What this study adds: This study provides important insights into the percentage of specimens with a count ≤ 200 cells/µL at district level, masked by overall national and provincial proportions. Specific districts in four provinces were identified that warrant further investigation and intervention.

South Africa has the world’s largest HIV epidemic globally, with 7.7 million people living with HIV (PLHIV) in 2023.^1,2 An HIV prevalence of 12.7% was reported in 2023.³ Several challenges remain, such as gaps in antiretroviral treatment (ART) coverage and sub-optimal virological suppression.^1,3,4,5 Provinces such as KwaZulu-Natal and Mpumalanga experience hyper-epidemics, with prevalence exceeding 15%.³ Antiretroviral treatment coverage has increased substantially, reaching 77% of PLHIV by 2023.¹ Data from the Thembisa model showed that between 2013 and 2023, ART coverage increased by 35.6% at the national level, while the number of new HIV infections decreased fivefold.⁶ In 2022, among PLHIV aged 15 years and older, 90% were aware of their status, 91% were on ART, and 94% were virally suppressed.³

Despite improvements in ART coverage and viral suppression, levels of advanced HIV disease (AHD) remain a concern.^7,8,9 Advanced HIV disease is defined as a CD4 cell count < 200 cells/µL or WHO clinical stage 3 or 4.¹⁰ Reports suggest that almost 30% of people that initiate ART are either ART-naïve patients or re-entering after a period of disengagement.⁹ Carmona et al. reported that, between 2005 and 2011, the proportion of PLHIV aged ≥ 15 years entering care with a first CD4 count < 200 cells/µL declined from 46.8% to 35.6%.⁷ However, between 2011 and 2016, the proportion of patients entering care with CD4 counts < 200 cells/µL remained relatively unchanged, at 32.9%.⁷ A subsequent study reported that between 2017 and 2023, the percentage of specimens with a count ≤ 200 cells/µL ranged from 19.5% and 20.8%, despite a consistent increase in the median CD4.⁸ Between 2013 and 2023, the percentage of specimens with a CD4 count ≤ 200 cells/µL decreased minimally, by 0.3%.⁸ Of more notable concern is the increase in the percentage of specimens with a count ≤ 200 cells/µL between 2013 and 2023 in the Western Cape province.⁸

Re-entry into care by patients, after being on ART, at an advanced stage of HIV disease is also an important consideration.⁹ It has been reported that many PLHIV start and stop ART multiple times in a cascade through which they can re-enter HIV care.^11,12 A Western Cape observational cohort study reported that 10% of PLHIV linked to HIV care disengaged after early ART initiation, and a further 12% disengaged after long-term ART.^2,13 Similarly, Osler et al. reported that in 2016, across 10 sites, 51.8% of the PLHIV with first CD4 < 50 cells/µL were ART-experienced.¹⁴ Furthermore, 76% could be confirmed to either be off ART or to have had viraemia at the time of the index CD4 test.¹⁴ This highlights the dual challenges of re-entry into care as well as ART-experienced PLHIV presenting with AHD.^9,14

Substantial local data have been generated and reported on the burden of PLHIV who have a CD4 count ≤ 200 cells/µL at the national and provincial levels, with limited data for districts.^{1,8,15,16,17,18}

Data are reported for 5 821 932 CD4 specimens (Table 1). The test volumes declined from 3 685 032 in 2013 to 2 136 900 by 2023. There was a decline in both absolute numbers of specimens with a CD4 count ≤ 200 cells/µL between 2013 (n = 762 868) and 2023 (n = 435 753), as well as the percentage (20.7% in 2013 to 20.3% by 2023). The national median CD4 was 396 cells/µL (interquartile range [IQR]: 233–581) in 2013, compared to 465 cells/µL (IQR: 243–706) by 2023, showing an increase of the median CD4 count by 69 cells/µL over 10 years.

This study aimed to evaluate the proportion of CD4 specimens with counts ≤ 200 cells/µL across 52 health districts in South Africa, comparing data from 2013 and 2023. Our findings reveal a slight improvement in the percentage of specimens with counts ≤ 200 cells/µL. However, a substantial increase in the median CD4 count was noted. The improvement in the percentage of specimens with counts ≤ 200 cells/µL as well as the median CD4 was predominantly noted in the KwaZulu-Natal province, followed by selected districts in the Eastern Cape, Limpopo, Mpumalanga, North West, and the Northern Cape provinces. Districts in the Western Cape (Cape Winelands, Overberg, and Garden Route), Eastern Cape (Sarah Baartman, and Nelson Mandela Bay), Gauteng (Ekurhuleni, and Sedibeng) and Limpopo (Sekhukhune) provinces, reported proportions of specimens with counts ≤ 200 cells/µL exceeding 27% by 2023. This is high compared to national values of around 20% in both 2013 and 2023. This was supported by the decrease observed in the median CD4 noted between 2013 and 2023 in these eight districts.

