The mediation and interaction of body mass index between night shift frequency and hypertension among nursing staff

Background Night shift work is associated with obesity and hypertension prevalence among nursing staff, and Body Mass Index (BMI), as an important indicator of obesity, may be involved in this association. This study examined the mediation and interaction components of BMI in the association between night shift frequency and hypertension prevalence. It also explored variations across gender, parental hypertension status, and hospital level. Methods A cross-sectional survey conducted from November 2016 to July 2018 included 143,772 nursing staff from 512 hospitals in 11 cities in Hebei Province, China. Using the four-way decomposition method in R software, mediation and interaction components were evaluated, with effect sizes reported as estimates and 95% confidence intervals (95% CI). Results Analysis of 127,715 participants showed significant associations between night shift frequency per month (“>0 and ≤ 5”, “>5 and ≤ 10”, “>10”, and “irregular”) and hypertension prevalence [Estimates and 95% CI: 1.09 (1.06, 1.11), 1.18 (1.12, 1.23), 1.28 (1.20, 1.37), and 1.38 (1.27, 1.51), respectively; P < 0.001]. Larger controlled direct effect (CDE) estimates were observed among males, whereas significant interaction components between BMI and night shift frequency were identified among females. Nursing staff with parental hypertension and those working in primary-level hospitals showed larger CDE and interaction component estimates, particularly among those with frequent or irregular night shifts. Conclusion Higher night shift frequency was significantly associated with higher hypertension prevalence among nursing staff, with direct associations accounting for most of the observed associations. Significant BMI-related mediation and interaction components were mainly observed among nursing staff in primary-level hospitals with frequent or irregular night shifts, suggesting that these BMI-related pathways may be context-dependent.