserted in Review Manager Software (Review Manager 4.2). When the estimate of variance of (ΔI) and (ΔC) was not given, it was calculated from the outcome measures in each study group using the formula Vpre+ Vpost ‐ 2r(SDpre*SDpost), where Vpre is the variance of the mean baseline outcome, Vpost is the variance of the mean follow‐up outcome, r is the correlation between the baseline and follow‐up values, and SDpre and SDpost are the standard deviations of the baseline and follow‐up groups, respectively. Since most studies do not report r, and its true value is unknown, data are presented with r = 0.75, and a sensitivity analysis was performed as described below.

Regression analyses

We performed a meta‐regression to determine whether various study‐level characteristics affect weight change and GHb. The meta‐regression was also weighted by the inverse of the variance of (ΔI ‐ ΔC). Interaction terms were examined for all models. The study‐level variables examined in the meta‐regression model included follow‐up interval, the number of contacts between the care provider and participants, and the percentage attrition in the intervention group. SAS was used to perform the meta‐regression (version 8.01, SAS Institute Inc., Cary, NC).

Subgroup analysis and investigation of heterogeneity

We planned analyses by the following subgroups if there was a significant change in weight and the amount of data would allow meaningful analyses:

overweight (25.0 ≤ BMI ≤ 30.0), obese (BMI >30.0), normal weight (BMI <25.0);
age: young (<40 years), middle‐aged (40 to 65 years), old (>65 years);
treatment: on insulin, oral agents, diet only;
sex;
race / ethnicity;
time frame over which the intervention was delivered.

Sensitivity analysis

Analyses were planned to examine the effect of internal validity on study results, using the categories of low, moderate, and high risk of bias. Sensitivity analyses were also used to compare the fixed and random effects model and using different values of the correlation between pre‐ and post‐ measures (0.25, 0.5, and 1.0).

Results

Description of studies

Figure 1 presents the review flow diagram. No studies were identified that fit inclusion criteria for amphetamine and its derivatives, benzocaine, bromocriptine, bupropion, ephedrine, ephedra, pseudoephedrine, sertraline, threachlorocitric acid, topiramate, and yohimbine. Data were sufficient to perform a quantitative analysis of fluoxetine, orlistat, and sibutramine, and thus this review focuses initially on these three drugs. We then provide narrative summaries of the results for cimetidine, diethylpropion, mazindol, phenmetrazine, and phentermine.

Study flow diagram   CRCT, Cochrane Controlled Trials Register   IPA, International Pharmaceutical Abstracts   WOS, Web of Science

Fluoxetine, orlistat, and sibutramine

Characteristics of the 22 eligible RCTs examining fluoxetine, orlistat, and sibutramine are shown in Appendix 3, Appendix 4, Appendix 7, Appendix 9 and Appendix 12. These studies included 296 participants who received fluoxetine, 2036 orlistat, and 1047, sibutramine. Follow‐up intervals ranged from 8 to 52 weeks for fluoxetine, 12 to 57 weeks for orlistat, and 12 to 52 weeks for sibutramine. Most studies used a run‐in period lasting 1‐5 weeks, where a placebo was given and dietary counselling started. Generally the duration of drug treatment was the same as the follow‐up interval, although in three studies weight change was recorded from the beginning of the run‐in period (Hollander 1998; Hanefeld 2002; Lindgarde 2000). Only one study (Gray 1992) examined weight maintenance after discontinuation of the study drug. Study participants' mean age was between 48 and 66 years across studies and somewhat more than half were female. There were insufficient data to draw conclusions from the funnel plots.

Mean weight of the control group at baseline was 95 kg (SD 18.5 kg) for fluoxetine, 95.9 kg (11.1 kg) for orlistat, and 97 kg (17.3 kg) for sibutramine. BMI was presented in only 14 of the studies (range 31 to 37). Participants generally had poor glycemic control by current treatment standards (ADA 2003). Most studies excluded patients who were taking insulin, although two studies examined insulin‐using subjects exclusively (Gray 1992; Kelley 2002). Drug dosages were very consistent among studies, except for one study of sibutramine that used a twice‐daily dosage regime (Gokcel 2001). All studies examined continuous therapy. All except one study of fluoxetine (O'Kane 1994) involved a dietary intervention for both the treatment and control groups, and the comparison groups all received a placebo. Average contacts ranged from 2 to 18, an average of 1.1 per month. Attrition during the run‐in period ranged from 1.5% to 22% in the studies where it was reported (Hollander 1998; Lindgarde 2000; Gray 1992; Finer 2000; Hanefeld 2002) In three studies (Hanefeld 2002; Hollander 1998; Daubresse 1996; Daubresse 1996) participants were randomized only if they had high rates of compliance for visits or pill consumption during the run‐in period.

Risk of bias in included studies

Fluoxetine, orlistat and sibutramine

The sampling frame and subject recruitment methods were rarely described. Only two studies described the randomization process (Zelissen 1992; Redmon 2003) and one discussed allocation concealment (Redmon 2003). In 18 of the 22 trials the drug's manufacturer supported the study. Attrition varied considerably; for the intervention group it ranged from 0% to 49%; for the control group from 0% to 52%. In seven of 20 studies where attrition rates were reported, the control group had a higher rate than did the intervention group, including four of seven studies of orlistat (Hollander 1998, Kelley 2002; Miles 2002). Most studies were described as double‐blinded (16 of 22), but none reported exactly which two parties were blinded. One study was open label (Tankova 2003). In nine studies LOCF measures were used in the event of attrition (Kelley 2002; Kutnowski 1992; Miles 2002; Fujioka 2000; Serrano‐Rios 2001; Hanefeld 2002; Redmon 2003; Kaukua 2004; Bloch 2003). Most studies reported using intention‐to‐treat methods of analysis, but several excluded participants for protocol violation (Hanefeld 2002; Fujioka 2000; Kutnowski 1992; Miles 2002; Kelley 2002; Wang 2003), noncompliance (Hanefeld 2002; Miles 2002; Hollander 1998; Lindgarde 2000), or treatment failure (Miles 2002). A study examining sibutramine (Halpern 2003) fit our inclusion criteria, but because of numerous inconsistencies noted in the presentation of data in the paper, this study was not included in the meta‐analysis.

Effects of interventions

Fluoxetine, orlistat, and sibutramine

Change in weight (kg) and GHb (%) for full‐text studies of fluoxetine, orlistat, and sibutramine are shown in Figure 2 and Figure 3,and the meta‐analysis results are presented in Appendix 15, Appendix 17, and Appendix 20. A summary of pooled effects for fluoxetine is found in Appendix 27, and for orlistat and sibutramine in Appendix 28. Weight loss ranged from 10.5 kg for sibutramine at 26 weeks follow‐up (95% CI, 7.6 ‐ 13.4) (Gokcel 2001) to 1.4 kg for fluoxitine at 8 weeks of follow‐up (95% CI, 0.2 to 2.6) (Kutnowski 1992). The pooled effects were the following: orlistat over all follow‐up periods demonstrated a loss of 2.0 kg (95% CI, 1.3 to 2.8); sibutramine over all follow‐up periods produced a loss of 5.1 kg (95% CI, 3.2 to 7.0); loss for fluoxitine at 8 to 16 weeks was 3.4 kg (95% CI, 1.7 to 5.2), 24 to 26 weeks was 5.1 kg (95% CI, 3.3 to 6.9), and one study examining fluoxetine at 52 weeks produced a loss of 5.8 kg (95% CI, 0.8 to 10.8).

Net change in weight (kg)   Pooled estimates are represented by boxes.

Net change in hemoglobin A1c (Hb A1c) (%)   Pooled estimates are represented by boxes.

Reduction of GHb ranged from 2.2% for sibutramine at 26 weeks follow‐up (Gokcel 2001) to 0.1% for three studies of sibutramine at follow‐up intervals of 24 and 52 weeks (Fujioka 2000; McNulty 2003; Serrano‐Rios 2002). Gokcel and colleagues used a dose of sibutramine of 10 mg twice daily, unlike other studies which used 15 or 20 mg in a single daily dose. One study published only as an abstract utilized up to 30 mg as a single daily dosage (Vargas 1994). The pooled reduction for GHb was 0.5% (95% CI, 0.3 to 0.6) for orlistat (follow‐up between 24 and 57 weeks); sibutramine 0.5% (95% CI, ‐0.2 to 1.3) (follow‐up 12 to 52 weeks); fluoxetine 1.0% (95% CI, 0.4 to 1.5) at 8 to 16 weeks, 1.0% (95% CI, 0.6 to 1.4) at 24‐26 weeks, and one study with a follow‐up of 52 weeks demonstrated a reduction of 1.8% (95% CI, ‐0.2 to 3.8) (O'Kane 1994).

Several studies examined more than one follow‐up interval (O'Kane 1994; Gray 1992; Connolly 1995); these results are shown in Figure 2 and Figure 3. Fluoxetine had sufficient studies to allow stratification by treatment duration. Weight loss was slightly greater with longer follow‐up, but differences were small and only one study examined fluoxetine for longer than 30 weeks of treatment (O'Kane 1994).     We identified 22 studies published only as abstracts for fluoxitine (four total, all RCTs), orlistat (14 total, 9 RCTs), and sibutramine (four total, three RCTs) that fulfilled our inclusion criteria: six for fluoxetine, 20 for orlistat, and seven for sibutramine. Pooled effects obtained by combining abstracts of RCTs and full‐text RCTs did not produce significant changes in the direction or significance of results of the full‐text studies only.

We performed a sensitivity analysis making two different assumptions about the behavior of intervention group dropouts. After excluding studies that used last outcome carried forward data‐reporting techniques, we had three sibutramine, two orlistat, and five fluoxetine studies remaining. We therefore performed the sensitivity analysis only for fluoxetine. When we assumed that none of the dropouts had a weight change, the pooled reduction in weight was 3.0 kg (95% CI, 1.4 to 4.6), and when we assumed that dropouts had weight changes equivalent to those of the control group, the pooled reduction was virtually identical 3.0 kg (95% CI, 1.5 to 4.6). These estimates did not differ a great deal from the pooled estimate of 4.0 kg (95% CI, 2.0 ‐ 5.9) for the five studies eligible for this analysis (O'Kane 1994; Zelissen 1992; Daubresse 1996; Connolly 1995; Gray 1992).

Using the between‐group change for each study, we performed a meta‐regression to investigate potential interactions of weight loss with study‐level variables, including follow‐up interval, number of contacts between care provider and participant, and percentage attrition in the intervention group. None of the interactions was significant.

Because all but one study involved a dietary intervention combined with the drug or placebo, we could not investigate whether the addition of a lifestyle or behavioral intervention added to the efficacy of a drug. Nor did we have enough studies in different strata to examine the relationship between patient characteristics (e.g., age, race, baseline weight, GHb) and change in weight.

We attempted to explore the relationship between the score for risk of bias and change in weight, but we did not have enough studies to stratify into categories (A, B and C) for each drug. We examined the relationship between risk of bias and weight change in a regression model, however, and found no significant interaction.

We did not have sufficient data to perform subgroup analyses for weight, age, sex, diabetes treatment, race / ethnicity, or duration of treatment. In particular, we were not able to examine the effect of metformin or acarbose on weight loss, as only two full‐text studies reported participants using metformin (Fujioka 2000; Gokcel 2001), and none using acarbose.

Since we had statistical heterogeneity for our pooled estimates of weight change ,we present the random effects model in our summary pooled effects. Both fixed and random effects models are presented in Appendix 21, Appendix 22, Appendix 23, Appendix 24, Appendix 25, and Appendix 26. These tables also contain the pooled effects using different values of the correlation between pre‐ and post‐ measures (0.25, 0.5, and 1.0); in no case was there a significant change in the results compared to using a value of 0.75.

Sibutramine studies showed significant heterogeneity for both weight (Chi‐squared test for heterogeneity, P < 0.0001) and GHb (p < 0.0001). The study by Gokcel and colleagues (Gokcel 2001) utilized twice daily dosing for sibutramine and had more marked improvements in both weight and GHb. The pooled effect excluding this study was a reduction in weight of 4.3 kg (95% CI 2.7 to 6.0) and in GHb of 0.2% (95% CI 0.4 to 0.04). Heterogeneity remained significant for weight (p < 0.0001), but was no longer significant for GHb (p = 0.84)

There were few data available for fluoxetine on other outcomes (Appendix 15). Orlistat was associated with statistically significant improvements in total cholesterol, LDL, and triglycerides, that were sustained at 52 weeks follow‐up. Several studies examined the effects of sibutramine on blood pressure (Finer 2000; Fujioka 2000; McNulty 2003; Serrano‐Rios 2002; Redmon 2003; Kaukua 2004) and lipids (Fujioka 2000; Gokcel 2001; McNulty 2003; Serrano‐Rios 2002; Kaukua 2004; Redmon 2003), and a decrease in systolic blood pressure of 0.8 mm Hg, 95% CI, 0.02 to 1.65) and in triglycerides (0.3 mmol/L (95% CI 0.04 to 0.50)).

