To assess whether the CYP2C19 polymorphism modified the effect of cyclophosphamide on ovarian toxicity in Thai patients with SLE. We performed a case-control study of female patients with SLE who were treated with cyclophosphamide at Ramathibodi Hospital, Bangkok, Thailand. Cases were patient who had ovarian toxicity (sustained amenorrhoea >12 months or lack of menstruation for >4 months). CYP2C19 polymorphism was genotyped using PCR-RFLP method. Logistic regression was applied to assess CYP2C19 polymorphism as an effect modifier of cyclophosphamide. Seventy-one patients with SLE were enrolled, of which 36 (59.7%) had ovarian toxicity. CYP2C19*2 allele frequencies were 27.8 and 21.4% in the ovarian and non-ovarian toxicity groups. Patients with CYP2C19*1/*1 genotype and higher cumulative dose of cyclophosphamide (>23.75 g) had the highest odds of ovarian toxicity, i.e. 11.0 (95% CI: 1.2-99.1) times higher than patients with the CYP2C19*1/*2 or *2/*2 genotypes who received less cyclophosphamide (<23.75 g). After adjusting for age at start of treatment, this risk increased to 13.6 (95% CI: 1.1-162.2). Our results suggest that a cumulative cyclophosphamide dose of 23.75 g or higher carries a twofold higher risk of ovarian toxicity and the CYP2C19*1/*1 genotype increases the risk of toxicity a further fivefold.