Secondary Logo

Journal Logo


Zidovudine impairs immunological recovery on first-line antiretroviral therapy: collaborative analysis of cohort studies in southern Africa

Wandeler, Gillesa,b; Gsponer, Thomasa; Mulenga, Lloydc; Garone, Danielad; Wood, Robine; Maskew, Mhairif; Prozesky, Hansg; Hoffmann, Christopherh; Ehmer, Jocheni; Dickinson, Dianaj; Davies, Mary-Annk; Egger, Matthiasa,k; Keiser, Oliviaafor the IeDEA Southern Africa Collaboration

Author Information
doi: 10.1097/QAD.0b013e328362d887


In low-income countries, the WHO recommends the use of standardized first-line antiretroviral therapy (ART) consisting of one nonnucleoside reverse transcriptase inhibitor (NNRTI) and two nucleoside reverse transcriptase inhibitors (NRTIs) [1]. At present, the two preferred backbones are lamivudine (3TC) or emtricitabine (FTC) + tenofovir (TDF) or 3TC + zidovudine (ZDV). Although the efficacy and toxicity of these regimens have been studied extensively in clinical trials in high-income countries, data on long-term clinical outcomes in sub-Saharan Africa are scarce.

A concern with the use of ZDV is its well known myelosuppressive effect, sometimes leading to severe hematologic toxicity [2,3]. ZDV-related anemia has been described in many studies, including in reports from sub-Saharan Africa [4,5]. As a consequence, most ART guidelines have precluded its use in patients with severe anemia. The impaired immunological recovery related to the use of ZDV has also been reported in clinical trials and observational studies, both in industrialized countries [6–11] and sub-Saharan Africa [12,13]. However, these studies were small and had limited follow-up time. In southern Africa, most of the HIV-infected patients are treated in settings in which viral load monitoring is not available and are, thus, switched to second-line regimens according to clinical and immunological criteria. Several studies, including analyses from the International epidemiologic Databases to Evaluate AIDS (IeDEA) [14–16], have described risk factors for impaired immunological response to ART, but only few publications have examined the role of specific antiretroviral drugs in sub-Saharan Africa. To avoid unnecessary switching to expensive protease inhibitor-based second-line ART, a better understanding of the factors impairing immunological recovery is needed.

We compared 5-year CD4+ cell count trajectories after ART initiation between patients on first-line regimen containing ZDV and those not receiving ZDV and assessed the association between ZDV and severe impairment of immunological recovery in a large collaborative analysis of nine cohorts in four countries in southern Africa.


Antiretroviral treatment programs

The International epidemiological Databases to Evaluate AIDS in southern Africa (IeDEA-SA) is a large collaboration of ART programs in southern Africa [17]. Data are collected at ART initiation (baseline) and each follow-up visit, using standardized instruments, and transferred to data centers at the Universities of Cape Town, Republic of South Africa (RSA) and Bern, Switzerland. All sites have ethical approval to collect data and to participate in IeDEA-SA.

We included all cohorts with at least 100 adult patients in both treatment groups (with and without ZDV). Six ART programs in RSA and one in Botswana monitored viral load and CD4+ cell counts every 6 months during the first year of ART, and then yearly: Aurum Institute (community and workplace), Themba Lethu, Khayelitsha, Gugulethu, Tygerberg, in South Africa, as well as Gaborone private clinic in Botswana. Two cohorts with routine CD4+ cell count, but no viral load monitoring were also included: the Ministry of Health, Centre for Infectious Disease Research in Zambia (MoH-CIDRZ) and the SolidarMed ART program in rural Lesotho. All treatment programs trace patients lost to follow-up.

Eligibility criteria

All patients aged 16 years and older who started a first-line ART regimen and had at least two CD4+ cell counts performed (one at ART start and at least one during follow-up) were included. We defined first-line regimens according to the WHO treatment guidelines as a regimen including an NNRTI and two NRTIs. Patients with prior exposure to ART before inclusion in the treatment program were excluded.