Adverse events for fluoxetine, orlistat, and sibutramine are summarized in Appendix 2. Adverse events were common in all three drugs, both in the intervention and control groups. Rates of gastrointestinal side effects with orlistat were about 20 percentage points higher in the treatment groups than in control groups. Tremor, somnolence, and sweating were common with fluoxetine, and palpitations with sibutramine. We included a study by Bach and colleagues (Bach 1999) which did not fulfil our inclusion criteria as no weight outcomes were presented, but the study examined cardiac value dysfunction among persons with diabetes using sibutramine, and we felt it was important to include this study in our narrative presentation of adverse events.

Narrative synthesis of other drugs

There were studies in the literature examining the efficacy of five other drugs for weight loss in adults with type 2 diabetes: cimetidine, diethylproprion, mazindol, phenmetrazine, and phentermine (Appendix 5, Appendix 6, Appendix 8, Appendix 10; Appendix 11, Appendix 13, Appendix 14, Appendix 16, Appendix 18, Appendix 19). There were insufficient data on these drugs for quantitative syntheses, therefore the results will be described in a narrative fashion.

One study examined the efficacy of cimetidine for weight loss in a double blind RCT (Stoa‐Birketvedt 1998) with 12 weeks of treatment. They noted a nonsignificant decrease in weight of 3.7 kg associated with a small improvement in glycemic control. Side effects included diarrhoea (10%), and one patient each with arthralgia, abdominal pain, and vomiting. No other literature was located on cimetidine that fit our inclusion criteria.

Three RCTs (Bratusch‐Marrian1979; Silverstone 1966; Williams 1968) and two pre versus post design studies (Montenero 1964; Hendon 1962) examined the efficacy of diethylpropion for weight loss. These were mostly older studies with sample size between 40 and 58 and follow‐up from 8 to 40 weeks. Weight change was the only outcome examined in these studies, and 2 RCTs demonstrated significant weight loss of 1.6 kg (95% CI 0.2 to 3.0) (Bratusch‐Marrian1979), 8.8 kg (95% CI 6.9 to 10.7) (Hendon 1962). In a pre versus post design study, Montenero and colleagues (Montenero 1964) demonstrated a loss of 5.3 kg (95% CI 4.1 to 6.4). Side effects were noted in two studies: dry mouth (13%) (Silverstone 1966) and headache and nausea (rate not given) (Hendon 1962).

Mazindol was examined in three full‐text RCTs (Bandisode 1975; Crommelin 1974; Slama 1978), and one abstract (Boshell 1974), as well as one study of uncertain design (Sanders 1976), one pre versus post design study (Dolecek 1976), and one cohort study with a comparison group (Felt 1977). These are all studies from the 1970's, with sample sizes ranging from 10 to 64, and follow‐up between 6 and 12 weeks. Significant weight loss was noted in three studies: Sanders and colleagues (Sanders 1976) 3.3 kg (95% CI 2.5 to 4.1), Slama et al. (Slama 1978) 12.5 kg (95% CI 5.5 to 19.5), and Boshell et al. 1.9 kg (Boshell 1974) (95% CI 3.1 to 0.8). Three other studies also demonstrated favorable changes in weight (Dolecek 1976; Bandisode 1975; Crommelin 1974; Felt 1977). None of these studies noted significant changes in fasting blood sugar. Constipation was not infrequent (Felt 1977; Dolecek 1976); other side effects included dry mouth (Crommelin 1974; Dolecek 1976), nervousness or the sensation of stimulation (Dolecek 1976; Bandisode 1975; Sanders 1976), and headache (Sanders 1976; Felt 1977; Bandisode 1975).

Phenmetrazine was only examined in one small study (Buckle 1966) which compared participants taking phenmetrazine hydrochloride to those taking a combination of phenmetrazine theoclate and phenbutrazate hydrochloride. This was a cross‐over study, and the results were presented for both groups combined. A significant decrease in weight was noted (2.9 kg (95% CI 2.3 to 3.6)) and no other outcomes were measured. Side effects included dizziness (20%), abdominal discomfort and nausea (15%), and dry mouth (5%).

Two studies examined phentermine (Campbell 1977; Gershberg 1977). It was unclear whether an abstract (Gershberg 1972) overlapped with the full text paper (Gershberg 1977). Follow‐up intervals were 16 to 26 weeks and a weight loss of 3.8 kg (95% CI 2.3 to 5.3) (Campbell 1977) and 5.7 kg (95% CI, 1.9 ‐ 7.9) were noted. Small, favorable changes in fasting blood sugar (p > 0.05), total cholesterol (P < 0.05), triglycerides (p < 0.05), and blood pressure (p > 0.05) were noted. Irritability and insomnia were noted in the first week of treatment in one study (Gershberg 1977), and dry mouth and a minor sleep disturbance in the other (Campbell 1977).

Discussion

This meta‐analysis provides evidence that fluoxetine, orlistat and sibutramine can achieve modest but statistically significant short‐term weight loss when used as a primary weight reduction strategy among adults with type 2 diabetes. Since treatment duration was up to 57 weeks for these three drugs, the long‐term effects of these drugs on weight and health outcomes in persons with type 2 diabetes remain uncertain. Across studies, participants were middle aged, were for the most part not using insulin, and were in moderately poor glycemic control. BMI was infrequently reported, making it difficult to characterize the degree of overweight of participants. Since study populations might be highly selected and run‐in periods eliminated noncompliant participants in some studies, our findings should be considered generalizable only to similar populations and not, for example, to the elderly.

There were few studies examining other drugs used for weight loss in populations with diabetes. Significant weight loss was seen in a small number of studies examining mazindol, phenmetrazine, and phentermine.

Weight loss from pharmacotherapy in nondiabetic populations is generally also modest, ranging from 2 kg to 10 kg; weight is usually regained after discontinuation of the drug; and generally there is no difference between treatment and placebo groups several months after treatment ends (National Task Force). The rather small reductions in weight noted in the current review may reflect the difficulty persons with diabetes have in losing weight (Wing 2000). Greenway (Greenway 1999) compared weight loss with orlistat and sibutramine in populations with and without diabetes and noted that weight loss was 52% and 69% greater for the subjects without diabetes.

This review has important limitations. Publication bias is possible in weight loss intervention studies (Allison 1996) and pharmacotherapy trials, which are often sponsored and financed by drug manufacturers. We attempted to obtain unpublished studies from the manufacturers of each of the included drugs as well as from researchers in this field, but received no data. We tried to minimize language bias by not excluding studies based on language of publication. Published drug trials funded by for‐profit organizations have been shown to have more positive conclusions about the drug than studies funded by nonprofit organizations (Als‐Nielson 2003). Although the causes for this association are not known, possible explanations include biased interpretation of trial results and reporting.     The quality of individual studies in this review was fairly consistent and common deficiencies were noted. Methods for concealing allocation (Clarke 2003) were described in only one study, and randomization method was described in only two. Most studies were described as double‐blind, but it was unclear which two parties were blinded. As Devereaux and colleagues have discussed (Devereaux 2002), the term double‐blind can have various definitions and interpretations among clinicians and researchers. Blinding may be difficult due to drug specific adverse events, for example, gastrointestinal side effects with orlistat. The reported data were too homogeneous to explore the effects of allocation concealment and blinding on outcomes. The quality of descriptive information on study population, setting, and the intervention was generally adequate. Sampling frame and the method of recruitment and selection of participants were rarely described, however, making it difficult to conclude from individual and pooled studies, to whom the interventions can be applied.

Attrition is an important issue in weight‐loss studies because selective loss to follow‐up has been demonstrated; higher attrition occurs among those who do not achieve a weight‐loss goal (Kaplan 1987). Attrition was often very significant in the control group, particularly for orlistat, perhaps because control participants became unblinded due to fewer gastrointestinal adverse events and had weight loss expectations that were not being fulfilled. Last‐outcome‐carried‐forward data were presented in a number of studies, which could have variable effects on measured outcomes depending on when the participant dropped out. If drug treatment was effective and the participant dropped out early after achieving minimal weight loss, final outcomes would be biased toward the null effect. If participants dropped out after 4 to 6 months in the longer follow‐up studies, however, their departure weight might have been lower than it would have been had they completed the study, as other researchers have noted that weight loss with pharmacotherapy tends to plateau at 6 months (Goldstein 1994a; National Task Force). We had to exclude from our sensitivity analysis of the effect of attrition in intervention groups, studies that used Last‐outcome‐carried‐forward techniques. Ideally, researchers would provide complete data on all subjects, including last measured weight and time and reason for attrition, particularly in studies of longer duration. The sensitivity analysis for fluoxetine demonstrates that with conservative assumptions for weight loss in the intervention dropouts, weight loss is smaller but remains statistically significant.

Orlistat, sibutramine and fluoxetine were generally well tolerated, and produced a low incidence of serious adverse events. Participants who took orlistat noted a high incidence of minor gastrointestinal side effects, as would be expected from the drug's mechanism. The use of orlistat has been associated with lower levels of fat soluble vitamins and supplementation (O'Meara 1998), although this was only evident in one study in this review (Hollander 1998). A variety of minor gastrointestinal and other side effects were noted with fluoxetine, a selective serotonin reuptake inhibitor (Yanovski 2002) and no cases were reported of withdrawal due to major adverse events.

Sibutramine produced palpitations and a nonsignificant increase in pulse rate consistent with its mechanism as a reuptake inhibitor of serotonin, norepinephrine, and dopamine (Yanovski 2002). Palpitations led to withdrawal from one study in two of 69 patients (Serrano‐Rios 2002). Major adverse cardiovascular events were not noted and rates of rhythm disturbances were similar in the intervention and control groups (Finer 2000). We found no significant blood pressure increase with sibutramine, however, only four studies reported this outcome (Serrano‐Rios 2002; Finer 2000; Fujioka 2000; McNulty 2003). Concerns have been raised about the safety of sibutramine after review of post‐marketing data (Wolfe 2002). Health Canada and a number of European countries are reviewing the safety of sibutramine, and Italy temporarily suspended marketing of the drug in March 2002 after adverse events (tachycardia, hypertension, arrhythmias) and two deaths were associated with use of the drug (Health Canada).

In nondiabetic populations, comprehensive, intensive group behavioral programs without pharmacotherapy produce mean losses of 8 kg to 10 kg at six months with a regain of 30% to 35% of weight loss at one year; 50% of participants have returned to baseline weight at 3 to 5 years (Kramer 1989; Wadden 2000). Brown and colleagues reviewed the effectiveness of weight‐loss interventions in persons with diabetes (Brown 1996), and noted that dietary interventions produced a weight loss of 9 kg and behavioral programs 3 kg, but few studies examined outcomes beyond six months. Pharmacotherapy in persons with diabetes thus appears to be no more efficacious than behavioral therapy at 1 year. Padwal 2004 recently reviewed the efficacy of long‐term pharmacotherapy for weight loss among general populations, including persons with diabetes. They noted a pooled weight change at one year follow‐up of ‐2.7 kg (95% CI 2.3 to 3.1) (11 studies) for orlistat, and 4.3 kg (95% CI 3.6 to 4.9) (five studies) for sibutramine. Similar results were observed in weight maintenance trials at up to tow year follow‐up.

Authors' conclusions

Implications for practice.

Although the weight loss demonstrated in this review is small, evidence in general populations suggests that modest loss may have health benefits. There are positive associations between weight loss and blood pressure, blood glucose, and serum lipid levels over a range of weight loss (Anderson 2001). Although few of our studies examined blood pressure, the magnitude of weight loss demonstrated in this review is equivalent to weight changes that have been efficacious in managing and preventing hypertension in high‐risk individuals over 2 to 4 years (Valdez 2002).

Fluoxetine and orlistat had statistically significant effects on GHb. The reduction of 1.0% for fluoxetine at 8 to 26 week follow‐up was sustained at 52 weeks in one study (1.8%) (O'Kane 1994). This reduction in GHb is encouraging given that the magnitude achieved was similar to that in the United Kingdom Prospective Diabetes Study (UKPDS 1998) and in the Diabetes Control and Complications Trial (DCCT 1993), where 1% absolute reductions in HbA1c resulted in significant reductions in microvascular complications from diabetes.

Orlistat was associated with statistically significant improvements in total cholesterol, LDL, and triglycerides, that were sustained at 52 weeks follow‐up. These changes in lipid levels have been noted by others (NHLBI 1998) and although modest improvements, they correspond to changes associated with a decrease in the incidence of ischemic heart disease (Law 1994). It remains unclear whether the improved glycemic control and lipid levels noted in this review could be maintained over the long‐term to influence the risk of complications as demonstrated in large trials.

The populations in the studies reviewed were generally self‐ or researcher‐selected, and often noncompliant patients were excluded from analyses. Therefore the efficacy of these drugs as delivered in a real‐world setting, will likely be less than that noted in these studies.

Concerns have been raised about adverse cardiovascular effects of sibutramine. Since this review was confined to populations with diabetes, we were not able to present a lot of information on adverse effects. Since persons with diabetes are at particularly high risk of cardiovascular events, the safety of sibutramine is of critical importance in this population, particularly if this drug is used in the long‐term.

Implications for research.

No studies in this review examined the efficacy of pharmacotherapy combined with a comprehensive lifestyle or behavioral‐modification program. In general populations, drugs have been combined with various lifestyle interventions, but most trials include relatively weak lifestyle programs, perhaps in part to better reveal the medication effects (Bray 1999a). There is some evidence that adding a lifestyle intervention improved treatment with pharmacotherapy in general obese populations (Craighead 1981; Wadden 2001a). Because moderate physical activity (Lee 1999) and improved lipid levels (Law 1994) can reduce the risk of cardiovascular disease independent of weight change, combined interventions can likely achieve improved health outcomes.

