With the availability of generic HAART at low cost, an increasing number of HIV-infected patients in resource-constrained settings are now receiving therapy.1 The safety, tolerability, and efficacy of generic HAART at increasing the CD4 cell count and suppressing the viral load in patients in the developing world have been demonstrated.2-5 Data on patients' long-term experience on therapy and why and when patients may modify or discontinue therapy are limited, however. Studies on patients taking proprietary HAART from the developed world indicate that 36%6 to 44%7 of patients may modify their regimen and a quarter might discontinue therapy in the first 12 months of treatment.8 In these studies, younger female patients, those with a lower CD4 cell count and higher viral load at the start of therapy, those who had previously been treated, and those who were on 4-drug regimens were more likely to modify their HAART regimens.6-8 The main reason for discontinuing or modifying HAART in these studies was treatment-associated toxicity.
The reasons for discontinuing and modifying therapy in resource-constrained settings could differ from those in the developed world, because of the following:
- Patients are on generic rather than proprietary HAART.
- Nonnucleoside reverse transcriptase inhibitor (NNRTI)-based regimens are uniformly used in first-line treatment, and only a small percentage of patients initiate therapy on protease inhibitors (PIs).
- Viral load monitoring, which is standard of care in the developed world, is rarely performed because of its high cost. Therefore, physicians rely on immunologic and clinical criteria to identify treatment failure.
- Most patients in the developing world pay “out of pocket” for health care, making cost of treatment a major issue. For instance, the least expensive triple-drug regimen costs approximately $20 per month in India and poses a substantial burden for patients in a country in which nearly half the population lives on less than $1 per day.9
In this study, we describe the profile of patients initiating generic HAART in southern India and the reasons for and determinants of modifying and discontinuing antiretroviral regimens in this cohort.
The YRG Center for AIDS Research and Education (YRG CARE) in Chennai (Madras) is a large tertiary HIV care and research institution in India. Since 1996, it has provided a continuum of care for more than 8000 HIV-infected individuals. All patients are treated according to the World Health Organization (WHO) treatment guidelines. Patients are advised to initiate antiretroviral therapy (ART) when CD4 counts are <200 cells/μL or if they develop an AIDS-defining illness and if they can afford treatment. Nevirapine (NVP)-based regimens, which are the least expensive and most widely available, are most commonly used as first-line therapy. Patients are seen every 3 months or as clinically indicated.
Patients who were older than 18 years of age, attending the YRG CARE clinic between January 1996 and October 2004, and naive to ART before initiation of triple ART (HAART) were included in the analysis. Study participants also had to have at least 1 follow-up visit after initiating HAART. Analysis was done using the YRG CARE Chennai HIV Observational Database. This database, updated daily by research nurses who are trained to use a validated prospective data collection instrument, includes demographics; clinical assessments, including data related to the occurrence of new opportunistic infections; and current treatment regimens and adverse events (AEs) as well as laboratory data, including hemoglobin, liver, and renal function tests; CD4 cell counts; plasma viral load (PVL); and genotypic testing, if available.
Patients coinfected with HIV and tuberculosis (TB) are treated per WHO recommendations. Current WHO guidelines recommend starting HAART between 2 weeks and 2 months after initiating TB treatment in patients with a CD4 count <200 cells/μL. For patients with a CD4 count between 200 and 350 cells/μL, HAART can be started after the initiation phase of TB treatment and deferred if the CD4 count is greater than 350 cells/μL. If the patient needs to be treated simultaneously for HIV and TB, efavirenz (EFV)-based regimens are recommended. When patients cannot afford EFV, they are placed on NVP and closely monitored for hepatotoxicity.
Treatment “modification” was defined as any alteration of 1 or more components of a participant's regimen. “Switching” was used to refer to those who modified therapy because of treatment failure, and “substitution” was used to refer to those who modified therapy because of the occurrence of an AE. Modification or discontinuation of therapy was evaluated in association with immunologic or HIV-related clinical indicators of treatment failure, treatment-associated toxicities, and inability to pay. Immunologic failure was defined as a failure to demonstrate a rise in CD4 cell count after at least 3 months on treatment, return of CD4 cell count to baseline, or a greater than 50% decline in CD4 cell count while on HAART. The clinical criterion for failure was the development of an AIDS-defining illness, including TB, occurring at a CD4 count <200 cells/μL after at least 3 months on HAART. “Cost” was used to categorize patients who stopped because they were unable to afford the cost of therapy. “Second-line regimen” was used to refer to any change from the original regimen. “Third-line regimen” was used to refer to any change from the second-line regimen. “Persistent symptomatic peripheral neuropathy” was used to refer to progressive neuropathy that did not respond to nutritional or vitamin supplementation.
