The remarkable recent roll-out of antiretroviral therapy (ART) in low-income countries has brought great hope as well as concern for complications, its continued success depends on understanding and responding to a spectrum of adverse events, one of which is hepatotoxicity. The reported risk of severe hepatotoxicity, usually defined as an elevation in either alanine aminotransferase (ALT) or aspartate aminotransferase (AST) > 5 times the upper limit of normal, ranges from 4 to 32% during the first year of therapy, depending on the ART regimen and patient population [1–7]. Factors associated with hepatotoxicity include nevirapine versus efavirenz use, chronic viral hepatitis, alcohol use, low CD4 lymphocyte counts (except in the case of nevirapine where high CD4 cell count is associated with hepatotoxicity), and elevated baseline liver enzymes [1–4,7–12].
Multiple potential additional causes of hepatotoxicity exist in Africa. For example, chronic co-infection with hepatitis B virus (HBV) is more common in some parts of Africa than in high-income countries [13–16]. In addition, tuberculosis is endemic in Africa and is increased in incidence among HIV-infected populations [17,18]. Tuberculosis disease and anti-tuberculosis therapy have not been specifically evaluated as risks for hepatotoxicity among individuals on ART but an increased risk of side effects among individuals on both tuberculosis therapy and ART versus tuberculosis therapy alone has been reported . Furthermore, herbal and traditional remedies, some of which are known hepatotoxins [20,21], are commonly used among ART patients in Africa . Given the frequency of these risks, we evaluated the incidence of hepatotoxicity, associated risk factors, and the clinical impact in a South African industrial worker HIV treatment cohort.
Patients and methods
Patients included in this study were enrolled in a workplace HIV care program in South Africa, received combination ART between November 2002 and December 2005, and were on their first ART regimen. The program is described in detail elsewhere . ART eligibility was based on modified World Health Organization (WHO) criteria: CD4 lymphocyte count < 250 cells/μl, WHO stage 3 and CD4 lymphocyte count < 350 cells/μlmm3, or WHO stage 4. ART initiation is deferred for patients with a pre-ART elevation in liver enzymes > 5 times the upper limit of normal. First-line ART was efavirenz, lamivudine, and zidovudine. Women of child bearing potential were given nevirapine instead of efavirenz. A questionnaire including demographics, alcohol use history, and prior illnesses was completed at enrollment. Routine baseline testing included chemistries, serum aminotransferases (ALT and AST), hemogram, quantitative plasma HIV RNA, and CD4 lymphocyte counts. Baseline laboratory values were defined as follows: baseline HIV RNA was the value from the last test before the ART start date, nadir CD4 count was the lowest CD4 count during the 12 months preceding ART initiation, and baseline liver enzymes were the ALT or AST obtained most closely before ART initiation. Standardized forms were used at clinic visits to record recent hospital admissions, side effects, symptoms, physical exam, and ART changes. Hospitalizations were further evaluated via abstraction of the hospital chart onto a standardized form. A separate tuberculosis clinic database was maintained with clinical data on patients diagnosed with tuberculosis. Tuberculosis disease was treated using standard WHO recommended therapy with new cases treated for 2 months using rifampicin, isoniazid, pyrazinamide, and ethambutol followed by 4 months of rifampicin and isoniazid. Retreatment cases received 3 months of rifampicin, isoniazid, pyrazinamide, ethambutol, and streptomycin followed by 1 month of four drugs and 5 months of three drugs. In keeping with local guidelines, no dose adjustment was made for either efavirenz or nevirapine during concomitant tuberculosis therapy. All demographic information, clinic visit findings, hospital admission data, and laboratory data were maintained on an electronic database with clinical information entered by trained data enterers and laboratory data captured through direct electronic download. Clinical outcomes were evaluated using HIV clinic, hospital, and employment records. All deaths, hospital admissions, or regimen starts, stops, and changes that occurred within 28 days of severe hepatotoxicity were classified as potentially related to hepatic injury.
Hepatotoxicity was examined up to 12 months after starting ART for all subjects who met the following enrollment criteria: 18 years of age or older, previously ART naive, baseline ALT or AST values available, and at least one follow-up ALT or AST during the first 6 weeks of ART or two ALT or AST tests during later follow-up periods. In addition, we limited analysis to one company within the workplace program because we were better able to capture all laboratory tests. Opportunity to complete 12 months of follow-up was not an entry criterion, thus subjects were included who initiated ART less than 12 months before the end of the study.
