Exercise capacity in obstructive sleep apnoea patients: An individual patient data and aggregate meta-analysis

A previous aggregate data (AD) meta-analysis showed that obstructive sleep apnoea patients had lower exercise capacity compared to controls (Mendelson et al. ERJ 2018). However, these authors reported a high level of heterogeneity, which may be explained by methodological limitations including an OSA definition apnoea-hypopnoea index (AHI) which varied from 5 to 30. The aim of this meta-analysis was to confirm whether exercise capacity is altered in moderate to severe OSA patients using individual patient data (IPD) and AD meta-analysis. A two-stage IPD meta-analysis model was used, allowing to standardize the apnea cutoff (AHI ≥15) and adjust for confounders. Nonrandomized trials performing sleep recording and exercise test with a measured of maximal oxygen consumption (VO2peak) were searched using PubMed, Embase and Web of Science. From the 666 studies identified, 18 met inclusion criteria. IPD were sought for this 18 studies and obtained from 5 trials (28% of all studies, N=289). For the 13 other studies, AD were used in sensitivity analyses (N=605). IPD meta-analysis demonstrated that OSA patients had similar VO2peak (mean difference (MD): -1.0 mL.kg-1.min-1; 95% CI: -3.8 to 1.8; p=0.47) with low heterogeneity compared to mild or non-OSA patients while AD meta-analysis revealed that VO2peak was reduced in OSA patients (MD: -2.3 mL.kg-1.min-1; 95% CI: -4.0 to -0.6; p<0.001) with high heterogeneity (I²=77%; p<0.001). Subgroup AD analysis revealed significant difference in severe OSA. In conclusion, IPD meta-analysis suggests that exercise capacity is preserved in moderate to severe OSA patients while AD meta-analysis suggests lower VO2peak in severe OSA.