Dietary counselling for weight loss - moderate effects, but the evidence is not strong
Dietary counselling is modestly effective for encouraging weight loss, but its effects seem to wane with time, according to this meta-analysis: the available evidence is, however, mostly of only poor to fair quality. The authors of this study note that obesity-related health problems are a major health-issue: in the US, around 65% of adults are overweight and around half of these are obese (defined as a BMI >30kg/m2). Previous reviews have indicated that dietary counselling can produce appreciable weight loss, however they are not clear on the extent of weight loss achievable, especially long term. This analysis aimed to update the evidence.
The authors carried out a literature search for randomised trials that investigated the effect of dietary counselling on BMI, compared to a controls (generally usual care or a minimal intervention). The starting point for the search was a previous major systematic review published in 1998. Exclusion criteria included studies in children, those where mean baseline BMI was less than 25, those lasting less than 12 weeks, and those not reporting effects at 16 weeks or more. For analysis, the net change in BMI (change from baseline in study group less change from baseline in controls) and its SE were extracted from the published data. Study quality was assessed as good, fair, or poor.
The initial literature search identified over 13,000 citations, of which 102 were potentially eligible: 56 were excluded because they did not fit the defined eligibility criteria, leaving 46 for analysis. There were 63 study groups, 26 involving diet only and 37 diet and exercise; these involved about 6,400 participants who received dietary counselling and about 5,500 controls. A wide range of different methods of delivering the interventions were used, including group meetings, individual meetings, a combination of these, and the internet. Only four studies (9%) were considered to be good quality, 63% were fair quality, and 28% were poor quality (mainly due to very high rates of withdrawal, incomplete reporting, and unclear analyses).
Analysis indicated a maximum net treatment effect of just under 2 BMI units over 12 months. This was equivalent roughly to 6% body-weight loss or 5kg at one year. Effects were shown during the active phases of interventions, with slow weight regain during maintenance. There was some indication that combining dietary counselling with exercise improved outcomes, but the difference was not statistically significant in most studies. The data available suggests that overall, after the intervention is finished, patients will return to their previous weight after about five to six years. The authors comment that their analysis adds new information to previous reviews, however there are significant limitations because of deficiencies in the underlying data. The trials were statistically and clinically heterogeneous, making analysis and interpretation difficult; there were also problems with missing data, which may be accounted for in ways that could either overestimate or underestimate the treatment effect. They conclude that their analysis shows dietary counselling-based weight loss shows modest effects relative to usual care, with diminishing returns over the course of the intervention. While the changes may appear modest, these may still have clinically significant effects: good quality trials to investigate this are needed.
Ann Intern Med 2007; 147: 41-50 (link to abstract)