After ART initiation, laboratory testing of liver enzymes was done at 2 and 6 weeks, every 3 months, and when clinically indicated. Serum HIV RNA and CD4 cell count were measured at 6 weeks and every 6 months. Baseline ALT and AST were obtained before ART initiation; additional pre-ART ALT and AST were obtained during acute visits and hospital admissions and were not part of scheduled testing. Serum remaining after HIV RNA and CD4 assays was stored at −70°C.
Stored serum from a visit prior to ART initiation was used for HBV surface antigen (HBsAg) testing and hepatitis C (HCV) antibody testing (both using ADVIA Centaur automated immunoassay system and kits, Bayer Diagnostics, Tarrytown, New York, USA). HBsAg was initially assayed in a randomly selected subgroup of 133 subjects selecting every fifth sequential subject based on unique identifiers. We did this to estimate HBsAg prevalence before testing a larger number of samples. Subsequently we tested samples from an additional 254 subjects. A subgroup of 53 subjects was also tested for HCV antibody selecting every tenth sequential subject based on unique identifiers. Because the prevalence of HCV antibody was 0% in this subgroup we did not do additional testing.
Hepatotoxicity was classified based on the following modified AIDS Clinical Trials Network definitions: subjects with pretreatment mean ALT and AST in the normal range were categorized based on elevation from the upper limit of normal (ULN): grade 0, < 1.25 × ULN; grade 1, 1.25–2.5 × ULN; grade 2, 2.6–5 × ULN; grade 3, 5.1–10 × ULN; and grade 4, > 10 × ULN. To avoid inappropriately classifying hepatotoxicity among subjects with elevated baseline liver enzymes, these subjects were categorized based on changes relative to baseline: grade 0, less than 1.25 × baseline; grade 1, 1.25–2.5 × baseline; grade 2, 2.6–3.5 × baseline; grade 3, 3.6–5 × baseline; and grade 4, > 5 × baseline . Severe hepatotoxicity was defined as a grade 3 or 4 elevation in either ALT or AST. Only the first episode of severe hepatotoxicity was used for the purpose of calculating incidence and evaluating associations. Follow-up AST and ALT values were evaluated for all subjects with grade 3 or 4 hepatotoxicity to identify repeat episodes and clinical outcomes.
Statistical analysis was completed using Intercooled Stata software version 9 (Stata Corp. College Station, Texas, USA). Group comparisons for categorical variables were made with the chi-square test and for continuous variables with the Student's t test. Incidence and variables associated with hepatotoxicity were evaluated using survival time analysis with subjects censored at the time of a grade 3 or 4 hepatotoxicity episode, death, discontinuation or change of ART, leaving the workforce, end of follow-up time if less than 12 months, or after 12 months of follow-up on ART. Risk factors associated with the development of severe hepatotoxicity were assessed with univariate and multivariate Cox proportional hazards modeling. Time dependent variable modeling was used for Cox-proportional hazards modeling of anti-tuberculosis therapy. The multivariate model was generated using variables with P < 0.05 and the following individual characteristics: age strata, sex, weight strata, and elevated baseline ALT. We stratified the HBsAg status by ‘positive’, ‘negative’, and ‘not tested’ for survival time analysis. Because the hazard ratio for the ‘not tested’ group was similar to the negative group we combined these two groups in a Kaplan–Meier plot. Change in regimen, hospital admission, and death were presented as proportions of the total number developing hepatotoxicity.
Ethical approval for this study was obtained from human subjects committees of the Nelson Mandela School of Medicine, University of Kwa-Zulu Natal, the Johns Hopkins University School of Medicine, AngloGold Health Services, and the London School for Hygiene and Tropical Medicine.
