Untitled

Safety of nevirapine-containing antiretroviral tripletherapy regimens to prevent vertical transmission in anAfrican cohort of HIV-1-infected pregnant women MC Marazzi,1 P Germano,2 G Liotta,3 G Guidotti,4 S Loureiro,5 A da Cruz Gomes,6 MC Valls Blazquez,5 P Narciso,7CF Perno,8 S Mancinelli3 and L Palombi31LUMSA University, Rome, Italy, 2Community of Sant’Egidio, DREAM Programme, Rome, Italy, 3Department of PublicHealth, University of Tor Vergata, Rome, Italy, 4National Institute of Health, ISS, Rome, Italy, 5Dream Program-Maputo,Mozambique, 6Catholic University of Beira, Beira, Mozambique, 7National Institute of Infectious Disease, INMI‘L. Spallanzani’ Rome, Italy, 8Department of Experimental Medicine, University of Tor Vergata, Rome, Italy ObjectiveTo assess the incidence and consequences of adverse reactions among African HIV-positive pregnantwomen treated with fixed-dose combinations of a nevirapine-containing antiretroviral (ARV) tripletherapy.
MethodsA retrospective analysis of the clinical files of 703 HIV-1-positive pregnant women treated with anevirapine-containing regimen between May 2002 and July 2004 was conducted. Selection criteriafor inclusion in the analysis were: (a) taking ARV for more than 14 days; (b) baseline values oftransaminases below the threshold of 2.5 times the upper limit of normal (ULN). The women were ona nevirapine-containing regimen for a median of 127 days [interquartile range (IQR) 86–190 days],starting on average at the 27th week of gestation (standard deviation Æ 9.5) and continuing up to amaximum of 6 months after delivery. All women were offered formula milk to feed the babies.
Highly active antiretroviral therapy (HAART) was continued beyond 6 months only if the patientqualified on the first visit. The main outcome measures were incidence of hepatotoxicity, skin rashesand Stevens–Johnson syndrome. Multivariate analysis to assess the impact of several factors on theadverse reaction rate was performed.
ResultsAs of 1 August 2004, 554 pregnancies reached term, 96 women were still pregnant, and 53 womendropped out of the programme before giving birth. After 2 months of therapy the percentage ofpatients with a viral load less than 1000 HIV-1 RNA copies/mL increased to 78.6%; average CD4 cellcounts increased from 490 cells/mL before therapy to 630 after therapy. The incidence of grade 3–4adverse reactions (hepatotoxicity, skin rashes and Stevens–Johnson syndrome) was 6.5, 2.4 and1.1%, respectively. Five women died during pregnancy (0.88%). Only one of the deaths could beassociated with ARV treatment.
ConclusionNevirapine-containing regimens in pregnant woman, at all CD4 cell count levels, appear to be safein African settings.
Keywords: hepatotoxicity, HIV mother-to-child transmission, nevirapine-containing regimen, publichealth, resource-limited settings Received: 7 July 2005, accepted 28 November 2005 Correspondence: Dr Giuseppe Liotta, Via Montpellier 1, Facolta` diMedicina–Laboratorio di Epidemiologia, Universita` di Tor Vergata, Torre A number of studies have recently raised the issue of the E-Sud, 00137 Rome, Italy. Tel: 1 39 06 72596613 or 1 39 06 8992225; toxicity of nevirapine (NVP) administered during preg- fax: 1 39 06 20427263 or 1 39 06 89922525;e-mail: [email protected] or [email protected] nancy to HIV-1-positive women [1–7]. The findings of Safety of ARV triple therapy containing NVP 339 these studies have not, however, been consistent. Some Table 1 Baseline characteristics and toxicities for the pregnant studies found an increased incidence of NVP-related adverse reactions, especially hepatotoxicity, in pregnant women with high CD4 cell counts. However, other studies did not confirmthese observations. One of the reasons for these contradictory results may be the limited power of the analyses because of the small number of patients enrolled in each study.
Drug Resources Enhancement against AIDS and Mal- nutrition (DREAM) is a programme designed and run by the Community of Sant’Egidio [8–11] in Mozambique as well as in other sub-Saharan African countries.
