OBJECTIVE: The aim of this study was to investigate the effect of categorical strategies based on subject-specific characteristics of dietary restraint and physical activity, on weight maintenance (WM) in overweight and obese subjects, after a very low calorie diet (VLCD). Furthermore, find predictors of WM that can be important in the context of obesity treatment. METHODS: In all, 120 subjects (age: 49.0+/-9.8 y, BMI 31.0+/-3.8 kg/m(2)) followed a VLCD (2.1 MJ/day) for 6 weeks in a free-living situation, followed by a period of 1 y WM. Body weight (BW), body composition, leptin concentration, dietary restraint and physical activity were determined right before (t0) and after (t1) the VLCD, after 3 months (t2) and after 1 y (t3). During WM, subjects were divided into four categories of WM guidelines (dietary, activity, diet+activity, placebo), taking their capability measured during weight loss and their preference for particular guidelines into account. RESULTS: BW loss during VLCD was 7.0+/-3.1 kg. After 1 y follow-up, BW regain was 56.3+/-55.0%, without significant differences between the four groups. With respect to measured capability during weight loss, subjects with an increased dietary restraint (F1 of the Three Factor Eating Questionnaire) had less BW regain than subjects with an increased physical activity (Baecke questionnaire) (35.5+/-53.2 vs 68.5+/-46.4%, P<0.05). Moreover, activity guidelines promoted WM in dietary disciplined subjects compared to activity-related disciplined subjects (% regain: 25.2 vs 74.3%, P<0.05). Subjects receiving guidelines that were opposite to their preference showed a better WM than subjects receiving preferred guidelines. After 1 y WM, 21 subjects were successful (<10% BW regain) and 99 unsuccessful (> or =10% BW regain). At baseline, these groups were significantly different in BMI (resp. 32.7+/-4.9 vs 30.7+/-3.5 kg/m(2), P<0.05), waist circumference (106.5+/-14.0 vs 100.6+/-11.2 cm, P<0.05) and fat mass (FM) (35.2+/-10.6 vs 32.1+/-6.6%, P=0.06). Finally, successful subjects appeared to spare fat-free mass (FFM) to a greater extent than unsuccessful subjects. CONCLUSION: After weight loss, type of guidelines (dietary, activity, placebo) is not related to the magnitude of WM, whereas guidelines opposite to the subject's capability and preference are related. These only reach successful WM (>90%) in originally dietary disciplined subjects who are supported by characteristics such as a relatively high baseline BMI, waist circumference and FM, together with the ability to spare FFM.