We compared CD4+ trajectories between patients receiving a first-line regimen including ZDV with those on other ART combinations over the first 5 years of treatment. Differences in CD4+ cell counts after 1 and 5 years of ART were compared between patients receiving ZDV and those on other first-line ART regimens. Follow-up was censored at the date of the first treatment change, the last visit, or the end of the study period, whichever occurred first. In a second analysis, we focused on patients who started ART with severe immunodeficiency, defined as a baseline CD4+ cell count below 100 cells/μl. WHO included three specific criteria in the definition of immunological treatment failure, one of which being the persistence of a CD4+ cell count below 100 cells/μl. We assessed the association between ZDV and impairment of immunological recovery, defined as a CD4+ cell count remaining below 100 cells/μl 1 year after the initiation of ART.

Statistical analyses

Baseline characteristics of patients on ZDV-containing regimens were compared to those of patients not on ZDV using χ2 and Mann–Whitney tests for categorical and continuous variables, respectively. We used multivariable linear mixed-effect models to analyze CD4+ cell count trajectories over the first 5 years of ART, as described in detail elsewhere [16]. Differences in CD4+ cell counts after 1 and 5 years of ART were compared between patients on ZDV and not on ZDV. The CD4+ cell counts were square-root transformed and trajectories modeled using fractional polynomials. Results are presented as marginal CD4+ estimates and trajectories. Additionally, differences in CD4+ cell counts between the two treatment groups within each cohort were described in a forest plot and overall estimates obtained using random-effect meta-analysis. We repeated all analyses in patients who remained virologically suppressed during follow-up in South Africa and Botswana and within each baseline CD4+ cell count category. Predictors of impairment of immunological recovery during the first year of ART were evaluated using multivariable logistic regression. For the latter analysis, only patients who started a first-line ART regimen with a CD4+ cell count below 100 cells/μl and with an available CD4+ measurement between 8 and 15 months were included.

All multivariable analyses were adjusted for sex, age (16–29, 30–39, or 40 years and over), CD4+ cell count at baseline (0–49, 50–99, 100–199, ≥200 cells/μl), calendar year of starting first-line ART (before 2007, 2007, 2008, 2009, and 2010), degree of anemia at baseline, type of NNRTI (efavirenz or nevirapine), and viral load monitoring (yes or no). Anemia was defined as severe (hemoglobin <5.0 mmol/l), moderate (5.0–<6.2 mmol/l in women and 5.0–<6.8 mmol/l in men), mild (6.2–<7.4 mmol/l in women and 6.8–<8.1 mmol/l in men), or none (≥7.4 mmol/l in women and ≥8.1 mmol/l in men). Multiple imputation was used to impute missing hemoglobin values at baseline, with analyses run on each of 20 datasets and results combined with Rubin's rules [18]. All analyses were performed using Stata software version 11 (College Station, Texas, USA).


Antiretroviral therapy program and patients characteristics

Table 1 shows the composition of cohorts. A total of 72 597 patients on first-line ART, including 19 758 (27.2%) on an ZDV-containing regimen were included in the analyses. The majority of patients were women in all cohorts except for the AURUM workplace cohort in South Africa, which was dominated by male miners. The median age ranged from 33 years in Gugulethu and Khayelitsha to 43 years in the AURUM workplace cohort. In South Africa, the proportion of patients on ZDV ranged from 2.8% in Thembalethu to 82.9% in the AURUM workplace cohort, and outside South Africa, this proportion ranged from 28.3% in the CIDRZ program in Lusaka to 77.1% in Gaborone, Botswana. Median CD4+ cell count and hemoglobin levels at initiation of ART ranged from 90 cells/μl [interquartile range (IQR): 41–158)] and 10.9 g/dl (9.6–12.1) in Khayelitsha to 160 cells/μl (90–228) and 13.3 g/dl (11.8–14.5) in the AURUM miners cohort, respectively. Data on baseline hemoglobin level were overall missing for 13.1% of patients.

Table 1
Table 1:
Baseline characteristics of patients.

Patients receiving ZDV were less likely to be women, to have started ART in recent years, and to be treated in the RSA or Botswana (Table 2). Median baseline CD4+ cell count (150 vs. 128 cells/μl, P < 0.001) and hemoglobin level (12.0 vs. 11.0 g/dl, P < 0.001) were higher in patients on ZDV compared to those in the other treatment group (Table 2). Severe anemia was present in 8.0% of patients initiating ART not containing ZDV and in 0.8% of patients receiving ZDV (P < 0.001). The most common ART combinations were 3TC/ZDV/NVP (68.0%) and 3TC/ZDV/EFV (31.8%) in the ZDV group, and 3TC/D4T/NVP (34.3%), 3TC/D4T/EFV (29.3%), FTC/TDF/EFV (20.5%), and FTC/TDF/NVP (11.5%) in the non-ZDV group.

