RESUMO
BACKGROUND: Meta-analyses of randomized controlled trials (RCTs) have shown the beneficial effect of exercise on HDL-cholesterol (HDL-C) levels. However, systematic reviews are not free of bias, and this could call into question their results. OBJECTIVES: The aim of this work was to conduct a critical assessment of meta-analyses of RCTs that analyze the association between exercise and HDL-C levels, evaluating their results and the risk of bias (RoB). METHODS: This systematic review of MEDLINE and EMBASE included meta-analyses of RCTs that studied the effects of exercise on HDL-C levels in healthy adults or patients at cardiovascular risk. The RoB was determined using AMSTAR-2, and information was obtained on exercise and the variation in HDL-C levels. RESULTS: Twenty-three meta-analyses were included. Great variability was found in exercise (different types, frequencies or intensities in the studied interventions). All the analyses found an improvement in HDL-C levels, ranging from 0.27 to 5.41 mg/dl, in comparison with the control group (no exercise). The RoB was very high, with 18 reviews obtaining a critically low confidence level and the remaining works obtaining the highest confidence level. CONCLUSIONS: Only one meta-analysis showed good quality, in which HDL-C levels increased by 3.09 mg/dl in healthy adults and patients at high cardiovascular risk who practiced yoga. The rest had high RoB. Therefore, new systematic reviews with low RoB are needed to apply the results to clinical practice. Register: CRD42020158471 (PROSPERO).