Cyclical Variation in Type 1 Childhood Diabetes

To the Editors: We investigated the temporal patterns and trends in incidence of childhood type 1 diabetes with the aim of better understanding its etiology. We specifically aimed to test predictions of increasing incidence and other temporal patterns that might arise from environmental drivers. We analyzed all children (aged 0–14 years) who were diagnosed with type 1 diabetes between 1 January 1990 and 31 December 2007, and who resided in a geographically-defined area of northeast England (Northumberland, Newcastle upon Tyne and North Tyneside) at the time of diagnosis. The population of the area is ethnically homogeneous and very stable, with low rates of inward or outward migration. Cases were identified from several independent sources, ensuring a high level of ascertainment. There were no known temporal or geographical biases. Age-specific incidence rates per 100,000 persons per year were calculated based on midyear population estimates for the study region obtained from the Office for National Statistics. We calculated age-standardized incidence rates based on the standard world population. Poisson regression models were used to analyze incidence rates. A total of 526 cases (261 boys, 265 girls) were diagnosed. The overall age-standardized incidence rate over the study period was 20.5 per 100,000—19.9 for boys and 21.2 for girls. The overall incidence increased over time from 14.6 per 100,000 in 1990–1995 to 27.4 per 100,000 in 2002–2007. A Poisson regression model showed that the incidence rate increased over the study period with a sinusoidal 6-year cyclical variation of 25% (rate ratio = 1.25 [95% confidence interval (CI) = 1.11–1.42]) and an overall increase of 5.3% per annum (95% CI = 3.4–7.1) (Figure).FIGURE.: Cyclical variation in the annual incidence rates (per 100,000 persons per year).Cyclical patterns of this type may be generated by delayed feedbacks operating internally within a system, or by external drivers. A delayed feedback (for example between incidence now and in the past) seems implausible for type-1 diabetes. A cyclical pattern in incidence is, however, consistent with the involvement of an infectious agent that itself displays a regular pattern in the environment. Both epidemiologic and experimental evidence supports the role of infection in the development of type 1 diabetes.1 The involvement of infections in our locality in Northern England is further supported by the identification of space-time clustering.2 Peaks in incidence have also been reported from the neighboring Yorkshire region, although the cycle followed a 4-year periodicity and was less striking than the pattern observed in our region further north.3 This analysis suggests a direct role for an environmental agent such as infection. A number of candidate viruses have been implicated, including enteroviruses, rotavirus, mumps, cytomegalovirus, rubella, and Ljungan virus.4,5 The increasing trend would suggest that the putative agent is becoming more prevalent, though the hygiene hypothesis has also been proposed to explain the association between greater disease incidence and improved sanitation. The cyclical pattern suggests that the prevalence of the agent exhibits a natural “epidemicity,” which may reflect the size of a reservoir host population. High levels of exposure to infectious agents in the population as a whole may modulate immune responses.6 We believe that a rising incidence of type 1 diabetes together with a 6-year cyclical pattern, supports the role of an infectious agent in the evolution of the disease in northeast England. While other noninfectious environmental factors could be partly responsible for the rising incidence, further research should focus on identifying candidate etiologic agents. Richard J. Q. McNally Simon Court Peter W. James Raymond Pollock Karen Blakey Institute of Health and Society Newcastle University England, United Kingdom [email protected] Mike Begon School of Biological Sciences University of Liverpool England, United Kingdom Tim D.