It is clear that obesity in persons with diabetes must be treated aggressively in the long‐term, as one would treat any other cardiovascular disease risk factor. Various potential approaches need to be examined in the future. Although pharmacotherapy has been used in nondiabetic populations for treatment lasting longer than one year (Hauner 1999), further research is needed with long‐term follow‐up of large populations with diabetes. More data are needed on health outcomes such as cardiovascular events, in addition to risk factors. Populations with broad ranges of BMI, age, and ethnicity need to be studied. Research is needed on the efficacy and safety of over‐the‐counter drugs that persons with diabetes are using for weight loss, and additional research is also needed on other drugs that appear promising in populations without diabetes. Goldstein has suggested that a targeted approach may be useful and that further research is needed to identify subsets of patients who can safely achieve and maintain long‐term weight loss with initial pharmacotherapy (Goldstein 1994a). Several years ago, Blackburn 1987 suggested an incremental approach with repeated goal‐setting for small amounts of weight loss; perhaps intermittent pharmacotherapy could be used with this approach. Future research must address reporting deficiencies noted in this literature, particularly descriptions of the sampling frame, methods of participant recruitment, and details of accompanying dietary interventions. Ideally, an analysis of individual patient data should be performed to examine relationships between weight loss and patient‐level characteristics such as age and initial weight.

Further work is needed to examine whether the combination of lifestyle modification and pharmacotherapy improves the efficacy of drug therapy (Phelan 2002), whether such combinations are synergistic or additive (Phelan 2002), and what dosage schedules and sequencing of the two interventions are optimal. The incidence of adverse events must be carefully monitored over the long term in diabetic populations, which already have multiple risk factors for major cardiovascular and neurologic events. The advancement of research in these areas will help reduce cardiovascular disease risk factors and events for persons with type 2 diabetes.

Feedback

Clarification about references, 27 February 2009

Summary

I could not find this reference in Diabetes ‐ on that date, page and volume is another paper. A pub med search did not reveal the true source of this: Guy‐Grand B, Valensi P, Joubert JM, Eschwege E, Amouyel P, Fagnani F. Modelisation of the 10‐year incidence reduction of coronary events in obese Type 2 diabetes patients treated with Orlistat. Diabetes 2002;51:1938. Can you help me find the correct link?   I have just sent a request stating that one of the articles had an incorrect link. On continuing to go through the references I have found another problem: Hanefeld M, Platon J, Sachse G. Orlistat promotes weight loss and improves glycaemic control in overweight patients with type 2 diabetes. Diabetologia 2001;44:889 ‐ the link goes to another article altogether.   THIRD reference with incorrect link and not found at journal web site/pubmed or any other place: Hawkins F, Duran S, Vilardell E, Soriguer F, Cabezas J, Escobar F, Milalles JM, Faure E, Bellido D, Herrera JL, Serrano‐Rios M, Tebar J, Freijane J, Armero F. Orlistat promotes glucemia control and other cardiovascular risk factors lowering in obese patients with type 2 diabetes. Randomised clinical trial. Diabetologia 2000;43:658. I am now questioning both my own searching but seriously worried about this paper .......

Reply

Thank you for picking up our errors. Abstract numbers were confused with page numbers. The correct citations are:

Guy‐Grand et al: Diabetes 2002; vol 51 (suppl 2): page A471

Hanefeld et al: Diabetologia 2001; vol 44 (suppl 1): page A231

Hawkins et al: Diabetologia 2000; vol 43 (suppl); page 171

Contributors

Comments made by Martin Dawes, occupation doctor (martin.dawes@mcgill.ca).

Susan Norris replied to the comments on behalf of the review authors for the review.

What's new

Date	Event	Description
15 May 2009	Feedback has been incorporated	Clarification about references

Adverse events	Orlistat	Sibutramine	Fluoxetine
Gastrointestinal	Minor   GI events: range 65% to 80% I, 27% to 62% C , most mild to   moderate, transient (Hollander 1998, Lindgarde 2000, Kelley 2002,   Miles 2002, Shi 2001, Halpern 2003, Hanefeld 2002, Kelley 2004)   34% GI effects (Allie 2004)	Minor   Constipation: 9% to 55% I, 6% to 8% C(Gokcel 2001, Fujioka 2000, Serrano‐Rios 2002,   Chaisson 1989); 4% (Tankova 2003)	Minor   Various: NSD between I and C(Connolly 1995)   Nausea: range 15% to 35% I, 6% to 20% C(Daubresse 1996, Kutnowski 1992, Chaisson 1989)   Diarrhea: 6% I, 2% C (p>0.05)(Daubresse 1996); 8% I, 4% C (p>0.05)(Gray 1992)   Anorexia: 12% I, 3% C (p<0.05)(Chaisson 1989)   Nausea, vomiting, diarrhea: 66% I, 60% C(O'Kane 1994)
Cardiovascular		Major   Rhythm disturbances: NSD between groups(Finer 1994)   Chest pain not suggestive of angina: 7% (2/27)(Sircar 2001)   Palpitations (moderate to severe): 41% I, 29% C(Serrano‐rios 2002)   Minor   Increased pulse rate: mean 2.4 beats/minute I (p>0.05)(Serrano‐Rios 2002);   mean 6 beats/minute I (p<0.01) (McNulty 2003)   Increased systolic blood pressure (4 mmHg) and diastolic blood pressure (3 mmHg) in 15mg   qd group; systolic blood pressure >=10 mmHG higher at endpont than baseline in 36% and   29% of patients receiving 15 and 20 mg (MuNulty 2003)   Palpitations: 7.4% I(Chaisson 1989)
Neurologic		Minor   Headache: 22% to 32% I, 40% C(Finer 2000, Sircir 2001)   Dizziness: 9 to 14% I, 5% to 13% C(Finer 2000, Sircir 2001)   Anxiety: 9% I, 0% C(Serrano‐Rios 2002)   Sleeplessness: 7% (Tankova 2003)	Minor   Tremor: 5% to 15% I, 0% to 3% C(Daubresse 1996, Kutnowski 1992, Chaisson 1989, Wise 1989)   Somnolence: 11% to 22% I, 4% to 7% C(Daubresse 1996, Chaisson 1989)   Headache: 13% I, 8% C(Gray 1992)   Asthenia: 37% I, 20% C (p>0.05)(Chaisson 1989)   Sweating: 28% I, 11% C (p<0.05)(Chaisson 1989)   Abnormal dreams: 12% I, 4% C (p<0.05)(Chaisson 1989)   Sweating, somnolence, nausea, tremor, anorexia: I > C(no statistics)   (Goldstein 1992)
Withdrawal due to   adverse effects	Minor   Various: 13% I, 8% C (Kelley 2002); 10% I, 5% C (p<0.05)(Miles 2002)   Deterioration in glycemic control: 15% I, 28% C(Kelley 2002)   GI: 4.3% I, 1.2% C(Hollander 1998); 2.6% I, 0.5% C(Lindgarde 2000);   4.7% I, 2.9% C(Halpern 2003); 0.3% I(Shi 2001); 13% I, NR for C (Kelley 2004)   22% I (Allie 2004)	Major   Palpitations: 3% I, 0% C(Serrano‐Rios 2002)   Hypertension: 3% (one patient) developed (Gokcel 2001)     Minor   Insomnia, nervousness: 6% (Redmon 2003)   Dizziness, insomnia, or diarrhea: 7% I(Finer 2000)   Chest pain not suggestive of angina: 4% I(Sircar 2001)   Dizziness, hyperglycemia, nausea: 3% I(Fukuika 2000)	Major   Chest pain: 8% I, 0% C (p>0.05)(Gray 1992)     Minor   GI: 22%(O'Kane 1994)   Nausea, lethargy, or excessive sweating: 20%(Connolly 1995)   Unspecified: 1% to 9% I, 1% to 2% C (Daubresse 1996, Kutnowski 1992)   Connolly 1995)
Other	Minor   Hypoglycemia: 7% to 17% I, 3% to 10% C(Kelley 2002, Miles 2002, Hanfeld 2002)   No gallstones, no renal stones(Hollander 1998)   Normal plasma concentrations vitamin A,D,E, beta‐carotene(Hollander 1998)   Decrease in vitamin E and beta‐carotene concentrations in I vs C (p<0.001)(Hollander 1998)   No significant difference in adverse events I and C (p=0.75)(Serrano‐rios 2001)	Major   Serious AE: 6% I, 1% C (1/ 5 in I possibly drug‐related (somnolence, dizziness, confusion))   (Fujioka 2000)   Minor   Dry mouth: 38% I, NR C(Gokcel 2001); 23% I, 11% C(Finer 2000); "common"(McNulty 2003);   reported in Redmon 2003 (no data); 6% (Tankova 2003)   Infection (not specified): 18% to 26% I, 2% to 24% C(Finer 2000, Fujioka 2000)   Increased platelet count and increased serum sodium in I (Serrano‐Rios 2002)   AE unspecified: 61% I, 52% C(Serrano‐Rios 2002)	Minor   Infections: 50% I, 55% C(Breum 1995); NSD between groups(Connolly 1995)   Decreased libido: 13% I, 0% C (p=0.07)(Gray 1992)

Study	Number	Follow‐up (weeks)	Age (years)	Sex (%female)	*Weight kg)**	*GHb (%)**	Diet only (%)**	Using insulin (%)	Diet
Fluoxetine   Connolly 1995   Daubresse 1996   Gray 1992   Kutnowski 1992   O'Kane 1994   Zelissen 1992	30   82   48   97   19   20	16   8   24   9   52   26	66   52   NR   51   57   51	38   NR   54   47   68   69	85.1(12.0)   90.9(16.4)   107.3(24.5)   92.3(16.7)   97.8(NR)   106.1(25.0)	8.7(2.5)   8.6(3.3)   10.2(3.0)   NR   8.8(NR)   9.0(1.6)	100   40   0   NR   37   NR	0   0   100   0   0   0	Low calorie   Low calorie   1200 Kcal/d   Low calorie   Usual   Low calorie
	296     Total	8‐52     range	54     mean	51.23     mean	94.9 (18.5)     mean(SD)	9.1 (3.0)     mean(SD)
Orlistat   Bloch 2003   Hollander 1998   Kelley 2002   Kelley 2004   Hanefeld 2002   Lindgarde 2000   Miles 2002   Wang 2003	76   322   550   39   383   99   504   63	12   57   52   26   52   54   52   24	56   55   58   51   56   54   53   41	83   49   56   67   51   64   48   48	87.5(17.9)   99.6(14.5)   102.0(1.0)   102 (16.9)   98.4(18.5)   NR   102.1(1.1)   83(9)	NR   8.5(1.0)   9.0(0.1)   8.1(1.2)   8.6(1.2)   10.0(NR)   8.9(1.0)   8.2(1.2)	79   0   0   NR   NR   NR   0   0	13   0   100   0   0   NR   0   0	30% fat   500‐600 Kcal/d deficit or low fat   500‐600 Kcal/d deficit   500‐600 Kcal/d deficit   600 Kcal/d deficit   500‐600 Kcal/d deficit + behavioral modification   500‐600 Kcal/d deficit   NR
	2036	12‐57	53	58.3	95.9 (11.1)	8.8 (0.9)
Sibutramine   Finer 2000   Fujioka 2000   Gokcel 2001   Kaukua 2004   McNulty 2003   Redmon 2003   Serrano‐Rios 2002   Tankova 2003	91   175   60   236   195   61   134   95	12   24   26   52   52   52   24   13	54   54   48   54   49   54   54   46	53   41   100   70   56   46   68   54	82.5 (NR)   98.2 (14.6)   95.5 (14.2)   100.8(17.4)   100.7(20.8)   112.4(21.0)   94.2 (19.9)   91.7(8.8)	9.2 (1.3)   8.3 (1.2)   9.8 (0.1)   NR   9.7(0.3)   8.2 (1.1)   9.5 (2.1)   NR	14   17   0   100   0   NR   0   30	24   0   NR   0   0   0   0   0	500‐600 Kcal/d deficit or low fat   500‐600 Kcal/d deficit or low fat   Low calorie   700 Kcal/d deficit   Standard diet advice   500‐1000 Kcal/d deficit; some meal replacem.   Low calorie   Low calorie
d, day   Ghb< glycated hemoglobin   NR, not reported   SD, standard deviation	1047	12‐52     * Weight and glycated hemoglobin (GHb) for control group at baseline   ** % of the study population treated with diet only	52	61	97.0 (17.3)	9.3 (1.3)