A major focus of the programme is the prevention of mother-to-child transmission during pregnancy, delivery and breast-feeding. Mothers are kept on highly active WHO clinical stage classification [n (%)] antiretroviral therapy (HAART) if they qualify for this treatment at the time of diagnosis. Antiretrovirals (ARVs), nutritional supplements and laboratory tests are provided To assess the potential toxicity of, and damage caused by, NVP-containing regimens, a cohort of 703 pregnant women enrolled in DREAM were analysed retrospectively.
ALT, alanine aminotransferase, AST, aspartate aminotransferase; ARV, The clinical files of all 999 pregnant women enrolled in the antiretroviral; ZDV, zidovudine; 3TC, lamivudine; NVP, nevirapine; d4T,stavudine; IQR, interquartile range; TB, tuberculosis; WHO, World Health programme between 1 May 2002 and 31 July 2004 were Organization; ULN, upper limit of normal.
reviewed. The baseline characteristics of the patients arepresented in Table 1.
Of these patients, 28 had taken HAART for less than 14 The protocol provided ARV administration to all the days; 109 had not yet begun HAART because of the stage of women irrespective of their CD4 cell count and viral load their pregnancy; 84 refused to continue treatment after the starting from the 25th week of pregnancy [8] or later, if the first visit, and 53 stopped treatment before delivery [the first ante-natal visit occurred later in the pregnancy. If, on overall refusal/lost-to-follow-up rate was 13.7% (137 their first visit, the patient’s clinical condition was patients out of 999)]; and 22 had liver enzyme measure- classified as stage 3–4 using the World Health Organization ments 2.5 times higher than the upper limit of normal (WHO) Clinical Classification for Resource-Limited Coun- (ULN) before starting HAART. Therefore, the total number tries [12], or if her CD4 cell count was o200 cells/mL or her of women included in this analysis was 703.
viral load exceeded 55 000 copies/mL, ARV treatment wasstarted in the 15th week.
Patients with a CD4 count o200 cells/mL were also given The CD4 cell count was determined using a Beckman- cotrimoxazole. For asymptomatic patients with a CD4 Coulter EPICS-XL MCL flow cytometer (Beckman-Coulter, count 4200 cells/mL and a viral load o55 000 copies/mL, Inc., Fullerton, CA, USA) equipped with an argon ion laser ARV treatment was continued for a maximum of 6 months (488 nm). The lymphocyte subset count was determined in after birth. The ARV drugs administered consisted of two dual platform mode using the haematology analyser generic fixed-dose combinations, both including NVP: (a) SYSMEX KX21 (Sysmex Co, Kobe, Japan). The antibodies 629 patients (89.6%) were given zidovudine (ZDV) 300 mg used were CD45-FITC and CD4-PE, and CD8-Pcy5 if twice daily, lamivudine (3TC) 150 mg twice daily and NVP required (Beckman-Coulter, Inc.). Viral load tests were 200 mg twice daily (once daily during the first 14 days of performed with System 340 (Bayer Diagnostic, Tarrytown, treatment); (b) 74 patients (10.4%) with haemoglobin levels NY, USA) using branched-DNA technology (version 3.0, o8 g/100 mL received stavudine (d4T) 30 or 40 mg twice detection limit 50–500 000 HIV-1 RNA copies/mL).
r 2006 British HIV Association HIV Medicine (2006) 7, 338–344 The protocol included monitoring the levels of haemo- After 60 days of treatment, median viral load sharply globin and transaminases before starting ARVs, and then decreased from 4.05 log HIV-1 RNA copies/mL (IQR: 3.41– every 2 weeks for the first month and every 4 weeks 4.49 log copies/mL) to 1.7 log copies/mL (IQR: o1.60–2.51 thereafter. Adverse reactions were defined according to the log copies/mL). A viral load o3.0 log copies/mL was Aids Clinical Trial Group (ACTG) Adult Adverse Experi- achieved in more than 75% of women. Similarly, CD4 cell ences Grading Scale [13]. The mean number of transami- counts increased from a median of 496 cells/mL (IQR: 308– nase measurements taken was 4.1 [median 4; interquartile 697 cells/mL) before therapy to 630 cells/mL (IQR: 418– range (IQR): 3–5; range 2–10]. The number of measure- 874 cells/mL) after 2 months of therapy, and to 694 cells/mL ments was significantly higher in the subsample of patients after 4 months of therapy (IQR: 494–918 cells/mL).