Data entry, database management, and statistical analyses were performed with SPSS software (version 10.0.5; SPSS, Chicago, IL). Descriptive statistics were used to calculate the frequency, mean, median, and standard deviation (SD). Diagnostic tests were performed, and the mean ± SD was used for variables that were normally distributed; and the median, interquartile range (IQR), and range, if required, were calculated for variables that were influenced by extreme values. To compare proportions, χ2 statistics were used, and the Mann-Whitney U test was used to compare median durations. Logistic regression was performed to identify the independent predictors of modification and discontinuation of therapy. A P value of less than 0.05 was considered statistically significant.
Of the 6821 adult patients attending the YRG CARE between January 1, 1996 and October 31, 2004, 3736 (55%) met immunologic, virologic, or clinical criteria for needing therapy, as defined by the WHO.10 More than half of these patients (n = 2033) were able to afford medication and initiated ART. Of these, 1443 patients were previously ART naive, initiated triple HAART, and had at least 1 follow-up visit at the YRG CARE (Figure 1). Only these 1443 patients are included in the subsequent analyses (Table 1). Eighty-two percent of this cohort was male, with a mean age of 35 ± 8 years. The major mode of transmission was heterosexual (95%). The median CD4 count (n = 1393) at time of initiating first-line HAART was 108 cells/μL (IQR: 50-187 cells/μL).
As seen in Table 1, the 4 most common first-line regimens were stavudine (d4T) plus lamivudine (3TC) plus NVP (63%), zidovudine (AZT) plus 3TC plus NVP (19%), d4T plus 3TC plus EFV (9%), and AZT plus 3TC plus EFV (4%). Over a follow-up of 2330 person-years, 285 (20%) patients modified therapy and 126 (9%) discontinued their first-line regimen (Table 2). Of those who discontinued therapy, almost half subsequently restarted the same regimen.
Two hundred eighty-five patients (20%) modified their first-line regimen. The following were the main reasons for those modifying therapy: AE (64%), cost (19%), and treatment failure (14%). Older patients were more likely to modify their regimen (odds ratio [OR] = 2.0, 95% confidence interval [CI]: 1.3 to 3.1); however, gender, CD4 cell count, presence of an AIDS-defining illness, hemoglobin, and body mass index (BMI) at time of initiation of therapy were not significantly associated with modifying therapy.
Overall, 13% (n = 183) patients substituted therapy because of the occurrence of an AE. The most common AEs in this group of patients were itching and/or skin rash (66%; not including mild allergic rash) and hepatotoxicity (27%). The median CD4 counts at initiation of therapy for those who developed a rash or hepatotoxicity were 113 cells/μL and 65 cells/μL, respectively. Other AEs included nausea and vomiting (25%) and anemia (23%). Twenty-seven (15%) of the 183 patients who substituted therapy because of AEs did so because of d4T-related toxicities (persistent symptomatic peripheral neuropathy [13%] and lipoatrophy [2.2%]). The median time on therapy before substitution because of any AE was 40 days (range: 1-2319 days).
As seen in Table 3, study patients were stratified into 2 groups: those on NVP-containing regimens and those on “other” (non-NVP) regimens. Eighty-three percent (n = 1195) of patients were on NVP-containing regimens. Of those on NVP-containing regimens, 156 (13%) substituted therapy because of an AE. Rash and hepatotoxicity accounted for 41 (26%) and 33 (21%) AEs, respectively. Those on NVP were more likely to substitute therapy because of AEs (P < 0.001), and the other regimens group was more likely to change therapy because of cost (P < 0.001; see Table 3). Those on NVP substituted therapy sooner than those on other regimens when the reason was an AE (median number of days: 34 vs. 74; P = 0.29), whereas those on other therapies changed therapy sooner when the reason for modifying therapy was cost (median number of days: 128 vs. 228; P = 0.16). The median time to development of hepatotoxicity on NVP was 34 days (range: 2-575 days). Of the 33 patients with hepatotoxicity, 27 (82%) were on rifampicin-containing anti-TB medications at the time of development of the AE.
Of the 183 patients who initially substituted therapy because of an AE, 15 (1%) discontinued their second-line regimen and 60 (33%) modified their second-line regimen, moving to a third-line regimen in a median time of 97 days. Of the 60 patients previously cited, 22 (37%) substituted therapy because of an AE, 11 (18%) switched because of treatment failure, and 10 (17%) changed therapy because of the cost of the second-line regimen.
