Weight Changes in General Practice.

INTRODUCTION: This PhD thesis is about weight changes. What determines long-term weight changes in the adult general population? Is it possible that weight loss may not always be healthy? The present clinical guidelines for general practice advice most overweight persons and patients with type 2 diabetes to lose weight. Are the guidelines based on firm evidence?   METHODS: The back-bone of the thesis is constituted by three scientific articles based on three different population based cohort studies. Multivariable modeling and other epidemiological methods were used.   RESULTS: Article 1 examined weight changes in the general population in relation to smoking status, and proposed a graphical 'smoking cessation weight change model', demonstrating the importance of time, age and smoking status in relation to long-term weight changes. Article 2 suggested new methods to improve the processing of dietary data. It was demonstrated how median imputation for missing values and assumptions about standard portion sizes were inferior to stochastic methods conditioning on information about physiology of the individual. Article 3 evaluated the influence of prospectively planned intentional weight loss on long-term morbidity and mortality in patients with type 2 diabetes. Therapeutic intentional weight loss supervised by a medical doctor was not associated with reduced morbidity or mortality. In the general population the dietary intake of fructose and soft drinks sweetened with sugar was not associated with weight change over 9 years. Weight gain rates were large in young adults and incrementally smaller in middle aged adults. Subjects more than 60 years lost weight on average. Historical weight data suggest that the body weight increases throughout life to the age of 60-65years. A study with simulated data indicates that bias in baseline BMI may misleadingly have favored weight loss in earlier cohort studies of intentional weight loss and mortality.   DISCUSSION: The findings regarding weight loss and mortality in patients with type 2 diabetes are in opposition to the prevailing observational literature. Harrington's meta-analysis of intentional weight loss and the underlying studies are evaluated along with the Look AHEAD trial and a number of diabetes prevention studies. Difficulties in conducting and interpreting weight change studies are discussed.   CONCLUSIONS: Surprisingly, intentional therapeutic weight loss in patients with type 2 diabetes, supervised by a medical doctor, did not seem to reduce the long-term risk for CVD, CVD-mortality or all-cause mortality. The contradictions between our results and the prevailing observational evidence may be explained by methodological weaknesses favoring weight loss in earlier studies. Consequently, there is no good evidence to support that intentional weight loss will reduce the risk of CVD or mortality in any group of patients in general practice or in the general population. Age was a powerful determinant of weight changes and the 'normal weight development' can be taken into consideration when evaluating weight studies, and when general practitioners are following their patients over time. Compared with age, sex, education, and comorbidity, lifestyle factors like the dietary intake and physical activity seemed to be of less importance for long-term weight development. An exception to this was smoking or smoking cessation. Based on the scientific literature in the field and on the results of article 3, it seems uncertain whether weight loss is beneficial or harmful in terms of mortality and cardiovascular morbidity in patients with diabetes and in overweight people in general. Improvements in for instance psychosocial factors and diabetes prevention may well be short term as only few are able to a maintain weight loss. Rather than going for weight loss in overweight high risk patients, it seems more rational for general practitioners to focus on other lifestyle changes like for instance Mediterranean diet and increased exercise.