Study	Drug dosage	Int. attrition(%)	Control attrition(%)
Fluoxetine   Connolly 1995   Daubresse 1996   Gray 1992   Kutnowski 1992   O'Kane 1994   Zelissen 1992	60 mg qd   60 mg qd   60 mg qd   60 mg qd   60 mg qd   60 mg qd	27.0   20.5   33.0   14.9   22.0   0.0	15   13.9   17.0   10.0   10.0   0.0
		20.1 (0‐33.0)     mean(range)	12.1 (0‐17.0)     mean(range)
Orlistat   Bloch 2003   Hollander 1998   Kelley 2002   Kelley 2004   Hanefeld 2002   Lindgarde 2000   Miles 2002   Wang 2003	120 mg tid   120 mg tid   120 mg tid   120 mg tid   120 mg tid   120 mg tid   120 mg tid   120 mg bid‐tid	6.7   14.7   49.0   34,6   33.0   NR   35.0   3.2	22.4   27.7   52.0   15.4   29.2   NR   44.0   0
		25.2 (3.2‐49.0)	27.2 (0‐52.0)
Sibutramine   Finer 2000   Fujioka 2000   Gokcel 2001   Kaukua 2004   McNulty 2003   Redmon 2003   Serrano‐Rios 2002   Tankova 2003	15 mg qd   5‐10 mg qd   10 mg bid   15 mg qd   15 or 20 mg qd   10‐15 mg qd   15 mg qd   10‐15 mg qd	9.0   33.0   3.0   8.0   24.6   10.0   23.2   NR	9.0   29.0   17.0   11.0   28.1   6.9   12.0   NR
bid, twice daily   Int, intervention   NR, not reported   qd, daily		15.8 (3.0‐33.0)     mean(range)	16.0 (6.9‐29.0)     mean(range)

Study ID	Methods	Participants	Intervention	Outcomes	Notes
Stoa‐Birketvedt 1988Multiple pub: No	Study design: RCTRandomization procedure: Randomized according to BMI; details unclearAllocation concealment: UnclearFollow‐up: 12w	Country: NorwaySetting: Hospital clinicNumber: 62Age: 48YSex: 33%FMedications: 49% on oral agentsBL wt: I 103.9, C 102.0BL BMI: I 33.8, C 34.0BL GHb: NR	Drug: CimetidineDosage: 400mg tidDuration: 12wDiet: Usual diet and activityComparison: Placebo + usual diet and activity	Weight: YesBMI: Yes>5% loss (%):FBS: YesGHb: YesCholesterol: YesLDL:HDL: YesTG: YesSBP: YesDBP: YesSide effects: Yes; 10% diarrhea, 5% each of abdominal pain, vomiting and arthralgia	Funding: Norwegian Research council, The Novo Nordic Foundation, The Norwegian Diabetes AssociationAbstract/full text: FTLOCF: NRITT: Yes, with attritionAttrition: 19%Blinding: Double blindBlinding assessor: UnclearBL comparable: YesJadad Score: 1,1,1,BRisk of bias: B
A, abstract; BMI, body mass index (kg/m2); C, comparison group; CHO, carbohydrate; F, female; FBS, fasting blood sugar; d, day;	FT, full text; GHb, glycated hemoglobin; I, intervention group; ITT, intention to treat; LOCF, last outcome carried forward; NA, not applicable; NR, not reported; qd, daily; RCT, randomized, controlled trial; y, year; w, weeks

Study	Weight	Glycemic Control	Lipids	Blood pressure
Stoa‐Birketvedt 1998   Study design: RCT   Follow‐up interval: 12 weeks     C, control group   I, intervention group   SE, standard error   RCT, randomized controlled trial   SBP, systolic blood pressure   DBP, diastolic blood pressure	1. Weight change (kg)   Delta I (SE): ‐5 (0.5)   Delta C (SE): ‐1.3 (0.2)   Delta (I‐C) (SE): ‐3.7 (2.0)   2. BMI (kg/m2)   Delta I (SE): ‐1.6 (0.5)   Delta C (SE): ‐0.4 (0.6)   Delta (I‐C) (SE): ‐1.2 (0.8)   3. % of weight loss   Delta I (SE): ‐4.8 (0.5)   Delta C (SE): ‐1.3 (0.2)   Delta (I‐C) (SE): ‐3.5 (0.5)	1. GHb (%)   Delta I (SE): ‐0.5 (0.2)   Delta C (SE): ‐0.3 (0.2)   Delta (I‐C) (SE): ‐0.2 (0.3)   2. Fasting blood sugar (mml/L)   Delta I (SE): ‐1.3 (0.4)   Delta C (SE): ‐0.5 (0.4)   Delta (I‐C) (SE): ‐0.8 (0.5)	1. Total cholesterol (mmol/L)   Delta I (SE): ‐0.1 (0.2)   Delta C (SE): ‐0.3 (0.2)   Delta (I‐C) (SE): 0.2 (0.2)   2. HDL cholesterol (mmol/L)   Delta I (SE): 0.1 (0.0)   Delta C (SE): ‐0.1 (0.0)   Delta (I‐C) (SE): 0.2 (0.1)   3. Triglycerides (mmol/L)   Delta I (SE): ‐0.5 (0.2)   Delta C (SE): 0 (0.4)   Delta (I‐C) (SE): ‐0.5 (0.4)	1. SBP   Delta I (SE): ‐6.9 (2.6)   Delta C (SE): ‐7.0 (2.7)   Delta (I‐C) (SE): 0.1 (2.7)   2. DBP   Delta I (SE): ‐6.0 (1.5)   Delta C (SE): ‐3.0 (1.0)   Delta (I‐C) (SE): ‐3.0 (1.0)

Fluoxetine	rho = 0.25	rho = 0.5	rho = 0.75	rho = 1.0
8‐16 weeks (88 treated vs 104 controls)     fixed effects model	‐3.9 (95% CI ‐3.0 to ‐4.8)	‐3.9 (95% CI ‐3.0 to ‐4.8)	‐3.0 (95% CI ‐2.3 to ‐3.8)	‐1.9 (95% CI ‐1.5 to ‐2.2)
Heterogeneity (q, p value)	q = 4.64, p = 0.33	q = 4.85, p = 0.3	q = 75. 58, p = 0.004	q = 30.6, p = 0.00001
Random effects model	‐4.0 (95% CI ‐2.8 to ‐5.3)	‐4.0 (95% CI ‐2.7 to ‐5.3)	‐3.4 (95% CI ‐1.7 to ‐5.2)	‐3.6 (95% CI ‐1.5 to ‐5.7)
Heterogeneity (q, p value)	q = 4.64, p = 0.33	q = 4.85, p = 0.3	q = 15.58, p = 0.004	q = 30.6, p = 0.00001
24‐26 weeks (45 treated vs 52 controls)     Fixed effects model	‐5.1 (95% CI ‐3.4 to ‐6.8)	‐5.1 (95% CI ‐3.4 to ‐6.8)	‐5.1 (95% CI ‐3.4 to ‐6.8)	‐5.1 (95% CI ‐3.4 to ‐6.8)
Heterogeneity (q, p value)	q = 3.19, p = 0.36	q = 3.19, p = 0.36	q = 3.19, p = 0.36	q = 3.19, p = 0.36
Random effects model	‐5.1 (95% CI ‐3.3 to ‐6.9)	‐5.1 (95% CI ‐3.3 to ‐6.9)	‐5.1 (95% CI ‐3.3 to ‐6.9)	‐5.1 (95% CI ‐3.3 to ‐6.9)
Heterogeneity (q, p value)     CI, confidence interval	q = 3.19, p = 0.36	q = 3.19, p = 0.36	q = 1.26, p = 0.74	q = 3.19, p = 0.36

Outcome	Follow‐up 8‐16w				Follow‐up 24‐26w
	No. of studies	N	Point estimate	95% CI	No. of studies	N	Point estimate	95% CI
Weight (kg)	5	192	‐3.4	(‐5.2, ‐1.7)	4	97	‐5.1	(‐6.90, ‐3.26)
percent weight loss					1	20	‐2.5	(‐7.9, 3.0)
% participants with weight loss >5%
BMI (kg/m2)	1	47	‐0.5	(‐1.0, ‐0.1)
Waist circumference (cm)
GHb (%)	4	145	‐1.0	(‐1.5, ‐0.4)	4	97	‐1.0	(‐1.4, ‐0.6)
Fasting glucose (mmol/l)	5	192	‐0.9	(‐2.1, 0.4)	4	97	‐0.9	(‐2.0, 0.2)
SBP (mm Hg)
DBP (mm Hg)
Total cholesterol (mmol/l)	2	85	‐0.1	(‐0.3, 0.2)	1	17	0.1	(‐0.4, 0.6)
LDL cholesterol (mmol/l)
HDL cholesterol (mmol/l)	1	68	0	(‐0.1, 0.1)
Triglycerides (mmol/l)	2	85	‐0.5	(‐1.1, 0.1)	1	17	‐0.2	(‐1.0, 0.7)

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Weight (kg)	5	192	Mean Difference (IV, Fixed, 95% CI)	‐3.04 [‐3.75, ‐2.33]
2 BMI	1	47	Mean Difference (IV, Fixed, 95% CI)	‐0.52 [‐0.96, ‐0.08]
3 GHb	4	145	Mean Difference (IV, Fixed, 95% CI)	‐0.97 [‐1.52, ‐0.41]
4 Fasting glucose	5	192	Mean Difference (IV, Fixed, 95% CI)	‐1.05 [‐1.71, ‐0.40]
5 Total cholesterol	2	85	Mean Difference (IV, Fixed, 95% CI)	‐0.05 [‐0.33, 0.24]
6 HDL cholesterol	1	68	Mean Difference (IV, Fixed, 95% CI)	0.03 [‐0.05, 0.11]
7 Triglycerides	2	85	Mean Difference (IV, Fixed, 95% CI)	‐0.5 [‐1.11, 0.11]
8 Weight (kg)	5	192	Mean Difference (IV, Random, 95% CI)	‐3.43 [‐5.20, ‐1.66]
9 BMI	1	47	Mean Difference (IV, Random, 95% CI)	‐0.52 [‐0.96, ‐0.08]
10 GHb	4	145	Mean Difference (IV, Random, 95% CI)	‐0.97 [‐1.52, ‐0.41]
11 Fasting glucose	5	192	Mean Difference (IV, Random, 95% CI)	‐0.85 [‐2.07, 0.38]
12 Total cholesterol	2	85	Mean Difference (IV, Random, 95% CI)	‐0.05 [‐0.33, 0.24]
13 HDL cholesterol	1	68	Mean Difference (IV, Random, 95% CI)	0.03 [‐0.05, 0.11]
14 Triglycerides	2	85	Mean Difference (IV, Random, 95% CI)	‐0.5 [‐1.11, 0.11]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Weight (kg)	4	97	Mean Difference (IV, Fixed, 95% CI)	‐5.10 [‐6.83, ‐3.37]
2 Percent weight loss	1	20	Mean Difference (IV, Fixed, 95% CI)	‐2.48 [‐7.94, 2.98]
3 GHb	4	97	Mean Difference (IV, Fixed, 95% CI)	‐1.03 [‐1.42, ‐0.63]
4 Fasting glucose	4	97	Mean Difference (IV, Fixed, 95% CI)	‐0.87 [‐1.96, 0.22]
5 Total cholesterol	1	17	Mean Difference (IV, Fixed, 95% CI)	0.10 [‐0.40, 0.60]
6 Triglycerides	1	17	Mean Difference (IV, Fixed, 95% CI)	‐0.16 [‐1.02, 0.70]
7 Weight (kg) random	4	97	Mean Difference (IV, Random, 95% CI)	‐5.08 [‐6.90, ‐3.26]
8 Percent weight loss	1	20	Mean Difference (IV, Random, 95% CI)	‐2.48 [‐7.94, 2.98]
9 GHb	4	97	Mean Difference (IV, Random, 95% CI)	‐1.03 [‐1.42, ‐0.63]
10 Fasting glucose	4	97	Mean Difference (IV, Random, 95% CI)	‐0.87 [‐1.96, 0.22]
11 Total cholesterol	1	17	Mean Difference (IV, Random, 95% CI)	0.10 [‐0.40, 0.60]
12 Triglycerides	1	17	Mean Difference (IV, Random, 95% CI)	‐0.16 [‐1.02, 0.70]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Weight (kg)	1	17	Mean Difference (IV, Fixed, 95% CI)	‐5.8 [‐10.84, ‐0.76]
2 Percent weight loss	0	0	Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
3 % with wt loss > 5%	0	0	Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
4 BMI	0	0	Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
5 Waist circumference	0	0	Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
6 GHb	1	17	Mean Difference (IV, Fixed, 95% CI)	‐1.8 [‐3.78, 0.18]
7 Fasting glucose	1	17	Mean Difference (IV, Fixed, 95% CI)	‐0.8 [‐2.50, 0.90]
8 SBP	0	0	Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
9 DBP	0	0	Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
10 Total cholesterol	1	17	Mean Difference (IV, Fixed, 95% CI)	0.5 [‐0.31, 1.31]
11 LDL cholesterol	0	0	Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
12 HDL cholesterol	0	0	Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
13 Triglycerides	1	17	Mean Difference (IV, Fixed, 95% CI)	‐0.5 [‐1.15, 0.15]
14 Weight (kg) random	1	17	Mean Difference (IV, Random, 95% CI)	‐5.8 [‐10.84, ‐0.76]
15 Percent weight loss	0	0	Mean Difference (IV, Random, 95% CI)	0.0 [0.0, 0.0]
16 BMI	0	0	Mean Difference (IV, Random, 95% CI)	0.0 [0.0, 0.0]
17 GHb	1	17	Mean Difference (IV, Random, 95% CI)	‐1.8 [‐3.78, 0.18]
18 Fasting glucose	1	17	Mean Difference (IV, Random, 95% CI)	‐0.8 [‐2.50, 0.90]
19 SBP	0	0	Mean Difference (IV, Random, 95% CI)	0.0 [0.0, 0.0]
20 DBP	0	0	Mean Difference (IV, Random, 95% CI)	0.0 [0.0, 0.0]
21 Total cholesterol	1	17	Mean Difference (IV, Random, 95% CI)	0.5 [‐0.31, 1.31]
22 HDL cholesterol	0	0	Mean Difference (IV, Random, 95% CI)	0.0 [0.0, 0.0]
23 Triglycerides	1	17	Mean Difference (IV, Random, 95% CI)	‐0.5 [‐1.15, 0.15]

