This was a study of a 1-year group-based dietary/behavioural program combined with orlistat following 3 months of dietary counselling in patients with SRBD. Our main findings were that the program was associated with reduced body weight, but compliance to dietary recommendations deteriorated somewhat during treatment. Dietary energy density, E% from saturated fat and intakes of fatty dairy products increased during the 1-year period. In the multiple regression analyses, further loss of body weight was associated with increased intake of protein and lower E% saturated fat and fatty dairy product intakes. These variables explained almost 20% of the variation in weight reduction. The strengths of our study are that we have follow-up data on all subjects included in the dietary/behavioural program and comprehensive dietary data. A limitation is the lack of a control group however the study was designed to demonstrate the results of treatment in a usual clinical setting.
Despite behavioural and pharmacological support, compliance with the diet worsened in all subjects during the year of treatment. Deterioration of dietary compliance after an initial intervention during long-time follow-up is common . Our data suggest that even treatment with a pharmacological aid as orlistat and moderately intense behavioural support does not prevent this deterioration. However we did observe that the group with high attendance showed better dietary maintenance than the low attendance group.
The increase in the intake of fat during 1 year is surprising in light of the mechanisms of action of orlistat which may lead to gastrointestinal problems if dietary fat is not restricted. Gastrointestinal symptoms are thought to occur with a dietary intake of fat >30% of total energy , but individual tolerability may vary. Thus, subjects may increase their intake of fatty dairy products without experiencing gastrointestinal problems. It has also been speculated that subjects treated with orlistat learn to titrate their dietary intake of fatty foods . Another possibility is that some subjects did not use the medication as prescribed.
We found that energy density increased during the year of follow-up however, change in energy density was not associated with weight loss. On the other hand, the high attendance group maintained the reduction in energy density during treatment with orlistat more successfully than the low attendance group. A reduced energy density has been associated with weight loss in two previous studies [18, 19]. The PREMIER study showed improved weight loss in subjects with the largest reduction in energy density compared to the subjects with the lowest reduction . The study by Ello-Martin et al. showed that subjects who reduced fat intake with a concomitant increase in vegetables and fruit lost more in weight and reported less hunger compared to subjects that only reduced their fat intake  supporting the notion that energy density is a determinant of weight loss.
We found that increased protein intake was associated with improved weight loss, in line with previous data [20–22]. Dietary protein is associated with increased energy expenditure and less hunger . We speculate that less hunger may have particular importance for subjects with SRBD since reduced sleep quality may interfere with appetite .
The amount of weight loss during the year of treatment with orlistat was about 5%. However, the subjects had already lost about 3 kg of weight prior to the initiation of orlistat. This total change is in accordance with the weight reduction achieved in a sibutramine assisted weight loss program in obese males with SRBD . In this study, favourable changes in the frequency of disordered breathing events and cardio-metabolic risk factors were shown concomitant with the loss in weight [24, 25]. A limitation of our study is that we did not measure the apnoea - hypo apnoea index after weight reduction. However, most of the participants used continuous positive airway pressure for symptomatic treatment of their SRBD before weight loss was initiated .