Murphy, Richard A. MD, MPH*; Sunpath, Henry MBBS, MPH†; Castilla, Carmen MD‡; Ebrahim, Shameez BPharm†; Court, Richard MBChB§; Nguyen, Hoang MD, MPH‖; Kuritzkes, Daniel R. MD¶; Marconi, Vincent C. MD#; Nachega, Jean B. MD, MPH, PhD**
*Montefiore Medical Center and Albert Einstein College of Medicine, Bronx, NY
†Department of Medicine, McCord Hospital, Durban, South Africa
‡Department of Medicine, Brigham and Women's Hospital, Boston, MA
§Department of Medicine, University of Cape Town, Cape Town, South Africa
‖Department of International Health, Johns Hopkins Bloomberg School of Public Health, Baltimore, MD
¶Department of Medicine, Division of Infectious Diseases, Brigham and Women's Hospital and Harvard Medical School, Boston, MA
#Department of Medicine, Infectious Disease Division, Emory University School of Medicine, Atlanta, GA
**Department of Medicine, Centre for Infectious Diseases, Stellenbosch University, Cape Town, South Africa
Correspondence to: Richard A. Murphy, Richard A. Murphy, MD, MPH, 1621 Eastchester Road, Bronx, NY 10461 (e-mail: email@example.com).
Supported in part by the IDSA Medical Scholars Program, the Arnold P. Gold Foundation Medical Student Scholarship, Partners AIDS Research Center, McCord Hospital, and National Institutes of Health Grants K24-RR016482 and P30 AI060354.
Presented at the Conference on Retroviruses and Opportunistic Infections, February 2011, Boston, MA.
R. A. Murphy and H. Sunpath contributed equally to this work.
D. R. Kuritzkes is a consultant to, or has received research funding from Abbott, Boehringer-Ingelheim, Bristol-Myers Squibb, Gilead, GlaxoSmithKline, Merck and Roche.
The other authors have no conflicts of interest to disclose.
Received February 17, 2012
Accepted May 24, 2012
A growing proportion of patients in resource-limited countries who have initiated antiretroviral therapy (ART) have experienced first-line regimen failure.1–3 However, evidence shows that most patients identified early and switched to a boosted protease inhibitor (PI)–containing second-line regimen—most commonly consisting in sub-Saharan Africa of lopinavir/ritonavir plus 2 nucleoside reverse transcriptase inhibitors (NRTIs)—subsequently achieve early virologic suppression.4–6 Currently, boosted PI–containing regimens are the only option after first-line regimen failure available for patients treated in the public sector in most resource-limited settings. Because such patients have typically not had prior exposure to PIs, adherence to treatment will likely be the principle determinant of long-term virologic suppression.
South Africa has the largest HIV epidemic in the world. The demand for first-line and second-line ART continues to grow as ART coverage, now 37% nationally, is achieved in a greater proportion of eligible patients.7 This study was performed in KwaZulu-Natal, an epicenter of the HIV epidemic in South Africa, where HIV incidence remains substantial, and recent antenatal surveys continue to reveal a high prevalence of HIV among pregnant women, in some districts above 40%.8,9
At a single antiretroviral clinic in Durban, we estimated adherence to second-line ART by calculating medication possession ratios (MPRs) using pharmacy refill claims from an on-site pharmacy and explore the relationship between adherence and long-term clinical outcomes. MPR is a well-studied tool for estimating adherence; the validity of MPR in resource-limited settings has been supported by prior studies, where it has been shown to be associated with virologic outcomes and mortality among patients receiving first-line ART.10,11
An improved understanding of second-line ART in resource-limited settings will be critical in maximizing the durability of second-line regimens, preventing disease progression and reducing long-term AIDS-related mortality in high HIV-prevalence countries without access to third-line ART regimens.
The Sinikithemba Outpatient HIV/AIDS Clinic at McCord Hospital provides HIV care for patients from Durban and the surrounding area with financial support from the President's Emergency Plan for AIDS Relief and the South African Department of Health. Approximately, 8200 adults and 1200 children receive ART and related care, including tuberculosis treatment, at the clinic. ART initiation and monitoring follows South African Department of Health recommendations, including both HIV-1 viral load (assay with detection limit of <50 copies/mL) and CD4 count testing every 6 months. Clinic counselors provide adherence training before first-line ART initiation and at the time of virologic failure before regimen change.
