Statistics from Altmetric.com
Editor,—Harding and van Heyningen1have done me the honour not only of testing a tentative hypothesis relating to a possible link between the season of birth and the prevalence of [one type of] cataract, but also of quoting from a preprint I sent them. The latter showed that a statistically significant variation of the season of birth is to be observed in the prevalence of mixed nuclear and posterior subcapsular cataracts both in an immigrant Indian population and among natives of the British Isles.
The latter observation does not seem to have caught the authors’ eyes, and I made no claim for the conditions described in their paper. While my results may not be valid for Oxfordshire, they refrained from defining the types and distribution of cataracts on which their analysis is based. Consequently we do not know whether Oxfordshire patients with mixed nuclear and subcapsular cataracts may not perhaps share the observed characteristic, and their results are merely confounded by the addition of those patients who do not.
Editor,—We are grateful for Professor Weale’s comments on our attempt to test his hypothesis that the risk of cataract may depend on the season of birth. Our results indicated that season of birth was of no importance in an Oxfordshire population, whereas he reported an excess of cataract patients, with some types of cataract, born in the spring on the Indian subcontinent. The two studies differed in both design and the populations studied. Our results were based on the dates of birth of 723 cataract patients and 1217 controls who took part in two case-control studies of cataract in Oxfordshire. The controls were from both hospital sources and the community (age and sex registers of general practitioners). Each control group was age and sex matched to the cataract group. Dates of birth of all subjects were recorded. All subjects in Professor Weale’s study were outpatients at Moorfields Eye Hospital. There was an excess of births reported on New Year’s Day in those born on the Indian subcontinent which was dealt with by partial elimination. The Moorfields patients were divided into three ethnic groups: native British, Indian, and Caribbean. There were no controls so he attempted to account for the known seasonal variation in birth rate by using a published monthly birth index. Our use of controls from exactly the same area matched for age and sex would seem preferable. There was a seasonal variation in births of patients in Oxfordshire but it corresponded to that in controls. The different designs could explain the different results.
Professor Weale found an excess of pooled cataract patients born in the spring on the Indian subcontinent. This excess was due to an excess in those with a combination of nuclear and subcapsular cataracts. The same type of cataract was associated with birth in both the summer and December in “native British”, again compared with a birth index, assumed to represent a British control population. In his letter he suggests that we may have lost a significant effect because the critical subtype was diluted by all other cataracts, but that did not occur in his study, indeed the highest level of significance was found when he pooled all types of cataract (p=0.00032 compared with 0.0168 for the nuclear plus posterior subcapsular combination). This of course provides the greatest numbers (362). The major disadvantage of subdivision is that the numbers dwindle away and thus the power of the study is diminished. It is not surprising that novel risk factors have mostly been identified in studies on mixed cataracts.