Food intake patterns and development of obesity

This PhD dissertation was accepted by the Faculty of Health Sciences of the University of Copenhagen, and defended on October 21, 2003.

Official opponents: Erik Lykke Mortensen, graduate in psychology, Lars Ulrik Gerdes and Lauren Lissner Östlund, nutritional epidemiologist, Sweden.

Tutors: Merete Osler, Berit L. Heitmann, and Thorkild I.A. Sørensen.

Correspondence to: Per Togo, Kong Oscars Gade 9, 2. th., DK-2100 Copenhagen Ø.
E-mail: per.togo@dadlnet.dk

Dan Med Bull 2004;51:152.

ABSTRACT

The aim of the PhD-study was to investigate the association between food intake pattern and development of obesity in a Danish population. The dissertation was based on a review of related studies, and analyses of epidemiological data from the Copenhagen County Research Centre for Prevention and Health. The work was carried out at the Research Centre and the Research Unit for Dietary Studies/ Danish Epidemiology Science Centre, at the Institute of Preventive Medicine, Copenhagen University Hospital.

The material was collected in 1982-1994 as part of the DanMONICA surveys. Data from between 1796 and 3485 men and women were analysed using the different designs. Food intake patterns were identified by factor analyses of food-frequency questionnaire data from the subgroup of subjects who also completed 7-day diet records. Individual factor-scores were associated with concurrent body mass index (BMI) and subsequent 5- and 11-year BMI change, controlling for scores on the other food intake factors, age, education, physical activity, smoking, parity (initial BMI and preceding BMI change).

The review of 30 studies including analyses of the association between food intake pattern (identified by a diet index, cluster or factor analysis) and BMI/obesity, showed that similar food intake patterns were observed in different populations, but the association with obesity was inconsistent. Furthermore the studies were all cross-sectional and the comparison was hampered by the different methods of identification of food intake patterns. In the analysed Danish population, three food intake factors for men: `Green', `Sweet' and `Traditional' and two for women: `Green' and `Sweet-Traditional' were identified. The factors were similar to factors identified in other populations, reproducible in the remaining subgroup of subjects (without diet record data) and also identified in diet record data. For men, significant inverse associations between the `Sweet' factor score and concurrent BMI, and between the `Traditional' factor-score and subsequent 11-year BMI change were found. For women, significant inverse associations between the `Sweet-Traditional' factor-score and both concurrent BMI and subsequent 5-year BMI-change were observed. However, no significant associations were observed in the fully adjusted analyses of factor-scores, factor-score changes and changes in BMI/odds of being obese after 11 years follow-up.

These associations between food intake patterns and obesity were modest and gender specific. To establish a better knowledge of the relation with obesity development, analyses of food intake patterns should be further developed and applied on other data, including longitudinal assessments of diet and weight changes.