with liver toxicity (4.7 vs 3.2 for the whole sample; Considering triple therapy in relation to the prevention Po0.01). If one of the transaminase measurements was of mother-to-child HIV transmission, 331 babies, born greater than five times the normal value (grade 3–4 before February 2004, were tested at 18 months (Rapid Test toxicity) and did not decrease on the next monthly visit, or Determine, Abbott Laboratories, Chicago, IL, USA, plus if clinical symptoms suggested any liver toxicity, therapy Unigold, Trinity Biotech Plc, Bray, Ireland, for confirmation was suspended. In the case of severe adverse reactions in the case of a positive test result). The infection rate was (grade 3–4) attributed to NVP, it was replaced by nelfinavir 3% [10 of 331; 95% confidence interval (CI): 1.1–4.9].
The mortality rate for these children was 4.9% (95% CI: Diagnosis of malaria was supported by laboratory evidence and tuberculosis diagnosis was made on the basis The maternal post-partum drop-out rate was 7.8% (55 of of sputum examination, chest X-ray and clinical examina- 703 women). The maternal drop-out rate pre- and post- tion. After delivery, formula milk was provided to the mothers, as well as a bottle and a filter for preparation ofclean water. The entire package (including antiviral drugs and laboratory monitoring) was offered to the women freeof charge.
During the 27 months of observation, five deaths were Statistical analyses were performed using the SPSS registered during pregnancy (a maternal mortality rate of statistical package (version 11.3; SPSS, Inc., Chicago, IL, 0.8%), which is lower than the registered national average USA). Generic statistical tests were performed as well as a for all women in Mozambique (1%) [14]. Only one death Cox proportional hazard regression analysis. To assess the was preceded by an increase in liver enzymes, which was difference in the mean onset time of hepatic toxicity, an detected after 21 days of treatment. This patient had a CD4 analysis of variance (ANOVA) univariate analysis was cell count of 322 cells/mL before treatment.
performed using the least significant difference (LSD) test It is not always possible to obtain detailed information and the Bonferroni test if equal variances were assumed, or about the cause of death. Nevertheless, the available the Games–Howell test if equal variances were not clinical and laboratory data seem sufficient to allow us to assumed. A linear regression was set up to assess the exclude use of ARV drugs as the cause of death. Neither a relation between hepatic toxicity onset time and pre- pathological increase in transaminases nor skin rashes were HAART CD4 cell count in these patients.
observed in the four remaining cases. At her initial visit,the first patient presented with severe anaemia (7.1 g/mLhaemoglobin) and a respiratory pathology. ARV treatment with stavudine/lamivudine/NVP was begun along withantibiotic treatment and the oral administration of iron.
After 2 weeks, she did not show clinical signs of toxicity or an alteration of laboratory test values. The patient died 4 By the end of the study, 554 pregnancies reached term, days later of unknown causes. The second case involved a while 149 (21.2%) women had not yet delivered. On patient who died after 40 days of therapy without any average, therapy was started in the 27th week of pregnancy specific pathology observed. Again a severe anaemic state [standard deviation (SD) Æ 4.8] and the median time of (7.2 g/mL haemoglobin) was reported along with a high exposure to ARV before the end of the pregnancy (for the viral load of 5.6 log copies/mL. The third patient died after 554 women whose pregnancies reached term) was 81 days about 60 days of treatment; she was symptomatic on her (IQR: 55–102 days). The median time of exposure before first visit (oral and vaginal candidiasis) and had respiratory the end of ARV treatment or the end of the study period disturbances. This patient was also anaemic (7.1 g/mL was 118 days (IQR: 75–164 days; n 5 703).
haemoglobin), as was the fourth woman who died. This last r 2006 British HIV Association HIV Medicine (2006) 7, 338–344 Safety of ARV triple therapy containing NVP 341 Grade 3–4 hepatotoxicityby CD4 count (cells/mL) Onset time (days) for grade 2–4hepatotoxicity by CD4 count *Seventy-four patients had anaemia before starting highly active antiretroviral therapy (HAART).