Between November 2002 and December 2005, 922 subjects were started on ART at a single company in a workplace ART program in South Africa. Eight-hundred and sixty-eight subjects met inclusion criteria for this study and contributed 519 person-years of follow-up. Those excluded were either missing baseline or follow-up liver enzyme evaluations (52) or had declined participation in ART evaluation (2). Included subjects had a median of eight liver enzyme assays during the period studied. Of the 868 subjects, median age was 41 years [interquartile range (IQR), 36–46 years), 94% were male, median nadir CD4 cell count was 136/μl (IQR, 73–208/μl), median log10 HIV RNA was 4.7 (IQR, 4.3–5.2 log10) copies/ml, median baseline AST was 35 IU/l (IQR, 28–46 IU/l; ULN, 40 IU/l), 57% had a history of tuberculosis disease, 17% of a random sample of 133 tested were HBsAg positive, and 825 (95%) received an efavirenz-based regimen, while 39 (5%) received a nevirapine-based regimen. The median follow-up time on ART was 239 days (IQR, 92–365 days). Baseline characteristics are shown in Table 1.
Characteristics during ART
During the first year of ART, 48 (6%) subjects received isoniazid prophylaxis, 588 (68%) received cotrimoxazole prophylaxis, and 214 (25%) received tuberculosis therapy (either initial or continuation phase or both). Of subjects on both ART and tuberculosis therapy, 80% (172/214) started tuberculosis therapy before ART, receiving a median of 85 days (IQR, 33–112 days) of therapy before ART initiation.
Subjects had a median eight ALT/AST measurements within 1 year after ART initiation with 95% of subjects having at least one ALT/AST measurement and 89% with two or more measurements during the first 6 weeks of ART. After ART initiation, 97 subjects (11%) developed grade 2 or higher hepatotoxicity. The incidence rate of grade 2+ hepatotoxicity was 19.7 episodes per 100 person years [95% confidence interval (CI), 16.1–24.0]. A first episode of grade 3 or 4 hepatotoxicity occurred among 40 subjects (4.6%) for a rate of 7.7 events per 100 person years (95% CI, 5.6–10.5). Grade 3 or 4 hepatotoxicity occurred at a median of 57 days (IQR, 23–100 days) from ART initiation. The early occurrence of hepatotoxicity is reflected in a Kaplan–Meier plot (Fig. 1).
Table 2 shows rates of severe hepatotoxicity and associations with independent variables in univariate and multivariate Cox proportional hazards analysis. Tuberculosis therapy was treated as a time dependent variable for all analyses. HBsAg status was stratified by positive, negative, and untested. Nadir CD4 lymphocyte count < 100 cells/μl [hazard ratio (HR), 2.5; 95% CI, 1.3–4.6], anti-tuberculosis therapy (HR: 11; 95% CI: 3.5–32), HBs antigenemia (HR, 3.2; 95% CI, 1.4–7.1), and baseline ALT (HR, 2.1; 95% CI, 1.1–4.0) were all associated with an increased rate of hepatotoxicity in univariate analysis. The timing of ART after starting tuberculosis treatment did not appear to affect the risk of hepatotoxicity. Subjects who developed hepatotoxicity started ART a mean of 88 days (95% CI, 23–92 days) after starting tuberculosis therapy, similar to the 98 days (95% CI, 68–86 days) for those without hepatotoxicity (P = 0.2). Not associated with hepatotoxicity were sex, age, isoniazid prophylaxis, cotrimoxazole prophylaxis, alcohol use, and history of tuberculosis (excluding current tuberculosis therapy).
The final model included anti-tuberculosis therapy, HBs antigenemia, nadir CD4 lymphocyte count, sex, age, weight, and elevated baseline ALT. Tuberculosis chemotherapy (adjusted HR,8.5; 95% CI, 2.7–27), HBs antigenemia (adjusted HR, 3.0; 95% CI, 1.3–7.0) remained strongly associated with hepatotoxicity. Nadir CD4 < 100 cells/μl was marginally associated with hepatotoxicity (adjusted HR, 1.9; 95% CI, 0.96–3.8). The combined effect on hepatotoxicity of receiving tuberculosis therapy and having HBs antigenemia is demonstrated in Fig. 1.
