HIV article
Safety of nevirapine in pregnancy
  HIV Medicine
Volume 8 Issue 1 - January 2007
U Natarajan11Department of Genitourinary Medicine, Guy's and St Thomas' NHS Foundation Trust, ,
A Pym22Centre for the Study of Sexual Health and HIV, Homerton University NHS Foundation Trust, ,
C McDonald33Department of Sexual Health, Kings College Hospital, P Velisetty44Department of Genitourinary Medicine, St. George's Healthcare NHS Trust, ,
SG Edwards55Mortimer Market Centre, University College Hospital, and , P Hay44Department of Genitourinary Medicine, St. George's Healthcare NHS Trust, , J Welch33Department of Sexual Health, Kings College Hospital, ,
A de Ruiter11Department of Genitourinary Medicine, Guy's and St Thomas' NHS Foundation Trust, ,
GP Taylor66Department of GU Medicine & Communicable Diseases, Faculty of Medicine, Imperial College, London, UK and
J Anderson22Centre for the Study of Sexual Health and HIV, Homerton University NHS Foundation Trust,
1Department of Genitourinary Medicine, Guy's and St Thomas' NHS Foundation Trust, 2Centre for the Study of Sexual Health and HIV, Homerton University NHS Foundation Trust, 3Department of Sexual Health, Kings College Hospital, 4Department of Genitourinary Medicine, St. George's Healthcare NHS Trust, 5Mortimer Market Centre, University College Hospital, and 6Department of GU Medicine & Communicable Diseases, Faculty of Medicine, Imperial College, London, UK
Background: Nevirapine has been widely used in pregnancy for its efficacy, low pill burden, bioavailability and rapid transplacental transfer. Concern about nevirapine toxicity during pregnancy has emerged over recent years.
Objectives: The aims of the study were to document the frequency of cutaneous and hepatic toxicity secondary to nevirapine use during pregnancy and to compare rates in women starting nevirapine during the current pregnancy with those in women who had commenced nevirapine prior to the current pregnancy.
Design: This was a retrospective, comparative, five-centre study carried out in London, UK, in 1997-2003.
Methods: All HIV-1-infected women who received nevirapine as part of combination antiretroviral therapy (ART) during pregnancy were included in the study. Data on demographics, HIV infection risk, Centers for Disease Control and Prevention (CDC) status, surrogate markers at initiation of therapy, other medications hepatitis B and C virus coinfection and clinical data relating to potential toxicity were collated and analysed.
Results: Fifteen of 235 eligible women (6.4%) developed rash and eight (3.4%) developed hepatotoxicity, including four with coexistent rash, giving a combined incidence of 19 potential cases of nevirapine toxicity during pregnancy (8.1%). Alternative causes of rash/hepatotoxicity were suspected in seven cases and only 10 mothers (5.8%) discontinued nevirapine. Of the 170 women who commenced nevirapine during this pregnancy, 13 (7.6%) developed rash and eight (4.7%) hepatotoxicity, a combined incidence of 10%. Only two of 65 women with nevirapine exposure prior to this pregnancy developed rash (3.1%).
Conclusions: Nevirapine-containing ART was well tolerated in this cohort of pregnant women. Although pregnancy did not appear to increase the risk of nevirapine-associated toxicity compared to published adult data, CD4 count may be less predictive of toxicity in pregnancy.
The use of antiretroviral therapy (ART) during pregnancy to minimize mother-to-child transmission of HIV-1 infection is well established [1,2]. Ideally, ART should be safe for mother and baby, effective and simple. Nevirapine has been widely used in pregnancy both as single-dose therapy during labour in resource-restricted settings and as part of short- and long-term highly active antiretroviral therapy (HAART). Physiological changes in pregnancy can significantly alter pharmacokinetics but nevirapine concentrations during pregnancy have been found to be comparable to those in nonpregnant adults [3,4].
