Latent class growth analysis successfully identified subgroups of participants during a weight loss intervention trial

Introduction: Weight loss interventions often present small mean weight changes over time, despite the fact that a substantial proportion of the participants lost more weight. This effect is often leveled out by the substantial proportion of participants who gained weight during the trial. The aim of this study is to identify and describe distinct subgroups of participants with different weight change trajectories during and after a weight loss intervention. Methods: We used data from a weight loss intervention that was part of a randomized controlled trial on the preventive effect of a tailormade weight loss intervention and oral glucosamine sulfate on the incidence of knee osteoarthritis in 407 overweight women aged 50 to 60 years. Latent class growth analysis (LCGA) was used to identify subgroups of participants with different weight change trajectories over time. Results: Using LCGA, we identified three subgroups with different trajectories of weight change, one large group (n = 298) with almost no change over time, and two smaller groups (both n = 48), of which one represents participants who steadily gained weight over time, whereas the other represents participants who steadily lost weight over time. Participants that had relatively low body weight around their 40th year of life and that gained weight in the year preceding the study were most likely to belong to the group that lost weight. Conclusion: LCGA was a suitable method to identify three distinct groups of participants with different trajectories of weight change. Low body weight at age 40 and weight gain in the year preceding the study were associated with a higher chance of membership of the group that lost weight. It seems weight loss that occurred during this weight loss intervention was mostly recently gained weight. (C) 2014 Elsevier Inc. All rights reserved.