Decreasing Adult Prevalence of Advanced HIV Disease From 2 Population-Based HIV Impact Assessments in Eswatini, Lesotho, Malawi, Tanzania, Uganda, Zambia, and Zimbabwe, 2015–2023

Abstract

Background

Limited population data exist assessing trends in advanced HIV disease (AHD) in the test-and-treat era. We examined changes in AHD prevalence from 2015 to 2017 (round 1) and 2019 to 2023 (round 2) in 7 African countries.

Methods

We pooled data from 2 rounds of the Population-Based HIV Impact Assessment (PHIA) surveys from 7 countries among persons with HIV (PWH) aged 15 to 59 years. AHD was defined as having an HIV-positive result and CD4 <200 cells/mm³. Awareness of HIV diagnosis and treatment status were based on self-report and detection of antiretrovirals. Weighted estimates between survey rounds were compared through complex survey methods.

Results

Across all countries, 13.2% and 6.8% of PWH met criteria for AHD during rounds 1 and 2, respectively. AHD prevalence among those aware of their HIV status and not on antiretroviral therapy (ART) increased by 31.7% (20.8% to 27.4%) but decreased by 17.7% (17.5% to 14.4%) among those unaware and not on ART and by 53.3% (10.5% to 4.9%) among those aware and on ART. The adjusted odds ratio for AHD comparing rounds 1 and 2 among those aware and on ART was 0.46 (95% CI, 0.39–0.53). Those who were aware and on ART accounted for the majority of PWH with AHD in both rounds (51.6%, round 1; 58.8%, round 2).

Conclusions

These findings indicate substantial reductions in AHD prevalence among those on ART between 2015–2017 and 2019–2023. Continued efforts are warranted to ensure timely diagnoses and treatment adherence and reduce barriers to care reengagement, which represent an increasingly important driver of AHD.

Seven African household surveys demonstrated a 48.5% decline (13.2% to 6.8%) in adult advanced HIV disease (AHD; CD4 <200 cells/mm³) from 2015 to 2023. Most people with AHD (>50%) were on antiretroviral treatment. Among new diagnoses, AHD decreased 17.7% (17.5% to 14.4%).

Advanced HIV disease (AHD) is defined by the World Health Organization (WHO) for individuals aged >5 years as having a CD4 cell count <200 cells/mm³ or WHO clinical stage 3 or 4 disease [1–3]. Individuals presenting to care with stage 3 and 4 disease are characterized by severe immune deficiency and life-threatening opportunistic infections [4]. Persons with HIV (PWH) who have AHD are at a 40% increased risk of early mortality due to opportunistic infections such as tuberculosis and cryptococcal meningitis, even after starting antiretroviral treatment (ART); the risk increases with decreasing CD4 cell count [5]. AHD is also associated with poor health-related quality of life [6]. Many opportunistic infections are difficult to screen, treat, or prevent with the limited diagnostics and treatment capacities present in resource-limited settings in most high-burden countries in sub-Saharan Africa [2]. Given limited CD4 testing in many countries [7, 8], an additional challenge is to identify patients with asymptomatic AHD, who represent nearly half of all patients who present with AHD [3, 9]. Caring for individuals with AHD and associated opportunistic infections is resource intensive and represents an important challenge, especially if managing additional comorbidities in aging PWH [10, 11]. Early identification and rapid treatment with a specific package of care for PWH with AHD remain critical to reducing HIV-associated morbidity and mortality [12].