ANOVA, analysis of variance; CI, confidence interval; SD, standard deviation.
Grade 3–4 hepatic adverse reactions occurred in 46 of the 703 patients; 6.6% of the cohort (Table 2). Liver enzymevalues were checked until the end of 2004 for these cases.
In these 46 patients, the increase in liver enzymes peakedduring the first 2 months of therapy (median 74 days; IQR29–143 days), and in 36 cases values returned to baselinedespite continuation of ARV treatment. In 10 cases (1.4% ofthe overall sample) the women had skin rashes or Stevens–Johnson syndrome (SJS) and nausea or jaundice combinedwith liver toxicity, and NVP was discontinued. When theliver enzyme values returned to baseline or to below the2.5 times ULN threshold (grade 2 toxicity), the patients re-started ARVs without prejudice after indinavir/nelfinavirhas been substituted for NVP. Grade 3–4 hepatic toxicityincidence rates were 9.4% (12 out of 128 patients) forpatients with pre-ARV CD4 cell counts o250 cells/mL and5.9% (34 out of 573) for those with pre-ARV CD4 cellcounts 4250 cells/mL (P 5 0.15) (Table 2). In addition, 40cases of grade 2 liver toxicity were detected [7.0% and5.4% (P 5 0.7) in the groups with CD4 cell counts belowand above 250 cells/mL, respectively]. None of the patientswith grade 2 liver toxicity showed any clinical signs oftoxicity so the ARVs were not discontinued.
The observation period for patients in the subsample Fig. 1 Linear regression model for 2-4 grade hepatotoxicity and pre- with grade 3–4 liver toxicity was extended to the end of 2004. All but seven patients showed a return to normalvalues; six patients moved from grade 3 to grade 2 toxicity patient, who died after less than a month on HAART, showed and one patient died, as reported above. The return to no symptoms or signs of an adverse reaction involving the normal transaminase levels or at least to values below the threshold of 2.5 times ULN occurred in 32 days, on average, r 2006 British HIV Association HIV Medicine (2006) 7, 338–344 Table 3 Risk of grade 3–4 hepatic toxicity by CD4 cell count, adjusted accomplishing this goal [15–18]. The limited drop-out rate for malaria episodes, using Cox proportional hazard model for patients in this large programme run in a public healthsetting in a limited-resource environment provides reas- surance that this goal is possible even in developingcountries.
The incidence rate of toxicity of NVP-containing regi- mens was consistent with that reported in other studies [2,4,7,19], and toxicity was frequently self-limiting. Whilea grade 3–4 elevation of alanine aminotransferase (ALT)/ CI, confidence interval; SE, standard error; B, regression coefficient; Exp(B),hazard ratio.
aspartate aminotransferase (AST) level was detected in6.6% of women using a NVP-containing regimen, most of for the patients who suspended therapy, and in 101 days these elevations resulted in no significant clinical hepato- for those who did not suspend therapy (P 5 0.004).
toxicity. Most of the 3–4 grade toxicities (36 out of 46) A statistically significant association between higher resulted in transient elevations of transaminase levels, CD4 cell count and shorter grade 2–4 hepatotoxicity onset which had decreased again by the next check-up without time was observed (Fig. 1 and Table 2). The average number ARV treatment being stopped or changed. It is likely that of days before grade 2–4 hepatic toxicity onset was 217 the majority of these elevations in liver enzymes were the days (median 146 days; IQ 42–357 days) for patients with a result of the NVP in the regimen [1–7]. The incidence rate pre-ARV CD4 cell count o250 cells/mL, 107 days (median of hepatic toxicities was higher in the women undergoing 72 days; IQ 32–164 days) for patients with CD4 cell counts between 251 and 500 cells/mL, and 72 (median 63 days; IQ It is interesting that our data did not confirm an 28–91 days) for patients with counts 4500 cells/mL association between hepatic toxicity and high CD4 cell (Po0.001). The Cox proportional hazards survival ana- count. However, we observed an earlier onset of hepatic lyses, performed to assess the determinants of hepatic toxicity in the group with CD4 cell counts 4250 cells/mL.
toxicity, showed no association between the onset of This observation is compatible with the hypothesis that toxicity and high CD4 cell count levels, even after hepatic toxicity is caused by an immune-mediated toxic adjusting for malaria episodes. In fact, no significant effect on the liver, which is more rapid when the immune difference was found between the two strata (Table 3).
