Long-Term Diets and Changes in Body Composition and Bone Mineral Density: A Systematic Review and Meta-analysis

Abstract Context The long-term effects of diets on body composition (BC) and bone mineral density (BMD) remain unclear. Objective To examine the effect of weight-loss inducing dietary interventions on BC and BMD at ≥12 months. Data Sources We conducted a systematic search on Medline, Embase, PubMed, and the Cochrane Library databases from inception until March 2025. Data Extraction We included randomized controlled trials (RCTs) of ≥12 months evaluating dietary interventions in adults (body mass index BMI ≥25 kg/m2) on BC and BMD. We categorized diets into moderate macronutrients (MM), low-fat/high-carbohydrate (LFHC), high-fat/low-carbohydrate (HFLC), and usual diet (UD). We completed data screening/extraction in duplicate and independently. We used RevMan 5.3 for pooling and the Cochrane tool for risk of bias (ROB) assessment. Results Of 34 382 records, we included 23 trials reporting on BC and BMD. Most RCTs had high or unclear ROB. Baseline BMI was 28.6-37.7 kg/m2 and most interventions lasted 12 months. Dual-energy X-ray absorptiometry was the primary BC assessement tool used (72%). Nine trials compared MM with HFLC diets and 5 trials compared MM diets with UDs. HFLC diets reduced fat mass (FM) compared with MM diets (2 trials; mean difference MD, −2.88 kg; 95% CI, −5.41 to −0.34). Compared with a UD, the MM diet did not reduce FM (2 trials; MD, −11.36 kg; 95% CI, −26.36 to 3.63). There was no difference in the FM change between HFLC and LFHC diets. Similalry, there was no difference in the visceral adipose tissue and the subcutaneous adipose tissue parameters across different diets, with only few exceptions. Four RCTs evaluated BMD parameters, and one trial only showed a possible reduction in bone loss with a high-protein compared to a low-protein diet. Conclusion HFLC diets modestly reduced FM when compared to MM diets. Conversely, FM did not differ between MM diets and UDs. The impact of diets on other BC parameters did not differ. Data on BMD parameters are very scarce, with a possible benefit from a high-protein diet. More high-quality trials are needed. Systematic Review Registration PROSPERO no. CRD42018103116.