Abstract
INTERPRETATION: Hb and respiratory infection showed a nonlinear U-shaped association; such a relation is modified by the chronotype sleep behavior. OBJECTIVE: To examine the association between Hb and the incidence of hospitalized respiratory infection, and to explore potential modification effects of sleep behaviors. METHODS: Included were 292,568 individuals without respiratory disease, cancer, or anemia diagnosis in the United Kingdom Biobank . Hb (g/dL) was measured at baseline. The interaction between Hb and sleep behaviors, including sleep duration, insomnia, chronotype, and daytime sleepiness with respiratory infection, was tested. RESULTS: The cohort was followed up at a median 12.6 years, and 16,669 incident respiratory infections (9,334 in men, 7,335 in women) were identified. There was a nonlinear U-shaped association between Hb and respiratory infection in both men and women, where the risk increased markedly with Hb above 15.0 g/dL for men and 13.5 g/dL for women. In men, compared with the third quintile group, the hazard ratio (HR; 95% confidence interval [CI]) of respiratory infection in the Q1, Q2, Q4, and Q5 quintile groups was 1.28 (1.21-1.37), 1.07 (1.00-1.14), 1.06 (0.99-1.13), and 1.09 (1.02-1.17), respectively. In women, the HR (95% CI) was 1.20 (1.12-1.29), 1.09 (1.01-1.17), 1.01 (0.94-1.09), and 1.05 (0.98-1.13) in the Q1, Q2, Q4, and Q5 quintile groups of Hb, respectively, compared with the third quintile group. There was a significant interaction between Hb concentration and chronotype on the risk of respiratory infection (P for interaction = 0.005). The elevated risk of respiratory infection associated with Hb was more pronounced among participants with late chronotype. CONCLUSION: The study suggests that Hb and respiratory infection have a nonlinear U-shaped association and that such a relation is modified by chronotype.