BACKGROUND
As the number of patients receiving first-line antiretroviral therapy (ART) has expanded in South Africa, so too have the number experiencing first-line ART regimen failure 1–4. Previously, we reported that specific resistance mutations encountered in South Africa at first ART failure are M184V/I (64%), K103N (51%), thymidine analogue resistance mutations (TAMs; 32%), V106M (19%) and protease inhibitor (PI) resistance mutations (4%) 5. However, there are limited data describing the treatment response after first-line ART failure in resource-limited settings. We report the clinical and virologic outcomes of patients who experienced initial ART regimen failure in KwaZulu Natal, South Africa, after 24-weeks of second-line ART.
METHODS
The Sinikithemba Clinic at McCord Hospital and the iThemba Clinic at St. Mary’s Hospital in South Africa provide vertical HIV care for patients from impoverished peri-urban townships. Monitoring follows South African Department of Health recommendations including HIV-1 viral load (assay with detection limit of <50 copies/ml) and CD4 count monitoring six monthly. Clinic counselors provide adherence training before ART initiation and after an elevated viral load.
Study participants
Patients (n=115) were prospectively enrolled adults with a single episode of virologic failure (HIV-1 RNA viral load (VL) ≥1000 copies/mL) during initial combination ART who underwent genotypic resistance testing. Patients with a prior history of dual or monotherapy were not excluded. A subset of patients (n=26; 18% of overall cohort) had resistance testing performed prior to the inception of the prospective cohort in 2005 and were added to the overall cohort. The second-line agents available during the study were: lopinavir/ritonavir (LPV/r; available as gel formulation [Kaletra™]), lamivudine, didanosine (enteric-coated formulation), zidovudine, stavudine, nevirapine and efavirenz. The option to continue an NNRTI-based regimen after initial ART failure was available.
Data collection
Data collected at regimen failure included treatment history, CD4 count, HIV-1 RNA level, WHO stage, hemoglobin and weight. Data collected after 24 weeks of subsequent ART included plasma HIV-1 RNA, CD4 cell count and clinical outcome.
Genotypic resistance testing
Genotypic testing of virus samples was performed at the Nelson Mandela School of Medicine (Durban), using the TRUGENE HIV-1 Genotyping Test (Siemens). Major resistance mutations were previously defined in the initial report describing the cohort.5
Statistical analysis
Analyses were performed using SAS software, version 9.1.3 (SAS Institute, North Carolina, USA). All tests of significance were 2-sided; associations with P < 0.05 were considered significant. Continuous variables were compared with Wilcoxon rank-sum test; categorical variables with the χ2 test or Fisher’s exact test. An intent-to-treat (ITT) (missing=failure) analysis was performed for the primary outcome of virologic suppression (<50 copies/µl) 24 weeks from enrollment. All outcomes among patients with and without major drug resistance mutations were compared using the χ2 test and Fisher’s exact test. A multivariate logistic regression was performed to determine risk factors associated with mortality after regimen failure.
The study was approved by the ethics committees at McCord and St. Mary’s Hospital and by the IRB at Partners HealthCare and Harvard Medical School in Boston, Massachusetts.
RESULTS
Patient characteristics
Between August 2004 and August 2006, 141 patients experienced initial ART virologic failure and underwent genotypic testing. Table 1 shows patient characteristics at regimen failure. At least one major resistance mutation at regimen failure was found for 122 (87%) patients and in 19 (13%) patients had no major resistance mutation detected (“wild-type” genotype).
Table 1.