Study ID	Methods	Participants	Outcomes	Intervention	NOtes
Allie 2004     Multiple pub:   No	Study design: Pre vs post, retrospectiveRandomization procedure: NAAllocation concealment: NAFollow‐up: 26 weeks	Country: UDASetting: Endocrinology clinicNumber: 23Age: 53Sex: NRMedications: NRBL wt: 118.0(2.5)BL BMI: 40.5(7.0)BL GHb: 7.9(1.6)	Drug: Orlistat Dosage: 120mg tidDuration: 13 to 26 weeksDiet: NRComparison: NA	Weight: YBMI: Y>5% loss (%): YFBS: GHb: YCholesterol: YLDL: YHDL: YTG: YSBP: YDBP: YSide effects: Y	Funding: Abstract/full text: FTLOCF: NAITT: NAAttrition: NA (retrospective)Blinding: NA Blinding pt: No Blinding assessor: NABlinding provider: NoBL comparable: NA
Bloch 2003     Multiple pub:   No	Study design: RCTRandomization procedure: Central random number listAllocation concealment: AdequateFollow‐up: 12 weeks	Country: BrazilSetting: Hypertension clinicNumber: 204 total; 76 analyzed with diabetesAge: 56 yearsSex: 83% overallMedications: I: 68% oral agents, 8% insulin; C: 63% oral agents and 18% insulinBL wt: I 91.5, C 87.5BL BMI: I 36.6, C 35.4BL GHb: NRNote: Demographic information was given only for whole study group (39% with diabetes), including persons with diabetes and those without.	Drug: Orlistat Dosage: 120mg tidDuration: 12 weeksDiet: Low calorie diet, 30% fat; advised to increase activityComparison: Diet and activity as for intervention group	Weight: Y   BMI:   >5% loss (%): Y   FBS: Y   GHb: Y   Cholesterol: Y   LDL:   HDL: Y   TG: Y   SBP: Y   DBP:   Side effects: Y	Funding: University Hospital Abstract/full text: FTLOCF: YesITT: YesAttrition: 31% overallBlinding: NRBlinding pt: No Blinding assessor: NRBlinding provider: NRBL comparable: Yes
Bonnici 2002Multiple pub: No	Study design: RCTRandomization procedure: NRAllocation concealment: UnclearFollow‐up: 24w	Country: South AfricaSetting: Multicenter trial; no detailsNumber: 284Age: NRSex: NRMedications: Metformin and/or sulfonylureaBL wt: NRBL BMI: NRBL GHb: NR	Drug: Orlistat Dosage: 120mg tidDuration: 24wDiet: 600kcal/d deficitComparison: Placebo + diet	Weight: YesBMI:>5% loss (%): YesFBS: YesGHb: YesCholesterol:LDL: YesHDL:TG:SBP:DBP:Side effects:	Funding: NRAbstract/full text: ALOCF: NRITT: NRAttrition: NR Blinding: Double‐blindBlinding assessor: UnclearBL comparable: NRJadad score: 1,1,0,BRisk of bias: C
Deerochanawong 2001Multiple pub: No	Study design: RCTRandomization procedure: NRAllocation concealment: UnclearFollow‐up: 24w	Country: NRSetting: NRNumber: 252Age: NRSex: NRMedications: No insulin or acarboseBL wt: I 77, C 77BL BMI: NRBL GHb: NR	Drug: OrlistatDosage: 120mg tidDuration: 24wDiet: 600kcal/d deficitComparison: Placebo + diet	Weight: YESBMI:>5% loss (%): YesFBS: YesGHb: YesCholesterol:LDL:HDL:TG:SBP:DBP:Side effects:	Funding: NRAbstract/full text: ALOCF: NRITT: UnclearAttrition: NRBlinding: Double‐blindBlinding assessor: UnclearBL comparable: NRJadad score: 1,1,0,BRisk of bias: B
Dimitrov 2001Multiple pub:No	Study design: Pre‐versus‐postRandomization procedure: NAAllocation concealment: NAFollow‐up: 3m	Country: BulgariaSetting: Academic medical clinicNumber: 12Age: NRSex: NRMedications: NRBL wt: 103.6BL BMI: NRBL GHb: NR	Drug: OrlistatDosage: 120mg tidDuration: 3mDiet: NRComparison: Nondiabetic, obese persons	Weight: YesBMI:>5% loss (%):FBS:GHb:Cholesterol: YesLDL: YesHDL:TG:SBP:DBP:Side effects:	Funding: NRAbstract/full text: ALOCF: NRITT: NAAttrition: NRBlinding: NABlinding assessor: NoBL comparable: NAJadad score: NARisk of bias: NA
Guy‐Grand 2001aMultiple pub: Guy‐Grande 2002bGuy‐Grand 2002	Study design: RCTRandomization procedure: NRAllocation concealment: UnclearFollow‐up: 26w	Country: FranceSetting: Multicenter, details NRNumber: 193Age: 52Sex: NRMedications: Oral hypoglycemic agentsBL wt: NRBL BMI: 33.7BL GHb: 7.7	Drug: OrlistatDosage: 120mg tidDuration: 26wDiet: low calorieComparison: Placebo + diet	Weight: YesBMI:>5% loss (%):FBS: YesGHb: YesCholesterol:LDL:HDL:TG:SBP:DBP:Side effects:	Funding: NRAbstract/full text: ALOCF: NRITT: UnclearAttrition: NRBlinding: Double‐blindBlinding assessor: UnclearBL comparable: YesJadad score: 1,1,1,BRisk of bias: C
Halpern 2003Halpern 2001 (abstract)	Study design: Multicenter RCTRandomization procedure: Randomization list generated by sponsorAllocation concealment: UnclearFollow‐up: 26w	Country: Latin AmericaSetting: NRNumber: 338Age: 51Sex: 69%FMedications: No insulin or acarboseBL wt: 89.6BL BMI: 34.6BL GHb: 8.4%	Drug: OrlistatDosage: 120mg tidDuration: 24wDiet: 600kcal/d deficit; caloric content: 30% fat, 50% CHO, 20% proteinComparison: Placebo + diet	Weight: YesBMI:>5% loss (%): YesFBS: YesGHb: YesCholesterol: YesLDL: YesHDL: YesTG: YesSBP: YesDBP: YesSide effects: Yes	Funding: F. Hoffman‐La roche (Basel, Switzerland)Abstract/full text: FT LOCF: YesITT: No; 5 patients withdrawn (no reason stated) after at least one follow‐up measurement; some patients withdrawn for 'noncompliance'Attrition: 18.4%Blinding: Double‐blindBlinding assessor: UnclearBL comparable: YesOther: Must have >60% compliance with placebo during 2w lead‐in to enter studyJadad score: 1,1,0,BRisk of bias: C
Hanefeld 2002Multiple pub: Hanefeld 2001 (abstract)	Study design: RCT, multicenterRandomization procedure: NRAllocation concealment: UnclearFollow‐up: 52w	Country: GermanySetting: Outpatient clinicsNumber: 383Age: 51%FSex: 56yMedications: Diet or sulphonurea; no insulinBL wt: I 98.4, C 99.4BL BMI: I 33.7, C 34.5BL GHb: I 8.6, C 8.6	Drug: OrlistatDosage: 120mg tidDuration: 48wDiet: 600kcal/d deficit Comparison: Diet + Placebo	Weight: YesBMI: Yes>5% loss (%): YesFBS: YesGHb: YesCholesterol: YesLDL:HDL: YesTG: YesSBP: YesDBP: YesSide effects: Yes	Funding: Hoffman‐La Roche AGAbstract/full text: FTLOCF: NRITT: No; some patients withdrawn for failure to complyAttrition: 31%Blinding: Double‐blindBlinding assessor: UnclearBL comparable: NROther: 22% of study population were not randomized after lead‐in period as did not comply with study processesJadad score: 1,1,1,B Risk of bias: C
Hawkins 2000Multiple pub: No	Study design: RCTRandomization procedure: NRAllocation concealment: UnclearFollow‐up: 6m	Country: NRSetting: Multicenter trial, details unclearNumber: 307Age: NRSex: NRMedications: NRBL wt: NRBL BMI: >27BL GHb: NR	Drug: OrlistatDosage: 120mg tidDuration: 24wDiet: HypocaloricComparison: Placebo + diet	Weight: YesBMI:>5% loss (%):FBS: YesGHb: YesCholesterol: YesLDL: YesHDL:TG:SBP: YesDBP: YesSide effects:	Funding: NRAbstract/full text: ALOCF: NRITT: Yes, with attrition Attrition: 2.5%Blinding: Double‐blindBlinding assessor: UnclearBL comparable: NRJadad score: 1,1,0,BRisk of bias: C
Hollander 1998a     Multiple pub:   Hollander 1997, 1998, 1999	Study design: RCTRandomization procedure: NRAllocation concealment: UnclearFollow‐up: 57w	Country: USASetting: multicenter, academic medical centersNumber: 322Age: 55Sex: 49%FMedications: Oral sulfonureaBL wt: I 99.7, C 99.6 BL BMI: I 34.0, C 34.5BL GHb: I 8.2, C 8.5	Drug: OrlistatDosage: 120mg tidDuration: 52wDiet: 500kcal/d deficitComparison: Placebo + diet	Weight: YesBMI:>5% loss (%): YesFBS: YesGHb: YesCholesterol: YesLDL: YesHDL: YesTG: YesSBP:DBP:Side effects: Yes	Funding: Hoffman‐LaRocheAbstract/full text: FTLOCF: NRITT: Yes, with attritionAttrition: 21%Blinding: Double‐blindBlinding assessor: UnclearBL comparable: YesJadad score: 1,1,1,BRisk of bias: C
Hollander 2001Multiple pub: No	Study design: RCTRandomization procedure: NRAllocation concealment: UnclearFollow‐up: 1y	Country: USASetting: NRNumber: 503Age: NRSex: NRMedications: MetforminBL wt: NRBL BMI: >28BL GHb: NR	Drug: OrlistatDosage: 120mg tidDuration: 1yDiet: Mildly reduced caloric Comparison: Placebo + diet	Weight: YesBMI:>5% loss (%): YesFBS: YesGHb: YesCholesterol: LDL: YesHDL: YesTG: YesSBP: YesDBP: YesSide effects:	Funding: NRAbstract/full text: ALOCF: YesITT: CompleteAttrition: NRBlinding: Double‐blindBlinding assessor: UnclearBL comparable: UnclearJadad score: 1,1,0,BRisk of bias: C
Kelley 2004     Multiple pub:   No	Study design: RCTRandomization procedure: NRAllocation concealment: UnclearFollow‐up: 26 weeks	Country: USA Setting: Academic center; community recruitmentNumber: 39Age: 51Sex: 67Medications: Oral agents or diet; oral agents withdrawn 1 month prior to interventionBL wt: I 99, C 102BL BMI: I 34.0, C 35.9BL GHb: I 8.1, C7.8	Drug: Orlistat Dosage: 120mg tidDuration: 3 monthsDiet: 500 calorie deficit; <=30% fat; activity encouragedComparison: 500 calorie deficit; <=30% fat; activity encouraged	Weight: YBMI: Y>5% loss (%): FBS: YGHb: YCholesterol: YLDL: YHDL: YTG:SBP:DBP:Side effects: Y	Funding: Roche laboratoriesAbstract/full text: FTLOCF: NoITT: PartialAttrition: 25%Blinding: Double blindBlinding pt: YBlinding assessor: UnclearBlinding provider: UnclearBL comparable: Y