Among the 40 subjects identified with severe hepatotoxicity, the clinical impact was modest. Headache, cough, sputum production, abdominal pain, diarrhea, or muscle pain was reported by 35% at the clinic visit within 28 days of hepatotoxicity compared to 16% for clinic visits not occurring within 28 days of hepatotoxicity. ART regimen change within 28 days of hepatotoxicity occurred for only eight subjects (20%); 32 subjects remained on the first-line regimen despite hepatotoxicity. Among individuals who did not change ART, three experienced repeat episodes of severe hepatotoxicity more than 1 month later and after a fall in ALT or AST to grade 2 or lower. All were continued without interruption on the first-line regimen, even after the second episode of severe hepatotoxicity. Nine subjects (22%) with grade 3 or 4 hepatotoxicity were in hospital or admitted to hospital within 28 days of the severe hepatotoxicity episode, although the admission was not necessarily directly related to hepatotoxicity and in many cases the elevation in liver enzymes may have been the result of non-hepatic systemic illness. Two deaths occurred within 28 days of severe hepatotoxicity; one was attributed to liver failure, the other to renal failure without any evidence of significant liver disease. The patient who died of hepatic failure was being treated for disseminated tuberculosis at the time of death. Two deaths among the 40 patients with hepatotoxicity (5%) is a similar proportion to the 39 deaths occurring during the first 12 months of ART among the 828 subjects (4.7%) without any episodes of hepatotoxicity.
Background liver enzyme elevations
The frequency of hepatotoxicity during ART appeared to be similar to the pre-ART background of liver enzyme elevations. During the 12 months prior to initiating ART, subjects had a median 63 days (IQR, 16–160 days) of care. During this pre-ART period 38% of subjects had only a baseline ALT and AST and 62% had two or more measurements. During the 12 month pre-ART initiation period, enzyme elevations > 5 times the ULN occurred among 4% of the subjects (34 events). There was no statistical difference in the proportion of subjects developing hepatotoxicity before and after ART initiation (McNemar's chi-square, P = 0.55) and liver enzyme elevation > 5 × ULN before ART initiation did not predict later severe hepatotoxicity.
To our knowledge this is the first detailed analysis of hepatotoxicity in an African HIV treatment program. We found that the risk of hepatotoxicity during ART is similar to that reported from studies in Europe and North America that also used efavirenz-based regimens despite higher prevalence of some risk factors, particularly HBs antigenemia, history of tuberculosis, and concurrent tuberculosis therapy [2,8,12]. Hepatotoxicity occurred at a rate of 7.7 episodes per 100 person-years with HBs antigenemia increasing the risk 3-fold, consistent with previous reports of a 2.5–9.2-fold increased risk [2,4,7,8]. We present a new finding of the contribution of tuberculosis therapy to hepatotoxicity, which increased the adjusted risk 8.5-fold. Analysis of tuberculosis therapy and ART may have been limited in previous studies because of a lower fraction of patients receiving concurrent tuberculosis therapy. Importantly, we did not find associations between hepatotoxicity and isoniazid preventive therapy, cotrimoxazole preventive therapy, or age. Consistent with previous reports from East and Southern Africa, hepatitis C was not found to be common in this cohort [10,11,24]. Hepatitis C is an important risk for hepatotoxicity in high-income country cohorts, possibly increasing risk beyond either hepatitis B or tuberculosis therapy. However, since it was absent in the subset tested in our cohort we were unable to perform comparisons.
Our analysis did identify a higher incidence of hepatotoxicity than reported from two African clinical trials also using an efavirenz-based regimen. Specifically, the FTC-302 study conducted in South Africa found no cases of severe hepatotoxicity at 12 weeks follow-up among 83 subjects on efavirenz . The subjects included in the FTC-302 study differed from our cohort in that they were predominantly women with low prevalence of HBs antigenemia (3.8%), not on concomitant tuberculosis therapy, and had much higher CD4 cell counts. Further, the Trivacan study in the Ivory Coast reported hepatotoxicity among 0.81% of subjects during the first 6 months of follow-up (approximately 1.6 episodes per 100 person-years) . Subjects in that study were also primarily women with higher CD4 cell counts (median, 255 cells/μl). The lower risk of hepatotoxicity reported from these two studies may be explained by a healthier population with fewer hepatotoxicity risk factors, shorter follow-up, fewer transaminase tests per subject, and possibly exclusion of individuals with abnormal baseline transaminases. Finally, although not specifically assessed in our study, traditional medicine use is common in our cohort (32% of patients on ART) , including use of known hepatotoxins, which may also have contributed to a higher incidence of hepatotoxicity.
