Female sex but not ethnicity is a strong predictor of non-nucleoside reverse transcriptase inhibitor-induced rash

Cutaneous drug reactions are a major adverse effect limiting the use of the non-nucleoside reverse transcriptase inhibitors (NNRTI) nevirapine and efavirenz, and have been reported in 17–18% of nevirapine and efavirenz licensing studies [1,2]. Factors associated with the development of NNRTI-induced rash have not been well described, and few studies have examined the impact of sexr or ethnicity on the incidence of rash [3,4]. Our HIV clinic in south London comprises a large and ethnically diverse population. The aim of our study was to establish whether sex or ethnicity were independent predictors for the development of an NNRTI-induced rash. We conducted a retrospective review of case records of all patients initiating either nevirapine over a 3 year period (January 1997–January 2000), or efavirenz over a 6 month period (June 1999–January 2000). The main outcome measures were the occurrence of rash within 3 months of initiating an NNRTI, and whether or not therapy was discontinued. The diagnosis of an NNRTI-induced rash was made on the basis of the assessment of the attending physician, and was categorized into mild, moderate and severe, based on a modified ACTG 241 protocol [5]. A total of 337 patients were identified who had received either nevirapine (n = 285) or efavirenz (n = 52); 103 (30.6%) were women, 160 (47.5%) were of black ethnicity, and the median age at the initiation of NNRTI was 34 years [interquartile range (IQR) 29.8–38.7]. The median CD4 cell count and log viral load at NNRTI initiation was 215 × 106/l (IQR 92–313) and 4.39 log (IQR 1.7–6.3), respectively. Twenty-three patients (6.8%) had a history of a previous drug allergy, and 19 (5.6%) had a history of atopy. All patients took these drugs as part of a combination with two other antiretroviral drugs. All 285 patients on the nevaripine dose escalated, starting with 200 mg and increasing to 400 mg after 2 weeks. Nineteen (6.7%) out of 285 patients on nevirapine and three (5.8%) out of 52 patients on efavirenz developed a rash. The majority occurred within 6 weeks of initiating nevirapine, with a median time of 16 days (range 2–39 days). All three efavirenz-related rashes occurred within 2 weeks of drug initiation. Therapy was discontinued in 13 cases (68%) developing a rash on nevirapine and in one of the three patients with an efavirenz-related rash. On stopping nevirapine, nine patients switched to efavirenz and none had a recurrence of the rash. Of the 22 who developed a rash, six were taking cotrimoxazole, but the timing of the rash made cotrimoxazole an unlikely cause. Table 1 summarizes the characteristics of patients with and without NNRTI-induced rash. Female sex was significantly associated with the occurrence of a rash, with an incidence of 14.6% among women compared with 3% for men [odds ratio (OR) 5.53 95% confidence interval (CI) 2.18–14.01;P ≤ 0.001]. A documented history of previous allergy was also associated with the development of a rash, but the numbers were too small to draw any firm conclusions. Black patients were twice as likely to develop a rash when compared with white patients, with an incidence of 8.8 and 4.7%, respectively, although this was not significantly different (OR 1.93, 95% CI 0.78–4.73;P = 0.151). After adjustment for sex, black ethnicity was no longer associated with the development of a rash (OR 0.49, 95% CI 0.14–1.72;P = 0.266). In contrast, after adjustment for ethnic group, female sex remained a strong predictor of NNRTI-induced rash, with almost a ninefold increased risk of rash (OR 8.66, 95% CI 2.42–31.08;P = 0.001) and even after an additional adjustment for a history of allergy (OR 9.00, 95% CI 2.40–33.71;P = 0.001). No other variable was significantly associated with the development of a rash.Table 1: Characteristics of patients with and without non-nucleoside reverse transcriptase inhibitor-related rash. Although cross-reactivity between NNRTI is well described for delavirdine and nevirapine [4], none of the nine patients who reactacted to nevirapine and switched to efavirenz had a recurrence of rash. However, cross-reactivity between nevirapine and efavirenz is much less likely given their distinct chemical structures: nevaripine is a member of the dipyridodiazepinone class whereas efavirenz is a trifluoroderivate [6]. However, black ethnicity was not a risk factor for the development of rash after taking sex into account. Female sex was a strong independent risk factor for developing an NNRTI rash, and confirmed the findings from two other studies [7,8]. Factors that predispose women to this are unclear and warrant further exploration. There is need to explore further the role of steroid hormones [9], oral contraceptives, menstruation and pregnancy, as well as the potential role of sex-related differences in the cytochrome p450 system [10] given that both nevirapine and efavirenz are metabolized via this route. Plasma levels of nevirapine do not appear to correlate with the incidence of rash [11], and therefore the lower body mass index of women is unlikely to be an explanation. The higher risk of rash should be a consideration for all women initiating nevirapine therapy, and may have implications for the widespread use of nevirapine in mother-to-child vertical transmission. Charles Mazhudea Sarah Jonesa Shahed Muradb Chris Taylora Philippa Easterbrookb