Baseline characteristics of patients with virologic failure during first-line ART with and without evidence of genotypic drug resistance
| Characteristic | ≥1 Major Resistance Mutation | No Major Mutation Detected | |
|---|---|---|---|
| N=122 | N=19 | ||
| Median age (years) [IQR] | 36 (30–42) | 43 (35–47) * | |
| Women (%) | 51 | 47 | |
| WHO classification (%) | |||
| Class 1 | 20 | 21 | |
| Class 2 | 23 | 21 | |
| Class 3 | 37 | 42 | |
| Class 4 | 20 | 16 | |
| ART regimen at virologic failure (%) | |||
| D4T – 3TC – EFV | 40 | 63 | |
| D4T – 3TC – NVP | 7 | 0 | |
| ZDV – 3TC – EFV | 29 | 16 | |
| ZDV – 3TC – NVP | 12 | 10 | |
| D4T – DDI – EFV | 2 | 0 | |
| Other | 10 | 11 | |
| Prior dual- or mono-therapy (%) | 20 | 21 | |
| Median months of NNRTI-based ART [IQR] | 13 (7–20) | 8 (6–12) * | |
| Median CD4 count at virologic failure (cells/ul) [IQR] 1 | 176 (112–259) | 128 (103–221) | |
| CD4 cell count category (%) 1 | |||
| 0–49 cells/ul | 9 | 6 | |
| 50–99 cells/ul | 12 | 17 | |
| 100–199 cells/ul | 36 | 44 | |
| 200–349 cells/ul | 34 | 27 | |
| ≥350 cells/ul | 9 | 6 | |
| Median plasma viral load at virologic failure (copies/ml) [IQR] 2 | 17,000 (5500–68,264) | 26,766 (2500–250,000) | |
| Viral load category (copies/ml) (%) 2 | |||
| 400–4,999 | 22 | 32 | |
| 5,000–29,999 | 38 | 20 | |
| 30,000–99,999 | 23 | 11 | |
| ≥ 100,000 | 17 | 37 | |
| Median hemoglobin (g/dl) [IQR] 3 | 13 (11–14) | 12 (11–13) | |
| Resistance mutations (%) | |||
| TAM1 | 14 | NA | |
| TAM2 | 30 | NA | |
| K65R | 6 | NA | |
| Dual Class Resistance (≥1 major NRTI and NNRTI mutation) |
21 | NA | |
| ART regimen following virologic failure (%) | |||
| Lopinavir/ritonavir-based | 90 | 21 ** | |
| Non-protease-inhibitor-based | 7 | 63 | |
| No subsequent regimen | 3 | 16 | |
Wilcoxon, Chi-square, and Fisher’s tests used for two group,
p<0.05
p<0.001
Two patients were missing baseline CD4 cell count
One patient was missing baseline viral load
Eight patients were missing baseline hemoglobin
Virologic and immunological outcomes at 24 weeks
ITT analysis showed that 24 weeks after virologic failure, 99 (70%) patients achieved viral suppression to <400 copies/mL, and 91 (65%) patients to <50 copies/mL. Overall, 50% of patients achieved a 30% improvement in CD4 cell count at 24 weeks follow-up; the median increase in CD4 cell count was 88 cells/µl (IQR 7–168). After 24 weeks, the median 24-week CD4 count was 249 cells/µl (166–343) and only 33% of patients remained with a CD4 cell count of <200 cells/µl.
Mortality and loss-to-follow-up at 24 weeks
The overall mortality among patients 24 weeks after initial ART virologic failure was 6% [95% CI 2-9%], and loss-to-follow-up was 9% [95% CI 4–13%]. Causes of death were tuberculosis (3 patients), gastroenteritis (2), lactic acidosis (1), suspected central nervous system mass (1), and unknown (1). Using a univariate analysis, we compared the characteristics of patients who did not survive 24 weeks of follow-up with those who survived (Table 2). There was a significant (inverse) relationship between the CD4 cell count at regimen failure and 24-week mortality, such that patients with CD4 counts at failure <100 cells/µl experienced higher 24-week mortality compared to patients with CD4 count ≥100 cells/µl (P=0.005). The median CD4 cell count at initial regimen failure among those who died was 70 cells/µl (IQR 27–123) compared to a CD4 cell count of 182 cells/µl (114–260) among patients who survived (P=0.01). Patients who received a boosted PI-containing second-line ART after initial regimen failure experienced a lower mortality over 24 weeks (2%) compared to patients who received NNRTI-based ART (15%) (P=0.004). However, both CD4 count at failure and subsequent regimen type were of borderline significance in multivariate analysis.
Table 2.