One hundred and sixty-five HIV-infected adults who initiated second-line ART after documented first-line ART virologic failure (VF) (HIV-1 RNA viral load ≥1000 copies/mL) were identified. After excluding 29 patients who received second-line ART for less than 6 months, we enrolled 136 patients. We did not include patients who were switched to second-line ART without documented virologic failure.
During the period of the study, the standard South African second-line regimen consisted of lopinavir/ritonavir (fixed dose combination dosed twice daily) plus zidovudine and enteric-coated didanosine. Lamivudine and stavudine were available for use in a second-line regimen, in the setting of prior adverse effect or contraindication to the standard second-line NRTI backbone.
Data Collection and Adherence Measurement
Using a standardized instrument, we collected baseline demographic and clinical data from an electronic clinic database (TrakHealth). Values for CD4 and viral load within 2 months of initiation of second-line ART initiation were considered to be the re-switch baseline values.
To estimate adherence, an MPR was calculated by dividing the number of months that a patient submitted refill prescriptions by the months since second-line ART regimen initiation. The adherence estimates calculated for various time intervals were cumulative. For example, adherence at month 12 reflected adherence during the initial 12 months of second-line ART. The measurement of adherence at the Sinikithemba Clinic was facilitated by the use of refill claims from an on-site pharmacy and an electronic clinical database. The dates of pharmacy refill were captured for the duration of second-line ART and for the final 6 months of first-line ART before VF. Patients were routinely scheduled for clinic visits every 28 days, at which time, they were given a 30-day supply of medication, and patients accrued 2 extra days of pills (30 days of ART provided every 28 days) at each refill. Any missed refill visit resulted in a reminder phone call from the clinic within 7 days.
Study Design and Endpoints
We conducted an observational retrospective study of patients who initiated second-line ART after first-line ART virologic failure. The primary outcomes were 12-month virologic suppression (viral load <50 copies/mL) and 12-month adherence calculated using MPRs. We also sought to determine factors associated with second-line virologic outcome and factors for suboptimal second-line ART adherence.
We used a modified missing-equal-to-failure rule for estimating rates of VF; if a patient was in active follow-up but did not have a viral load measured at a 6-month interval, VF was assumed. Patients who died, transferred care, or were lost to follow-up were excluded from calculations of adherence, virologic, or immunological outcomes after the patient's date of death, transfer, or loss to follow-up, respectively. Patients who do not return to clinic were actively sought for 6 months and a disposition was ultimately determined (died, lost to follow-up, or transferred care).
We defined 3 strata of adherence based on medication refill pattern (>90%, 80%–90%, and <80%). The Wilcoxon rank sum test was used to compare median levels of adherence across time intervals. We analyzed the relationship between adherence and virologic suppression using a Kaplan–Meier analysis that estimated time to suppression by adherence strata.
Multivariable logistic regression modeling was used to explore the association between patient characteristics and virologic suppression at month 12 of ART. Covariates included in the logistic regression model included age, gender, CD4 cell count at switch, viral load at switch, second-line NRTI backbone, first-line ART adherence in the 6 months before switch, and second-line ART adherence. We also use multivariable logistic regression modeling to explore associations between patient characteristics and optimal month-12 second-line ART adherence of ≥90%. Analyses were performed using STATA software, version 10.
This retrospective study was approved by the research ethics committee at McCord Hospital in Durban, South Africa, and the requirement for informed consent was waived.
We analyzed 136 adult patients—65% female, median age 36 years (interquartile change, 31–43)—who initiated second-line ART with a lopinavir/ritonavir–containing second-line regimen after first-line ART virologic failure. The median pre-ART nadir CD4 cell count was 70 cells/mm3 (IQR, 20–118 cells/mm3) and median prior duration of first-line ART was 13 months (IQR, 7–20 months). At the time of first-line ART regimen failure, the median CD4 count and viral load at failure was 153 cells/mm3 (IQR, 89–232 copies/mm3) and 28,548 copies/mL (IQR, 7000–89,000/mL), respectively.
