The study, published in the 1 January issue of the Annals of Internal Medicine, examined the evidence for the major commercial and organized self-help weight loss programmes in the United States that provide in-person or online counselling.
It only included programmes supported by randomized trials of at least 12 weeks long that included a follow-up evaluation lasting at least one year. These were eDiets.com, Health Management Resources, Take Off Pounds Sensibly, OPTIFAST, and Weight Watchers.
"Of three randomized, controlled trials of Weight Watchers, the largest reported a loss of 3.2 per cent of initial weight at two years," write authors Adam Gilden Tsai and Thomas A. Wadden from the University of Pennsylvania.
Research on very-low-calorie diet programmes showed patients who completed treatment lost approximately 15-25 per cent of initial weight but these were associated with high costs, high attrition rates, and a high probability of regaining 50 per cent or more of lost weight in one to two years.
Commercial interventions available over the Internet and organized self-help programmes produced minimal weight loss, added the authors.
"Because many studies did not control for high attrition rates, the reported results are probably a best-case scenario."
"With the exception of one trial of Weight Watchers, the evidence to support the use of the major commercial and self-help weight loss programmes is suboptimal. Controlled trials are needed to assess the efficacy and cost-effectiveness of these interventions," they concluded.
Despite the lack of evidence for weight loss diets, the weight loss foods category is booming. Sales of the Weight Watchers food range in the UK increased 38 per cent from 2002 to 2003, and 20.9 per cent from 2003 to 2004, according to the firm.
Almost one third of people living in the European Union are overweight and more than one in ten is now obese, according to European Association for the Study of Obesity, and with this figure climbing, the trend in weight loss foods looks set to continue.