Short Communication: Lower Baseline CD4 Count Is Associated with a Greater Propensity Toward Virological Failure in a Cohort of South African HIV Patients

Worldwide, approximately 34 million individuals are living with HIV with close to 70% of these individuals residing in sub-Saharan Africa. ¹ In an attempt to help overcome challenges in accessing antiretroviral therapy (ART) for individuals in developing countries, various initiatives have been implemented. These include the World Health Organization (WHO) “3 by 5” campaign (with the goal of providing ART to 3 million individuals from low and middle-income countries by the year 2005) and the “Universal access by 2010” interagency initiative. ^2,3 Although these programs have resulted in dramatic increases in ART availability, accompanying these scale-up programs are the challenges of managing heavy case loads in the setting of resource constraints.

Within South Africa, KwaZulu-Natal is the district with the highest prevalence of HIV. ⁴ The antiretroviral service at Edendale Hospital in Pietermaritzburg, KwaZulu-Natal, South Africa has initiated more than 9,000 adults on therapy since 2004; however, virological outcomes among this patient cohort have not been systematically assessed. Therefore, we conducted a retrospective chart review of patients initiating ART in recent years of the antiretroviral roll-out to determine the efficacy of this program. A secondary objective was to identify any potential correlates associated with virological failure in this population.

Two hundred and twenty-eight clinic records were randomly selected from patients who had initiated ART in the period between January 2009 and December 2012. Virological failure was defined as failure to achieve a plasma viral load (VL) <25 copies/ml after 6–12 months of ART initiation or ≥2 consecutive HIV-RNA viral loads ≥400 copies/ml following suppression of <25 copies/ml (NucliSens Easy Q assay, bioMérieux Diagnostics). Individuals with consecutive VLs between 25 and 399 copies/ml were also considered as treatment failures. Individuals achieving periods of complete VL suppression interrupted by nonconsecutive VLs of 25–399 copies/ml were considered as experiencing treatment blips and were not considered as treatment failures. Lost to follow-up was defined as having missed a clinic appointment by >3 months after the scheduled date.

Twenty-one out of 228 (9%) individuals experienced virological failure necessitating a change in their antiretroviral regimen. The median [interquartile range, (IQR)] duration of antiretroviral exposure was 19 (11–31) months. Individuals experiencing virological failure did not differ from individuals experiencing virological success with regards to sex,age, race, or baseline hemoglobin, creatinine, alanine aminotransferase level, or weight (p>0.05; Table 1). No differences were observed in the type of antiretroviral regimen, WHO stage at time of HAART initiation, or tuberculosis infection status between groups. However, individuals experiencing virological failure were found to have significantly lower baseline CD4 counts than individuals experiencing virological success [median (IQR) 74 (31–94) versus 142 (61–211) cells/μl, p=0.0036 (Mann–Whitney U test)]. Peripheral neuropathy and lipodystrophy were other common causes for ART regimen change (6% and 4% of all patients, respectively; Table 1). There were no differences between individuals experiencing virological success versus individuals experiencing failure in terms of the conditions developing while on ART (Table 1). Overall, 25 (11%) individuals were lost to follow-up. Loss to follow-up was similar between individuals experiencing virological success and failure.

In previously published studies examining virological failure, patients had received antiretrovirals for a range of 3–48 months. ⁵ In a systematic review performed by Barth et al., 19 studies reported virological failure in adult patients in sub-Saharan Africa as an HIV-RNA rebound greater than 1,000 copies/ml. ⁵ Based on 17 of 19 studies, the median proportion (range) of individuals experiencing virological failure based on this definition was 14% (0–43%). ⁵ As discussed by Barth et al., 43% of virological failure was observed in a cohort of Rwandan patients in which many individuals were receiving mono or dual therapy. ^5,6 In our study individuals experiencing virological failure were found to have significantly lower baseline CD4 cell counts than individuals experiencing virological success. In line with our results, other studies on both African and non-African cohorts have found that baseline CD4 counts are associated with increased virological failure. ^{7 –13}

In a retrospective chart review performed between 2004 and 2009 for 3,862 randomly selected individuals from resource-limited settings including Kenya, Uganda, Zambia, and Nigeria, and using cross-sectional viral load data, Amoroso et al. found that baseline CD4 count <100 cells/μl was associated with a negative impact on virological outcome. ⁷ In a prospective study of 3,162 individuals in Capetown, South Africa starting ART between 2002 and 2008, Nglazi et al. also found that a baseline CD4 count ≤50 cells/μl was a risk factor for poor virological outcomes. ⁹ In addition, in an observational multicohort study involving 66,953 patients from 57 South African sites who initiated ART between 2004 and 2010, Fatti et al. found that baseline CD4 <200 cells/μl was strongly predictive of reduced virological suppression in addition to mortality. ¹⁰ In our study, we observed a significant difference in baseline CD4 with individuals experiencing virological failure as opposed to virological success.

Although other studies performed in African cohorts have found that individuals experiencing virological failure were more likely to be female, ¹⁴ younger (≤25 years ⁹ or <30 years ⁷ ), pregnant, ¹⁵ and have lower baseline hemoglobins (<12 g/dl), ¹⁶ we did not observe these differences between individuals experiencing virological success versus failure. We also did not observe any difference in type of antiretroviral regimen, WHO stage at time of HAART initiation, or tuberculosis infection status between groups. Other studies in African cohorts have found that protease inhibitor-based therapies were associated with increased virological failure, ^14,17 and tuberculosis coinfection has been shown to be an independent risk factor for adverse virological outcomes. ¹⁷ It is possible that we were unable to detect these differences due to our relatively small sample size.

The lost to follow-up rate in our study was comparable to the rate observed in another large study in a South African cohort from KwaZulu-Natal. ¹⁸ However, our study was limited by its retrospective nature and the relatively small sample size, which may have precluded us from identifying other predictors of virological failure. Furthermore, portions of data and entire charts were sometimes missing from the medical records office. It is possible that individuals with missing data or charts may have differed systematically from individuals whose charts were present and whose charts contained complete information. Nonetheless, using a relatively strict definition of virological failure, we demonstrated that virological success was achievable in over 90% of individuals initiating ART at Edendale Hospital and loss to follow-up was comparable to loss observed in another large study on patients from the same province. ¹⁸ Given the significant challenges of providing healthcare in resource-limited settings during periods of rapid expansion, it is encouraging that virological success appears to be achievable in a majority of individuals.