Factors Associated with 24-Week Mortality After Initial ART Virologic Failure
| - Univariate - | - Multivariate - | |||
|---|---|---|---|---|
| Characteristics | N | 24-Week Mortality no. (%) |
P* | Odds Ratio 95% CI |
| All patients | 141 | 8 (6) | ||
| Gender | ||||
| Female | 71 | 4 (6) | ||
| Male | 70 | 4 (6) | 0.98 | 1.5 (0.2 – 12.3) |
| History of suboptimal ART | ||||
| None | 113 | 7 (6) | ||
| Prior dual or monotherapy | 28 | 1 (4) | 0.6 | 0.4 (0.03 – 7.2) |
| HIV-1 drug resistance at initial ART failure | ||||
| ≥1 resistance mutation | 122 | 5 (4) | ||
| No resistance | 19 | 3 (16) | 0.06 | 2.1 (0.1 – 36.2) |
| Subsequent regimen type1 | ||||
| LPV/r-based ART | 114 | 2 (2) | ||
| NNRTI-based ART | 20 | 3 (15) | 0.02 | 6.3 (0.5 – 83.9) |
| CD4 cell count at initial ART failure (cells/ul)2 | ||||
| ≥100 | 110 | 2 (2) | ||
| <100 | 29 | 5 (17) | 0.005 | 7.9 (0.8 – 79.7) |
| HIV-1 RNA viral load at initial ART failure (copies/mL)3 | ||||
| ≥ 100,000 | 27 | 2 (7) | ||
| < 100,000 | 113 | 6 (5) | 0.7 | 4.1 (0.3 – 63.7) |
| WHO clinical stage at initial ART failure4 | ||||
| Stage III or Stage IV | 64 | 6 (9) | ||
| Stage I or Stage II | 47 | 2 (4) | 0.5 | 0.4 (0.04 – 5.8) |
P-values are for univariate logistic regression model; odds ratio refer to the multivariate model logistic regression model.
Seven patients did not initiate a regimen after virologic failure and three patients from this group died.
One patient who died did not have a CD4 count at first ART failure
One patient who survived did not have a viral load within 8 weeks of first ART failure
Thirty patients who survived did not have a recorded WHO staged at first ART failure
Drug resistance at first ART failure
Patients in whom one or more HIV-1 drug resistance mutations were found at virologic failure were compared to patients without resistance mutations detected (“wildtype” genotype). The two groups did not differ by age, gender, ART regimen at failure, or by history of prior dual- or mono-therapy. Patients with drug resistance had a longer median duration of prior initial ART compared to patients without resistance (13 months versus 8 months) (P<0.05). The median CD4 cell count at initial ART failure was 176 cells/µl (IQR 112–259) in patients with drug resistance and 128 cells/µl [103–222] in patients without resistance (P=0.34). The median HIV-1 RNA viral load at failure among patients with drug resistance was 17,000 copies/mL (IQR 5,500 – 68,264) and 26,766 copies/mL (2,500 – 250,000) in patients without resistance (P=0.4). There was no significant association between the level of viral load at regimen failure and the presence or absence of drug resistance. Patients with drug resistance at regimen failure were more likely to be started on a ritonavir-boosted PI-containing second-line regimen (90%) compared to patients experiencing virologic failure without drug resistance (21%) (P=0.001) as clinicians attempted to optimize regimens.
Viral suppression rates at 24 weeks differed among patients with and without evidence of drug resistance at initial virologic failure. At 24 weeks, 84 of 122 patients (69%) with at least one major mutation achieved viral suppression compared to 7 of 19 patients (37%) without resistant virus (ITT analysis; P=0.01). The median 24-week improvement in CD4 cell count was 89 cells/µl (IQR 12–168) in patients with baseline drug resistance and 34 cells/µl (0–160) in patients without resistant virus (P=0.67). After 24 weeks, 4% of patients with drug resistance and 16% of patients without drug resistance had died (P=0.02).
Effect of drug resistance on boosted PI-based second-line ART outcomes
A total of 107 patients received lopinavir/ritonavir-containing second-line ART. Among patients who initiated a lopinavir/ritonavir-containing regimen, viral suppression at week 24 was achieved in 31 of 39 patients (79%) with at least one TAM as compared to 69 of 102 patients (68%) with no baseline TAMs (P=0.20) and in 4 of 5 patients who initiated lopinavir/ritonavir with ≥3 TAMS. Viral suppression on a lopinavir/ritonavir-containing regimen was achieved in 4 of 5 patients with a K65R mutation. Viral suppression was also achieved in 4 of 5 patients with evidence of one or more major protease mutations. For a fifth patient with at least one major PI resistance mutations, a week 24 plasma HIV-1 RNA measurement was unavailable.