The most common second-line ART NRTI backbones were: AZT + DDI, n = 88 (65%); AZT + 3TC, n = 30 (22%); 3TC alone, n = 7 (5%), and D4T + 3TC, n = 6 (4%), and the median duration of second-line ART follow-up was 34 months (IQR, 20–44 months). The median adherence in the 6 months before switch to second-line ART switch (during NNRTI–based first-line ART) was 67% (IQR, 33%–67%). Median adherence improved significantly in the first 6 months of second-line ART (median adherence 6 months before switch, 67%; median adherence during initial 6 months of second-line ART, 100%; P = 0.001). The median adherence remained high at months 12, 18, and 24 at 92% (IQR, 92%– 100%), 94% (IQR, 89%–100%), and 96% (IQR, 88%–100%), respectively.
Outcomes During Second-Line ART
Virologic, immunological, and clinical outcomes during the initial 24 months of second-line ART including events are described (Table 1). The rate of virologic failure during the first 24 months of second-line therapy was moderate (month 6: 26%, n = 36/136; month 12: 25%, n = 32/126; month 18: 21%, n = 23/112; and month 24: 25%, n = 25/99). Median CD4 cell count rose to more than 200 cells/mm3 rapidly by 6 months of second-line ART (month 6: median CD4 cell count 228 cells/mm3; month 12: median CD4 cell count 276 cells/mm3) and was greater than 300 cells/mm3 by month 18 of second-line ART (month 18 median CD4 cell count, 315 cells/mm3; and month 24, 330 cells/mm3).
We evaluated factors associated with virologic suppression at 12 months (Table 2). A higher rate of adherence during the first 12 months of second-line ART was independently associated with an increased odds of month-12 viral suppression (odds ratio 2.5 per 10% increase in adherence, 95% CI 1.3 to 4.8, P = 0.01). There was no significant association identified between month-12 virologic suppression and patient age, gender, second-line NRTI backbone, viral load at first-line failure, or CD4 cell count at first-line failure.
Adherence and Virologic Outcome
We analyzed the relationship between adherence and virologic suppression using a Kaplan–Meier analysis that estimated time to virologic suppression by adherence strata (Fig. 1). Time to suppression was most rapid among patients with 91%–100% adherence (log rank test, P = 0.01). A plot of virologic suppression and adherence using 3 adherence strata suggested increasing rates of virologic suppression associated with progressively higher rates of adherence (adherence <80%; 80%–90%, and >90%) at 6, 12, 18, and 24 months of second-line ART (Fig. 2). Among patients with an adherence of >90%, the rate of virologic suppression at 6, 12, 18, and 24 months was 73% (IQR, 64–82), 87% (IQR, 80–93), 93% (IQR, 87–98), and 97% (IQR, 95–100), respectively. Among patients with adherence of <80%, the rate of virologic suppression to <50 copies per milliliter at 6, 12, 18, and 24 months was 20% (IQR, 2–38), 44% (IQR, 19–69), 67% (IQR, 44–89), and 83% (IQR, 61–100), respectively.
Multivariate Analysis of Factors Associated With Adherence of ≥90%
In multivariate analysis (Table 3), the median adherence during the final 6 months of first-line ART showed a borderline association with high-level adherence of ≥90 during the first 12 months of second-line ART (odds ratio: 2.5, 95% CI: 0.7 to 8.6, P = 0.15). There was no significant association noted between age, gender, second-line NRTI backbone, CD4 cell count at first-line failure, or with viral load at failure and adherence of ≥90 during the first 12 months of second-line ART.
Among patients who initiated a boosted PI–containing second-line ART in South Africa, median adherence improved after switch to second-line ART from less than 70% to greater than 90%. Not surprisingly, time to second-line ART virologic suppression was most rapid among patients with 91%–100% adherence (log rank test, P = 0.01) and was least rapid among patients with second-line adherence of <80%. Furthermore, a plot of virologic suppression according to adherence strata suggested a dose–response relationship between level of adherence and virologic suppression at each 6-month follow-up time point, with higher rates of suppression at progressively higher levels of adherence.