Recent safety data on the use of nevirapine in pregnancy has given rise to concern. In the Paediatric AIDS Clinical Trials Group 1022 study, five of 17 subjects (29.4%) were reported to have an adverse event related to nevirapine, including one fatality from fulminant hepatitis [5]. Furthermore, a number of deaths in pregnancy among women taking nevirapine-containing ART have been reported to the Food and Drug Administration [6]. Such concerns have led to decreased use of nevirapine in pregnancy in the UK.
The most common side effect associated with nevirapine is rash [7]. In controlled trials the frequency of rash has been found to be 14% and Stevens-Johnson syndrome (SJS) has been estimated to occur in 0.3% of patients [8]. Nevirapine may also cause asymptomatic or severe clinical hepatitis [9,10].
Skin complaints are common during normal pregnancy [11,12] and causes include cholestasis, herpes gestationalis, pruritic urticarial papules and erythema nodosum. Life-threatening complications of pregnancy such as acute fatty liver of pregnancy and the haemolysis, elevated liver enzymes, low platelets (HELLP) syndrome that occur during the third trimester [13] can potentially be confused with nevirapine or nucleoside reverse transcriptase inhibitor (NRTI)-related hepatotoxicity.
Risk factors for nevirapine-related adverse events include higher CD4 counts at the start of therapy [14] and female gender [15]. Whether pregnant women are more susceptible to these adverse events is uncertain, but pregnant women are more likely to commence ART, even for short courses, at higher CD4 counts than nonpregnant women.
The primary objective of this study was to document the experience of pregnant women in London taking nevirapine and to examine the relationships between adverse events and laboratory parameters, concomitant use of other therapy, use of alcohol and timing of therapy.
All pregnant women attending five centres from 1 January 1997 who delivered by 31 December 2003 and who were taking nevirapine as part of combination ART were identified retrospectively from HIV/antenatal clinic records. Women who had a single dose of nevirapine in labour only (one woman), women whose pregnancies did not continue into the second trimester as a result of either a spontaneous miscarriage or medical termination of pregnancy (three women), and women who were lost to follow up during pregnancy (four women) were excluded from the study. Case records and laboratory reports were individually studied and relevant details collated according to an agreed proforma. Demographic details, transmission risk group, Centers for Disease Control and Prevention (CDC) staging, hepatitis B and C virus coinfection, clinical side effects potentially related to nevirapine, changes in liver enzymes, ART history, gestation age at which nevirapine was started, total duration of nevirapine exposure, CD4 count and plasma HIV-1 viral load (measured using COBAS, Amplicor HIV1RNA Version 1.5 Monitor Assay, Roche Diagnostics or Bayer bDNA HIV-1, according to local practice) at the time of starting nevirapine and at the closest point to delivery were recorded. As nevirapine toxicity usually occurs during the first few months of therapy, we analysed separately mothers who commenced a nevirapine-containing regimen during pregnancy (group 1: 'during') and those who had commenced nevirapine before pregnancy (group 2: 'before'). In this comparator group, CD4 counts and plasma HIV-1 viral load measurements in the second trimester and closest to delivery were recorded.
Hepatotoxicity was defined as aspartate aminotransferase (AST) or alanine aminotransferase (ALT) >3 times the upper limit of normal (ULN). Rash was defined as any new cutaneous abnormality that might be caused by drug toxicity.
The groups were compared using t-tests (continuous variables) and 2 tests (categorical variables). It was not possible to conduct several of the 2 analyses, as there were fewer than five cases in more than 20% of the cells. In addition, CD4 count and log viral load were compared between those who did and those who did not develop a rash (t-tests). Analyses were carried out using the Statistics Package for Social Sciences (SPSS) version 12.0 program.
Two hundred and thirty-five pregnant women met the inclusion criteria. One hundred and seventy started nevirapine during the pregnancy, all of whom were treatment naive except two who switched from nelfinavir to nevirapine because of nelfinavir-associated side effects and four who had initially started zidovudine monotherapy but switched to HAART during the pregnancy (group 1). Sixty-five commenced nevirapine prior to this pregnancy (group 2). Group demographic details are shown in Table 1.
