(OR31-02-24) Descriptive Systematic Review and Meta-Analysis of Adherence to Randomized Dietary Intervention Weight Loss Trials

Objectives: Poor adherence is often cited as a limitation of long-term diet trials; however, the extent of adherence and its impact on weight loss has not been rigorously evaluated. We conducted a systematic review to estimate adherence to randomized dietary interventions and weight loss outcomes.

Methods: Our protocol was prospectively registered (PROSPERO CRD42021264367). We searched Medline and other databases for randomized controlled trials (RCTs) through 6/2021. We included RCTs ≥6 months duration testing effects of diet reporting weight loss as the primary outcome and any measures of diet adherence (N≥100). We summarized adherence outcomes across treatment groups via meta-analysis of means with inverse-variance weighting and calculated correlation coefficients between mean adherence vs. weight change.

Results: We identified 47 RCTs comparing weight loss effects of 108 dietary interventions among 8952 participants with overweight or obesity (96% adults). Most only reported adherence as group-level mean nutrient intakes from 24hr recalls, food records, or food frequency questionnaires; less common were self-assessment scores and objective biomarkers. Among trials reporting energy intake, there was no difference in total kcal for energy-restricted vs. ad libitum intervention groups (mean intakes 1507 vs. 1608 kcal, respectively; p-heterogeneity=0.55). For carbohydrates, average intake during carb-restricted diets was 24.8% kcal, in contrast with 45.7% among non-carb-restricted diets (p < 0.001). For total fat, intake during low-fat interventions was 28.5% kcal compared with 36.9% kcal without fat restriction (p < 0.001). Despite macronutrient reductions, none of the low-carb and only 10% of low-fat treatment groups’ average intakes were within the intervention goals. Among the carb-restricted interventions, groups achieving lower %kcal from carbohydrate had less weight loss (r=-0.41). Conversely, among fat-restricted diets, lower %kcal from fat was modestly correlated with greater weight loss (r=0.21).

Conclusions: Overall, adherence to dietary interventions is poorly defined and reported. Indirect inference from group-level intakes suggests inadequate adherence to macronutrient restriction and minimal adherence to energy restriction by end of follow-up.

Funding Sources: American Diabetes Association Research Award