Several mechanisms may account for hepatotoxicity among patients either on tuberculosis therapy or with HBs antigenemia. First, three of the four first-line anti-tuberculosis agents are known hepatotoxins: isoniazid, rifampicin, and pyrazinamide. In addition, tuberculosis may lead to immune reconstitution inflammatory syndrome (IRIS) if ART was started before, or concurrently with, tuberculosis therapy. IRIS often leads to hepatic inflammation and transaminase elevation due to immune surveillance of mycobacterial antigens in the liver. Hepatotoxicity can be related to HBV through underlying liver damage and HBV mediated acute hepatic inflammation. In addition, during lamivudine therapy, some individuals with suppressed HBV replication may develop breakthrough HBV replication following development of point mutations in the HBV genome that confer resistance to lamivudine; this may lead to hepatic inflammation and transaminase elevation . It is unknown whether use of a HBV suppressive agent, such as lamivudine, during ART impacts early HBV-related hepatotoxicity in co-infected patients. Tuberculosis disease and hepatitis B may also lead to ALT and AST elevation, even in the absence of ART.
Remarkably, the fraction of subjects developing severe hepatotoxicity on ART appeared similar to the background rate of liver enzyme elevations (4.6% versus 4.0%) suggesting that many of the hepatotoxicity episodes that we observed may be related to pre-existent factors. We have found this despite obtaining fewer liver enzyme measurements pre-ART than on ART. Aside from a single baseline assay, ALT and AST measurement was not part of scheduled routine testing before starting ART. Thus, had routine transaminase monitoring occurred with similar frequency before starting ART and during ART, we may have identified an even higher incidence of liver enzyme elevations before ART. No other studies from low-income settings have reported frequency of liver enzyme elevation before subjects started ART, possibly because of a shorter period of pre-ART management and more restricted access to laboratory monitoring.
Importantly, we observed overall minimal morbidity from hepatotoxicity even among patients with HBsAg or on anti-tuberculosis therapy who developed hepatotoxicity. Of those individuals developing hepatotoxicity, only nine individuals were admitted to hospital, for any reason, within 28 days of hepatotoxicity, eight had a change in ART regimen, and two deaths were potentially related to hepatotoxicity. Of these two deaths only one was clearly liver related and that individual had severe concomitant disease.
These findings support an approach of treating patients with ART even if they also require tuberculosis therapy or have a positive HBsAg test. When clinically appropriate, deferring ART until after completion of tuberculosis treatment should be considered to further reduce the risk of hepatotoxicity, consistent with WHO recommendations . As diagnosis of advanced tuberculosis disease among HIV-infected individuals in Africa is common , increased emphasis on early and thorough screening of individuals for tuberculosis could increase identification of cases for treatment before those patients have immunologic indication for ART. Our findings also support close early follow-up of patients with HBs antigenemia.
Importantly, we did not observe increased severe hepatotoxicity among subjects receiving both ART and isoniazid preventive therapy, further supporting the safety of preventive therapy among this population at high risk for reactivation of latent tuberculosis.
Strengths of this study are the use of a standard protocol for managing and monitoring patients and the completeness of comprehensive follow-up records, which included frequent liver function assays (median of eight per subject), tuberculosis clinic data, hospital admission records, and mortality reports. However, there are potential limitations to our study. The most important limitation is the generalizability of findings. Although the patients had advanced HIV before starting ART, similar to many African ART treatment cohorts, the population studied was predominantly male (94%) and employed (100%). However, it is unclear what effect sex and employment may have on hepatotoxicity with an efavirenz-based regimen. In addition, the number of subjects on nevirapine was insufficient to power a comparison of hepatotoxicity between efavirenz and nevirapine. A second limitation is the lack of repeat HBsAg testing to confirm the diagnosis of chronic HBV infection among subjects with a single positive test; in addition, not all the subjects were tested for HBsAg. We feel that the former is unlikely to have resulted in a significant overestimation of the prevalence of chronic hepatitis B because most hepatitis B in South Africa is transmitted during childhood; the risk of hepatotoxicity associated with HBsAg that we calculated was within the range reported by other cohorts, and, for the latter, we stratified HBsAg status into the three categories of positive, negative, and not tested.
