Dietary Fat Intake Predicts 1-Year Change in Body Fat in Adolescent Girls With Type 1 Diabetes

Twenty-six girls who had type 1 diabetes for at least 2 years were recruited in this study. Twenty-three of these were followed up after 1 year. Among the girls completing the study, two were using insulin pumps at baseline and four at follow-up; 17% were in Tanner stage 2–3 and 83% were in Tanner stage 4–5 at baseline. Four were being treated for celiac disease and one for hypothyroidism. The girls who were lost to follow-up had a mean age of 15.5 years, mean BMI of 25.4 kg/m², mean HbA_1c (A1C) of 8.5% and mean daily dosage of insulin of 1.0 unit · kg⁻¹ · day⁻¹ at baseline. The control girls were identified as the females of the same age and from the same residential area appearing next to the diabetic patients in the register of the county taxation authorities. Twenty-three control girls were recruited, 19 of whom agreed to participate in the follow-up. Of the control girls completing the study, 31% were in Tanner stage 2–3 at baseline. Control girls who did not participate in the follow-up measurements had a mean age of 17.6 years and a mean BMI of 23.9 kg/m². Clinical characteristics of participants completing the study are presented in Table 1. All subjects and their parents gave informed consent. The study was approved by the Ethics Committee of Örebro County Council.

Pubertal stage assessments were made on the basis of a self-estimating manual (“Flickboken”), using schematic drawings of different Tanner stages (11).

Dual-energy X-ray absorptiometry (Lunar, Madison, WI) was used to assess body composition. The coefficients of variation of body composition measurement, determined by three measurements on each of three phantoms with a fat content of 10, 20 and 40 kg, were 10.4, 1.7, and 0.3%, respectively.

The measurement and analysis of physical activity and food intake have been described in detail previously (10). Physical activity and dietary intake were measured simultaneously over a 7-day period with a uniaxial accelerometer (model WAM 6471; Computer Science and Applications, Schalimar, FL) and a 7-day food diary (“Menyboken,” Swedish National Food Administration), respectively.

A1C was measured by high-pressure liquid chromatography with the Mono-S standard (12). The reference level for healthy persons is 3.5–5.3%.

Differences among the groups for clinical characteristics, body composition, physical activity, and energy intake were assessed by an unpaired t test or Mann-Whitney U test. Associations between changes in body composition variables (fat mass, percent body fat, and lean body mass) with energy intake and physical activity were assessed by ANCOVA, adjusting for study group. Further adjustments were made for Tanner stage and body composition (fat mass or lean body mass) at baseline. Interactions terms between study group and Tanner group, with energy-adjusted fat intake were included in the models to examine whether case-control group or Tanner stage modified the associations between the main exposures with change in body composition. The results of the ANCOVA are presented as a regression slope (β coefficient). In the diabetic group the associations of daily dosage of insulin and A1C with change in body composition (i.e., fat mass or lean body mass) variables were analyzed by multiple regression analysis, with change in body composition as the dependent variable, adjusting for the body composition variable of interest at baseline. StatView, version 5.0.1, was used for all statistical analyses. The level of statistical significance was set at P < 0.05.

Data on body composition are presented in Table 1. Lean body mass, total fat mass, and percent body fat increased significantly in both groups. There was no significant difference between girls with diabetes and control girls regarding change in any of the body composition variables.

Physical activity did not differ significantly between the groups (type 1 diabetes 463 ± 128 cpm, control girls 541 ± 153 cpm; P = 0.10). Neither did we observe any significant difference between the groups in total daily energy intake (type 1 diabetes 8.4 ± 1.8 MJ/day, control girls 9.6 ± 2.8 MJ/day). Energy-adjusted carbohydrate intake was significantly lower in diabetic girls (50 ± 6 vs. 53 ± 5%; P < 0.05), whereas energy-adjusted protein intake was significantly higher (16 ± 2 vs. 13 ± 2%; P < 0.0001) than in the healthy control girls. No significant difference was observed between groups for energy-adjusted fat (34 ± 6 vs. 33 ± 5%) and fiber intake (20 ± 7 vs. 17 ± 6 g/day).

Changes in percent body fat and body fat mass were positively associated with energy-adjusted fat intake (P < 0.01 and P < 0.05, respectively) and inversely associated with energy-adjusted carbohydrate intake (P < 0.01 and P < 0.01, respectively) (Fig. 1). Total physical activity (counts per minute) and total energy intake were not associated with changes in percent body fat or in body fat mass.

Changes in lean body mass were not associated with energy-adjusted carbohydrate or fat intake but were positively associated with the total amount of physical activity (β = 0.006 [kilograms per counts per minute], P < 0.01). No significant interactions were observed for any of the analyses.

There were no significant associations between daily dosage of insulin or A1C and changes in percent body fat or in lean body mass in multivariate models. However, a significant difference was found between the 1-year increase in percent body fat between those treated with four daily dosages of insulin (n = 9) compared with those treated with six doses (n = 6) (0.8 ± 2.6 and 3.6 ± 2.1%-units body fat, respectively, P < 0.05).

We found that high dietary fat intake adjusted for total energy intake significantly predicted gain in body fat in adolescent girls with type 1 diabetes and in healthy control girls. The total amount of physical activity was not associated with changes in body fat mass, but high levels of physical activity significantly predicted gain in lean body mass. No significant interactions were observed, indicating that these associations did not differ between type 1 diabetic and healthy adolescent girls.

The role of dietary fat intake in the development of obesity in childhood is controversial (13–17). However, it is possible that higher intakes of dietary fat contribute to an increase in total energy intake because dietary fat has a weaker effect on satiation compared with carbohydrates (18,19) and is more energy dense. The fact that we did not observe any association between total energy intake and change in percentage body fat may be explained by under-reporting of total dietary intake, which is common in adolescents (20). It is also plausible that dietary fat has a specific role in the accumulation of body fat, because excess energy from dietary fat intake is more easily stored as body fat than is excess energy derived from carbohydrates (21).

Insulin is a potent inhibitor of lipolysis (22) and stimulates lipogenesis (23), which could result in greater fat accumulation in girls with diabetes due to the effect of subcutaneous insulin therapy, resulting in peripheral hyperinsulinemia (24). This may then contribute to the development of overweight often observed in diabetic girls, and it could be hypothesized that the association between fat intake and percent body fat would be steeper in girls with diabetes compared with control girls (interaction). However, our results indicated that higher dietary fat intake predicted a similar increase in percentage body fat in both diabetic girls and control girls.

Our small sample size and relatively short follow-up time may be limiting factors to detect a differential association between fat intake and change in body fat. However, our precise measure of body composition at baseline and follow-up and the detailed characterization of activity patterns and dietary intake suggest that our results are not due to chance or measurement error.

In summary, our results suggest that high dietary fat intake predicts gain in percent body fat similarly in adolescent girls with diabetes and in healthy control girls. Furthermore, physical activity is associated with a gain in lean body mass in adolescent girls. Our findings emphasize the need for dietary advice to decrease fat intake and increase levels of physical activity in adolescent girls with and without type 1 diabetes to avoid unhealthy gains in fat mass.

The study was supported financially by the Research Committee of Örebro County Council and the Swedish Child Diabetes Foundation (Barndiabetes fonden).

The authors thank Linda Sjölander and Ingegerd Karlsson for contributing to the data collection and Lolita Mörk for analyzing the food records.

Part of this article were presented in abstract form at the annual meeting of the International Society for Pediatric and Adolescent Diabetes, St. Malo, France, 3–6 September 2003.