The authors concluded that the percentage of weight lost as fat-free mass in medical interventions is influenced by the level of calorie restriction, exercise and rate of weight loss, and that diversionary surgery is associated with a greater percentage fat- free mass loss than biliopancreatic laparoscopic adjustable banding. Limitations in reporting and review methods weaken the robustness of the conclusions.

To evaluate the changes in fat-free mass (FFM) of weight loss interventions associated with significant changes in weight.

MEDLINE and EMBASE were searched from inception to April 2006 for studies published in English language journals; the search terms were reported. The reference lists of retrieved publications were also checked. In addition, the journals Obesity Research, International Journal of Obesity, and Obesity Surgery were handsearched for the years 1990 to 2005.

Study designs of evaluations included in the review

Studies had to record the period of follow-up to be eligible for inclusion; no other inclusion criteria were defined. The included studies were randomised controlled trials (RCTs), clinical controlled trials and observational studies.

Specific interventions included in the review

Studies of dietary, behavioural, pharmaceutical and surgical interventions were eligible for inclusion. The included studies assessed the following: low calorie diets (LCD) with and without exercise (including aerobic and resistance exercise), orlistat or sibutramine; very low calorie diets (VLCD) with and without exercise; biliopancreatic diversion (BPD); roux en Y bypass (RYGB) and laparoscopic adjustable banding (LAGB).

Participants included in the review

Studies of adults who were obese or significantly overweight at baseline (body mass index, BMI, greater than 28) were eligible for inclusion. The baseline BMI ranged from 28.7 to 36.9 in the included medical intervention studies, and from 40.5 to 50.1 in surgical intervention studies.

Outcomes assessed in the review

Studies that recorded a significant weight loss (greater than 10 kg) and reported a change in FFM (in kg), percentage fat mass or fat mass (in kg), or sufficient data for their calculation, using dual-energy X-ray absorptiometry, underwater weighing, air displacement plethysmography or deuterium-oxide dilutometry for total body water were eligible for inclusion. Studies that measured FFM using bioelectrical impedance or total body potassium were excluded. The included studies used dual-energy X-ray absorptiometry, underwater weighing and deuterium-oxide dilutometry for total body water. The review also included studies that measured FFM using multislice whole body magnetic resonance imaging (MRI).

How were decisions on the relevance of primary studies made?

The authors did not state how the papers were selected for the review, or how many reviewers performed the selection.

The authors did not state that they assessed validity.

The authors did not state how the data were extracted for the review, or how many reviewers performed the data extraction. Data were extracted on each cohort timepoint. For each cohort, changes in weight and FFM were extracted and used to calculate the percentage of weight loss consisting of FFM (%FFML).

How were the studies combined?

The authors stated that surgical and medical weight loss studies could not be generally compared due to the range in baseline BMI, time to follow-up and magnitude of weight loss, therefore the studies were grouped by intervention. Studies that measured FFM using MRI were considered separately. Data from controlled studies and observational studies were pooled using a fixed- effect model without weighting by the sample size. Where two or more cohort timepoints for an intervention were reported the median and interquartile range (IQR) or mean and standard deviation were calculated. Comparisons between interventions were evaluated using the Mann-Whitney U-test or Student's t-test. Log transformation was used on skewed data.

How were differences between studies investigated?

In studies of dietary and behavioural interventions, linear regression analysis was used to assess the following as predictors of %FFML: degree of calorie restriction, exercise, weight loss and gender. In studies of surgical interventions, linear regression was used to assess the following as predictors of %FFML: type of surgery (LAGB versus BPD), level of weight loss, initial BMI and gender. Additional pooled analysis was calculated for men and women separately, and the number of cohorts in each intervention that reported above average gender- weighted %FFML was presented and compared using chi-squared analysis (details of methods used for gender weighting were described).

Sixteen studies (n=683) with 26 cohort timepoints evaluated dietary, behavioural and pharmaceutical interventions. Seventeen studies (n=995) with 29 cohort timepoints evaluated surgical interventions. Three RCTs (n=127) evaluated FFM loss measured using MRI.

Dietary and behavioural intervention.

The median %FFML was 14.0 (IQR 10) for LCD, 23.4 (IQR 8) for VCLD and 22.5 (IQR 11) for VCLD plus exercise. Linear regression showed that %FFML was greater with VLCD (p=0.006), and that there was a tendency for higher %FFML in men compared with women.

RCTs evaluating exercise.

Three RCTs reported that LCD plus supervised exercise for 16 weeks was associated with a consistent reduction in %FFML. Women lost significantly less %FFML than men (p<0.001).

Pharmaceutical studies.

Two observational studies reported that sibutramine plus LCD was associated with 31% and 38% FFML, respectively. One RCT compared orlistat plus LCD with LCD alone and reported no difference in %FFML.

Surgical treatment.

The median %FFML was 25.6 (IQR 11) for BPD, 31.3 (IQR 12.2) for RYGB and 17.5 (IQR 3.7) for LAGB. Both BPD and RYGB were associated with significantly greater %FFML than LAGB (p=0.002). Linear regression analysis suggested that the %FFML was significantly greater with BPD and RYGB than LAGB (p<0.001) and that the difference was not influenced by baseline BMI or the extent of weight loss.

%FFML and association with gender.

Dietary and behavioural interventions were associated with 27% FFML in men and 20% FFML in women. The proportion of cohort timepoints reporting greater than average %FFML loss was 100% for VLCD and RYGB, 75% for BPD, 40% for VLCD plus exercise, 8% for LCD and LAGB, and 0% for LCD plus exercise.

The percentage of weight lost as FFM in medical interventions was influenced by the level of calorie restriction, exercise and rate of weight loss. BPD and RYGB were associated with greater %FFML than LAGB.

The review addressed a clear question that was defined in terms of the participants, interventions and outcomes; the inclusion criteria for study design were broad. Several relevant sources were searched but no attempts were made to minimise either publication or language bias. The methods used to select the studies and extract the data were not described, so it is not known whether any efforts were made to reduce reviewer error and bias. Study validity was not assessed, therefore results from these studies and any synthesis might not be reliable.

Data were pooled without weighting and data from each cohort (gender or timepoint) were treated as separate studies, which might undermine the reliability of the analysis and subsequent review findings. However, the authors did note this limitation. A further limitation is that the majority of evidence was based on before-and-after data; this reliance upon what are effectively observational data weakens the strength of the evidence. Comparisons between surgical interventions were based on indirect comparisons and might not be reliable. Overall, there appear to be shortcomings in the available evidence, and limitations in the reporting of the review methodology and some of the methods of analysis mean that the results might not be reliable.

Practice: The authors did not state any implications for practice.

Research: The authors stated the need for further research to evaluate the long-term effect on body composition of weight loss due to VLCD and the association between the extent of weight loss and %FFML after BPD. They recommended that studies evaluating all novel medical and surgical weight loss interventions should include measurement of changes in fat and FFM.

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.