Kelley 2002Multiple pub: Kelley 2001Bray 2001	Study design: RCTRandomization procedure: NRAllocation concealment: UnclearFollow‐up: 52w	Country: USASetting: Multicenter; academic medical centersNumber: 550Age: 58Sex: 57%FMedications: Insulin +/‐ oral agent (excluding thazolidindiones)BL wt: I 101.8, C 102.0 BL BMI: I 35.6, C 35.8BL GHb: I 9.0, C 9.0	Drug: OrlistatDosage: 120mg bidDuration: 52wDiet: 500kcal/d deficitComparison: Placebo + diet	Weight: YesBMI:>5% loss (%): YesFBS: YesGHb: YesCholesterol: YesLDL: YesHDL: YesTG: YesSBP: YesDBP: YesSide effects: Yes	Funding: Hoffman‐LaRocheAbstract/full text: FTLOCF: YesITT: CompleteAttrition: 52%Blinding: Double‐blindBlinding assessor: UnclearBL comparable: YesJadad score: 1,1,1,BRisk of bias: B
Kelly 1997Multiple pub: No	Study design: RCTRandomization procedure: NRAllocation concealment: UnclearFollow‐up: 57w	Country: USASetting: MulticenterNumber: 322Age: NRSex: NRMedications: SulfonureasBL wt: NRBL BMI: NRBL GHb: NR	Drug: OrlistatDosage: 120mg tidDuration: 52wDiet: 500kcal/d deficitComparison: Placebo + diet	Weight: YesBMI:>5% loss (%):FBS: YesGHb: YesCholesterol: YesLDL: YesHDL: TG: YesSBP:DBP:Side effects: Yes	Funding: Hoffman‐LaRocheAbstract/full text: ALOCF: NRITT: Yes, with attritionAttrition: I 15%, C 28%Blinding: Double‐blindBlinding assessor: UnclearBL comparable: NRJadad score: 1,1,0,BRisk of bias: B
Le Roux 2001Multiple pub: No	Study design: Pre‐versus‐postRandomization procedure: NAAllocation concealment: NAFollow‐up: 6m	Country: EnglandSetting: NRNumber: 7Age: NRSex: NRMedications: NRBL wt: NRBL BMI: 40.2BL GHb: 8.7	Drug: OrlistatDosage: 120mg tidDuration: 6mDiet: UnclearComparison: NA	Weight:BMI: Yes>5% loss (%):FBS:GHb: YesCholesterol: YesLDL: YesHDL:TG: YesSBP:DBP:Side effects:	Funding: NRAbstract/full text: ALOCF: NAITT: NAAttrition: NRBlinding: NABlinding assessor: NoBL comparable: NAJadad score: NARisk of bias: NA
Lindgarde 2000Multiple pub: No	Study design: RCT; 26% of total study population had type 2 diabetesRandomization procedure: NRAllocation concealment: UnclearFollow‐up: 54w	Country: SwedenSetting: 33 primary care centersNumber: 99Age: 54y (whole population)Sex: 64% (whole population)Medications: NRBL wt: NR for diabetic populationBL BMI: NR for diabetic populationBL GHb: I 8.7, C 10.0	Drug: OrlistatDosage: 120mg tidDuration: 52wDiet: 600kcal/d deficitComparison: Placebo + diet	Weight: YesBMI:>5% loss (%): YesFBS: YesGHb: YesCholesterol:LDL:HDL:TG:SBP:DBP:Side effects: Yes	Funding: Roche AB, Stockholm, SwedenAbstract/full text: FTLOCF: NRITT: Yes, with attritionAttrition: 14%Blinding: Double‐blindBlinding assessor: UnclearBL comparable: Yes (for whole population)Jadad score: 1,1,1,B Risk of bias: B
Martin SF 2001     Multiple pub:   No	Study design: Cohort with comparison groupRandomization procedure: NAAllocation concealment: NAFollow‐up: 6m	Country: Northern IrelandSetting: Obesity clinicNumber: 55Age: NRSex: 51%FMedications: NRBL wt: I: 102.8, C 101.1BL BMI: NRBL GHb: I 37.8, C 42	Drug: OrlistatDosage: NRDuration: 26wDiet: Dietary adviceComparison: No orlistat	Weight: YesBMI:>5% loss (%): YesFBS:GHb:Cholesterol:LDL:HDL:TG:SBP:DBP:Side effects:	Funding: NRAbstract/full text: A LOCF: NoITT: Yes, with attritionAttrition: 59%Blinding: NRBlinding assessor: NRBL comparable: NoOther: Intervention group was persons who lost >‐2kg in 4w lead‐in periodJadad score: NARisk of bias: C
Mendoza‐Guadarrama 2000Multiple pub: No	Study design: RCTRandomization procedure: NRAllocation concealment: UnclearFollow‐up: 26w	Country: MexicoSetting: obesity clinicNumber: 30Age: 51Sex: 60%FMedications: NRBL wt: NRBL BMI: I 31.3, C 30.6BL GHb: NR	Drug: OrlistatDosage: 120mg tidDuration: 26wDiet: 500kcal/d deficitComparison: Placebo + diet	Weight: BMI: Yes>5% loss (%):FBS:GHb:Cholesterol:LDL:HDL:TG:SBP:DBP:Side effects:	Funding: NRAbstract/full text: ALOCF: NRITT: UnclearAttrition: NRBlinding: Double‐blindBlinding assessor: UnclearBL comparable: NRJadad score: 1,1,0,BRisk of bias: C
Miles 2002Multiple pub: Miles 2001	Study design: RCTRandomization procedure: NRAllocation concealment: UnclearFollow‐up: 52w	Country: USASetting: Multicenter; UnclearNumber: 505Age: 53ySex: 48%FMedications: Metformin +/‐ sulfonureaBL wt: I 101.1, C 102.1BL BMI: I 35.2, C 35.6BL GHb: I 8.8, C 8.9	Drug: OrlistatDosage: 120mg tidDuration: 52wDiet: 500kcal/d deficitComparison: Placebo + diet	Weight:BMI:>5% loss (%):FBS:GHb:Cholesterol:LDL:HDL:TG:SBP:DBP:Side effects:	Funding: Hoffman‐LarocheAbstract/full text: FTLOCF: YesITT: CompleteAttrition: 40%Blinding: Double‐blindBlinding assessor: UnclearBL comparable: YesJadad score: 1,1,1,BRisk of bias: B
Segal 2000Multiple pub: No	Study design: RCTRandomization procedure: NRAllocation concealment: UnclearFollow‐up: 52w	Country: USASetting: NRNumber: 245Age: NRSex: NRMedications: Oral sulfonureasBL wt: NRBL BMI: NRBL GHb: NR	Drug: OrlistatDosage: 120mg tidDuration: 52wDiet: low calorieComparison: Placebo; unclear if dietary intervention	Weight: YesBMI:>5% loss (%):FBS:GHb:Cholesterol:LDL:HDL:TG:SBP:DBP:Side effects:	Funding: Hoffman La Roche, NJ, USAAbstract/full text: ALOCF: NRITT: NRAttrition: NRBlinding: Double‐blindBlinding assessor: UnclearBL comparable: NRJadad score: 1,1,0,BRisk of bias: C
Serrano‐Rios 2001Multiple pub:No	Study design: RCTRandomization procedure: NRAllocation concealment: UnclearFollow‐up: 26w	Country: SpainSetting: Multicenter; no other detailsNumber: 237Age: NRSex: NR Medications: Sulfonureas and/or metforminBL wt: NRBL BMI: >27BL GHb: NR	Drug: OrlistatDosage: 120mg tidDuration: 24wDiet: HypocaloricComparison: Placebo + diet	Weight: YesBMI: Yes>5% loss (%): YesFBS: YesGHb: YesCholesterol: LDL:HDL:TG:SBP: YesDBP: YesSide effects: Yes	Funding: NRAbstract/full text: ALOCF: NRITT: NRAttrition: NRBlinding: Double‐blindBlinding assessor: UnclearBL comparable: NRJadad score: 1,1,0,BRisk of bias: C
Tong 2002     Multiple pub:   Sea 2002 ;   unclear if related to Chan 2001 and Sea 2001	Study design: Pre‐versus‐postRandomization procedure: NAAllocation concealment: NAFollow‐up: 26w	Country: ChinaSetting: NRNumber: 27Age: 36Sex: 61%FMedications: NRBL wt: 93.2BL BMI: 34.2BL GHb: 8.5	Drug: OrlistatDosage: 120mg tidDuration: 26wDiet: NoneComparison: NA	Weight: YesBMI: Yes>5% loss (%):FBS: YesGHb: YesCholesterol: YesLDL: YesHDL: YesTG: YesSBP: YesDBP: YesSide effects: Yes	Funding: NRAbstract/full text: FTLOCF: NRITT: NAAttrition: NRBlinding: NABlinding assessor: NABL comparable: NAJadad score: NARisk of bias: NA
Vesari 2000Multiple pub: No	Study design: Pre‐versus‐postRandomization procedure: NAAllocation concealment: UnclearFollow‐up: 45d	Country: NRSetting: NRNumber: 21Age: 55ySex: 80%FMedications: 48% on oral agentsBL wt: NRBL BMI: 36.3BL GHb: NR	Drug: OrlistatDosage: 120mg bid to tidDuration: 45dDiet: 1500kcal/dComparison: NA	Weight:BMI: Yes>5% loss (%):FBS: YesGHb: YesCholesterol: YesLDL: YesHDL: YesTG: YesSBP:DBP:Side effects:	Funding: NRAbstract/full text: ALOCF: NRITT: NAAttrition: NRBlinding: NABlinding assessor: NoBL comparable: NAJadad score: NARisk of bias: NA
Wang 2003     Multiple pub:   No	Study design: RCTRandomization procedure: Randomization tableAllocation concealment: Unclear Follow‐up: 24w	Country: China Setting: ClinicNumber: 63Age: 41Sex: 47.6Medications: 100% oral agentsBL wt: I 85.0, C 83.0BL BMI: I 30.0, C 31.0 BL GHb: I 8.3, C 8.2	Drug: Orlistat Dosage: 120mg bid to tidDuration: 24wDiet: NRComparison: Placebo + diet	Weight: YBMI: Y>5% loss (%): YFBS: YGHb: YCholesterol: YLDL: YHDL: YTG: YSBP: YDBP: YSide effects: NR	Funding: NRAbstract/full text: FTLOCF: NRITT: 2 patients withdrawn (no reason stated)Attrition: 3.2%Blinding: NRBlinding pt: YesBlinding assessor: UnclearBlinding provider: UnclearBL comparable: yes
Zaletel 2002Multiple pub:No	Study design: Pre‐versus‐postRandomization procedure: NAAllocation concealment: NAFollow‐up: Unclear; second phase was 6m	Country: SloveniaSetting: UnclearNumber: 31Age: 54Sex: 58Medications: NRBL wt: NRBL BMI: 38.1BL GHb: NR	Drug: OrlistatDosage: 120mg tidDuration: UnclearDiet: UnclearComparison: NA	Weight: YesBMI:>5% loss (%):FBS: YesGHb: YesCholesterol: YesLDL:HDL:TG: YesSBP: YesDBP:Side effects:	Funding: NRAbstract/full text: ALOCF: NAITT: NAAttrition: 6%Blinding: NABlinding assessor: NoBL comparable: NAJadad score: NARisk of bias: NA

