This review concluded that diet/lifestyle and pharmacologic weight loss interventions provided modest weight loss and may improve markers of cardiovascular risk factors in at-risk groups. Limitations of the included studies meant that applicability of the findings to clinical practice was restricted. The conclusion is appropriately cautious given the limitations of the available data.

To assess the long-term efficacy of weight loss methods in reducing weight and the effects of weight loss on cardiovascular risk factors.

MEDLINE, HealthSTAR, and Cochrane Central Register of Controlled Trials were searched from inception to September 2003. No language restrictions were applied. Search terms were specified. Relevant bibliographies were reviewed and experts consulted to identify other relevant articles.

Randomised controlled trials (RCTs) and non-randomised studies of dietary/lifestyle, pharmacological and surgical interventions for overweight and obese adults (body mass index≥25 kg/m²) with at least two years of follow-up were eligible for inclusion. Duration of follow-up required for studies of pharmacological interventions was reduced to one year due to the limited number of studies available. Studies were required to have at least 100 consecutively recruited participants. Only interventions approved for long-term clinical use were eligible for inclusion.

Where reported: most participants in the included studies were female; mean age ranged from 34.9 years to 59 years; baseline weight appeared to vary from 77.7kg to 142kg; duration of (nonsurgical) intervention ranged from three months to 48 months; loss to follow-up ranged from 3% to 66%; and duration of follow-up ranged from one year (pharmacologic interventions) to over four years. Within pharmacologic therapies, all trials were placebo controlled; most controls also involved low calorie diets and many involved counselling. Many participants had comorbid conditions including type 2 diabetes, impaired glucose tolerance and hypertension.

Dietary/lifestyle interventions included low-calorie diets, behaviour counselling, phone contact and exercise. Surgical interventions included gastroplasty, gastric bypass and gastric banding. Most pharmacologic interventions used orlistat (120mg) or sibutramine (15mg to 20mg); two studies used metformin (850mg).

Two reviewers independently assessed eligibility. Disagreements were resolved by consensus.

Studies were assessed for: adequacy of randomisation (randomised trials) or adequate adjustment for confounders (non-randomised trials); comparability between groups; loss to follow-up; reliable and valid measurements; interventions clearly defined; outcomes considered; and appropriate analysis. Trial quality was assessed as good if all criteria were met, fair if not all criteria were met fully but no major flaws and poor if major flaws were identified that substantially reduced trial quality.

Two reviewers extracted data on weight loss efficacy: body weight (kg) at baseline/one year/two years/three years/four years and proportion of participants with at least 5% weight loss. Percentage change in cardiovascular risk factors were extracted (low-density and high-density lipoprotein cholesterol, triglycerides, fasting blood glucose, systolic and diastolic blood pressure). For pharmacologic interventions, odds ratios were extracted for number of patients in the treatment compared to the placebo arm who achieving weight loss of at least 5% of baseline either 12 or 24 months after the start of the trial.

Data were extracted separately for all participants (last observation carried forward) and for participants who completed the trial.

A narrative synthesis was used for dietary/lifestyle and surgical interventions due to clinical heterogeneity within these categories.

For pharmacologic interventions, meta-analysis was performed for orlistat and sibutramine for the outcome of at least 5% weight loss. Both fixed-effect and random-effects pooled estimates were calculated using inverse variance weighting. Χ² test was performed to assess heterogeneity and considered significant if p<0.10, in which case the reported results were based on the random-effects rather than fixed-effect estimate.

Forty-four studies were included in the systematic review: 16 studies (n=5,698 participants, range 100 to 2,382) of dietary/lifestyle therapy; 19 studies (n=9,953 participants, range 194 to 2,155) of pharmacologic therapy; and nine studies (n=91 participants, range 91 to 1,478) of surgical therapy. Thirty-two studies were RCTs and 12 were cohort studies. Thirty-four studies were assessed as fair quality, eight were assessed as good quality and two as being of poor quality.

Pharmacologic therapy: Where reported, mean weight at baseline ranged from 87kg to 102kg, loss to follow-up ranged from 7% to 57% and weight loss after 12 months varied from 3.9kg to 10.3kg and after 24 months from 4.9kg to 8.9kg. Orlistat therapy was associated with a higher odds of achieving a weight loss of at least 5% after 12 months (OR 2.54, 95% CI 2.17 to 2.90; 10 studies), and after 24 months (OR 4.55, 95% CI 1.99 to 10.4; two studies). This was also true for Sibutramine therapy after 12 months (OR 5.03, 95% CI 2.27 to 11.1; two studies) and 24 months (OR 4.44, 95% CI 1.87 to 10.5; two studies). The results of Χ² tests for heterogeneity were not reported, but the results quoted all used random-effects rather than fixed-effect models, which indicated that there was statistically significant heterogeneity. The two trials that assessed metformin therapy indicated modest weight-loss reduction.

Dietary/lifestyle therapy: Mean baseline weight varied from 78kg to 116kg. Loss to follow-up ranged between 0% and 66%. Weight loss was usually less than 5kg between two and seven years from the start of intervention.

Surgical therapy: Where reported and applicable, mean baseline weight varied from 110kg to 142kg and loss to follow-up or withdrawal ranged from 5% to 42%. Weight loss was reported as between 28kg and 76kg after two years and between 17kg and 73kg after three years.

Effects on cardiovascular risk factors were reported in a minority of studies (markers reported and effects varied between studies).

The authors concluded that dietary/lifestyle therapy provided less than 5kg weight loss after two to four years, pharmacologic therapy provided 5kg to 10kg of weight loss after one to two years and surgical therapy provided between 25kg and 75kg weight loss after two to four years. The effect of weight loss of at least 5% body weight on cardiovascular risk factors seemed greatest for high-risk groups and to be intervention specific. Weight loss studies have methodological limitations that limit applicability to clinical practice.

This review addressed a clear research question supported by relevant and clearly defined inclusion criteria. A number of databases were searched without language restrictions. Attempts made to identify further studies. Study selection and data extraction were conducted in duplicate, which reduced risks of error and bias; it was unclear whether the same process was used for quality assessment. Study quality was assessed and reported in aggregate. Most studies were assessed as being of fair quality. Relevant details of the primary studies were provided, but information about diet and lifestyle interventions was fairly limited. There was a lack of investigation of clinical heterogeneity within the narrative syntheses and a lack of investigation of statistical heterogeneity within the meta-analyses. The authors' overall conclusions appeared reasonable, but their generalisability was unclear due to a lack of investigation into variation in the interventions and populations, as well as statistical heterogeneity, within the three groups of interventions. The conclusions about the effect of the interventions on cardiovascular risk factors are appropriately cautious given the limited data available.

Implications for practice: The authors stated that there was a need for additional interventions (such as lipid-lowering and antihypertensive drug therapy) in obese people with cardiovascular risk factors to attain long-term and clinically significant improvements in these risk factors. For obese people without cardiovascular risk factors, maintaining a stable weight was a reasonable outcome. There should be increased focus on preventing obesity.

Implications for research: The authors stated that additional studies that used standardised criteria to assess long-term weight loss efficacy and clinical significance were needed.

Not stated.

This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.