Starting in 2016, HIV treatment coverage increased substantially with the adoption of WHO’s “test and treat” guidelines, which recommended ART initiation for patients following HIV diagnosis, without requiring a baseline CD4 count or clinical staging at HIV diagnosis [1]. Expanded HIV treatment resulted in substantial progress in reducing mortality due to HIV, but reduction in AIDS-related deaths has slowed and plateaued in the last 10 years. The proportion of PWH with AHD remains significant. In 2025, Stelzle et al. estimated that nearly 1.9 million PWH in sub-Saharan Africa were living with AHD [13]; the Joint United Nations Programme on HIV and AIDS estimated 390 000 AIDS-related deaths in Africa, >60% of the global total [14]. When strategies are systematically implemented to address AHD in the community, this is followed by an accompanied reduction in AIDS-related mortality. In addition, there has been a shift in country investments over the past 10 years from CD4 to viral load testing to monitor the response to ART and as an indicator of ongoing transmission [7, 8]. Virologic control has been the focus of the HIV response and is considered the key indicator of individual and program outcomes [2, 15]. However, an analysis of population-based data in Zimbabwe showed that 30% of PWH with AHD were virally suppressed, highlighting that PWH may experience immunosuppression despite ART and virologic response [16]. Accurate, timely, and available CD4 testing remains essential for diagnosis of AHD at baseline and when patients return to care; it has also been shown to be an important predictor of mortality, regardless of sex, age, or type and number of comorbidities [1, 6, 7, 17].

Limited population-level data exist that describe the prevalence of AHD across sub-Saharan Africa in the era of test and treat. Recent country-specific estimates of PWH with AHD found an 11% prevalence of AHD in Rwanda and 17% in Zimbabwe [16, 18]. A systematic review of studies conducted in South Africa reported a pooled AHD prevalence of 44% among patients who were treatment naive and 59% among those who were treatment experienced [19]. Between 2004 and 2015, a 10-country modeling study showed a decline in the burden of AHD among PWH on treatment [20]. However, these modeled estimates were based on limited data, and declines occurred before the shift toward test-and-treat guidance in 2016. To understand recent AHD trends over time at a population level to guide service planning and to address factors that lead to AHD, we used data from 14 Population-Based HIV Impact Assessments (PHIAs) from 7 countries in sub-Saharan Africa where 2 rounds of surveys were conducted between the years 2015 and 2023. We compared, by survey round, the proportion and associations of PWH with AHD and related demographic and clinical factors.

METHODS

We conducted a secondary data analysis of 14 PHIA surveys from 7 sub-Saharan Africa countries: Eswatini, Lesotho, Malawi, Tanzania, Uganda, Zambia, and Zimbabwe. Each country conducted 2 separate cross-sectional surveys that included CD4 data approximately 6 years apart. Data collection was conducted between 2015 and 2017 for round 1 surveys and between 2019 and 2023 for round 2. PHIA surveys provide data that are used to evaluate the impact of HIV programs and progress toward the 95-95-95 goals of the Joint United Nations Programme on HIV and AIDS [21]. Each PHIA survey used a 2-stage cluster sampling design to obtain a nationally representative sample of households, stratified by designated geographic zones; PHIA survey design and implementation have been described in detail elsewhere [22]. Household interviews were conducted to capture demographic, behavioral, and clinical information, including self-reported knowledge of HIV status and testing and treatment history. Individuals were eligible to participate if they slept in a selected household the night before the interview and provided informed consent.

Survey procedures included home-based HIV testing based on each country's national algorithm, counseling, return of results, active linkage to care where possible, and laboratory-based confirmation of HIV-positive and indeterminate results [23]. Blood samples from PWH underwent CD4 testing in the household (round 1) or nearby satellite laboratory (round 2) with the Pima CD4 point-of-care assay (Abbott) [24], as well as viral load testing and qualitative testing for the presence of select antiretroviral (ARV) drugs based on the first- and second-line HIV treatment regimens in each country [25].

Variable Definitions

AHD was defined as a CD4 result <200 cells/mm³. Viral load suppression (VLS) was categorized as <1000 HIV-1 RNA copies/mL. Demographic factors included country, sex, age, marital status, education, and household wealth. Age was categorized in 5- and 10-year bands. Household wealth was defined as upper 60% for the top 3 quintiles of the results and lower 40% for the lower 2 quintiles. Awareness of HIV status and treatment status was determined by responses provided in the survey questionnaires, final HIV classification status, and ARV detection. Participants with a positive HIV test result were categorized into 3 categories based on awareness of HIV status and treatment status: unaware–not on ART, aware–not on ART, or aware–on ART (Table 1).

