Corrigendum to “Sleep duration and metabolic syndrome: An updated systematic review and meta-analysis” [Sleep Med. Rev. 59 (2021) 101451] (Sleep Medicine Reviews (2021) 59, (S1087079221000368), (10.1016/j.smrv.2021.101451))

The authors regret that used pre-converted values of 95% confidence intervals (95%CI) in the studies by Yang et al., 2016 and Kim et al., 2015 to calculated pooled risk ratios (RRs) of metabolic symptom. As stated in the manuscript, “among studies that ORs were provided, we calculated RRs by using the ORs and control event rates (CERs) in normal duration sleepers reported in the original studies", we converted the odds ratios (ORs) to RRs in the studies by Yang et al., 2016 and Kim et al., 2015. However, because the values of pre and post-converted 95%CI were very close we did not convert the corresponding 95% CIs and reported the pre-converted values of 95% CIs for these two studies. These 95%CIs should have been converted. The converted values of the 95% CIs for Kim et al. and 2015 and Yang et al., 2016 in Fig. 2 should be 1.05–1.58 and 0.72–1.53 respectively, and the pooled RR for the short sleep duration group should be 1.27 (1.08–1.50) instead of the reported 1.28 (1.07–1.53) (Fig. 2 and Table 2). In Fig. 3, the converted values of the 95% CIs for Kim et al., 2015 and Yang et al., 2016 should be 0.45–1.13 and 1.06–1.35 respectively, and the pooled RR for long sleep duration group should be 1.15 (0.95–1.38) instead of the reported 1.16 (0.95–1.41) (Fig. 3 and Table 2). The corresponding results for the sensitive analyses were also corrected (Fig. S1). While making the above corrections we found additional errors which required correction. Choi et al., 2011 (male group) was mis-grouped into age >60 years subgroup and Chang et al., 2015 was mis-grouped into sleep duration<6 hours subgroup, respectively (Table 1). In Table 2, regarding short sleep duration, the pooled RR for age≤60 years subgroup in longitudinal studies should be 1.30 (1.07–1.57) instead of the reported 1.33 (1.09–1.61). Regarding long sleep duration, the pooled RR should be 1.16 (0.90–1.48) instead of the reported 1.14 (0.88–1.46) for the age ≤60 years subgroup. The pooled OR for age >60 years subgroup for short sleep duration in cross-sectional studies should be 0.99 (0.94–1.03) instead of the reported 0.99 (0.93–1.07) (Table 2). Finally, the pooled OR for sleep duration <5 hours subgroup in cross-sectional studies should be 1.33 (1.04–1.71) instead of the reported 1.36 (1.04–1.78) (Table 2 and Fig. 4). The corrected Tables 1 and 2, Figs. 2–4 and Fig. S1 were shown in supplementary materials. Correcting these marginal errors did not change any of the overall findings and conclusions of the review. The authors would like to apologise for any inconvenience caused.[Formula