Of the overall 3% of patients who switched because of treatment failure, (Table 4), 18 switched after the occurrence of an opportunistic infection while on therapy, 29 switched because of immunologic failure, and 3 switched because of virologic failure. Median time to switching therapy because of treatment failure was 406 days (range: 93-1008 days). Of the 40 patients who switched because of treatment failure, 3 eventually discontinued therapy and 18 modified therapy again to move on to a salvage regimen.
Cost was the primary reason (n = 81) for discontinuing any ART, followed by an AE (21%) and treatment failure (1.6%). The median time on therapy before discontinuing was 124 days (range: 1-1050 days). Those with higher CD4 counts (>200 cells/μL) at the time of initiation of therapy more frequently discontinued therapy (OR = 2.5, 95% CI: 1.5 to 4.3). Age, gender, presence of an AIDS-defining illness, hemoglobin, and BMI at the time of initiation were not significantly associated with discontinuing therapy.
Cost was also the third most common reason for modifying therapy. 3TC plus d4T plus NVP, the most commonly used regimen, costs approximately $20 per month. 3TC plus AZT plus NVP, the next most commonly used regimen, costs $25 per month. Combinations with EFV are almost twice as expensive, at $50 to $58 dollars per month. PI-based regimens cost almost $140 per month.
At the YRG CARE, a tertiary HIV care and research center in a resource-constrained setting, more than half the patients who needed ART according to WHO guidelines10 had initiated therapy by 2004. This is in stark contrast to the mere 8.2% of patients who initiated therapy at this center between 1996 and 1999.11 This change is attributable to the introduction of generic ART at a lower cost in the year 2000. Since this time, the number of people initiating ART has drastically increased and HIV-related morbidity and mortality have significantly declined.1
It is interesting to note that although 95% of infections were most likely acquired heterosexually, 82% of those on therapy were male. Part of this gender disparity might be explained by the fact that we included only those patients who were ART naive before initiation of triple HAART, effectively excluding women who previously received monotherapy with AZT or single-dose NVP during pregnancy. It is likely that this disparity between the number of male and female patients on therapy has more to do with the social and economic standing of women within the family unit and in society at large, however.
In this cohort, antiretroviral regimens were modified for a number of reasons, but as has been the case in many studies from the developed world,6-8 occurrence of AEs was the most common reason, accounting for 13% of the entire cohort and 63% of patients modifying therapy. As expected, those who substituted drugs because of an AE substituted earlier than those who switched because of treatment failure or cost.
Most patients at the YRG CARE were on NVP-containing regimens and were more likely to substitute therapy because of AEs than patients on other therapies. Rash and hepatotoxicity were the most common AEs noted in the NVP cohort. Median time to development of hepatotoxicity was 34 days. WHO guidelines recommend laboratory monitoring every 6 months after initiating therapy.10 In light of our data, however, earlier laboratory monitoring for toxicity is advisable. In addition, 87% of patients were on rifampicin-containing anti-TB medications at the same time. The cost of alternative regimens that are less hepatotoxic limits their use in resource-constrained settings. Further research is needed on HIV-TB drug toxicities and interactions and on how best to avoid them.
The most common and least expensive ($20 per month) first-line regimen used in this context is d4T plus 3TC plus NVP. EFV-based therapies are almost twice as expensive, and PI-based regimens are almost 7 times more expensive. Despite the high incidence of d4T-related toxicities, including lipoatrophy and peripheral neuropathy, d4T continues to be widely used in resource-constrained settings because of its low cost. Efforts should be made to replace d4T with alternative affordable drugs. In addition, more than 80% of our cohort was on NVP-containing regimens, also known to have a significant risk of causing hepatotoxicity. Once again, availability at low cost necessitates the use of this drug. Although patients were more likely to substitute earlier from NVP because of toxicity, they changed therapy earlier because of cost if they were on other regimens. The fine balance between tolerability and affordability needs to be considered when making treatment decisions in resource-constrained environments.
Most patients in India pay out of pocket for their HIV treatment. Almost 9% of patients discontinued therapy after approximately 4 months, usually because of the cost of therapy. Even though the cost of ART has fallen drastically in the past few years, it still remains out of the reach of many Indians with HIV. Approximately half of the patients who stopped therapy restarted on the same therapy, presumably after their financial situation improved. Inconsistent therapy can be harmful and can put the patient at risk for developing resistance to the less expensive regimens and force the patient to eventually go off therapy or to move to second- or third-line agents, which are often more expensive and less likely to be affordable to a patient who already stopped taking a first-line regimen because of cost. This underlines the importance of financial counseling and long-term financial planning that might help patients to remain on therapy without interruptions. International donor programs or government ART roll-out programs, which provide free or low-cost ART, can play an important role in removing cost as a barrier to treatment success.
