In this research paper by Little and colleagues (BMJ 2023;380:e074224, doi:10.1136/bmj‑2022‑072924, published 9 February 2023), incorrect data for ERR (excess relative risk)/Gy for one study (Chen et al 2019) were used. The corrected entries use grade ≥3 cardiac events, rather than all cause mortality, resulting in ERR/Gy using a heart dose of 0.756 (95% confidence interval (CI) −0.397 to 4.108) and using a left ventricle dose of 0.044 (−0.679 to 2.399). We also overlooked ERR/Gy for ischaemic heart disease and other heart disease in another study (Kashcheev et al 2017). These entries (ischaemic heart disease, 0.42 (95% CI 0.25 to 0.60); heart disease other than ischaemic heart disease, 0.36 (95% CI 0.14 to 0.59)) should replace the ERR/Gy data used in the meta-analysis for the same endpoints in this cohort, derived from an earlier analysis of this cohort (Ivanov et al 2006, 2007).
The analyses have been updated accordingly, with minimal changes to the estimates reported in the original publication. Corrections have been made to tables 2-6 in the main article and in web appendices 1 (supplement S3 tables S3.1, S3.3-S3.7) and 2 (supplement S5). The conclusions of the paper are unaltered.