Table 2
Table 2:
Characteristics of patients at the start of first-line antiretroviral therapy.

CD4+ trajectories

The total follow-up time was 34 964 and 68 353 person-years for the ZDV and non-ZDV groups, respectively. Figure 1 shows CD4+ trajectories over the first 5 years of ART, by first-line treatment group. During follow-up, CD4+ cell counts were similar during much of the first year on ART, but substantial differences emerged from year 1 onward. After 1 and 5 years of ART, patients on ZDV reached estimated CD4+ cell counts of 301 cells/μl [95% confidence interval (CI) 299–302] and 386 cells/μl (383–389), whereas those not on ZDV reached CD4+ cell counts of 317 cells/μl (316–318) and 442 cells/μl (440–444), respectively. Estimated differences were −16 CD4 cells/μl (95% CI −18 to −14) at 1 year and −56 CD4 cells/μl (52–59) at 5 years in favor of the group of patients not receiving ZDV (Web table 1, Adjusted estimates of difference in CD4+ cell count between treatment groups for each cohort are shown in Figure 2. One and 5 years after the initiation of ART, most cohorts showed a higher absolute CD4+ cell count in patients not receiving ZDV compared to those on ZDV. At each time point, only three out of nine cohorts had marginally higher CD4+ cell counts in patients on ZDV, but these differences were not statistically significant. However, there was substantial heterogeneity between cohorts in both analyses (χ2 test for heterogeneity: P = 0.01 at 1 year and P < 0.001 at 5 years).

Fig. 1
Fig. 1:
Estimated CD4+ trajectories in patients starting first-line antiretroviral therapy with and without zidovudine.The dotted lines indicate 95% confidence intervals. ART, antiretroviral therapy; ZDV, zidovudine.
Fig. 2
Fig. 2:
Adjusted differences in CD4+ cell counts between patients on zidovudine and those not on zidovudine.Adjusted differences in CD4+ cell counts between patients on zidovudine (ZDV) and those not on ZDV after 1 (a) and 5 years (b) of antiretroviral therapy (ART), by cohort. I2 (P value from test of heterogeneity): 1 year (a): 58.7% (P = 0.013); 5 years (b): 87.8% (P < 0.001). Analyses were adjusted for sex, age, CD4+ and anemia at baseline, calendar year, type of nonnucleoside reverse transcriptase inhibitor (NNRTI), and viral load monitoring.

Several additional analyses were performed to examine immunological recovery in subgroups of patients (Web table 1, The difference between the treatment groups was most pronounced in patients who started ART with a CD4+ cell count below 100 cells/μl: −66 cells/μl (95% CI −72 to −61) at 5 years in comparing ZDV with no ZDV. In RSA and Botswana, where virological monitoring was available, a higher proportion of patients on ZDV had at least one detectable viral load during follow-up compared to those not on ZDV (39 vs. 20%, P < 0.001). However, in analyses restricted to patients who were fully suppressed 6 months after the initiation of ART and who remained so during the whole follow-up period, we found no difference in CD4+ cell count at 1 year between patients on ZDV and those not on ZDV and a significant difference at 5 years (difference −21 cells/μl, 95% CI −29 to −14) (Web table 1,

Predictors of impairment of immunological recovery

This analysis was based on 14 529 patients who started ART with CD4+ cell counts below 100 cells/μl. The number of imputed hemoglobin values were 938 (15.1%) in the ZDV group and 1137 (10.4%) in the other group. A total of 407 patients (11.4%) on ZDV and 914 patients (8.3%) on other regimens remained below 100 CD4+ cells/μl at 1 year (P < 0.001). In multivariate logistic regression, patients on ZDV were more likely to experience severe impairment of immunological recovery than those on other first-line regimens [adjusted odds ratio (aOR) 1.40, 95% CI 1.22–1.61] (Table 3). When this analysis was repeated on the complete-case dataset, the result was similar (aOR 1.38, 95% CI 1.19–1.60, data not shown). Male patients, those older than 40 years as well as patients with very low CD4 cell counts (<50 cells/μl) at start of ART were also at increased risk of impaired immunological recovery (Table 3).

