Seasonal and spatial trends of rickettsioses in Sri Lanka (2009–2023): Spatio-temporal analysis and climate-linked modeling
Background Rickettsioses are caused by obligate intracellular bacteria of the order Rickettsiales, which includes several genera such as Rickettsia, Orientia, Anaplasma, Ehrlichia, Neoehrlichia, and Neorickettsia, and are significant yet under-recognised vector-borne febrile illnesses in Sri Lanka. Despite being a notifiable disease, gaps in understanding its spatial and temporal distribution hinder effective control and prevention. Therefore, the present study aimed at determining long-term epidemiological trends and environmental determinants of rickettsioses incidence in Sri Lanka between 2009 and 2023. Methods A retrospective ecological analysis was conducted using weekly rickettsioses case reports obtained from the Epidemiology Unit of the Ministry of Health, Sri Lanka, covering all 25 districts over the period 2009–2023. Mid-year district population data and district land area were retrieved from the Department of Census and Statistics to calculate incidence rates per 100,000 population. Meteorological variables (mean, maximum, and minimum temperatures; total rainfall; and mean relative humidity) were retrieved using NASA POWER satellite-derived datasets. A Zero-Adjusted Gamma (ZAGA) model within the Generalized Additive Model for Location, Scale, and Shape (GAMLSS) framework was utilized to estimate the influence of climatic and spatial variables on incidence rates, with environmental predictors selected based on a priori biological hypotheses regarding vector ecology. Results Between 2009 and 2023, a total of 18,486 rickettsioses cases were reported. The highest case burden was detected in Jaffna (n = 7,080), followed by Kandy (n = 1,363), Badulla (n = 1,319), Nuwara Eliya (n = 1,141), Hambantota (n = 1,066), and Monaragala (n = 1,008). The higher incidence rates were concentrated in the Northern and Central highland districts compared to low-incidence regions in the Western and Southern provinces. Temporal analysis indicated seasonal peaks coincide with monsoonal rainfall, with a rise of cases during January to February and September to November. The GAMLSS model demonstrated that mean temperature, relative humidity, and rainfall significantly influenced incidence rates (P < 0.05), with high humidity and post-monsoonal rainfall correlating with increased risk. Spatial predictors (latitude and longitude) also contributed to explaining distribution patterns, highlighting localized ecological drivers. Conclusions This study provides the first comprehensive nationwide analysis of rickettsioses incidence in Sri Lanka, highlighting firm spatial heterogeneity and clear climatic associations. The findings emphasize the necessity of targeted surveillance and early diagnostic capacity in high-burden districts, particularly Jaffna and the central highlands. Integrating climatic data into surveillance systems may improve early warning and outbreak preparedness. Improved diagnostic availability at regional hospitals and enhanced physician awareness are essential to reduce under-diagnosis and guide timely interventions.