DISCUSSION
This is the first prospective study of second-line ART outcomes in a resource-limited setting. In ITT analysis, after 24 weeks of subsequent ART, 65% of patients achieved viral suppression to <50 copies/mL with a median CD4 cell count improvement of nearly 90 cells/µl. The experience of patients in our cohort compared favorably to that of ART-naïve patients in clinical trials of lopinavir/ritonavir-containing regimens conducted in high-income settings 6.
The use of genotypic drug resistance testing at first ART failure provided important insights. The subgroup of patients in whom no major resistance mutations (“wild-type” genotype) were detected at initial regimen failure experienced higher mortality and greater subsequent loss-to follow-up compared to patients with evidence of drug resistance. A possible explanation for this paradoxical observation is the role of poor adherence, which may not have been resolved before the salvage regimen was initiated. Risk factors for suboptimal adherence have been identified in resource-poor contexts including clinic fees, stigma, regimen complexity, and drug supply interruptions 3, 7–9.
We examined the impact of specific resistance mutations seen at initial virologic failure on salvage outcomes. In this cohort, NRTI resistance mutations (including K65R and TAMs) had minimal impact on 24-week outcomes on boosted PI-based ART. The high pharmacological and genetic barriers to resistance to ritonavir-boosted lopinavir may have allowed patients to overcome the deleterious effects of significant NRTI resistance mutations. However these results should be interpreted with caution given the relatively short follow-up.
The study has several limitations. Second-line ART after virologic failure was informed by genotypic resistance testing and the nucleoside backbone was optimized based upon resistance mutations at initial ART failure. Because this study was not conducted as a controlled clinical trial of the impact of drug resistance testing, we cannot estimate the direct contribution of genotype testing on the outcomes. Second, to maximize the benefit to patients of genotype testing, with limited exceptions (n=4), lopinavir/ritonavir was not given to patients with “wild-type” genotypes, a pattern known to be associated with suboptimal adherence. If suboptimal adherence to initial ART is linked to adherence to subsequent regimens, we may have overestimated the efficacy of lopinavir/ritonavir-containing second-line ART.
In summary, virologic monitoring linked to resistance testing helped demonstrate the efficacy of lopinavir/ritonavir-containing regimens as second-line ART in South Africa. Resistance testing identified a high-risk group without drug resistance who might benefit from increased medication access and/or adherence support. Although regimens that include LPV/r remain more expensive than first-line regimens, even at local access prices, our results suggest that switching to second-line regimens in patients with virologic failure and resistance has substantial and rapid immunological and clinical benefits. Models predict that the prevalence of acquired HIV drug resistance in sub-Saharan Africa will grow substantially over the next decade 10. Early detection of regimen failure and reductions in the price of boosted PI-based regimens must be prioritized if this patient population is to be effectively treated.
Acknowledgements
We would like to express our admiration for the work of the iThemba HIV/AIDS Clinic at the St. Mary’s Hospital and the Sinikithemba Clinic at McCord Hospital. Important contributions were made by K. Nixon, J. Naidoo and S. Pertab.
Financial support
Grant support from Gilead, NIH (P30 AI60354 to Harvard CFAR and K24 RR16482 to D.R.K.), Harvard Program on AIDS, CDC Cooperative Agreement (U62/CCU123541-01), Schwartz Global Health Fellowship and Elizabeth Glaser Pediatric AIDS Foundation (Project HEART).
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Potential conflicts of interest
D.R.K. is a consultant to, or has received research funding from Abbott, Boehringer-Ingelheim, Bristol-Myers Squibb, Gilead, GlaxoSmithKline, Merck and Siemens.
Conferences
Presented at CROI 2009
Contributor Information
Richard A. Murphy, Division of Infectious Diseases, Massachusetts General Hospital, Boston, MA, [Manuscript preparation and data collection]
Henry Sunpath, McCord Hospital, Durban, South Africa, [Study design and implementation].
Zhigang Lu, Massachusetts General Hospital, Boston, MA, [Statistical analysis].
Neville Chelin, McCord Hospital, Durban, South Africa, [Study implementation and data collection].