Table 3
Table 3:
Predictors of severe impairment in immunological recovery (N = 14 529).


This study of nine ART programs and over 70 000 patients from four countries in southern Africa found that ZDV influenced the magnitude of the CD4+ cell count increase over the first 5 years of treatment. Patients receiving ZDV had lower CD4+ cell counts at 1 and 5 years and were more likely to remain severely immune suppressed during the first year of treatment. The difference in immunological recovery between patients on ZDV and those on other backbones was consistent across baseline CD4+ cell count categories and, to a lesser extent, was also seen in the subset of patients who had an undetectable viral load throughout the follow-up period.

Numerous studies from high-income countries have assessed predictors of CD4+ cell count increase following ART initiation [19–22]. Even though some individual characteristics, including older age, lower baseline CD4+ cell counts, and hepatitis C virus coinfections have consistently been shown to be associated with impaired immunological recovery, there remains controversy on the impact of specific ART components on CD4+ cell recovery [11]. In accordance with the results of two small studies from Botswana and Cameroon [12,13], we showed that patients on ZDV reached significantly lower CD4+ cell counts on ART compared to those on other regimens. The reasons for the association of ZDV with impairment of total CD4+ lymphocyte recovery are poorly understood: the most probable explanation is related to the well known bone marrow suppression caused by ZDV [2–5]. Data from the Swiss HIV Cohort Study showed reduced absolute but not relative CD4+ cell counts in patients on ZDV, suggesting a general effect on total lymphocyte count, rather than a CD4+-specific effect [8].

Impaired immunological recovery in patients on ART may not have important clinical consequences when it concerns patients with high baseline CD4+ cell counts. However, in those with very low CD4+ cell counts at treatment initiation, failure to reach adequate cellular immunity not only leads to an increased mortality and HIV-associated morbidity, but also has important implications on clinical decision-making: a CD4+ cell count persistently below 100 cells/μl is one of the three criteria used to diagnose immunological failure [1]. Thus, in settings without access to viral load testing, many patients who remain under this threshold during ART are switched to a protease inhibitor-based regimen, despite possible virological efficacy of the first-line regimen. In our study, patients on ZDV were more likely to remain severely immune suppressed after 1 year of ART compared to those not receiving ZDV. Even though in our study the observed inferior virological efficacy of ZDV-containing regimens in RSA might partially explain the impaired immunological recovery in this group, we found a difference in long-term CD4+ recovery between the two groups when the analyses were restricted to virologically suppressed patients. This suggests that the impact of ZDV on immunological recovery is, to some extent, independent of its virological efficacy.

Only few studies have compared outcomes between different ART regimens in sub-Saharan Africa [23,24]. Our study is unique regarding its sample size and length of follow-up: it involved a large number of patients from a wide range of settings in southern Africa and described immunological recovery over 5 years after the initiation of ART. The main limitation of observational cohort data comparing treatments in different countries lies in the lack of randomization and the heterogeneity between the treatment sites. Confounding by indication may have affected our results: patients with advanced clinical disease and low CD4+ cell counts are also more likely to have severe anemia, which is the main reason for not being prescribed ZDV. As a consequence, sicker patients are less likely to receive ZDV in many settings. This might have led to an underestimation of the negative impact of ZDV on immunological recovery. Unfortunately, due to the lack of data on pre-ART follow-up, causal modeling was not possible. Furthermore, baseline hemoglobin levels were missing in a substantial number of patients: although we addressed this issue by using multiple imputation, the missing data may have biased our results. A further limitation of our study was the wide variation in the proportion of patients on a ZDV-containing regimen across countries and calendar time. This issue was described in detail in a previous multiregional IeDEA analysis [5] and most probably reflects national treatment guidelines. Although the association between ZDV and impaired immunological recovery was fairly consistent across most cohorts in our study, the results from a few treatment programs differed substantially from the overall estimates. Finally, we had no data on relative CD4+ cell counts (CD4+ percentage), which could have given more information on the mechanism of impaired immunological recovery under ZDV-containing ART.