Table 1.

Awareness of HIV Status and Treatment Status Definitions

	HIV Status		Treatment Status
Category	Self-report		Self-report		ARV Detectiona
Unaware–not on ART	Reported negative or unknown HIV status	AND	Reported not on treatment	AND	Not detected
Aware–not on ART	Reported HIV positive	AND	Reported not on treatment	AND	Not detected
Aware–on ART	Reported HIV positive	AND	Reported on treatment	OR	Detectedb

Abbreviations: ART, antiretroviral therapy; ARV, antiretroviral.

^aARVs were detected in dried blood spots by the liquid chromatography–mass spectrometry method.

^bIncludes 20 persons with HIV classified as aware–on ART who had ARVs detected in blood but were missing self-report HIV status.

Statistical Methods

The analysis was conducted by pooling the following 14 PHIA datasets: the 2015–2017 and 2019–2023 datasets across the 7 countries by survey round for participants with a HIV-positive test result. The analysis dataset was restricted to those with an HIV-positive test result, valid CD4 test result, and an age of 15 to 59 years, an age group common to all 14 surveys. We used pooled weights for the 2 surveys by country via the jackknife method with replicate weights to account for an unequal probability of household selection, as well as nonresponse and noncoverage [26]. We reported unweighted frequencies and weighted percentages. We estimated overall weighted AHD prevalence, by awareness and treatment status, VLS status, country, and survey round. We compared AHD prevalence overall and by various demographic characteristics and awareness/treatment status between survey rounds using 95% CIs generated via jackknife variance estimation, percentage differences, and χ² tests to generate P values. We calculated weighted frequency estimates by 5-year age groups for PWH and AHD to compare by round, sex, and awareness and treatment status. We modeled AHD by awareness and treatment status using logistic regression to compute unadjusted and adjusted odds ratios between survey rounds (P < .05 was considered statistically significant). Variable selection for inclusion in the model was based on a priori evidence that demographic variables such as age, sex, residence, country, and socioeconomic status represent likely confounders. Weighted analyses were performed in SAS version 9.4 (SAS Institute).

Patient Consent Statement

PHIA surveys were approved by institutional review boards at the US Centers for Disease Control and Prevention, Columbia University Medical Center, Westat, and University of Maryland, Baltimore, and in the following countries: Eswatini Health and Human Research Review Board, Eswatini; Ministry of Health Research and Ethics Committee, Lesotho; National Health Sciences Research Committee, Malawi; National Institute for Medical Research, Tanzania and Zanzibar Health Research Institute, Zanzibar; Uganda Virus Research Institute and Uganda National Council for Science and Technology, Uganda; Tropical Disease Research Centre and National Health Research Authority, Zambia; and Medical Research Council of Zimbabwe and Research Council of Zimbabwe, Zimbabwe (see 45 CFR part 46; 21 CFR part 56). All participants provided written or verbal informed consent prior to participating.

RESULTS

From the round 1 and round 2 surveys, there were 133 085 and 129 658 eligible participants, respectively, aged 15 to 59 years who had valid HIV and CD4 test results in the pooled sample from the 7 countries (Figure 1). During round 1, 8.2% (n = 17 180) tested positive for HIV, and of these 13.2% (n = 2015) met the criteria for AHD. During round 2, 7.4% (n = 15 774) tested positive for HIV, and of these 6.8% (n = 980) met the criteria for AHD, an overall decrease of 48.5% in AHD from round 1 to round 2 (P < .05). For round 1, the proportion of PWH with AHD was similar among men and women (men, 50.6%; women, 49.4%), whereas for round 2 (P = .092), there was a significantly higher proportion of men with AHD as compared with women (men, 54.7%; women, 45.3%; P < .05).

