[Eur Heart J 2019;40:1149–1157].
During a subsequent analysis of one of the cohorts analyzed and reported in the above paper, the authors discovered that some of the respiratory events were mislabeled by the polysomnography system in one of the two cohorts (MrOS) and were not included in the analysis of hypoxic burden. The analysis for the MrOS cohort was repeated using the corrected hypoxic burden index (Sleep Heart Health Study data were unaffected by this). The point estimates for the paper’s main models got somewhat stronger in the re-analyzed data, but did not materially change the interpretation of the paper. The paper has now been amended accordingly in print and online. The supplementary material online has also been amended.
The authors apologise for the error.