Influence of Intensive Diabetes Treatment on Body Weight and Composition of Adults With Type 1 Diabetes in the Diabetes Control and Complications Trial


Diabetes Care. 2001;24(10):1711-172. 

In This Article

Research Design and Methods

Experimental Design

The DCCT design and methods have been described in detail[3]. Randomization of the first patient occurred in August 1983, and the last person began assigned therapy in June 1989. Follow-up ranged from 3.5 to 9 years, with the study ending in May 1993. A total of 29 clinical centers and 1,441 patients participated. A coordinating center, central biochemistry laboratory, central nutrition coding unit, and four other central units analyzed and managed data.


DCCT eligibility criteria included insulin dependence for up to 15 years, as evidenced by deficient C-peptide secretion; diabetes duration for 1-15 years; age 13-39 years; and HbA1c >6.55% (>3 SD above the mean of a sample of nondiabetic people aged 13-40 years[4]). Other eligibility requirements included good general health and no more than moderate nonproliferative retinopathy. Candidates were excluded if they were obese, defined as body weight >130% of the ideal[5]. Subjects were also excluded if they required >2 units of insulin per kilogram of body weight per day. Detailed eligibility criteria and baseline characteristics of the entire cohort have been published[2,3]. The present report is restricted to the 1,246 subjects who were adults (age 18 years or older) when randomized.

Selected baseline characteristics are shown by treatment group in Table 1 . Males randomized to intensive treatment weighed significantly less and had lower BMIs than those in the conventional treatment group. They also had a 9-month longer average duration of type 1 diabetes. There were no other significant differences between groups in these variables at baseline.

Treatment Groups

The details of the two treatment regimens have been published[6]. Only those aspects relevant to this report will be reviewed here.

Conventional treatment. The clinical goals of conventional therapy included 1) absence of symptoms attributable to glucosuria or hyperglycemia, 2) absence of ketonuria, 3) maintenance of normal growth and development and ideal body weight, and 4) freedom from frequent or serious hypoglycemia. Insulin was administered by one or two injections per day and included mixtures of short-, intermediate-, or long-acting insulin. Self-monitoring was with urine or blood glucose testing, with the majority of patients performing daily blood glucose monitoring. Subjects were given individualized meal plans that specified amounts of food and meal times. There were no specific exercise protocols, but exercise was encouraged. Subjects in the conventional group were seen by the health care team every 3 months.

Intensive treatment. Intensive treatment had the same clinical goals as the conventional treatment group, with the additional goal of maintaining blood glucose control as close to the nondiabetic range as possible while minimizing hypoglycemia. Target ranges for glycemic control were: fasting and preprandial blood glucose 70-120 mg/dl; postprandial blood glucose <180 mg/dl; 3:00 A.M. blood glucose, tested weekly, >65 mg/dl; and monthly HbA1c ≤6.05% (mean + 2 SD of a sample of nondiabetic people aged 13-40 years). With the advice of the health care team, intensive therapy subjects could choose either multiple daily injections or continuous subcutaneous insulin infusion using an external pump. Insulin doses were guided by the results of self-monitoring of blood glucose performed at least four times per day, and they were further adjusted based on meal content and composition and anticipated exercise. Intensive treatment group subjects were hospitalized, usually for 3-4 days, to initiate therapy, and they were seen at least monthly thereafter. Although the intensive group followed a dietary protocol similar to that of the conventional group, greater attention was given to adjusting insulin and diet to achieve intensive therapy's glycemic targets.


Dietary guidelines for the DCCT. Meal plans and educational strategies were tailored to each individual, with a caloric allowance to achieve and maintain ideal body weight and with goals of 15-20% of the energy from protein, 30-35% from fat, and 50-55% from carbohydrates[7]. In response to the National Cholesterol Education Program (NCEP), the study protocol was amended in 1988 to provide all participants with counseling in the NCEP Step 1 diet[8]. Instruction in the Step 2 diet was given to subjects who did not attain target levels of LDL-cholesterol, despite adherence to the Step 1 diet[7].

Weight management for intensive therapy. Once the potential for weight gain with intensive therapy was recognized, dietitians increased their emphasis and counseling on weight management. To achieve target glycemic levels without hypoglycemia or weight gain, even greater attention was given to the relation between nutrient intake and insulin. Meal plans for patients assigned to intensive therapy often subtracted 250-300 kcal from estimated daily energy needs before the DCCT to maintain weight targets. Dietary advice also focused on a healthy eating approach, emphasizing low-fat choices and recognition of hunger needs versus appetite wants, and addressing life factors that influence food choices[9]. Other preventive strategies included teaching portion control and appropriate treatment of hypoglycemia, using food records, weight graphs, exercise programs, behavior modification, contracting, and goal-setting. Focus groups and weight management sessions were held to identify eating problems and discuss strategies to resolve them. Social and family pressures, eating away from home, food preparation methods, and preventing relapse were discussed[10].


