We read with interest the article by Weinreich and colleagues.1 They showed that transgenic mice born to mothers aged 8 months or older had a significantly lower frequency of murine ankylosing enthesopathy than mice born to younger mothers. They speculated that an age related increase in maternal antibody levels resulted in increasing protection of offspring against a ubiquitous, potentially arthritogenic, micro-organism. In humans, as in the mouse model, environmental factors must influence the development of ankylosing spondylitis (AS). We sought evidence that the age of conception in women with AS influenced the risk of their offspring developing disease.

We collected data from 3473 patients with AS, regarding the age, and AS status, of relatives. Nine hundred and sixty two were female, male: female ratio 2.6:1. Six hundred and sixty five were recruited from the Royal National Hospital for Rheumatic Diseases (RNHRD), 2779 from the National Ankylosing Spondylitis Society (NASS), and 29 were relatives of RNHRD patients or NASS members, but were not RNHRD patients or NASS members themselves. Those from the RNHRD have been confirmed, by a rheumatologist, as having AS according to the New York criteria.2 The diagnosis of AS in the overall patient population has been validated in three separate cohorts in which 100% of 146 patients had AS (personal communications M Brown, Oxford), 92.7% of 330 patients had AS (personal communications K Gomez, Bath), and 96% of 50 patients had AS.3 Data were available for 39 pairs in which both mother and offspring had AS (21 mother:daughter pairs, 18 mother:son pairs). There were no families with data available for more than one mother:child pair. The higher than expected ratio of mother:daughter to mother:son pairs is the subject of an ongoing investigation. From our data base we chose 39 control pairs, where the mother had AS and the offspring did not, by matching the offspring for sex and age to the nearest year. We were able to match the offspring for birth rank in 37 cases (20 mother:daughter pairs, 17 mother:son pairs). The offspring were not matched for sibship size. The maternal ages at birth of offspring are shown (see table 1). Using pairedt tests, we compared maternal ages at childbirth in the groups where the offspring had AS, and where the offspring did not have AS. The difference in the maternal age at childbirth was significant only when comparing mothers of sons with AS and mothers of sons without AS (p = 0.04; difference in mean maternal age at childbirth = 2.78 years). Mothers with AS, whose sons had AS, had given birth at a older age than those whose sons did not have AS.

Unlike the situation in the murine model, we find no evidence that increasing maternal age in humans is protective against the development of AS. However, the number of pairs available for our analysis was small. The influence of age at conception, on the development of AS in the offspring, remains an interesting area that deserves further attention.