Quitting smoking and mortality : a population-based mortality case-control study in Hong Kong

Abstract

Background: Our knowledge of the health benefits of quitting smoking is mainly derived from prospective cohort studies in developed countries/regions, but such evidence is scarce and urgently needed in developing countries/regions. Mortality case-control studies could be a quick and low cost alternative to cohort study design. Smoking-attributable mortality was estimated using mortality case-control studies, but the effects of quitting have not been reported. Evidence on whether the benefits of quitting can be observed in mortality case-control studies is needed.

Objectives: 1. To critically review the methods used in previous mortality case-control studies. 2. To examine the association between quitting smoking and total- and cause-specific mortality in Hong Kong, using the Hong Kong LIfestyle and MORtality (HK LIMOR) data. 3. To analyse and compare the effects of using different definitions of cases and controls on the association between the health benefits of quitting and total- or cause-specific mortality using the HK LIMOR data.

Methods: The HK LIMOR study, a large population-based case-control study, included 81% of all registered deaths at age 30+ years from all four Hong Kong death registries in 1998, with information on duration of quitting collected in the second half of the year. Different definitions of cases and controls were used in previous mortality case-control studies. The main analysis followed the definitions in Sitas et al 2013. Cases were deaths from diseases that could be affected by smoking (N=19,526); controls were deaths from all non-smoking related causes (N=6,076). Two other definitions (Liu et al 1998 and Alam et al 2013) were used for sensitivity analyses. Unconditional logistic regression was used to calculate adjusted odds ratios of total-cause and cause-specific mortality by smoking and quitting status, duration of quitting, and age at quitting, adjusting for sex and age at death.

Results: 1. Mortality case-control study design was used to estimate smoking-attributable mortality only in developing regions, except Hong Kong. There were considerable variations in data collection methods and definitions of cases and controls. 2. Compared with continuing smokers, sex- and age-specific (middle age at death [35-64 years] and old age [65-94 years]) risk reductions from quitting were observed for total deaths of interest and 9 other smoking-related diseases. Dose-response relations across mortality risk reductions with quitting duration and age at quitting were found. 3. Varied definitions of cases and controls in mortality case-control studies yielded different results in middle age at death by quitting status, duration of quitting and age at quitting for total deaths of interest, lung cancer and cardiovascular disease mortality, but the results were similar in old age.

Conclusion: The HK LIMOR is the first mortality case-control study that has shown health benefits (mortality risk reduction) of quitting smoking. Our results are consistent with the findings from cohort studies. In addition, the mortality case-control study design seems to be more practicable to estimate and monitor the harms of smoking and the benefits of quitting in both developed and developing regions.