Figure 1.

Analytic study population consisting of persons with HIV aged 15 to 59 years in the PHIA surveys: 7 African countries, 2015–2023. Data are presented by survey round, advanced HIV disease status (CD4 <200 cells/μL), and sex. The figure reports unweighted frequencies and weighted percentages. Round 1 PHIA surveys: 2015–2017. Round 2 PHIA surveys: 2019–2023. Abbreviations: AHD, advanced HIV disease; PHIA, Population-Based HIV Impact Assessment. ^aMissing CD4 testing: round 1, n = 90 (0.52%); round 2, n = 22 (0.14%).

AHD prevalence by sex significantly decreased by survey round overall and within each survey country, except for men in Uganda (Figure 2). Men had a significantly higher prevalence of AHD (18.1%) when compared with women (10.3%) overall and in most countries, except Uganda during round 1 and Zambia for round 2. Zimbabwe had the highest prevalence of AHD in men during round 1 (24.8%) and round 2 (15.4%). AHD prevalence decreased significantly by round for all age groups and all educational, marital, and wealth status categories (P < .05; Table 2). In round 1, AHD prevalence was significantly lower among those aged 15 to 24 years (8.1%) than those aged 25 to 34 years (12.5%, P < .05). However, by round 2, there was no significant difference in AHD by age group. AHD prevalence decreased by 17.7% among those unaware–not on ART and increased by 31.7% among those aware–not on ART, although the latter was not significant (P = .0502). Among those aware–on ART, AHD prevalence decreased 53.3% from 10.5% to 4.9%; a similar significant decline was noted among those with VL <1000 copies/mL (45.0% decline, P < .05).

Figure 2.

Prevalence of advanced HIV disease (CD4 <200 cells/μL) by country, sex, and survey round among persons with HIV aged 15 to 59 years in the Population-Based HIV Impact Assessment surveys: 7 African countries, 2015–2023. Round 1 surveys: 2015–2017. Round 2 surveys: 2019–2023.

Table 2.

AHD Prevalence and Percentage Difference by Select Demographic Characteristics and Clinical Indicator Status by Survey Round Among PWH Aged 15–59 Years in the PHIA Surveys: 7 African Countries, 2015–2023