The measurements described below were obtained at the eligibility, baseline, and quarterly examinations, unless otherwise noted. Height (in centimeters) was measured with a stadiometer. Weight (in kilograms) was measured with the subject in light clothing and stocking feet on the same balance-beam scale for the duration of the trial[7]. BMI was calculated by dividing weight (kilograms) by height (meters) squared.

The ratio of the natural waist-to-hip measurements (WHR) provides insight into the distribution of body mass[11]. This procedure was added to the study protocol in March 1992 as part of the assessment of body composition and was performed once during the last year of the trial. Circumference measurements of hip and waist were obtained in duplicate by study-certified dietitians using inelastic tapes. If they differed by >0.5 cm, they were repeated. The waist measurement was taken at the narrowest part of the torso in a horizontal plane when viewed from behind; the hip measurement was taken at the maximum extension of the buttocks, with the subject in a relaxed standing posture[7]. Fat-free body mass was estimated by tetrapolar bioelectrical body impedance analysis (BIA) using proximal electrode placement. To validate the use of this technique in type 1 diabetes, lean body mass was measured in a subset of 46 DCCT patients via dual-energy X-ray absorptiometry; a regression model specific to type 1 diabetes was developed from these data and applied studywide[12]. Percent body fat was calculated as the difference between total body weight and fat-free mass, expressed as a percentage of total weight.

Subjects' daily diets were assessed using a standardized dietary history at baseline, at years 2 and 5, and upon exit from the study[13]. Dietitians were trained and certified in the collection of these data[7]. Dietary data were analyzed by the DCCT Central Nutrition Coding Unit[14]. Activity levels at school, work, and during leisure time were estimated using a standardized questionnaire at baseline and annually thereafter. HbA1c measurements were determined at the Central Biochemistry Laboratory with high-performance liquid chromatography[4].

Statistical methods

Major weight gain was defined as an increase in BMI of at least 5 kg/m2 from baseline, which is equivalent to 20% weight gain, or 14 kg for most subjects[15]. Overweight was defined using the National Center for Health Statistics definition for adults: BMI ≥27.8 kg/m2 for men and ≥27.3 kg/m2 for women[16]. Finally, WHRs >0.85 for women or >0.9 for men were considered "increased"[17].

All data analyses were performed according to the original randomized treatment groups. Adherence to assigned therapy was 97 and 98% of study time for conventional and intensive therapy, respectively[2]. Data from all women who became pregnant during the trial were censored from the time of their first conception through study end, unless otherwise noted. All P values are two-sided and are reported at their nominal levels, i.e., without formal adjustment for multiple comparisons.

The Wilcoxon's rank-sum test[18] was used to test whether the distributions of continuous variables (including changes from baseline in continuous measures) differed between groups. A linear covariance adjustment was made to those outcome variables proving to be highly correlated with some baseline characteristics. The significance of within-group changes from baseline was assessed using Wilcoxon's signed-rank test for paired differences[18]. Fisher's exact test[19] was used to test binary variables for association with treatment group at a single point in time.

Consistent differences in the distributions of continuous variables over time were evaluated with the nonparametric test of stochastic ordering proposed by Wei and Lachin[20], weighting each of the univariate Mann-Whitney U test differences in proportion to the corresponding sample sizes[21], after adjusting for covariance with the baseline value of the outcome. Consistent group differences in repeated dichotomous measures, such as major weight gain, were examined using the test of stochastic ordering in the multiple nonindependent 2 x 2 tables of Lachin and Wei[22].

Two-stage random-effects models[23,24] were fit via restricted maximum-likelihood estimation[25] to the consecutive measurements of BMI to obtain concise summaries of its typical rates of change. Separate models were initially fit to each sex within each treatment group, including baseline BMI and time as covariates. To minimize the cohort effects associated with staggered entry, the corresponding empirical rates of change were computed as unweighted averages of the average annual changes during the period in question.

The growth-curve fittings were carried out by Program 5V of the BMDP statistical software package, version 1990[26]. All other analyses and all data management were carried out using SAS software, Version 6[27,28].


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: