Mortality after partner's cancer diagnosis or death: A population-based prospective cohort study in Japan

Takeshi Makiuchi,Masako Kakizaki,Tomotaka Sobue,Tetsuhisa Kitamura,Hiroshi Yatsuya,Taiki Yamaji,Motoki Iwasaki,Manami Inoue,Shoichiro Tsugane,Norie Sawada
DOI: https://doi.org/10.2188/jea.JE20240114
2024-08-24
Abstract:Background: The health statuses of closely connected individuals are interdependent. Little is known about mortality risk associated with partner's cancer diagnosis and cause-specific mortality risk associated with partner's death. Methods: Relative risks for all-cause and cause-specific mortality following a partner's cancer diagnosis or death compared to the period when the partner is cancer-free and alive were investigated in the population-based prospective cohort study that enrolled 140,420 people at the age between 40-69 in 1990-1994. Results: 55,050 participants (27,665 men and 27,385 women) who were identified as married couples were followed-up for 1,073,746.1 (518,368.5 in men and 555,377.6 in women) person-years, during which 9,816 deaths (7,217 in men and 2,599 in women) were observed. After a partner's cancer diagnosis, the mortality rate ratio (MRR) of all-cause mortality was not increased among both men and women, while an increase of externally-caused MRR was observed. The suicide MRR significantly increased among men (MRR = 2.90 [95% CI, 1.70-4.93]) and it persisted for more than 5 years. After a partner's death, the MRRs of all-cause, cardiovascular disease (CVD), respiratory disease (RD), and externally-caused mortality significantly increased only among men. Stratified analysis by smoking status among men showed significantly increased MRRs of CVD and RD mortality among former/current smokers, but not among never-smokers. Conclusion: Partner's cancer diagnosis did not increase all-cause mortality risk, but increased externally-caused mortality risk, especially suicide among men. The impact of partner's death on mortality risk differed by the mortality causes and sex, and smoking affected some of cause-specific mortality risk.
What problem does this paper attempt to address?