Methods	Study design: Pre vs post, retrospectiveRandomization procedure: NAAllocation concealment: NAFollow‐up: 26 weeks
Participants	Country: USA   Setting: Endocrinology clinic   Number: 23   Age: 53   Sex: NR   Medications: NR   BL wt: 118.0(2.5)   BL BMI: 40.5(7.0)   BL GHb: 7.9(1.6)
Interventions	Drug: Orlistat Dosage: 120mg tidDuration: 13 to 26 weeksDiet: NRComparison: NA
Outcomes	Weight: YBMI: Y>5% loss (%): YFBS: GHb: YCholesterol: YLDL: YHDL: YTG: YSBP: YDBP: YSide effects: Y
Notes	Funding: Abstract/full text: FTLOCF: NAITT: NAAttrition: NA (retrospective)Blinding: NA Blinding pt: No Blinding assessor: NABlinding provider: NoBL comparable: NA
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Unclear risk	D ‐ Not used

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Weight (kg)	7	1363	Mean Difference (IV, Fixed, 95% CI)	‐2.12 [‐2.63, ‐1.60]
2 Percent weight loss	4	1008	Mean Difference (IV, Fixed, 95% CI)	‐2.43 [‐2.95, ‐1.90]
3 % with wt loss > 5%	5	1273	Mean Difference (IV, Fixed, 95% CI)	21.39 [15.16, 27.62]
4 BMI	2	100	Mean Difference (IV, Fixed, 95% CI)	‐0.71 [‐1.52, 0.10]
5 Waist circumference	6	1111	Mean Difference (IV, Fixed, 95% CI)	‐1.02 [‐1.33, ‐0.70]
6 GHb	7	1373	Mean Difference (IV, Fixed, 95% CI)	‐0.45 [‐0.58, ‐0.31]
7 Fasting glucose	8	1449	Mean Difference (IV, Fixed, 95% CI)	‐0.70 [‐0.89, ‐0.51]
8 SBP	5	740	Mean Difference (IV, Fixed, 95% CI)	‐2.19 [‐3.84, ‐0.55]
9 DBP	4	441	Mean Difference (IV, Fixed, 95% CI)	‐3.94 [‐5.18, ‐2.71]
10 Total cholesterol	6	1324	Mean Difference (IV, Fixed, 95% CI)	‐0.41 [‐0.52, ‐0.30]
11 LDL cholesterol	6	1287	Mean Difference (IV, Fixed, 95% CI)	‐0.32 [‐0.42, ‐0.23]
12 HDL cholesterol	6	994	Mean Difference (IV, Fixed, 95% CI)	‐0.02 [‐0.05, 0.00]
13 Triglycerides	6	994	Mean Difference (IV, Fixed, 95% CI)	‐0.23 [‐0.40, ‐0.05]
14 Weight (kg)	7	1363	Mean Difference (IV, Random, 95% CI)	‐2.03 [‐2.82, ‐1.25]
15 Percent weight loss	4	1008	Mean Difference (IV, Random, 95% CI)	‐2.34 [‐2.97, ‐1.70]
16 % with wt loss > 5%	5	1273	Mean Difference (IV, Random, 95% CI)	21.39 [15.16, 27.62]
17 BMI	2	100	Mean Difference (IV, Random, 95% CI)	‐0.71 [‐1.52, 0.10]
18 Waist circumference	6	1111	Mean Difference (IV, Random, 95% CI)	‐1.84 [‐2.99, ‐0.68]
19 GHb	7	1373	Mean Difference (IV, Random, 95% CI)	‐0.45 [‐0.58, ‐0.31]
20 Fasting glucose	8	1449	Mean Difference (IV, Random, 95% CI)	‐0.82 [‐1.14, ‐0.50]
21 SBP	5	740	Mean Difference (IV, Random, 95% CI)	‐2.99 [‐6.29, 0.32]
22 DBP	4	441	Mean Difference (IV, Random, 95% CI)	‐4.21 [‐7.82, ‐0.61]
23 Total cholesterol	6	1324	Mean Difference (IV, Random, 95% CI)	‐0.41 [‐0.52, ‐0.30]
24 LDL cholesterol	6	1287	Mean Difference (IV, Random, 95% CI)	‐0.32 [‐0.43, ‐0.21]
25 HDL cholesterol	6	994	Mean Difference (IV, Random, 95% CI)	‐0.02 [‐0.05, 0.00]
26 Triglycerides	6	994	Mean Difference (IV, Random, 95% CI)	‐0.23 [‐0.40, ‐0.05]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Weight (kg)	10	2045	Mean Difference (IV, Fixed, 95% CI)	‐2.08 [‐2.39, ‐1.77]
2 Percent weight loss	10	2833	Mean Difference (IV, Fixed, 95% CI)	‐2.55 [‐2.91, ‐2.20]
3 % with wt loss > 5%	10	2871	Mean Difference (IV, Fixed, 95% CI)	21.59 [17.08, 26.09]
4 BMI	3	130	Mean Difference (IV, Fixed, 95% CI)	‐0.20 [‐0.33, ‐0.06]
5 Waist circumference	8	1647	Mean Difference (IV, Fixed, 95% CI)	‐1.13 [‐1.42, ‐0.85]
6 GHb	13	2898	Mean Difference (IV, Fixed, 95% CI)	‐0.42 [‐0.50, ‐0.34]
7 Fasting glucose	13	2737	Mean Difference (IV, Fixed, 95% CI)	‐0.73 [‐0.88, ‐0.58]
8 SBP	6	977	Mean Difference (IV, Fixed, 95% CI)	‐2.54 [‐3.95, ‐1.13]
9 DBP	5	678	Mean Difference (IV, Fixed, 95% CI)	‐3.54 [‐4.59, ‐2.49]
10 Total cholesterol	6	1324	Mean Difference (IV, Fixed, 95% CI)	‐0.38 [‐0.47, ‐0.30]
11 LDL cholesterol	7	1571	Mean Difference (IV, Fixed, 95% CI)	‐0.29 [‐0.36, ‐0.22]
12 HDL cholesterol	6	994	Mean Difference (IV, Fixed, 95% CI)	‐0.02 [‐0.04, ‐0.01]
13 Triglycerides	6	994	Mean Difference (IV, Fixed, 95% CI)	‐0.24 [‐0.38, ‐0.09]
14 Weight (kg)	10	2045	Mean Difference (IV, Random, 95% CI)	‐2.12 [‐2.67, ‐1.57]
15 Percent weight loss	10	2833	Mean Difference (IV, Random, 95% CI)	‐2.55 [‐2.91, ‐2.20]
16 % with wt loss > 5%	10	2871	Mean Difference (IV, Random, 95% CI)	21.59 [17.08, 26.09]
17 BMI	3	130	Mean Difference (IV, Random, 95% CI)	‐0.25 [‐0.56, 0.06]
18 Waist circumference	8	1647	Mean Difference (IV, Random, 95% CI)	‐1.71 [‐2.48, ‐0.94]
19 GHb	13	2898	Mean Difference (IV, Random, 95% CI)	‐0.42 [‐0.50, ‐0.34]
20 Fasting glucose	13	2737	Mean Difference (IV, Random, 95% CI)	‐0.81 [‐1.01, ‐0.60]
21 SBP	6	977	Mean Difference (IV, Random, 95% CI)	‐3.22 [‐5.93, ‐0.51]
22 DBP	5	678	Mean Difference (IV, Random, 95% CI)	‐3.85 [‐6.53, ‐1.18]
23 Total cholesterol	6	1324	Mean Difference (IV, Random, 95% CI)	‐0.38 [‐0.47, ‐0.30]
24 LDL cholesterol	7	1571	Mean Difference (IV, Random, 95% CI)	‐0.30 [‐0.38, ‐0.21]
25 HDL cholesterol	6	994	Mean Difference (IV, Random, 95% CI)	‐0.02 [‐0.04, ‐0.01]
26 Triglycerides	6	994	Mean Difference (IV, Random, 95% CI)	‐0.24 [‐0.38, ‐0.09]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Weight (kg)	8	845	Mean Difference (IV, Fixed, 95% CI)	‐3.60 [‐4.18, ‐3.01]
2 Percent weight loss	3	426	Mean Difference (IV, Fixed, 95% CI)	‐4.03 [‐5.52, ‐2.55]
3 % with wt loss > 5%	2	204	Mean Difference (IV, Fixed, 95% CI)	21.16 [12.48, 29.83]
4 BMI	6	517	Mean Difference (IV, Fixed, 95% CI)	‐1.29 [‐1.53, ‐1.04]
5 Waist circumference	5	475	Mean Difference (IV, Fixed, 95% CI)	‐4.13 [‐5.16, ‐3.10]
6 GHb	7	612	Mean Difference (IV, Fixed, 95% CI)	‐0.82 [‐0.97, ‐0.66]
7 Fasting glucose	5	434	Mean Difference (IV, Fixed, 95% CI)	‐1.27 [‐1.73, ‐0.82]
8 SBP	6	673	Mean Difference (IV, Fixed, 95% CI)	‐0.91 [‐1.61, ‐0.20]
9 DBP	4	480	Mean Difference (IV, Fixed, 95% CI)	1.43 [0.06, 2.79]
10 Total cholesterol	6	529	Mean Difference (IV, Fixed, 95% CI)	‐0.02 [‐0.12, 0.09]
11 LDL cholesterol	6	529	Mean Difference (IV, Fixed, 95% CI)	0.00 [‐0.11, 0.11]
12 HDL cholesterol	5	419	Mean Difference (IV, Fixed, 95% CI)	0.07 [0.03, 0.11]
13 Triglycerides	6	529	Mean Difference (IV, Fixed, 95% CI)	‐0.22 [‐0.35, ‐0.08]
14 Weight (kg)	8	845	Mean Difference (IV, Random, 95% CI)	‐5.10 [‐5.00, ‐3.20]
15 Percent weight loss	3	426	Mean Difference (IV, Random, 95% CI)	‐4.03 [‐5.52, ‐2.55]
16 % with wt loss > 5%	2	204	Mean Difference (IV, Random, 95% CI)	21.16 [12.48, 29.83]
17 BMI	6	517	Mean Difference (IV, Random, 95% CI)	‐1.87 [‐2.64, ‐1.10]
18 Waist circumference	5	475	Mean Difference (IV, Random, 95% CI)	‐4.68 [‐7.36, ‐1.99]
19 GHb	7	612	Mean Difference (IV, Random, 95% CI)	‐0.54 [‐1.32, 0.24]
20 Fasting glucose	5	434	Mean Difference (IV, Random, 95% CI)	‐1.35 [‐3.68, 0.99]
21 SBP	6	673	Mean Difference (IV, Random, 95% CI)	‐0.84 [‐1.65, ‐0.02]
22 DBP	4	480	Mean Difference (IV, Random, 95% CI)	1.43 [0.06, 2.79]
23 Total cholesterol	6	529	Mean Difference (IV, Random, 95% CI)	‐0.11 [‐0.37, 0.15]
24 LDL cholesterol	6	529	Mean Difference (IV, Random, 95% CI)	‐0.07 [‐0.31, 0.16]
25 HDL cholesterol	5	419	Mean Difference (IV, Random, 95% CI)	0.07 [0.03, 0.11]
26 Triglycerides	6	529	Mean Difference (IV, Random, 95% CI)	‐0.27 [‐0.50, ‐0.04]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Weight (kg)	9	863	Mean Difference (IV, Fixed, 95% CI)	‐3.53 [‐4.10, ‐2.96]
2 Percent weight loss	4	662	Mean Difference (IV, Fixed, 95% CI)	‐4.21 [‐5.54, ‐2.88]
3 % with wt loss > 5%	3	440	Mean Difference (IV, Fixed, 95% CI)	23.41 [15.10, 31.71]
4 BMI	6	517	Mean Difference (IV, Fixed, 95% CI)	‐1.29 [‐1.53, ‐1.04]
5 Waist circumference	5	475	Mean Difference (IV, Fixed, 95% CI)	‐4.13 [‐5.16, ‐3.10]
6 GHb	7	612	Mean Difference (IV, Fixed, 95% CI)	‐0.82 [‐0.97, ‐0.66]
7 Fasting glucose	5	434	Mean Difference (IV, Fixed, 95% CI)	‐1.27 [‐1.73, ‐0.82]
8 SBP	6	673	Mean Difference (IV, Fixed, 95% CI)	‐0.91 [‐1.61, ‐0.20]
9 DBP	4	480	Mean Difference (IV, Fixed, 95% CI)	1.43 [0.06, 2.79]
10 Total cholesterol	6	529	Mean Difference (IV, Fixed, 95% CI)	‐0.02 [‐0.12, 0.09]
11 LDL cholesterol	6	529	Mean Difference (IV, Fixed, 95% CI)	0.00 [‐0.11, 0.11]
12 HDL cholesterol	5	419	Mean Difference (IV, Fixed, 95% CI)	0.07 [0.03, 0.11]
13 Triglycerides	6	529	Mean Difference (IV, Fixed, 95% CI)	‐0.22 [‐0.35, ‐0.08]
14 Weight (kg)	9	863	Mean Difference (IV, Random, 95% CI)	‐4.77 [‐6.50, ‐3.04]
15 Percent weight loss	4	662	Mean Difference (IV, Random, 95% CI)	‐4.21 [‐5.54, ‐2.88]
16 % with wt loss > 5%	3	440	Mean Difference (IV, Random, 95% CI)	25.86 [13.25, 38.47]
17 BMI	6	517	Mean Difference (IV, Random, 95% CI)	‐1.87 [‐2.64, ‐1.10]
18 Waist circumference	5	475	Mean Difference (IV, Random, 95% CI)	‐4.68 [‐7.36, ‐1.99]
19 GHb	7	612	Mean Difference (IV, Random, 95% CI)	‐0.54 [‐1.32, 0.24]
20 Fasting glucose	5	434	Mean Difference (IV, Random, 95% CI)	‐1.35 [‐3.68, 0.99]
21 SBP	6	673	Mean Difference (IV, Random, 95% CI)	‐0.84 [‐1.65, ‐0.02]
22 DBP	4	480	Mean Difference (IV, Random, 95% CI)	1.43 [0.06, 2.79]
23 Total cholesterol	6	529	Mean Difference (IV, Random, 95% CI)	‐0.11 [‐0.37, 0.15]
24 LDL cholesterol	6	529	Mean Difference (IV, Random, 95% CI)	‐0.07 [‐0.31, 0.16]
25 HDL cholesterol	5	419	Mean Difference (IV, Random, 95% CI)	0.07 [0.03, 0.11]
26 Triglycerides	6	529	Mean Difference (IV, Random, 95% CI)	‐0.27 [‐0.50, ‐0.04]

Methods	Study design: RCT; some Pre‐versus‐post without comparison group   Randomization procedure: NR   Allocation concealment: Unclear   Follow‐up: 32w   Note: This study did not fit inclusion criteria as did not present weight outcomes, however it presented adverse event data among persons with diabetes, and is therefore presented here.
Participants	Country: UKSetting: Multicenter; details unclearNumber: 210Age: 54Sex: 59Medications: None (diet only)BL wt: NRBL BMI: NRBL GHb: NR
Interventions	Drug: Sibutramine dosage: 15‐20mg qdDuration: 32wDiet: NRComparison: Placebo
Outcomes	Weight:BMI:>5% loss (%):FBS:GHb:Cholesterol:LDL:HDL:TG:SBP:DBP:Side effects: Yes
Notes	Funding: Knoll Pharmaceutical Co.,US and UKAbstract/full text: FTLOCF: NRITT: Yes, with attritionAttrition: 11%Blinding: Double‐blindBlinding assessor: UnclearBL comparable: YesJadad score: 1,1,1,BRisk of bias: B
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Unclear risk	B ‐ Unclear

