A prospective operational study of a community-based antiretroviral treatment (ART) programme in Cape Town, South Africa has reported a very high rate of mortality among treatment-naive patients in the first few months after enrolling into the site — most notably among patients waiting to go on treatment. According to the study, published in AIDS this month, nearly half of the observed deaths occurred in patients who had recently enrolled in the programme but who were not yet on treatment. The vast majority of deaths, whether on treatment or waiting for treatment, were in patients with CD4 cell counts below 50 and advanced symptomatic disease (WHO stage 3 and 4).
A number of studies have previously reported on the survival benefit observed after the ART rollout in resource-limited settings, but this is the first to report on the mortality rates among patients during the time between enrolment into the programme and the actual start of treatment. The “data highlight substantial ‘unseen’ in-programme mortality,” according to the authors, who also stressed that the wait for ART at their clinic was relatively brief compared to many other sites in South Africa. The in-programme mortality at these other ART sites is likely to be even greater.
The Study
The trial was conducted at the Gugulethu Community Health Centre, which services an economic disadvantaged population in Gugulethu township, on the outskirts of Cape Town. Between September 2002 and February 2005, 758 individuals were referred to clinic for ART — although only 712 patients were included in the final analysis because some were too young, or had previous experience on ART, or were ineligible for treatment.
After referral to the ART service, patients had to make at least three visits to the clinic before they could actually receive treatment.
During the first basic screening/enrolment visit, a ‘treatment readiness’ evaluation was completed, subjects were clinically assessed for symptoms of disease (opportunistic infections etc.) and given four weeks of cotrimoxazole prophylaxis. Each patient was also appointed a therapeutic counsellor from their community whose role was to use clinic and home visits to provide ongoing counselling, adherence support, and to provide weekly updates to the clinic team on the status (including the treatment readiness status) of the patient.
Two weeks later, the study participants had blood tests for viral load and CD4 cell counts, and their cotrimoxazole pills were counted to see whether they were adhering to the treatment.
Four weeks after the first clinic visit, cotrimoxazole pill counts were taken again, and if patients were considered ready to adhere to treatment, they were prescribed d4T/3TC plus efavirenz (Sustiva) or nevirapine (Viramune).
Analysis
The authors conclude that “using on-treatment analyses greatly underestimates early mortality” and that a better way to gauge a treatment programme’s success would be to include the deaths of patients who don’t receive treatment in time.
While the death among those waiting for treatment was high, it is difficult to fault the Gugulethu clinic. The waiting time at the Gugulethu clinic was short compared to other clinical trial sites in South Africa — some of which report a minimum of four months waiting time before ART begins. And during the pre-treatment period at Gugulethu, there was ongoing adherence counselling to produce better and more sustainable results in patients who go onto ART. Also, the community-based adherence counsellors doubtlessly improved follow-up (and ability to determine outcomes) in the people enrolled into this clinic’s programme.
A certain amount of delay after enrolment is inevitable in any programme, and treatment readiness training appears essential for long-term success. Given the very advanced stage of disease that is seen in patients enrolling into ART programmes in developing world, it may not be possible to pre-vent all early in-programme mortality.
The study team writes “minimisation of the in-programme pre-treatment interval may well decrease mortality in this and other programmes. However, a balance needs to be established between minimising the pre-treatment interval (potentially reducing early mortality risk) and allowing adequate time to prepare patients for treatment (promoting high rates of treatment adherence and reducing long-term mortality rates)."
They go onto suggest that a system could possibly be developed to “fast-track” patients at the highest risk of death (those with stage 4 disease, a CD4 count
A final way of preventing these deaths would be to redouble the efforts to get people with HIV onto treatment earlier in the course of disease (before developing symptomatic disease).
Reference
Lawn SD et al.Early mortality among adults accessing a community-based antiretroviral service in South Africa: implications for programme design. AIDS. 19(18): 2141-2148, 2005.
Results
Of the 712 patients included in the final analysis, the median CD4 count was 94 cells and the median plasma viral load was 72 349 copies/ml (IQR, 30 721–191 000). The vast majority of patients had advanced disease, WHO clinical stage 3 for 354 (50%) and stage 4 for 215 (30%).
A total of 578 patients (81%) started ART, a median of 29 days after enrolment (96% within 90 days). The most frequent reasons that the remaining 134 patients (19%) did not go on ART were 1) death, 2) decision to access treatment elsewhere, 3) failure to attend follow-up clinic appointments, 4) moving out of the area and 5) psychosocial reasons, such as denial of HIV infection status. The median period of observation for the patients who didn’t go on ART was 28 days (IQR, 19–50; maximum, 125 days).
Sixty-eight (9.5%) of the patients who enrolled into the programme died during the course of the study. The high mortality rate of 35.6 deaths/100 person years before treatment fell to 2.5/100 person-years at one year among those on ART. Overall, 37 patients died on treatment and 31 deaths were recorded for the 134 patients who did receive ART.
But within the first three months of enrolment, 29 of 44 (66%) deaths occurred among patients not yet on ART. In a multivariate analysis the risk of death (for all enrolled patients) was independently associated with baseline CD4 cell count and stage 4 disease. The major causes of death were wasting syndrome, tuberculosis, acute bacterial infections, malignancy and immune reconstitution disease (most of which were due to cryptococcal meningitis).