Seasonality of birth of patients with childhood diabetes in Britain

BMJ 1996; 312 doi: (Published 08 June 1996) Cite this as: BMJ 1996;312:1456
  1. P M Rothwell,
  2. A Staines, research fellowa,
  3. P Smail, consultant paediatricianb,
  4. E Wadsworth, research associatec,
  5. P McKinney, epidemiologistd
  1. a Leukaemia Research Fund Centre of Clinical Epidemiology, University of Leeds, Leeds LS2 9NG
  2. b Royal Aberdeen Children's Hospital, Aberdeen AB9 2ZG
  3. c Institute of Child Health, Bristol BS2 8BJ
  4. d Paediatric Epidemiology Group, University of Leeds, Leeds LS2 9LN
  5. Department of Clinical Neurosciences, Western General Hospital, Edinburgh EH4 2XU PM Rothwell, research fellow.
  1. Correspondence to: Dr Rothwell.
  • Accepted 18 December 1995

Indirect evidence suggests that early environment is important in the causation of insulin dependent diabetes mellitus.1 Viral infection is thought to be the most likely trigger. Although biochemical and immunological abnormalities develop several years before the clinical onset of insulin dependent diabetes, the age at which exposure to infection might induce the disease is unclear. The high prevalence of diabetes mellitus in patients with the congenital rubella syndrome shows that the process may be initiated in utero.1 As most viral infections are seasonal, the pattern of dates of birth of people with diabetes should differ from that of the general population if a significant proportion of cases was caused by intrauterine or perinatal viral infection. We therefore examined seasonality of birth in three large independent populations of children with insulin dependent diabetes in Britain.

Patients, methods, and results

The details of the three registers have been reported previously.2 3 4 To ensure that the populations were independent, cases incident in Yorkshire2 and Scotland3 were excluded from the population in the British Paediatric Association's study.4 The analysis was restricted to births during the time common to all three populations—that is, 1974-88. The method of Walter and Elwood was used for analysis of seasonality of month of birth.5 We adjusted for the seasonality of live births in the general population by constructing pseudocohorts of births on the basis of the number of births during each month of the study. The numbers of live births in England and Wales published by the Office of Population Censuses and Surveys were used for analysis of the data from Yorkshire and the British Paediatric Association's study.2 4 The numbers of live births obtained from the registrar general for Scotland were used for analysis of the Scottish data.3

The Scottish, Yorkshire, and British Paediatric Association registers contained respectively 2258, 1142, and 1265 patients with diabetes born during 1974-88. For each register the monthly pattern of births differed significantly from that in the general population (Scotland, χ2 = 9.7, P = 0.002; Yorkshire, χ2 = 8.7, P = 0.005; England and Wales, χ2 = 3.5, P = 0.04). For each register more patients were born during the spring and early summer and fewer during the winter months compared with the general population (fig 1).

Fig 1
Fig 1

Observed divided by expected number of births each month in three populations with diabetes compared with respective general populations. Numbers are births each month in each diabetic population


The seasonal pattern of birth of patients with diabetes mellitus has not, to our knowledge, been reported to differ from that of the general population. Our study has shown abnormal seasonality of birth in three large independent populations with childhood onset diabetes. This is unlikely to be due to chance, and it is difficult to conceive of a bias that might account for the results.

Several important conclusions can be drawn from our observation. Firstly, as seasonality of birth cannot be accounted for by genetic mechanisms, environmental factors must be important in the aetiology of diabetes. Secondly, to induce seasonality of birth these environmental factors must exert their influence in utero or in the first year of life. Thirdly, these environmental factors must, of course, be seasonal in nature.

Our observations are consistent with the hypothesis that the disease process resulting in childhood diabetes is initiated by viral infection early in life. Further studies of diabetic patients' environmental exposures during intrauterine and early life, together with differences in the pattern of seasonality of birth with place and time, may identify environmental exposures specific to the initiation of the diabetogenic process.

We thank the following people for help and advice: Dr J Bodansky (Leeds), Mr G Law (Leeds), Professor AS Douglas (Aberdeen), Dr J Shield (Bristol), and Professor D Baum (Bristol).


  • Funding None.

  • Conflict of interest None.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
View Abstract