Sustainability of antiretroviral therapy in HIV-infected adult urban Kenyans: lessons for other resource-poor countries

Frequent antiretroviral therapy (ART) switches in HIV-infected Kenyan urban adults might limit the efficacy of ART, according to the findings of an observational study published in the July 1^st issue of the Journal of Acquired Immune Deficiency Syndromes. The authors sound the alarm that this is a potentially serious threat to the sustainability of HIV treatment programmes in Kenya and other developing countries.

The availability of ART in developing countries has become more widespread and by June 2005, an estimated 500,000 Africans were receiving ART. Several studies have reported clinical and immunologic outcomes in sub-Saharan Africa that are comparable or superior to those in developed countries.

These robust responses might not be maintained in resource-poor countries with limited access to alternative ART regimens. It is imperative that ART-related toxicities and failure be monitored and prevented and to ensure that ART is efficacious, safe, accessible and affordable. These considerations are the underpinnings of a sustainable HIV treatment programme.

Such a programme must pre-empt the development of viral resistance due to treatment interruptions. This can be achieved through the provision of adequate patient follow-up, constant access to drugs, and monitoring patients on ART.

Few studies have reported on the long-term sustainability of HIV treatment programmes in Africa. A team of Kenyan and US investigators have addressed this gap in an observational study of ART treatment outcomes in HIV-infected Kenyans in an urban, resource-limited setting.

The study took place at St Mary’s Mission Hospital in Nairobi which provides basic HIV care to local residents. All HIV-infected patients who were at least 18 years old, except pregnant women, visiting the outpatient clinic were eligible for enrolment. Analysis included HIV-infected patients who were ART-naive prior to initiating treatment. Patients were required to have at least one follow-up visit while on ART.

Patients were followed up by longitudinal monitoring of clinical, immunologic, and treatment parameters from September 2004 until August 2006. A rigorous quality assurance policy followed to ensure reliability of the study findings. The main immunological outcome measured was the change in CD4 count over time while the clinical outcomes assessed included opportunistic infections and HIV/AIDS-related symptoms. Patients showing these clinical outcomes were followed up over time.

A switch in ART was recorded if the ART regimen at follow-up differed from the ART provided at recruitment. A toxicity occurred when a symptom directly due to ART was recorded during clinic visits. Laboratory-based liver toxicity occured if the serum level of the liver enzyme GLT was significantly higher than normal. Loss to follow-up was defined as missed clinic visits and failure to collect ART refills for three or more months. Deaths were based on physician or family notification.

A total of 1,286 patients were analysed with a median age of 36 years, 59.1% were female, and 77.6 % had baseline CD4 cell counts of 200 cells/mm³ or less. About 40 % had an active opportunistic infection or a history of opportunistic infections and 34.3 % had a history of TB before starting ART. The majority of the patients (98.9%) started non-nucleoside reverse transcriptase inhibitor (NNRTI)-based ART that in 62.1% of the subjects were primarily stavudine (d4T, Zerit), lamivudine (Epivir), and nevirapine (Viramune). Median ART duration was 350 days (11.6 months).

Significant improvements in clinical and immunologic status were noted after twelve months of therapy. These included significant increases in CD4 counts, weight, haemoglobin, and reductions in HIV/AIDS-related symptoms. The most frequent opportunistic infections were pulmonary TB and candidiasis seen in over 60% of the cases. CD4 cell counts of 100 cells/mm³ or less was the only significant predictor of a new opportunistic infections.

ART switches occurred in 701 (54.5%) patients. The cumulative incidence of ART switch at twelve months was 78.4%. Concurrent ART-related toxicities (40.6%) and tuberculosis treatment interactions (28.1%) were the most frequent reasons for ART switch. Baseline AIDS symptoms (P = 0.01) and a CD4 count of 100 cells/mm³ or less (P = 0.04) were independent predictors of ART switch. Switching was itself an important predictor of subsequent toxicity, a new opportunistic infection, or withdrawal from the program.

ART-related clinical toxicity occurred in 341 (26.5%) patients. Peripheral neuropathy was reported most frequently (20.7%). A CD4 count of 100 cells/mm³ or less and being older than 40 years were independent predictors of clinical toxicity. No ART-related laboratory liver toxicity was observed.

The findings echo those of studies presented recently at the 2007 HIV Implementers' Meeting in Rwanda, which showed high rates of drug switching, primarily attributable to d4T-related toxicity. Treatment programmes are increasingly having to grapple with the question of whether they can afford to use first-line drugs other than d4T in order to reduce rates of treatment switching. However the findings reported from Kenya show perhaps the highest rate of treatment switches yet recorded by an African treatment site using a d4T-based regimen.

"Because of the likelihood of stavudine toxicity affecting initial ART regimen durability in this setting, future HIV programs should reconsider using stavudine as part of front-line therapy," the authors conclude..

They also underline the problems faced in treating HIV and TB together. In this study, the authors point out, more than one in four patients had to change drugs because of interactions between antiretroviral drugs and TB medications, and 14% of patients developed a new opportunistic infection within two months of starting ART. The most common opportunistic infection in people taking antiretrovirals was TB, perhaps, say the authors, because of relatively insensitive screening methods.

The study also reported a very high rate of drop-outs: 40% of patients withdrew from the programme, which the authors attribute to the nominal clinic fee charged. However, 85% of those who withdrew were lost to follow-up, and the presence of AIDS-defining illness was a significant predictor of withdrawal, suggesting that a large proportion may have withdrawn from the programme due to death or illness rather than attempting to obtain free treatment elsewhere.