Elena Losina, Massachusetts General Hospital, Boston, MA, [Statistical analysis].
Michelle Gordon, Nelson Mandela School of Medicine, Durban, South Africa, [Drug resistance testing and study design].
Douglas Ross, St. Mary’s Hospital, Mariannhill, South Africa, [Study design and implementation].
Aba D. Ewusi, Harvard Medical School, Boston, MA, [Study data collection and critical review of manuscript]
Lynn T. Matthews, Division of Infectious Diseases, Beth Israel Deaconess Medical Center, Boston, MA, [Study data collection and critical review of manuscript]
Daniel R. Kuritzkes, Section of Retroviral Therapeutics, Brigham and Women’s Hospital, Boston, MA, [Study design and critical review of manuscript].
Vincent C. Marconi, Infectious Disease Service, San Antonio Military Medical Center, Brooke Army Medical Center, Fort Sam Houston, TX, [Study design and critical review of manuscript]
References
- 1.Novitsky V, Wester CW, DeGruttola V, et al. The reverse transcriptase 67N 70R 215Y genotype is the predominant TAM pathway associated with virologic failure among HIV type 1C-infected adults treated with ZDV/ddI-containing HAART in southern Africa. AIDS Res Hum Retroviruses. 2007 Jul;23(7):868–878. doi: 10.1089/aid.2006.0298. [DOI] [PubMed] [Google Scholar]
- 2.Orrell C, Bangsberg DR, Badri M, Wood R. Adherence is not a barrier to successful antiretroviral therapy in South Africa. AIDS. 2003 Jun 13;17(9):1369–1375. doi: 10.1097/00002030-200306130-00011. [DOI] [PubMed] [Google Scholar]
- 3.Mills EJ, Nachega JB, Buchan I, et al. Adherence to antiretroviral therapy in sub-Saharan Africa and North America: a meta-analysis. JAMA. 2006 Aug 9;296(6):679–690. doi: 10.1001/jama.296.6.679. [DOI] [PubMed] [Google Scholar]
- 4.Nachega JB, Hislop M, Dowdy DW, Chaisson RE, Regensberg L, Maartens G. Adherence to nonnucleoside reverse transcriptase inhibitor-based HIV therapy and virologic outcomes. Ann Intern Med. 2007 Apr 17;146(8):564–573. doi: 10.7326/0003-4819-146-8-200704170-00007. [DOI] [PubMed] [Google Scholar]
- 5.Marconi VC, Sunpath H, Lu Z, et al. Prevalence of HIV-1 drug resistance after failure of a first highly active antiretroviral therapy regimen in KwaZulu Natal, South Africa. Clin Infect Dis. 2008 May 15;46(10):1589–1597. doi: 10.1086/587109. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Walmsley S, Bernstein B, King M, et al. Lopinavir-ritonavir versus nelfinavir for the initial treatment of HIV infection. N Engl J Med. 2002 Jun 27;346(26):2039–2046. doi: 10.1056/NEJMoa012354. [DOI] [PubMed] [Google Scholar]
- 7.Nachega JB, Stein DM, Lehman DA, et al. Adherence to antiretroviral therapy in HIV-infected adults in Soweto, South Africa. AIDS Res Hum Retroviruses. 2004 Oct;20(10):1053–1056. doi: 10.1089/aid.2004.20.1053. [DOI] [PubMed] [Google Scholar]
- 8.Hardon AP, Akurut D, Comoro C, et al. Hunger, waiting time and transport costs: time to confront challenges to ART adherence in Africa. AIDS Care. 2007 May;19(5):658–665. doi: 10.1080/09540120701244943. [DOI] [PubMed] [Google Scholar]
- 9.Laniece I, Ciss M, Desclaux A, et al. Adherence to HAART and its principal determinants in a cohort of Senegalese adults. AIDS. 2003 Jul;17 Suppl 3:S103–S108. doi: 10.1097/00002030-200317003-00014. [DOI] [PubMed] [Google Scholar]
- 10.Blower S, Bodine E, Kahn J, McFarland W. The antiretroviral rollout and drug-resistant HIV in Africa: insights from empirical data and theoretical models. AIDS. 2005 Jan 3;19(1):1–14. doi: 10.1097/00002030-200501030-00001. [DOI] [PubMed] [Google Scholar]