The increase in the percentage of specimens with counts ≤ 200 cells/µL, specifically in the Western Cape, is consistent with a local analysis of CD4 data for the same period.⁸ What is surprising is that the district analysis in this study reported a percentage of specimens with counts ≤ 200 cells/µL 27% also being seen in other provinces, such as the Eastern Cape, Gauteng, and Limpopo provinces. This finding was not expected and could in part be explained by lower ART coverage rates, compared to a national ART coverage of 76.6% identified by the Thembisa model.³¹ In 2023, ART coverage rates for Western Cape (73.2%), Eastern Cape (71.6%), Gauteng (72.1%), and Limpopo (71.1%) provinces were lower than other provinces.⁸ The KwaZulu-Natal province, with the lowest percentage of specimens with counts ≤ 200 cells/µL, reported an ART coverage of 83.4%.⁶ Despite the national increase in ART coverage by 35.6% between 2013 and 2023, selected districts reported a high level of the percentage of specimens with counts ≤ 200 cells/µL, that could in part be explained by lower ART coverage.⁶

There are multiple factors that could contribute to the differences noted in the percentage of specimens with counts ≤ 200 cells/µL at the district level. One possible explanation may relate to some districts having a much higher HIV/tuberculosis co-infection rate, reported to exceed 50%.³² An analysis of 2023 laboratory data for the percentage of specimens with counts ≤ 200 cells/µL and the Mycobacterium tuberculosis complex detection rates at the district level, reported a correlation coefficient of 0.62.³³ This indicates that HIV/tuberculosis coinfection might explain some of the differences in the percentage of specimens with counts ≤ 200 cells/µL.³³ This was repeated for reflexed cryptococcal antigenaemia detection rates offered for counts < 100 cells/µL, surprisingly reporting a negative correlation coefficient of −0.209.³³ Cryptococcal antigenaemia is a common and clinically significant opportunistic infection in individuals with AHD, often serving as a precursor to cryptococcal meningitis, and associated with increased morbidity and mortality in the absence of timely screening and pre-emptive antifungal therapy.^23,24 However, factors such as other co-infections are outside the scope of this study, and may have played a contributing role.^10,24

Additionally, socioeconomic factors, as well as access to healthcare services, may also play a role.^34,35 Delays in diagnosis and treatment may also contribute, including poor communication and long waiting times at public health facilities that could discourage patients from seeking timely care.³⁶ Integration of local clinical and laboratory data is urgently needed, specifically to understand the underlying reasons for the worrying trends documented here in the Cape Winelands, Overberg, Garden Route, Sarah Baartman, and Sekhukhune districts, where a median CD4 decrease is reported in this study between 2013 and 2023. Importantly, integration of this database would enable further work to identify what percentage of testing was done for ART-naïve patients versus patients re-entering care. Furthermore, the disaggregation of data by facility type would help to elucidate where these patients with a count ≤ 200 cells/µL are presenting for care (at hospitals or primary healthcare facilities). Finally, a geospatial analysis should be conducted using facility coordinates to identify hotspots in these districts. These findings could be used to focus programmatic interventions.

Despite advancements in the accessibility of HIV services as well as improved ART coverage, the wide gap in the median CD4 at the district level across South Africa is of great concern.^6,37 This work is the first known study to identify South Africa districts with higher than expected or, in some instances, rising proportions of patients with AHD that would have been otherwise masked by only examining provincial trends. In 2023, despite a national median exceeding 460 cells/µL, there were three districts where the median CD4 was below 350 cells/µL. The decline in median CD4 values in these districts warrants special programmatic investigations. Of interest was that by 2023, the Sekhukune district reported a percentage of specimens with a count ≤ 200 cells/µL above 30%, yet the rate per 100 000 population had declined substantially. There was a notable reduction in the absolute number of specimens with a count ≤ 200 cells/µL between 2013 and 2023, with a corresponding increase of 13.3% in population estimates.³⁰ These findings confirm that, in addition to the percentage of specimens and the rate per population, the absolute number of specimens with a count ≤ 200 cells/µL is an important factor to consider in programmatic interventions. The increase in the median CD4 is also reflected as a continued increase in the percentage of specimens with a count ≥ 500 cells/µL (associated with HIV wellness).^8,16

The analysis of the 29 districts that reported a percentage point increase in specimens with a count ≤ 200 cells/µL between 2013 and 2023 indicates that over a third of specimens originate from three metropolitan districts. Furthermore, the data outcomes reported here could enable policymakers and programme stakeholders at provincial level to optimise allocation of public funds to target 95-95-95 interventions towards specifically identified populations at the district level with a higher burden of AHD.³⁸ However, focusing only on districts with high numbers of specimens with a count ≤ 200 cells/µL would be a challenge in terms of universal coverage.³⁸

Despite a consistent increase in the national median CD4 count since 2013, persisting high levels of specimens with counts ≤ 200 cells/µL continue to be seen at specific health districts in 2023 that include the Western Cape (Cape Winelands, Overberg, and Garden Route), Eastern Cape (Sarah Baartman, and Nelson Mandela Bay), Gauteng (Ekurhuleni, and Sedibeng) and Limpopo (Sekhukhune) provinces. This warrants urgent attention and further investigation. To our knowledge, the is the first South African study to document the variation in the percentage of specimens with counts ≤ 200 cells/µL at the district level, with an earlier study reporting variation at the provincial level.⁸ Despite lack of de-duplication, this analysis provides important information about specific districts and highlights areas of concern with rising burden of specimens with counts ≤ 200 cells/µL that would otherwise be masked by provincial-level data.⁸ This work provides the impetus for the development of better monitoring systems, including the integration of laboratory and clinical data for meaningful analysis and interpretation at the patient-level to provide a framework for patient-centric data approaches for interventions.