In conclusion, patients on first-line ART regimens including ZDV are at risk of inferior CD4+ cell recovery compared to those treated with other backbones. Treatment failure in patients remaining severely immune suppressed during the first year on ZDV should be confirmed, wherever possible, with viral load testing before switching them to a second-line regimen. In patients with impaired immunological recovery but adequate virological response with an ZDV-containing regimen, the replacement of ZDV by another NRTI might be an adequate therapeutic approach. Further studies evaluating the role of specific ART components in immunological recovery are needed, especially from settings in which virological monitoring and data on relative CD4+ cell counts (or percentage) are available.


The authors would like to thank all study participants and staff of all participating sites.

G.W., O.K., and M.E. designed the study. G.W. and O.K. performed the statistical analyses and wrote the first draft of the article. All authors contributed to the interpretation of the results and to the final version of the article. G.W. had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. M.E. and M.-A.D. are the principal investigators of IeDEA Southern Africa.

This study was supported by the National Institute of Allergy and Infectious Diseases (grant 1 U01AI069924–01) and a fellowship to O.K. from the Swiss National Science Foundation (Grant 32333B_131629).

Conflicts of interest

The authors declare no conflict of interest.


1. World Health Organization. Antiretroviral Therapy for HIV Infection in Adults and Adolescents: recommendations for a Public Health Approach. 2010 Revision. Geneva, Switzerland: World Health Organization. [Accessed 30 January 2013].
2. Moyle G, Sawyer W, Law M, Amin J, Hill A. Changes in hematologic parameters and efficacy of thymidine analogue-based, highly active antiretroviral therapy: a meta-analysis of six prospective, randomized, comparative studies. Clin Ther 2004; 26:92–97.
3. Richman DD, Fischl MA, Grieco MH, Gottlieb MS, Volberding PA, Laskin OL, et al. The toxicity of azidothymidine (ZDV) in the treatment of patients with AIDS and AIDS-related complex. A double-blind, placebo-controlled trial. N Engl J Med 1987; 317:192–197.
4. Ssali F, Stohr W, Munderi P, Reid A, Walker AS, Gibb DM, et al. Prevalence, incidence and predictors of severe anaemia with zidovudine-containing regimens in African adults with HIV infection within the DART trial. Antivir Ther 2006; 11:741–749.
5. Zhou J, Jaquet A, Bissagnene E, Musick B, Wools-Kaloustian K, Maxwell N, et al. Short-term risk of anaemia following initiation of combination antiretroviral treatment in HIV-infected patients in countries in sub-Saharan Africa, Asia-Pacific, and central and South America. J Int AIDS Soc 2012; 15:5.
6. Pozniak AL, Gallant JE, DeJesus E, Arribas JR, Gazzard B, Campo RE, et al. Tenofovir disoproxil fumarate, emtricitabine, and efavirenz versus fixed-dose zidovudine/lamivudine and efavirenz in antiretroviral-naive patients: virologic, immunologic, and morphologic changes – a 96-week analysis. J Acquir Immune Defic Syndr 2006; 43:535–540.
7. Gallant JE, DeJesus E, Arribas JR, Pozniak AL, Gazzard B, Campo RE, et al. Tenofovir DF, emtricitabine, and efavirenz vs. zidovudine, lamivudine, and efavirenz for HIV. N Engl J Med 2006; 354:251–260.
8. Huttner AC, Kaufmann GR, Battegay M, Weber R, Opravil M. Treatment initiation with zidovudine-containing potent antiretroviral therapy impairs CD4 cell count recovery but not clinical efficacy. AIDS 2007; 21:939–946.
9. Eron JJ Jr, Murphy RL, Peterson D, Pottage J, Parenti DM, Jemsek J, et al. A comparison of stavudine, didanosine and indinavir with zidovudine, lamivudine and indinavir for the initial treatment of HIV-1 infected individuals: Selection of Thymidine Analog Regimen Therapy (START II). AIDS 2000; 14:1601–1610.