All the patients were on nevirapine 200 mg once daily for 2 weeks, after which the dose was escalated to 200 mg twice a day. Only two women were exposed to nevirapine for less than 6 weeks, both of whom were in group 1. Two patients in each group consumed alcohol, but at less than 2 units/week.
Group 1
The outcomes and adverse events are summarized in Table 2. Thirteen patients were documented to have developed a rash while on nevirapine, but in three cases this was thought to be caused by cotrimoxazole, which had been commenced concurrently. These patients had CD4 counts of 80 cells/μL, 150 cells/μL and 'not stated' at the time of starting nevirapine-based therapy.
Nevirapine was continued, and only cotrimoxazole was discontinued in all three of these patients. Among the 10 patients in whom the rash was considered to be nevirapine-associated, four had only a mild rash and continued nevirapine safely, including two patients concomitantly taking antituberculosis treatment. In these patients, the rash was not considered severe enough to stop either treatment or an association with nevirapine was not confirmed. Six patients with severe generalized rash discontinued nevirapine, including one patient (0.6%) with Stevens Johnson Syndrome (SJS) who made a full recovery with treatment.
Hepatotoxicity occurred in eight patients within 6 weeks of starting nevirapine (Table 3). ALT ranged from 161 to 807 IU/L. Six out of the eight patients had grade 3/4 toxicity (>5 times the upper limit of normal), four had a coexistent rash and all of them stopped nevirapine; they all had a previously normal ALT measurement and had no evidence of hepatitis B or C virus infection. All made a full recovery.
One patient, who started nevirapine at 21 weeks, developed abdominal pain, proteinuria, low platelet levels and deranged liver enzyme levels. She had evidence of intrauterine growth retardation (IUGR) and the clinical features suggested HELLP syndrome. Although her condition was not thought to be nevirapine-related, nevirapine was stopped in view of raised liver enzymes and she was delivered by emergency caesarean section at 30 weeks, following which she made a full recovery.
Group 2
Two out of 65 patients who had commenced nevirapine prior to their current pregnancy developed rash. One was on nevirapine for 20 weeks prior to this pregnancy, continued nevirapine through early pregnancy and developed a mild rash in her 20th week of pregnancy (after 40 weeks on nevirapine). Her liver enzymes were normal and the rash resolved spontaneously whilst she continued nevirapine. The other patient was on nevirapine in her previous pregnancy for 20 weeks without any side effects and stopped nevirapine after the previous childbirth. During the current pregnancy she again started nevirapine at 16 weeks and developed generalized maculopapular rash 5 weeks later; her liver enzymes were not raised, and she recovered after stopping nevirapine.
To examine the effect of CD4 count on the development of nevirapine toxicity, patients in group 1 were analysed using two separate CD4 count cut-offs, namely 200 and 250 cells/μL. In the analysis with 200 cells/μL as the cut-off, five of 51 patients (9.8%) developed rash and nine of 102 (8.8%) had rash and/or raised liver enzymes with counts lower than 200 cells/μL, and seven of 102 patients (6.8%) had rash and nine of 102 (8.8%) had rash and/or raised liver enzymes with counts higher than 200 cells/μL. In the analysis with 250 cells/μL as the cut-off, six of 75 patients (8%) developed rash and seven of 75 (9.3%) had rash and/or raised liver enzymes with counts below 250 cells/μL and six of 78 patients (7.6%) had rash and seven of 78 (8.9%) had rash and/or raised liver enzymes with counts above 250 cells/μL (Table 4). CD4 counts were not available for three patients who experienced adverse effects. Similarly, there was no significant difference in CD4 count or log viral load between those who did and those who did not develop a rash (CD4 count, P=0.14; viral load, P=0.5) or between those who did and those who did not develop hepatitis (CD4 count, P=0.8; viral load, P=0.33).