	Round 1		Round 2
Characteristic	No.	AHD, % (95% CI)	No.	AHD, % (95% CI)	Percentage Differencea
Sex
Men	5452	18.1 (16.6–19.6)	4737	10.7 (9.4–12.1)	−40.9*
Women	11 728	10.3 (9.5–11.2)	11 307	4.8 (4.2–5.4)	−55.1*
Age, y
15–24	1901	8.1 (6.2–9.9)	1426	5.4 (3.7–7.0)	−33.3*
25–34	5098	12.5 (11.0–14.0)	3875	6.7 (5.4–7.9)	−46.4*
35–44	5694	14.8 (13.5–16.1)	5244	7.7 (6.6–8.7)	−48.0*
45–54	2424	14.6 (12.7–16.4)	4027	6.7 (5.5–8.0)	−54.1*
55–59	1518	13.4 (10.7–16.1)	2068	6.2 (4.8–7.6)	−53.7*
Residence
Urban	5815	13.3 (12.2–14.5)	6920	6.4 (5.4–7.3)	−51.9*
Rural	10 260	13.1 (12.0–14.2)	9959	7.2 (6.4–8.0)	−45.0*
Marital status
Never married	2633	13.0 (10.7–15.3)	2333	6.6 (4.9–8.3)	−49.2*
Married	9677	12.6 (11.6–13.6)	8694	6.6 (5.9–7.4)	−47.6*
Divorced/separated	2478	14.8 (13.0–16.5)	2595	8.4 (7.0–9.9)	−43.2*
Widowed	2344	14.2 (11.9–16.6)	2069	5.5 (3.9–7.0)	−61.3*
Education
No education	978	13.1 (10.4–15.8)	1376	6.2 (4.4–8.0)	−52.7*
Primary	7194	11.7 (10.7–12.6)	7545	6.6 (5.7–7.5)	−43.6*
Secondary	6564	13.6 (12.4–14.8)	6108	7.5 (6.6–8.5)	−44.9*
More than secondary	722	15.5 (11.8–19.1)	730	5.9 (3.4–8.3)	−61.9*
Wealth
Lower 40%	5667	12.5 (11.3–13.8)	5843	7.4 (6.3–8.5)	−40.8*
Upper 60%	9790	12.8 (11.9–13.7)	8267	6.7 (5.8–7.5)	−47.7*
Country
Eswatini	2788	7.5 (6.5–8.6)	2611	3.8 (3.1–4.5)	−49.8*
Lesotho	3191	11.5 (10.2–12.7)	3237	6.7 (5.7–7.6)	−42.0*
Malawi	2132	12.8 (11.0–14.7)	2238	6.0 (4.8–7.2)	−53.0*
Tanzania	1702	14.8 (12.5–17.1)	1621	6.6 (5.1–8.1)	−55.4*
Uganda	1686	8.8 (7.4–10.3)	1384	5.7 (4.2–7.2)	−35.6*
Zambia	2446	13.9 (12.4–15.3)	2045	7.0 (5.1–8.9)	−49.6*
Zimbabwe	3235	16.8 (15.2–18.4)	2638	9.5 (8.2–10.7)	−43.7*
Clinical indicatorb
Unaware–not on ART	3771	17.5 (15.8–19.2)	1750	14.4 (12.1–16.7)	−17.7*
Aware–not on ART	1324	20.8 (17.9–23.7)	379	27.4 (20.6–34.3)	+31.7
Aware–on ART	12 076	10.5 (9.6–11.4)	13 643	4.9 (4.3–5.5)	−53.3*
VLS <1000 copies/mL	11 091	6.0 (5.3–6.7)	13 044	3.3 (2.8–3.8)	−45.0*
Total	17 180	13.2 (12.4–14.0)	15 774	6.8 (6.2–7.5)	−48.5*

Table reports unweighted frequencies and weighted percentages. Round 1 PHIA surveys: 2015–2017. Round 2 PHIA surveys: 2019–2023.

Abbreviations: AHD, advanced HIV disease (CD4 <200 cells/μL); ART, antiretroviral therapy; ARV, antiretroviral; PHIA, Population-Based HIV Impact Assessment; PWH, persons with HIV; VLS, viral load suppression.

^aRao-Scott χ² test. Percentage difference: (round 2 – round 1) / round 1 × 100.

^b Unaware–not on ART, self-reported as not previously diagnosed and not on ART and ARVs not detectable; aware–not on ART, self-reported as previously diagnosed and not on ART and ARVs not detectable; aware–on ART, self-reported as previously diagnosed and on ART or ARVs detectable.

^* P < .05.

The total weighted estimated number of PWH with AHD was 790 358 in round 1, which decreased to 415 864 in round 2 (Table 3). Among the population with AHD, the proportion of unaware–not on ART decreased from 37.4% to 31.4%, and the proportion of aware–not on ART decreased from 11.0% to 9.8% from round 1 to round 2. Those who were aware–on ART accounted for the largest proportion of PWH with AHD in both rounds (51.6% for round 1%, 58.8% for round 2). The overall number of PWH (Figure 3A, Supplementary Table 1) differed slightly between the rounds with an increase in PWH aged ≥45 years in round 2. The number of PWH with AHD (Figure 3B) decreased in round 2 among the 7 countries. AHD was more prevalent among men in both rounds of the survey, and the proportion of PWH aware–on ART with AHD increased in the second round, especially among women. Among PWH aged ≤34 years, AHD was more prevalent in women than PWH aged ≥35 years. AHD was more prevalent in men in both rounds.