Additional studies will be needed to clarify factors leading to overall higher level adherence after second-line switch and the durability of the effect. However, factors that may have been contributed include peer-based adherence counseling received at the time of first-line VF, informal doctor and nurse–driven adherence support during follow-up clinic visits and the role of improved regimen tolerability. Compared with first-line ART (at the time of the study, D4T + 3TC + EFV), patients may have found boosted PI–containing second-line ART to be associated with fewer adverse effects. D4T in particular has been associated in resource-limited settings with a high rate of metabolic complications, including painful peripheral neuropathy and lipoatrophy.12,13
During second-line ART, at each 6-month interval, VF was observed in about one-quarter of patients who remained in active follow-up. Although resistance testing was not routinely performed, based on prior studies, VF of second-line ART was unlikely to have been the result of acquired PI-associated drug resistance mutations. A large randomized trial showed that among patients without prior exposure to PIs–as is the case in most HIV-infected South Africans–VF with boosted PI regimens is rarely associated with emergence of new PI mutations.14 Furthermore, a recent study in South Africa found among patients failing lopinavir/ritonavir–containing second-line regimens, a very low prevalence of major lopinavir resistance mutations.15,16 Taken together, these data suggest that adherence—not resistance—is the primary cause in South Africa of second-line ART failure in adults. As a result, patients failing second-line ART would be expected to have a high likelihood of viral resuppression if adherence were to improve and may be an appropriate target for novel adherence interventions.
Another appropriate target for new adherence interventions may be patients identified to have poor adherence to first-line ART prior to switch. In evaluating for independent risk factors for suboptimal levels of second-line ART adherence, we found a borderline association suggesting that patients with first-line ART adherence below the median (in the 6 months before failure) were more likely, to second-line switch to subsequently demonstrate suboptimal adherence to second-line ART. Of note, we had previously observed that South African patients found by viral genotype testing to have wild-type virus at the time of first-line ART failure experienced inferior second-line ART virologic outcomes compared with patients with at least one major resistance mutation. We observed that 84 (69%) of 122 patients with at least one major mutation achieved second-line ART viral suppression 24 weeks after failure compared with 7 (37%) of 19 patients without a resistant virus (P = 0.01).5 It is likely that wild-type viral genotype results were a proxy for poor first-line ART adherence that eventually undermined second-line ART outcome. Therefore, another opportunity for novel adherence intervention may be among patients with lower levels of first-line ART adherence before switch to second-line ART.
This study has several limitations. Adherence measured using MPRs based on refill pattern is an indirect measure of behavior. It does not capture pill-taking behavior directly and as a result is not sensitive to adherence interruptions resulting from behaviors such as pill sharing or “pill dumping.”17 However, it is an inexpensive and readily accessible monitoring tool that has previously been shown to be associated with virologic outcome and mortality in sub-Saharan Africa.11,18–22 A second potential weakness was that our adherence estimates calculated at various time intervals were cumulative. For example, adherence at month 24 of ART reflected adherence during the entire first 24 months of second-line ART. If adherence declined substantially over time, it may not have been evident using a cumulative MPR. Last, we cannot exclude the potential impact of survivorship bias. Patients remaining in active follow-up for 12 or 24 months after initiation of second-line ART may reflect a more adherent and overall healthier population of second-line ART patients in South Africa. By excluding patients from the study who completed less than 6 months of second-line ART, we risked underestimated second-line ART loss to follow-up and mortality because near the time of switch, loss to follow-up and mortality tend to concentrate.5
The switch to second-line ART in South Africa was associated with an improvement in adherence and a rapid immunological recovery. However, a moderate ongoing rate of VF—among approximately 25% of patients receiving second-line ART patients at each follow-up interval—is a cause for concern. Median adherence was not uniformly ≥90% after initiation of second-line ART and differences in second-line ART adherence help explain why some patients in South Africa achieved virologic suppression after switch and other patients did not. Novel adherence interventions may usefully target patients with second-line ART failure who—given a low likelihood of failure with major PI drug resistance mutations—have a high likelihood of achieving viral resuppression.
The authors acknowledge the immeasurable efforts of the nurses, doctors, counselors, administrators, and volunteers at the Sinikithemba Clinic at McCord Hospital in Durban, South Africa. The clinic leadership offers vital ongoing support for studies to improve the lives of patients living with HIV/AIDS. We would like to note the contributions of Roma Maharaj in data collection.
1. 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;46:1589–1597.
2. Kamya MR, Mayanja-Kizza H, Kambugu A, et al.. Predictors of long-term viral failure among ugandan children and adults treated with antiretroviral therapy. J Acquir Immune Defic Syndr. 2007;40:187–193.
3. Hosseinipour MC, van Oosterhout JJ, Weigel R, et al.. The public health approach to identify antiretroviral therapy failure: high-level nucleoside reverse transcriptase inhibitor resistance among Malawians failing first-line antiretroviral therapy. AIDS. 2009;23:1127–1134.