These findings have important public health implications regarding use of ART: (i) hepatotoxicity with efavirenz-based ART is not a major cause of morbidity; (ii) hepatitis B infection is common and increases hepatotoxicity, but not to the extent that ART is compromised; (iii) co-administration of tuberculosis treatment and ART increases the risk of hepatotoxicity, but is well tolerated overall; and (iv) use of isoniazid preventive therapy during ART did not increase hepatotoxicity. Although we have found an efavirenz-based regimen to be relatively safe, we believe that even the limited morbidity of co-administration of tuberculosis therapy and ART should reinforce calls for more aggressive case-finding of tuberculosis disease and treatment of latent tuberculosis infection among HIV-infected individuals earlier in the natural history of HIV before ART is necessary. In endemic areas, individuals found not to have tuberculosis disease could be treated for latent tuberculosis infection.
In addition, as patients continue long-term ART, the distribution of side effects may differ considerably from those observed during the first year of therapy [29,30]. Given the potential for hepatotoxicity to manifest years after ART initiation and the high prevalence of chronic hepatitis B in much of sub-Saharan Africa, we believe that further research is important to describe the sequelae of chronic ART use and HIV-HBV co-infection in Africa.
The authors thank AngloGoldAshanti for allowing us to use their data and for their support for doing this study. In addition, we would like to thank the following individuals for their contributions to this study: Michael Eisenstein and Pule Seatlanyane, both from Aurum Institute for Health Research.
This work was supported by NIH AI5535901 and the Aurum Institute. CJH was supported by NIH DK074348, CLT by NIH AI60449, REC by NIH AI5535901 and AI016137, and ADG by a UK Department of Health Public Health Career Scientist Award.
1. Sulkowski MS, Thomas DL, Chaisson RE, Moore RD. Hepatotoxicity associated with antiretroviral therapy in adults infected with human immunodeficiency virus and the role of hepatitis C or B virus infection. JAMA 2000; 283:74–80.
2. Sulkowski MS, Thomas DL, Mehta SH, Chaisson RE, Moore RD. Hepatotoxicity associated with nevirapine or efavirenz-containing antiretroviral therapy: role of hepatitis C and B infections. Hepatology 2002; 35:182–189.
3. Bruno R, Sacchi P, Maiocchi L, Zocchetti C, Filice G. Hepatotoxicity nelfinavir: a meta-analysis. Clin Gastroent Hepatol 2005; 3:482–488.
4. den Brinker M, Wit FW, Wertheim-van Dillen PM, Jurriaans S, Weel J, van Leeuwen R, et al
. Hepatitis B and C virus co-infection and the risk for hepatotoxicity of highly active antiretroviral therapy in HIV-1 infection. AIDS 2000; 14:2895–2902.
5. Dieterich DT, Robinson PA, Love J, Stern JO. Drug-induced liver injury associated with the use of nonnucleoside reverse-transcriptase inhibitors. Clin Infect Dis 2004; 38(Suppl 2):S80–S90.
6. Nunez M, Soriano V. Hepatotoxicity of antiretrovirals. Drug Safety 2005; 28:53–66.
7. Wit FWM, Weverling GJ, Weel J, Jurriaans S, Lange JMA. Incidence of and risk factors for severe hepatotoxicity associated with antiretroviral combination therapy. J Infect Dis 2002; 186:23–31.
8. Law WP, Dore GJ, Duncombe CJ, Mahanontharit A, Boyd MA, Ruxrungtham K, et al
. Risk of severe hepatotoxicity associated with antiretroviral therapy in the HIV-NAT Cohort, Thailand, 1996–2001. AIDS 2003; 17:2191–2199.
9. Dieleman JP, Jambroes M, Gyssens IC, Sturkenboom MC, Stricker BH, Mulder WM, et al
. Determinants of recurrent toxicity-driven switches of highly active antiretroviral therapy. AIDS 2002; 16:737–745.
10. Lodenyo H, Schoub B, Ally R, Segal I. Hepatitis B and C virus infections and liver functions in AIDS patients at Chris Hani Baragwanath Hospital, Johannesburg. East Afr Med J 2000; 77:13–15.
11. Matee MIN, Magesa PM, Lyamuya EF. Seroprevalence of human immunodeficiency virus, hepatitis B and C viruses and syphilis infections among blood donors at the Muhimbili National Hospital in Dar es Salaam, Tanzania. BMC Public Health 2006; 6:21.
12. van Leth F, Phanuphak P, Ruxrungtham K, Baraldi E, Miller S, Gazzard B, et al
. Comparison of first-line antiretroviral therapy with regimens including nevirapine, efavirenz, or both drugs, plus stavudine and lamivudine. Lancet 2004; 363:1253–1263.