During the entire period of observation, eight cases Five deaths occurred in this ARV-treated cohort of (1.1%) of SJS were recorded. None was lethal, but in each patients. Only one of these deaths could be associated with case ARVs were suspended for the time required for the elevation of liver enzymes (incidence rate 0.18%; 95% patient to recover. Then treatment was resumed with a CI: 0.00–0.54). With regard to the other four deaths, none regimen without NVP being administered. In addition to of these patients had either elevated transaminases or skin the eight cases of SJS, 17 patients were observed to have rashes. The small percentage of deaths that could be related skin rashes of grade 3 severity (2.4%). In these cases also, to liver toxicity is consistent with that reported in other a regimen without NVP was substituted after a short Reliance on a retrospective analysis of clinical files The total number of patients who stopped NVP-contain- resulted in some limitations. In each case, the discontinua- ing regimens because of SJS, skin rashes or hepatic toxicity tion of NVP occurred in patients who experienced a skin Before beginning treatment, 74 women had haemoglo- The present study did not allow us to measure the impact bin levels 48 g/100 mL, so they started HAART with of several other related factors, such as the prevalence of stavudine; among the remaining 629 who started therapy hepatitis B virus and hepatitis C virus infection. Such with ZDV, 100 (15.9%) were placed on triomune because factors could have had important effects on liver toxicity their haemoglobin level fell below 8 g/100 mL.
associated with ARVs in our study [21].
The incidence of episodes of serious skin rashes and SJS was also low in this study (3.5%) and comparable to thatobserved by other authors [3].
When treating pregnant women, our goal should be to The incidence of anaemia highlights one of the decrease the viral load as quickly and effectively as disadvantages of using zidovudine, although this drug is possible. It is clear from this and other studies that one of the most effective in preventing mother-to-child ARV triple therapy represents the gold standard for r 2006 British HIV Association HIV Medicine (2006) 7, 338–344 Safety of ARV triple therapy containing NVP 343 The protocol did not include tests on the babies other with nevirapine- and efavirenz-containing regimens in HIV- than for diagnosing HIV infection using branched-DNA at infected patients. Int J STD AIDS 2003; 14: 776–781.
1, 6 and 12 months of age. This meant that we were not 5 Martin A, Cameron P, Nolan D et al. Predisposition to able to evaluate possible hepatic toxicity in the babies nevirapine hypersensitivity associated with HLA-DRB1*01 and resulting from the therapy taken by their mothers.
higher CD4 1 T cell counts. XV. International AIDS Conference.
However, the low mortality rate (4.9%) among the infants Bangkok, Thailand, July 2004 [Abstract LbOrB13].
compared with that of the overall paediatric population in 6 Hitti J, Frenkel L, Stek A et al. Maternal toxicity with Mozambique indicates that the presence of any toxicity continuous nevirapine in pregnancy: Results from PACTG 1022. J Acquir Immune Defic Syndr 2004; 36: 772–776.
The results obtained here, showing a high rate of 7 Dieterich DT, Robinson PA, Love J, Stern JO. Drug-induced prevention of mother-to-child transmission, suggest that liver injury associated with the use of non-nucleoside reverse- the use of triple therapy is effective in greatly reducing the transcriptase inhibitors. Clin Infect Dis 2004; 38 (Suppl. 2): number of HIV-infected newborns. The use of formula milk to avoid transmission after delivery could be limited by 8 Community of Sant’Egidio. DREAM. Treating AIDS in Africa.
social and economic difficulties. Triple therapy after Milan, Italy: Leonardo International, 2003.
delivery could therefore be an effective means of prevent- 9 Marazzi MC, Palombi L, Emberti Gialloreti L et al.
