Yellow fever mortality in Brazil: an age-period-cohort study

To analyze yellow fever mortality trends in Brazil, focusing on sexes differences and using an age-period-cohort model.

This ecological study analyzed yellow fever mortality data in Brazil from 1980 to 2019 sourced from Datasus. Population estimates were retrieved from the Instituto Brasileiro de Geografia e Estatística (Brazilian Institute of Geography and Statistics). Mortality data, including age, year of death, and cause (ICD-9: 060; ICD-10: A95), were analyzed using an age period cohort model. A Poisson distribution was assumed for mortality counts, and analyses were conducted using Holford's method and its adaptations on R.

The results show that the incidence rate peaked at younger ages, such as 30 years (0.010/100,000 individuals, 95%CI: 0.008/100,000 to 0.013/100,000), followed by a gradual declining trend with increasing age, reaching 0.007/100,000 individuals (95%CI: 0.005/100,000 to 0.010/100,000) at 50 years onward. Regarding period, a substantial increase in the adjusted hazard ratio occurred over time, especially in 2015 (13.923 [95%CI: 11.095 to 17.471]), suggesting a significant elevation when compared with previous periods. Cohort analysis showed a trend of increasing risk until 1960 (RR = 1.000), followed by a marked reduction for more recent cohorts, such as 2010: RR = 0.056 (95%CI: 0.028 to 0.112). Vaccination analysis showed alternating periods of significant increases and decreases in vaccination rates.

Younger individuals showed higher mortality rates, with a gradual decline with advancing age. Period effects highlighted a pronounced resurgence in recent years, particularly during the 2015 epidemic, underscoring the influence of temporal factors such as outbreaks and vaccination campaigns. Cohort analysis showed a progressive decline in mortality risk among more recent birth cohorts, likely reflecting the impact of expanded immunization programs and improved public health measures. The proposed yellow fever vaccination trends in Brazil may explain some of the observed patterns.