Sex difference in prescription of asthma drugs is smaller than previously found

EDITOR—We were surprised to see the relation between people's sex and, in those with asthma, their likelihood of being prescribed oral steroids, as found by Sexton et al.1 We analysed a larger dataset, as they suggest. We included computerised 1996 RNZCGP data from a consulting population of 201 954 patients (90 214 men and 109 079 women; 2661 sex not recorded). The nature and reliability of the RNZCGP database are well established.2 3 The target patients were in the age band 20-54 years (42 264 men and 55 670 women). Patients were identified as asthmatic if they received a prescription for inhaled corticosteroids, inhaled β agonists, inhaled sodium cromoglycate or nedocromil, or theophylline (table)

In our dataset, the raw odds ratio of receiving a prescription for oral corticosteroids is 1.39 (95% confidence interval 1.17-1.65) for women compared with men. This is significant but may have minor clinical importance. Logistic regression analysis of our data using sex and the number of asthma consultations as explanatory variables found the odds ratio for oral steroid use for women compared with men almost identical to 1.39 after controlling for the number of asthma consultations. This is because the increase in the proportion of patients taking oral steroids with a number of asthma consultations is similar for men and women, which contrasts with the dose response data of Sexton et al. The proportions prescribed oral steroids more than once were 120/2961 (4%) of women and 47/1922 (2.4%) of men, which is also significant.

The proportion of women with asthma who received at least one prescription for oral corticosteroids is less by a factor of more than two than that found in the study by Sexton et al. The sex odds ratio we found for prescriptions of oral corticosteroids is substantially lower (1.39 v 5.55). This difference could perhaps be related to the selection criteria employed by Sexton et al, with only questionnaire responders being included, or to our differing definitions of asthma. Of 253 patients possibly asthmatic on the basis of prescribing alone, only 16 were considered not asthmatic for the Sexton analyses. We therefore do not consider the different inclusion criteria for our own much larger analysis cohort to be crucial to the rather different results. We are unable satisfactorily to address actual dispensing or drug use from our computerised data. There may be other confounding factors. Overall our data support the view that if the difference in the rate of prescribing oral corticosteroids to women compared with men is not due to confounding, it is likely to be of lesser magnitude than described by Sexton et al.