ing HIV transmission after delivery in resource-limited DREAM Project early results. Ig Sanita Pubbl 2002; 58: In conclusion, the low frequency and mainly minor 10 Palombi L, Narciso P, Perno CF et al. One year of HAART in consequences of adverse reactions to a NVP-based regimen Mozambique: survival, virological and immunological results in poor women in this resource-limited environment of DREAM project in adults and children. Conference on suggest that such a regimen should be considered by Retroviruses and Opportunistic Infections. San Francisco, CA, policy makers and those involved in HIV programmes as the preferred treatment regimen for HIV-positive pregnant 11 Palombi L, Germano P, Mancinelli S et al. DREAM results from Mozambique. Predictors of mortality in a HAART treated adult cohort. XV. International AIDS Conference. Bangkok, Thailand, 12 World Health Organization. Scaling Up Antiretroviral Therapy in Resource-Limited Settings. Treatment Guidelines for a Public The DREAM programme – Prevention of Mother to Child Health Approach. Geneva, Switzerland: WHO, 2003.
Transmission (MTCT) Branch was funded by the following 13 AIDS Clinical Trial Group. Table of Grading Severity of Adult public and private bodies: Unicredit Bank – Unidea Adverse Experience. http://www.aactg.org/textmenu// Foundation, the Finnish Embassy in Mozambique, Mes- saggero di Sant’Antonio, Dopolavoro Ferrovie dello Stato, 14 Anonymous. United Nations Development Programme.
and the Rissho Kosei-Kai Foundation. The representatives Rapporto su lo Sviluppo Umano, Chapter 14. Le Azioni contro of the funding sources had no role in writing this paper or la Poverta`. Trento, Italy, Rosenberg & Sellier, 2003.
15 World Health Organization. Antiretroviral Drugs for Treating Pregnant Women and Preventing HIV-Infection in Infants: Guideline on Care, Treatment and Support for Women Living with HIV/AIDS and Their Children in Resource-Constrained 1 De Maat MM, ter Heine R, van Gorp EC, Mulder JW, Mairuhu Settings. Geneva, Switzerland: WHO, 2004.
AT, Beijnen JH. Case series of acute hepatitis in a non-selected 16 Cooper ER, Charurat M, Mofenson L et al. Combination group of HIV-infected patients on nevirapine-containing antiretroviral strategies for the treatment of pregnant antiretroviral treatment. AIDS 2003; 17: 2253–2256.
HIV-1-infected women and prevention of perinatal HIV-1 2 Martinez E, Blanco JL, Arnaiz JA et al. Hepatotoxicity in transmission. J Acquir Immune Defic Syndr 2002; 29: 484–494.
HIV-1-infected patients receiving nevirapine-containing 17 Mofenson LM, Lambert JS, Stiehm ER et al. Risk factors for antiretroviral therapy. AIDS 2001; 15: 1261–1268.
perinatal transmission of human immunodeficiency virus type 3 De Maat MM, ter Heine R, Mulder JW et al. Incidence and risk 1 in women treated with zidovudine. Pediatric AIDS Clinical factors for nevirapine-associated rash. Eur J Clin Pharmacol Trials Group Study 185 Team. N Engl J Med 1999; 341: 4 Ena J, Amador C, Benito C, Fenoli V, Pasquau F. Risk and 18 Garcia PM, Kalish LA, Pitt J et al. Maternal levels of plasma determinants of developing severe liver toxicity during therapy human immunodeficiency virus type 1 RNA and the risk of r 2006 British HIV Association HIV Medicine (2006) 7, 338–344 perinatal transmission. Women and Infants Transmission Study 20 Sanne I, Mommeja-Marin H, Hinkle J et al. Severe Group. N Engl J Med 1999; 341: 394–402.
hepatotoxicity associated with nevirapine use in 19 Stern JO, Robinson PA, Love J, Lanes A, Imperiale MS, Mayers HIV-infected subjects. J Infect Dis 2005; 191: DL. A comprehensive hepatic safety analysis of nevirapine in different populations of HIV infected patients. J Acquir 21 Liotta G, Palombi L, Riccardi F et al. HIV infection in northern Immune Defic Syndr 2003; 34: S21–S33.
Mozambique. S Afr Med J 2002; 92: 12.
r 2006 British HIV Association HIV Medicine (2006) 7, 338–344

Source: http://dream.santegidio.com/UserFiles/File/HIV_386.pdf