Methods	Study design: Cross‐over study comparing phenmetrazine hydrochloride with phenmetrazine hydrochloride plus phenbutrazate hydrochloride Randomization procedure: NR   Allocation concealment: Unclear   Follow‐up: 8w
Participants	Country: UKSetting: Hospital diabetes clinicNumber: 22Age: 58 from table 1Sex: 80%FMedications: NRBL wt: 78BL BMI: NRBL GHb: NR
Interventions	Drug: PhenmetrazineDosage: 25mg tidDuration: 8w (until first cross‐over)Diet: 1000 kcal/d Comparison: Filon® [phenmetrazine theoclate 30mg and phenbutrazate hydrochloride 20mg] tid with 1000 kcal/d diet
Outcomes	Weight: YesBMI:>5% loss (%):FBS:GHb:Cholesterol:LDL:HDL:TG:SBP:DBP:Side effects: Yes; dizziness (20%), abdominal discomfort and nausea (15%, and dry mouth 5%)
Notes	Funding: NRAbstract/full text: FTLOCF: NRITT: Yes, with attritionAttrition: 9%Blinding: Double‐blindBlinding assessor: UnclearBL comparable: NRJadad score: 1,1,0,BRisk of bias: B
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Unclear risk	B ‐ Unclear

Methods	Study design: RCT Randomization procedure: NRAllocation concealment: NRFollow‐up: 12w
Participants	Country: USASetting: Private practiceNumber: 10Age: Approximately 50Sex: Predominantly femaleMedications: NRBL wt: 85.0BL BMI: NRBL GHb: NR
Interventions	Drug: MazindolDosage: 1mg tidDuration: 12wDiet: Individual diet, no detailsComparison: Placebo + diet
Outcomes	Weight: YesBMI:>5% loss (%):FBS:GHb:Cholesterol:LDL:HDL:TG:SBP:DBP:Side effects: Yes; lightheadedness, dry mouth, vertigo; increased pulse rate noted with I group, not quantified.
Notes	Funding: NRAbstract/full text: FTLOCF: NRITT: Yes, with attritionAttrition: 10%Blinding: Double‐blindBlinding assessor: UnclearBL comparable: YesJadad score: 1,1,1,BRisk of bias: B
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Unclear risk	B ‐ Unclear

Methods	Study design: Unclear; 2 parallel groupsRandomization procedure: NRAllocation concealment: NRFollow‐up: 16w
Participants	Country: USASetting: NRNumber: 12Age: NRSex: NRMedications: NRBL wt: ave 143% ideal body weightBL BMI: NRBL GHb: NR
Interventions	Drug: PhentermineDosage: NRDuration: 16wDiet: 1000kal/dComparison: Placebo + diet
Outcomes	Weight: YesBMI:>5% loss (%):FBS: YesCholesterol: YesLDL:HDL:TG: YesSBP:DBP:Side effects:
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Unclear risk	B ‐ Unclear

Methods	Companion abstract to Gray 1992 and 1991
Participants
Interventions
Outcomes
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Unclear risk	D ‐ Not used

Methods	Study design: Two parallel groups, unclear if randomizedRandomization procedure: UnclearAllocation concealment: UnclearFollow‐up: 12w
Participants	Country: USASetting: NRNumber: 83Age: NRSex: NRMedications: NRBL wt: NRBL BMI: NRBL GHb: NR
Interventions	Drug: SibutramineDosage: 15mg qdDuration: 12wDiet: NRComparison: Placebo
Outcomes	Weight: YesBMI:>5% loss (%):FBS: YesGHb: YesCholesterol: YesLDL:HDL:TG: YesSBP:DBP:Side effects:
Notes	Funding: NRAbstract/full text: ALOCF: NRITT: NRAttrition: NRBlinding: Double‐blindBlinding assessor: UnclearBL comparable: NRJadad: 0,1,0,BRisk of bias: C
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Unclear risk	B ‐ Unclear

PERMALINK

Pharmacotherapy for weight loss in adults with type 2 diabetes mellitus

Susan L Norris

Xuanping Zhang

Alison Avenell

Edward Gregg

Christopher H Schmid

Joseph Lau

Abstract

Background

Objectives

Search methods

Selection criteria

Data collection and analysis

Main results

Authors' conclusions

Plain language summary

Background

Description of the condition

Description of the intervention

Why it is important to do this review

Objectives

Primary research question

Secondary research questions

Methods

Criteria for considering studies for this review

Types of studies

Length of follow‐up and timing of outcomes measurement

Full text and abstracts

Publication status

Types of participants

Age

Type 2 diabetes

Overweight or obese

Types of interventions

Centrally acting appetite suppressants:

Peripheral effect on appetite:

Nutrient partitioning:

Increase thermogenesis:

Combined drug therapy:

Comparison groups:

Types of outcome measures

Primary outcomes

Secondary outcomes

Exclusion criteria

Search methods for identification of studies

Electronic searches

Searching other resources

Data collection and analysis

Selection of studies

Data extraction and management

Assessment of risk of bias in included studies

Internal validity

Other design issues

Assessment of heterogeneity

Assessment of reporting biases

Data synthesis

Statistical pooling

Regression analyses

Subgroup analysis and investigation of heterogeneity

Sensitivity analysis

Results

Description of studies

1.

Fluoxetine, orlistat, and sibutramine

Risk of bias in included studies

Fluoxetine, orlistat and sibutramine

Effects of interventions

Fluoxetine, orlistat, and sibutramine

2.

3.

Narrative synthesis of other drugs

Discussion

Authors' conclusions

Implications for practice.

Implications for research.

Feedback

Clarification about references, 27 February 2009

Summary

Reply

Methods	Companion abstract to Hollander 1998a
Participants
Interventions
Outcomes
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Unclear risk	D ‐ Not used

Methods	Study design: Two study groups; pre‐versus‐postRandomization procedure: NAAllocation concealment: NAFollow‐up: 20‐240d
Participants	Country: ItalySetting: NRNumber: 50Age: 54Sex: 65%FMedications: 17% insulin; 67% oral agentsBL wt: I 97 , C 92 BL BMI: NRBL GHb: NR
Interventions	Drug: DiethylpropionDosage: 2‐3qd (dosage not specified)Duration: 20‐240dDiet: 1000‐1800kcal/dComparison: Both groups got same diet and dosage diethylpropion; group A was on hypoglycemic agents, group B was diet controlled
Outcomes	Weight: YesBMI:>5% loss (%):FBS: YesGHb:Cholesterol:LDL:HDL:TG:SBP:DBP:Side effects: Yes; per Pina: 4/50 quit for SE, including general malaise, epigastric disturbance, and dermatitis. No untoward effects in person with HT and CVD; normal LFT and renal function
Notes	Funding: NRAbstract/full text: FTLOCF: NRITT: Yes, with attritionAttrition: 8%Blinding assessor: NRBL comparable: NRJadad score: NARisk of bias: NA
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Unclear risk	D ‐ Not used

Methods	Study design: Two groups, unclear if randomized; cross‐over q6wRandomization procedure: NRAllocation concealment: NRFollow‐up: 6w
Participants	Country: AustraliaSetting: NRNumber: 18Age: 40‐65Sex: 80%FMedications: 11% diet, 61% oral agents, 28% insulinBL wt: NRBL BMI: NRBL GHb: NR
Interventions	Drug: MazindolDosage: 2mg qdDuration: 6wDiet: Dietary advice for 8w before onset of drug treatmentComparison: Placebo
Outcomes	Weight: YesBMI:>5% loss (%):FBS: YesGHb:Cholesterol:LDL:HDL:TG:SBP:DBP:Side effects: Yes; "stimulation", headache
Notes	Funding: NRAbstract/full text: FTLOCF: NRITT: Yes, with attritionAttrition: 17%Blinding: Double‐blindBlinding assessor: UnclearBL comparable: BLJadad score: NARisk of bias: B
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Unclear risk	D ‐ Not used

Methods	Sircar 2001Multiple pub: No
Participants	Study design: Pre‐versus‐postRandomization procedure: NAAllocation concealment: NAFollow‐up: 12w
Interventions	Country: IndiaSetting: UnclearNumber: 27Age: 44.7Sex: 89%Medications: NRBL wt: 75.4BL BMI: 32.1BL GHb: 9.6
Outcomes	Drug: SibutramineDosage: 10‐15mg qdDuration: 12wDiet: Prescribed; Unclear typeComparison: NA
Notes	Weight: YesBMI:>5% loss (%):FBS:GHb: YesCholesterol:LDL:HDL:TG:SBP:DBP:Side effects: Yes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Unclear risk	D ‐ Not used

Methods	Vargas 1994Multiple pub:No
Participants	Study design: RCTRandomization procedure: NRAllocation concealment: UnclearFollow‐up: 12w
Interventions	Country: USASetting: NRNumber: 18Age: NRSex: NRMedications: BRBL wt: NRBL BMI: NRBL GHb: NR
Outcomes	Drug: SibutramineDosage: 20‐30mg qdDuration: 12wDiet: NRComparison: Placebo
Notes	Weight: YesBMI:>5% loss (%):FBS: YesGHb:Cholesterol:LDL:HDL:TG:SBP:DBP:Side effects:
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Unclear risk	B ‐ Unclear

Study	Reason for exclusion
Anchors 1997	No diabetes population
Apfelbaum 1999	No diabetes population
Astrup 1985	No diabetes population
Astrup 1992	No diabetes population
Boneva 2002	No weight outcomes for diabetes subgroup
Bowen 2000	No diabetes population
Bray 1996	No diabetes population
Bray 1999	No diabetes population
Breum 1995	IGT and type 2 diabetes; can't separate the two populations
Broom 2001	No diabetes subgroup
Chengappa 2001	Goal is not weight loss
Conte 1973	No diabetes population
Daly 1993	No diabetes population
Darga 1991	Only 2 persons with diabetes
Davison 1999	No diabetes subpopulation outcome data
Derby 1999	No diabetes population
Drent 1995	No diabetes population
Duncan 1960	No diabetes population
Edmonds 1983	Goal is treatment of diabetic neuropathic edema, not weight loss
Egart 1979	No subgroup analysis
Enzi 1976	No diabetes population
Fanghanel 2000	No diabetes population
Faria 2001	No diabetes subgroup
Fava 1999	Not a weight loss study and no diabetes subgroup
Fernandez‐Soto 1995	No diabetes population
Finer 2000e	No diabetes population
Generali 2001	Review
Gokcel 2002a	No diabetes population
Gokcel 2002b	Only 10% with diabetes; no subgroup analysis
Goldstein 1993	No diabetes population
Goldstein 1994	No diabetes population
Greenway 1999e	No diabetes population
Hadler 1967	No diabetes population
Haller 2000	Review
Hanefeld 2002b	No weight outcomes
Hanotin 1998	No diabetes population
Hansen 2001	No diabetes population
Hauptman 1992	No diabetes population
Hauptman 2000	No diabetes population
Heal 1998	No diabetes patients
Heath 1999	Duplicate abstract with Rissanen
Heymsfield 2000	Meta‐analysis; no primary data
Hill 1999	No diabetes population
Hollenbeck 1987	No weight loss drug
Inoue 1992	No diabetes population
Inoue 1995	No diabetes population
Jacob 2002	No weight outcomes
James 1997	No diabetes population
James 2000	No diabetes population
Jones 1995	No diabetes‐specific data
Langlois 1974	No diabetes population
Lee 1999a	IGT population only; no diabetes population
Lee 1999b	IGT population; no diabetes population
Lustman 2000	Goal is not weight loss
Maetzel 2002	No weight outcomes, is an economic study
Maheux 1997	Goal is not weight loss
Malchow‐Moller 1981	No diabetes population
Marcus 1990	No diabetes population
McLaughlin 2001	No diabetes population
McMahon 2000	No diabetes population
Meier 1992	Goal is to decrease body fat, not weight loss
Michelson 1999	No diabetes subgroup analysis
Miles 2001	No weight outcomes
Miles 2002b	No weight outcomes
Pasquali 1987	No diabetes population
Pedrinola 1996	No diabetes population
Pijl 2000	Goal is to decrease body fat, not weight loss
Rasmussen 1993	No diabetes population
Rissanen 1999b	No weight outcomes
Rissanen 2000a	No weight outcomes
Rissanen 2000b	No weight outcomes
Rolls 1998	No diabetes population
Rosenfalck 2002	No diabetes population
Samsa 2001	No diabetes specific data
Sax 1991	No diabetes population
Seagle 1998	No diabetes population
Seedat 1974	No diabetes population
Shi 2001	No weight results for diabetes subgroup
Sirtori 1971	No diabetes population
Sjostrom 1998	No weight results for diabetes subgroup
Steel 1973	No diabetes subpopulation
Stoa‐Birketvedt 1993	No diabetes population
Tan 2002	Goal not weight loss; 8 hour follow‐up only
Thompson 1998	Not a weight loss drug
Toft‐Nielsen 1999	Not a weight loss study
Toplak 1998	No diabetes population
Torgerson 2001	No outcomes
Toubro 1993	No diabetes population
Van Gaal 1998a	No diabetes population
Van Gaal 1998b	No diabetes population
Vanloon 1992	Goal not weight loss
Vernace 1974	No diabetes population
Wadden 1995	No diabetes population
Wadden 1997	No diabetes population
Wadden 2001	No diabetes population
Walker 1977	No diabetes population
Wasada 2000	Goal is to decrease body fat, not weight loss
Wilding 1998	No weight outcomes
Wilding 1999	No diabetes population
Wilding 2001	No diabetes subgroup
Williams 1981	No diabetes subgroup
Wilson 1960	No diabetes subgroup
Wirth 2001	No population with diabetes
Woodhouse 1975	Experimental drug (AN448); no diabetes population
Yoshida 1994	No diabetes populaion
Zavoral 1998	No weight outcomes for diabetes subgroup
Ziegler 1971	Formula diet; not a weight loss drug