Consistent with studies from the developed world, treatment failure was not the major reason for switching therapy.6-8 A significant number of people did experience clinical or immunologic failure, however. Almost half of the patients who switched because of treatment failure did so after the occurrence of an opportunistic infection, and 70% switched after a decline in CD4 cell count; the median time to switching was approximately 13 months. This is in contrast to reports from the developed world, where most patients who switched because of treatment failure did so as the result of an unsuppressed viral load.6-8 Studies show that virologic failure, reflected by a rise in PVL, or the lack of adequate suppression of viral load occurs before immunologic or clinical failure.11 Monitoring viral load can facilitate early detection of treatment failure and avoid the unnecessary morbidity experienced by our patients. Few viral load levels were available for the patients in our study because of the high cost of monitoring them. Although routine viral load testing is part of the guidelines from the US National Institutes of Health12 and is used widely in the developed world, the WHO guidelines for the developing world do not recommend it because of concern about its high cost. Efforts should be made to develop cost-effective means of quantifying PVL suitable to resource-constrained settings.
In conclusion, a significant number of people substituted therapy because of toxicities. The most frequently used NVP-containing regimens, although less expensive, were more likely to be associated with toxicities. Because of lack of resources for early detection of treatment failure through viral load monitoring, most of those who experienced treatment failure had clinical or immunologic failure as the indication for switching therapy. Less expensive and more tolerable first-line regimens as well as cost-effective treatment monitoring tools are needed to achieve better treatment outcomes and improve access in resource-constrained settings. More studies are needed from the developing world to establish criteria for substituting and switching therapy, evaluate the best combinations for second-line therapy, and direct treatment guidelines for resource-constrained settings.
The authors are grateful to A. Priya Cristy Bai and the research nurses as well as all the clinical staff at the YRG CARE, Voluntary Health Services, Chennai, India for their facilitation of the study. The authors thank S. Anand, K. Jayakumar, and T. Guru for database management.
1. Kumarasamy N, Solomon S, Cecelia A, et al. Impact of HAART on opportunistic infections and on survival of persons with HIV disease in Southern India [abstract 61.024]. Presented at: 11th International Congress of Infectious Diseases; 2004; Cancun.
2. Kumarasamy N, Solomon S, Chaguturu SK, et al. The safety, tolerability and effectiveness of generic antiretroviral drug regimens for HIV-infected patients in south India. AIDS
3. Hosseinipour M, Namrika D, Magomero K, et al. The Malawian antiretroviral program: the first year experience with Triomune [abstract 172]. Presented at: 10th Conference on Retroviruses and Opportunistic Infections; 2003; Boston.
4. Laurent C, Kouanfack C, Koulla-Shiro S, et al. Effectiveness and safety of a generic fixed-dose combination of nevirapine, stavudine, and lamivudine in HIV-1 infected adults in Cameroon: open-label multicentre trial. Lancet
5. Kumarasamy N, Vallabhaneni S, Flanigan TP, et al. Rapid viral load suppression following generic HAART in Southern Indian HIV-infected patients. AIDS
6. d'Arminio MA, Cozzi LA, Rezza G, et al. Insights into the reasons for discontinuation of the first highly active antiretroviral (HAART) regimen in a cohort of antiretroviral-naïve patients. AIDS
7. Mocroft A, Youle M, Phillips AN, et al. Reasons for modification and discontinuation of antiretrovirals: results from a single treatment centre. AIDS
8. van Roon EN, Verzjil JM, Juttmann JR, et al. Incidence of discontinuation of highly active antiretroviral (HAART) and its determinants. J Acquir Immune Defic Syndr
9. World Bank. Attacking Poverty. The World Development Report
. Washington, DC: The World Bank; 2000.
10. World Health Organization. Scaling up antiretroviral therapy in resource-limited settings: treatment guidelines for a public health approach. 2003. Available at: http://www.who.int/3by5/publications/documents/arv_guidelines/en/
11. Deeks SG, Barbour JD, Grant RM, et al. Incidence and predictors of clinical progression among HIV infected patients experiencing virologic failure of protease inhibitor-based regimens [abstract 428]. Presented at: Eighth Conference on Retroviruses and Opportunistic Infections; 2001; Chicago.
12. National Institutes of Health. Guidelines for the use of antiretroviral agents in HIV-1 infected adult and adolescents. March 23, 2004. Available at: http://aidsinfo.nih.gov/guidelines/adult/AA_032304.html