4. Pujades-Rodriguez M, Balkan S, Arnould L, et al.. Treatment failure and mortality factors in patients receiving second-line HIV therapy in resource-limited countries. JAMA. 2010;304:303–312.
5. Murphy RA, Sunpath H, Lu Z, et al.. Outcomes after virologic failure of first-line ART in South Africa. AIDS. 2010;24:1007–1012.
6. Fox MP, Ive P, Long L, et al.. High rates of survival, immune reconstitution, and virologic suppression on second-line antiretroviral therapy in South Africa. J Acquir Immune Defic Syndr. 2010;53:500–506.
7. World Health Organization. World Health Statistics 2011, Geneva, Switzerland: World Health Organization; 2011.
8. Department of Health. The 2010 National Antenatal Sentinel HIV and Syphilis Prevalence Survey in South Africa. Pretoria, South Africa: National Department of Health; 2011.
9. Nel A, Mabude Z, Smit J, et al.. HIV incidence remains high in KwaZulu-Natal, South Africa: evidence from three districts. PLoS One. 2012;7:e35278.
10. Bisson GP, Gross R, Bellamy S, et al.. Pharmacy refill adherence compared with CD4 count changes for monitoring HIV-infected adults on antiretroviral therapy. PLoS Med. 2008;5:e109.
11. Nachega JB, Hislop M, Dowdy DW, et al.. Adherence to highly active antiretroviral therapy assessed by pharmacy claims predicts survival in HIV-infected South African adults. J Acquir Immune Defic Syndr. 2006;43:78–84.
12. Kumarasamy N, Vallabhaneni S, Cecelia AJ, et al.. Reasons for modification of generic highly active antiretroviral therapeutic regimens among patients in southern India. J Acquir Immune Defic Syndr. 2006:53–58.
13. van Oosterhout JJ, Bodasing N, Kumwenda JJ, et al.. Evaluation of antiretroviral therapy results in a resource-poor setting in Blantyre, Malawi. Trop Med Int Health. 2005:464–470.
14. Lathouwers E, De Meyer S, Dierynck I, et al.. Virological characterization of patients failing darunavir/ritonavir or lopinavir/ritonavir treatment in the ARTEMIS study: 96-week analysis. Antivir Ther. 2011;16:99–108.
15. Wallis CL, Mellors JW, Venter WD, et al.. Protease inhibitor resistance is uncommon in HIV-1 subtype C infected patients on failing second-line lopinavir/r-containing antiretroviral therapy in South Africa. AIDS Res Treat. 2011;2011:769627.
16. van Zyl GU, van Mens TE, McIlleron H, et al.. Low lopinavir plasma or hair concentrations explain second-line protease inhibitor failures in a resource-limited setting. J Acquir Immune Defic Syndr. 2011;56:333–339.
17. Acri T, TenHave TR, Chapman JC, et al.. Lack of association between retrospectively collected pharmacy refill data and electronic drug monitoring of antiretroviral adherence. AIDS Behav. 2010;14:748–754.
18. Rougemont M, Stoll BE, Elia N, et al.. Antiretroviral treatment adherence and its determinants in Sub-Saharan Africa: a prospective study at Yaounde Central Hospital, Cameroon. AIDS Res Ther. 2009;6:21.
19. Chi BH, Cantrell RA, Zulu I, et al.. Adherence to first-line antiretroviral therapy affects non-virologic outcomes among patients on treatment for more than 12 months in Lusaka, Zambia. Int J Epidemiol. 2009;38:746–756.
20. Toure S, Kouadio B, Seyler C, et al.. Rapid scaling-up of antiretroviral therapy in 10,000 adults in Cote d'Ivoire: 2-year outcomes and determinants. AIDS. 2008;22:873–882.
21. Ross-Degnan D, Pierre-Jacques M, Zhang F, et al.. Measuring adherence to antiretroviral treatment in resource-poor settings: the clinical validity of key indicators. BMC Health Serv Res. 2010;10:42.
22. Nachega JB, Trotta MP, Nelson M, et al.. Impact of metabolic complications on antiretroviral treatment adherence: clinical and public health implications. Curr HIV/AIDS Rep. 2009;6:121–129.