No abstract

The literature on presentation of seasonality of leukaemia fails to provide clear conclusions. Using the Data Collection Study organised by the Leukaemia Research Fund (LRF) Centre for Clinical Epidemiology, University of Leeds, this question was re-examined. This data has a high level of ascertainment and is population based. The results were mainly negative, there being no fit to a sinusoidal curve (Cosinor analysis) in any part of the data. Using normal approximation to Poisson there were individual months with positive findings. Such a method could not be specifically depended on to provide reliable conclusions, nevertheless, further work in this area is justified. Non-Hodgkin's lymphoma was also examined but again with mainly negative findings.

There is extensive literature describing the effect of season on mortality rates, especially in cardiovascular and respiratory disease. This study compares latitude with the extent of seasonal variation of monthly deaths from all causes. In developed countries, there is a peak of deaths in winter and a trough in summer. Monthly numbers of deaths were established in 89 countries in the Northern and Southern Hemisphere. Using cosinor analysis, the extent of seasonal variation (amplitude) was established and correlated with latitude. The amplitude of seasonality was greatest in mid-latitude around 35 degrees, but low or absent near the equator and subpolar regions. The amplitude can differ at the same latitude. The weather in equatorial regions and in habitations near the Arctic Circle is very different, but death has a similar seasonal rhythm. The purpose is to record this epidemiological finding even though no simple explanation is provided. Weather alone cannot explain it, and it is possible that day length (photoperiod) has an important, but complex, underlying role.

Environmental influences are thought to have an aetiological role in onset of diabetes in children. Month of onset in over 2000 children in Scotland was established and there was an excess in colder/darker months than in warmer/lighter months. A meta-analysis of 21 previous studies with over 13,000 patients gave the same result at a much higher level of significance. A mechanism is postulated based on previous viral induced islet cell damage with ongoing progressive auto-immune destruction. There may be physiological seasonal changes with winter stress on carbohydrate and lipid metabolism. The raised winter levels of pituitary, adrenal and thyroid hormones fail to be antagonised by falling level of insulin. A role for seasonal variation in exercise and nutrition is considered.

Summary.Earlier literature suggested there may be a seasonal rhythm of onset of Hodgkin's disease. This issue has been re-examined using population-based prospectivelycollected data with high ascertainment levels. The Data Collection Study (DCS) of the Leukaemia Research Fund (LRF) Centre for Clinical Epidemiology (University of Leeds) generated the information used, which was based on a population of 13·5 million -about one quarter of England and Wales -over 10 years. The RYE histopathological classification was employed. The findings show that in patients with nodular sclerosing histopathology there was a highly significant circannual rhythm with a low amplitude (extent of seasonal variation) and a peak in March. A significant, but different, rhythm with a high amplitude and a peak in August was found in lymphocyte predominant Hodgkin's disease. However, this finding is less certain, due to smaller numbers and a lower significance level. The main conclusion is that there is a highly significant seasonality in nodular sclerosing Hodgkin's disease. The findings provide further evidence that nodular sclerosing and lymphocyte predominant may be two different diseases. The differing seasonality rhythms may provide aetiological clues.