Table 3.

AHD Burden by Awareness and Treatment Status by Survey Round Among PWH Aged 15–59 Years in the PHIA Surveys: 7 African Countries, 2015–2023

	Round 1		Round 2
Awareness / Treatment Statusa	Weighted No.	AHD, % (95% CI)	Weighted No.	AHD, % (95% CI)	Percentage Differenceb
Unaware–not on ART	293 706	37.4 (34.2–40.6)	130 371	31.4 (26.9–35.9)	−15.9
Aware–not on ART	87 304	11.0 (9.3–12.7)	40 835	9.8 (7.1–12.5)	−10.9
Aware–on ART	409 348	51.6 (48.4–54.8)	244 658	58.8 (54.3–63.3)	+13.5
Total	790 358		415 864

Table reports weighted frequencies and weighted percentages. Round 1 PHIA surveys: 2015–2017. Round 2 PHIA surveys: 2019–2023.

Abbreviations: AHD, advanced HIV disease (CD4 <200 cells/μL); ART, antiretroviral therapy; ARV, antiretroviral; PHIA, Population-Based HIV Impact Assessment; PWH, persons with HIV.

^a Unaware–not on ART, self-reported as not previously diagnosed and not on ART and ARVs not detectable; aware–not on ART, self-reported as previously diagnosed and not on ART and ARVs not detectable; aware–on ART, self-reported as previously diagnosed and on ART or ARVs detectable.

^bPercentage difference: (round 2 – round 1) / round 1 × 100.

Figure 3.

A, Pooled 5-year age distribution by survey round, sex, and awareness and treatment status of PWH aged 15 to 59 years in the PHIA surveys: 7 African countries, 2015–2023. B, Pooled 5-year age distribution of AHD (CD4 <200 cells/μL) by survey round, sex, and awareness and treatment status among PWH aged 15 to 59 years in the PHIA surveys: 7 African countries, 2015–2023. Note that scales differ between panels A and B for the PWH and AHD populations. Figure reports weighted frequencies and weighted percentages. Round 1 PHIA surveys: 2015–2017. Round 2 PHIA surveys: 2019–2023. Awareness and treatment status: unaware–not on ART, self-reported as not previously diagnosed and not on ART and ARVs not detectable; aware–not on ART, self-reported as previously diagnosed and not on ART and ARVs not detectable; aware–on ART, self-reported as previously diagnosed and on ART or ARVs detectable. Abbreviations: ART, antiretroviral therapy; ARV, antiretroviral; PHIA, Population-Based HIV Impact Assessment; PWH, persons with HIV.

Table 4 provides the unadjusted and adjusted odds ratios of AHD comparing survey round by awareness and treatment status among PWH. For those unaware–not on ART, we found a 20% significant reduction in odds of AHD between survey rounds, but this was not significant in the fully adjusted model (adjusted odds ratio, 0.84; 95% CI, .68–1.04). There was a nonsignificant increase in the unadjusted and adjusted odds of AHD among those aware–not on ART. We found a significant reduction in odds of AHD among those aware–on ART between rounds, with a 54% decrease (adjusted odds ratio, 0.46; 95% CI, .39–.53) in the odds of AHD between rounds, adjusted for age, sex, residence, country, marital status, education, and wealth.

Table 4.