13. Menendez C, Sanchez-Tapias JM, Kahigwa E, Mshinda H, Costa J, Vidal J, et al
. Prevalence and mother-to-infant transmission of hepatitis viruses B, C, and E in Southern Tanzania. J Med Virol 1999; 58:215–220.
14. Sutcliffe S, Taha TE, Kumwenda NI, Taylor E, Liomba GN. HIV-1 prevalence and herpes simplex virus 2, hepatitis C virus, and hepatitis B virus infections among male workers at a sugar estate in Malawi. J Acquir Immune Defic Syndr 2002; 31:90–97.
15. Rouet F, Chaix M-L, Inwoley A, Msellati P, Viho I, Combe P, et al
. HBV and HCV prevalence and viraemia in HIV-positive and HIV-negative pregnant women in Abijan, Cote d'Ivoire: the ANRS 1236 Study. J Med Virol 2004; 74:34–40.
16. Danel C, Moh R, Anzian A, Abo Y, Chenal H, Guehi C, et al
. Tolerance and acceptability of an efavirenz-based regimen in 740 adults in west Africa. J Acquir Immune Defic Syndr 2006; 42:29–35.
17. Corbett EL, Steketee RW, ter Kuile FO, Latif AS, Kamali A, Hayes RJ. HIV-1/AIDS and the control of other infectious diseases in Africa. Lancet 2002; 359:2177–2187.
18. Grant AD, Charalambous S, Fielding KL, Day JH, Corbett EL, Chaisson RE, et al
. Effect of routine isoniazid preventive therapy on tuberculosis incidence among HIV-infected men in South Africa. JAMA 2005; 293:2719–2725.
19. Dean GL, Edwards SG, Ives NJ, Matthews G, Fox EF, Navaratne L, et al
. Treatment of tuberculosis in HIV-infected persons in the era of highly active antiretroviral therapy. AIDS 2002; 16:75–83.
20. Stewart MJ, Steenkamp V, van der Merwe S, Zuckerman M, Crowther NJ. The cytotoxic effects of a traditional Zulu remedy, impila. Hum Exp Toxicol 2002; 21:643–647.
21. Mills E, Cooper C, Seely D, Kanfer I. African herbal medicines in the treatment of HIV: Hypoxis and Sutherlandia. An overview of evidence and pharmacology. Nutr J 2005; 4:19.
22. Babb DA, Pemba L, Seatlanyane P, Charalambous S, Churchyard GJ, Grant AD. Use of traditional medicine by HIV-infected individuals in South Africa in the era of antiretroviral therapy
. Psychol Health Med
2006 (in press).
23. Charalambous S, Grant AD, Day JH, Pemba L, Chaisson RE, Kruger P, et al
. Establishing a workplace antiretroviral therapy programme in South Africa. AIDS Care 2007; 19:34–41.
24. Fang CT, Field SP, Busch MP, Heyns Ad. Human immunodeficiency virus-1 and hepatitis C virus RNA among South African blood donors. Vox Sang 2003; 85:9–19.
25. Sanne I, Mommeja-Marin H, Hinkle J, Bartlett JA, Lederman MM, Maartens G, et al
. Severe hepatotoxicity associated with nevirapine use in HIV-infected subjects. J Infect Dis 2005; 191:825–829.
26. Liaw Y-F, Chien R-N, Yeh C-T, Tsai S-L, Chu C-M. Acute exacerbation and hepatitis B virus clearance after emergence of YMDD motif mutation during lamivudine therapy. Hepatology 1999; 30:567–572.
27. TB/HIV: A Clinical Manual.
Geneva: World Health Organization; 2004.
28. Wood R, Middelkoop K, Myer L, Grant AD, Whitelaw A, Lawn SD, et al.Undiagnosed tuberculosis in a community with high HIV-prevalence: implications for TB control
. Am J Repir Crit Care Med
2006; Electronic release.
29. Bica I, McGovern B, Dhar R, Stone D, McGowan K, Scheib R, et al
. Increasing mortality due to end-stage liver disease in patients with human immunodeficiency virus infection. Clin Infect Dis 2001; 32:492–497.
30. Gebo KA, Diener-West M, Moore RD. Hospitalization rates differ by hepatitis C status in an urban HIV cohort. J Acquir Immune Defic Syndr 2003; 34:165–173.