In this study of 235 pregnant women, nevirapine was well tolerated, with a lower rate of adverse events than has previously been reported. The 6.4% frequency of nevirapine-related rash noted in this study is much lower than the 12-22% usually reported. However, a proportion of mothers had commenced nevirapine prior to pregnancy, introducing a bias in this subset towards women who had come through the early risk period without toxicity. Exclusion of these women from the analysis did not significantly change the result, with a rash frequency of 7.6%. Similarly, the rate of nevirapine-associated increase in transaminases was lower than the previously reported 8-10% [10]. This may be partly explained by the low rates of hepatitis B and C virus infection in our cohort. The rates of hepatitis B and C virus infection were four out of 235 (1.7%) and seven out of 235 (2.9%), respectively. None of the patients with adverse effects was known to be infected with either hepatitis B or C virus.
For comparison with other reports, we also analysed all patients with transaminase levels higher than 1.25 times the ULN but lower than our cut-off of 3 times the ULN. By these criteria, a further 12 patients in group 1 (7%) and four patients in group 2 (6.2%) would have been categorized as having hepatotoxicity. However, all patients had continued nevirapine therapy and, given the natural history of nevirapine toxicity and the similar pattern seen in the two groups, it seems unlikely that these cases represented nevirapine hepatotoxicty.
Although the usually accepted cut-off is 250 cells/μL, higher rates of nevirapine-associated toxicity have been found in women in the 250-400 cells/μL and >400 cells/μL ranges. In view of the fact that physiological haemodilution occurs in pregnancy, which tends to decrease the CD4 count [16], we reanalysed the data with an arbitrary lower cut-off of 200 cells/μL. This did not produce any significant difference in the frequency of adverse events.
The Paediatric AIDS Clinical Trials Group (PACTG) 1022 team [5] has questioned whether pregnancy per se poses an additional risk for nevirapine-associated hepatic toxicity beyond the risk associated with female gender. In our study, a lower than expected incidence of nevirapine-related adverse events was observed, which suggests that pregnancy does not predispose to nevirapine toxicity. This has also been suggested by recent studies from Thailand [17], Kenya [18] and Brazil [19], none of which found a high rate of adverse events despite a high proportion of patients with CD4 counts>250 cells/μL. These studies and our own data contrast with the very high rate (29.4%) reported by PACTG 1022. Although it is not possible to determine the incidence of fulminant hepatic failure and death associated with continuous nevirapine use in pregnancy, as the number of pregnant women who have received nevirapine-containing ART is unknown, certainly there are more retrospective data to suggest that nevirapine is safer than suggested by PACTG 1022.
These discrepant results raise the possibility that ethnicity or other factors contribute significantly to the risk of toxicity with nevirapine. Of the subjects in PACTG 1022, nine were African American (53%), seven were Hispanic (41%) and one was white (6%). Only 5% of our cohort is of Caucasian origin, with 93.5% being of black African or black Caribbean origin.
Concomitant administration of nevirapine with other therapies associated with a high incidence of rash was noted in this study, and these other therapies were thought to be the cause of the rash in at least five of the 15 patients (33%). Further, not all cases of nevirapine-associated rash necessitated discontinuation of therapy.
In summary, these data suggest the following.
-Nevirapine, with careful management, remains a useful component of antiretroviral combination therapy in pregnancy.
-Clinicians need to consider all potential causes of rash in pregnancy to avoid unnecessarily eliminating nevirapine from the patient's therapeutic options.
-Prescribers should try to avoid coinitiation of therapies with overlapping toxicities.
-A CD4 count cut-off of 250 cells/μL may not be applicable in pregnant women, and, taken together with similar data from Kenya and Thailand, our data suggest that toxicity rates may vary with ethnicity and that a blanket warning to avoid prescription to all women with CD4 counts of >250 cells/μL might have been premature and may have to be reconsidered. It may be useful to look at more data and the collective experiences of different workers before a final recommendation is made.
  icon paper stack View Older Articles Back to Top