10. Crespo M, Ribera E, Suarez-Lozano I, Domingo P, Pedrol E, Lopez-Aldeguer J, et al. Effectiveness and safety of didanosine, lamivudine and efavirenz versus zidovudine, lamivudine and efavirenz for the initial treatment of HIV-infected patients from the Spanish VACH cohort. J Antimicrob Chemother 2009; 63:189–196.
11. Byakwaga H, Zhou J, Petoumenos K, Law MG, Boyd MA, Emery S, et al. Effect of nucleoside reverse transcriptase inhibitors on CD4 T-cell recovery in HIV-1-infected individuals receiving long-term fully suppressive combination antiretroviral therapy. HIV Med 2009; 10:143–151.
12. Laurent C, Bourgeois A, Mpoudi-Ngole E, Ciaffi L, Kouanfack C, Mougnutou R, et al. Tolerability and effectiveness of first-line regimens combining nevirapine and lamivudine plus zidovudine or stavudine in Cameroon. AIDS Res Hum Retroviruses 2008; 24:393–399.
13. Bussmann H, Wester CW, Thomas A, Novitsky V, Okezie R, Muzenda T, et al. Response to zidovudine/didanosine-containing combination antiretroviral therapy among HIV-1 subtype C-infected adults in Botswana: two-year outcomes from a randomized clinical trial. J Acquir Immune Defic Syndr 2009; 51:37–46.
14. Zoufaly A, an der Heiden M, Kollan C, Bogner JR, Fatkenheuer G, Wasmuth JC, et al. Clinical outcome of HIV-infected patients with discordant virological and immunological response to antiretroviral therapy. J Infect Dis 2011; 203:364–371.
15. Tuboi SH, Brinkhof MW, Egger M, Stone RA, Braitstein P, Nash D, et al. Discordant responses to potent antiretroviral treatment in previously naive HIV-1-infected adults initiating treatment in resource-constrained countries: the Antiretroviral Therapy in Low-Income Countries (ART-LINC) collaboration. J Acquir Immune Defic Syndr 2007; 45:52–59.
16. Nash D, Katyal M, Brinkhof MW, Keiser O, May M, Hughes R, et al. Long-term immunologic response to antiretroviral therapy in low-income countries: a collaborative analysis of prospective studies. AIDS 2008; 22:2291–2302.
17. Egger M, Ekouevi DK, Williams C, Lyamuya RE, Mukumbi H, Braitstein P, et al. Cohort profile: the International epidemiological Databases to Evaluate AIDS (IeDEA) in sub-Saharan Africa. Int J Epidemiol 2012; 41:1256–1264.
18. Rubin DR. Multiple imputation for nonresponse in surveys. New York:John Wiley & Sons; 1987.
19. Kaufmann GR, Bloch M, Finlayson R, Zaunders J, Smith D, Cooper DA. The extent of HIV-1-related immunodeficiency and age predict the long-term CD4 T lymphocyte response to potent antiretroviral therapy. AIDS 2002; 16:359–367.
20. Miller MF, Haley C, Koziel MJ, Rowley CF. Impact of hepatitis C virus on immune restoration in HIV-infected patients who start highly active antiretroviral therapy: a meta-analysis. Clin Infect Dis 2005; 41:713–720.
21. Florence E, Lundgren J, Dreezen C, Fisher M, Kirk O, Blaxhult A, et al. Factors associated with a reduced CD4 lymphocyte count response to HAART despite full viral suppression in the EuroSIDA study. HIV Med 2003; 4:255–262.
22. Garcia F, de Lazzari E, Plana M, Castro P, Mestre G, Nomdedeu M, et al. Long-term CD4+ T-cell response to highly active antiretroviral therapy according to baseline CD4+ T-cell count. J Acquir Immune Defic Syndr 2004; 36:702–713.
23. Spaulding A, Rutherford GW, Siegfried N. Stavudine or zidovudine in three-drug combination therapy for initial treatment of HIV infection in antiretroviral-naive individuals. Cochrane Database Syst Rev 2010. CD008651.
24. Spaulding A, Rutherford GW, Siegfried N. Tenofovir or zidovudine in three-drug combination therapy with one nucleoside reverse transcriptase inhibitor and one nonnucleoside reverse transcriptase inhibitor for initial treatment of HIV infection in antiretroviral-naive individuals. Cochrane Database Syst Rev 2010. CD008740.

cohort study; first-line antiretroviral therapy; immunological recovery; southern Africa; zidovudine

Supplemental Digital Content

© 2013 Lippincott Williams & Wilkins, Inc.