Mortality and morbidity risk prediction for older former smokers based on a score of smoking history: evidence from UK Biobank and ESTHER cohorts

Abstract

Abstract Background Rapid population ageing has raised the proportion of older former smokers considerably, but a comprehensive assessment tool of former smoking-related health risks is absent. Objective We utilised the large-scale data of UK Biobank and ESTHER study to build a former smoking score (FSS) for older former smokers using three major former smoking traits: pack-years, smoking duration and time since smoking cessation. Design UK Biobank and ESTHER study are two cohorts of older adults with 502,528 and 9,940 participants from the UK and Germany, respectively. Methods Smoking history and covariates were retrieved from the self-administrated questionnaires and mortality and morbidity data were obtained through regular linkages to hospital records. Results We constructed the FSS based on the 94,446 former smokers of UK Biobank by retrieving the averaged effect estimates of each trait with a 100-time random sampling. This score was robustly associated with higher risks of mortality and incidence of major smoking-related diseases, outperforming each trait. In the validation panel of 2,683 former smokers from ESTHER study, the FSS was highly predictive of mortality and morbidities. Particularly, compared with the 1st quartile of the FSS group, the 4th quartile group had 114.1, 104.5 and 158.9% higher risks of all-cause, CVD and cancer mortality, respectively, and 41.9, 31.9, 52.4 and 831.3% higher risks of incident CVD, type 2 diabetes, any cancers and lung cancer, respectively. Conclusions Our study demonstrates the large potential of refined risk assessment of former smokers by more comprehensive consideration of the major traits of former smoking.