Unadjusted and Adjusted Odds Ratios of AHD Comparing Survey Round by Awareness and Treatment Status Among PWH Aged 15–59 Years in the PHIA Surveys: 7 African Countries, 2015–2023

	Odds Ratio (95% CI)
Awareness / Treatment Statusa	Unadjusted	P Value	Adjustedb	P Value
Unaware–not on ART
Round 1	1 [Reference]		1 [Reference]
Round 2	0.80 (.65–.98)	.0363	0.84 (.68–1.04)	.1105
Aware–not on ART
Round 1	1 [Reference]		1 [Reference]
Round 2	1.44 (.98–2.11)	.0601	1.48 (.99–2.20)	.0552
Aware–on ART
Round 1	1 [Reference]		1 [Reference]
Round 2	0.43 (.37–.51)	<.0001	0.46 (.39–.53)	<.0001

Round 1 PHIA surveys: 2015–2017. Round 2 PHIA surveys: 2019–2023.

Abbreviations: AHD, advanced HIV disease (CD4 <200 cells/μL); ART, antiretroviral therapy; ARV, antiretroviral; PHIA, Population-Based HIV Impact Assessment; PWH, persons with HIV.

^a Unaware–not on ART, self-reported as not previously diagnosed and not on ART and ARVs not detectable; aware–not on ART, self-reported as previously diagnosed and not on ART and ARVs not detectable; aware–on ART, self-reported as previously diagnosed and on ART or ARVs detectable.

^bAdjusted for age, sex, residence, country, marital status, education, and wealth.

DISCUSSION

We found a significant reduction in AHD prevalence consistently by country at the population level among those aware–on ART even when controlling for select demographics, signifying substantial progress in treatment scale-up over the last decade. Other studies have found a significant decline in the prevalence of AHD over a similar period [19, 27], but estimates vary depending on the setting in which the data were collected. A study using data from the AFRICOS cohort study in 3 East African countries and Nigeria reported that AHD prevalence, as defined by CD4 count <200 cells/mm³, decreased from 10.5% in 2013 to 3.1% in 2021, representing a 74% reduction in AHD prevalence [28]. This study also found that when compared with being ART naive and not having VLS, the odds of AHD were lowest for those who were on ART for ≥2 years and had VLS.

Although there was an overall decline in the proportion of PWH with AHD, over half the population with AHD in round 1 and 2 surveys were aware–on ART. This represents a significant burden of individuals on treatment with AHD. Our findings also indicate that a significant number of individuals from both rounds who were previously diagnosed either never initiated ART or interrupted treatment. These findings align with studies that estimated that up to 60% of patients presenting for care with AHD had previously disengaged from care [20, 29, 30], and they underscore the importance of strategies that encourage return to care for those who may cycle in and out of care [19]. In addition, these findings support the need for additional investment toward increased routine CD4 testing in programs for patients newly diagnosed and for those returning to care after treatment interruption. However, CD4 testing in care and treatment programs has declined over the years, with more focus on viral load testing [7].

PWH who were unaware of their HIV status and had a low CD4 count met the criteria for a late HIV diagnosis. This is defined as a person at first HIV diagnosis with a CD4 count <350 cells/mm³ or with an AIDS-defining event, regardless of CD4 cell count [31]. In an analysis of round 1 PHIA data from 3 countries, approximately half of those who were unaware of their status met the criteria for late diagnosis (CD4 <350 cells/mm³) and more than half were men [32]. Our findings, with other key indicators from PHIA surveys, continue to highlight the urgent need to close the gap for the first “95” target by increasing HIV testing and initiating treatment, especially among men [33]. Recent emphasis on tailored and differentiated HIV testing approaches across country programs, including active index testing and self-testing, have helped to diagnose individuals and their partners to initiate care earlier [34]. As more at-risk people are screened and receive testing services, this will further decrease the number of individuals who present late to care, reducing severe illness or death among PWH.

Strengths of this study are inclusion of data from 14 large population-based and nationally representative surveys with high participation and rigorously conducted biomarker testing, including CD4 testing. The consistency and completeness of the objective assay-based CD4 methodology across all 14 surveys made it possible to accurately measure changes in AHD prevalence over time and across multiple countries. In addition, defining AHD by CD4 criteria is considered the best measurement to determine immune and clinical status [19, 35] and likely captured most individuals with AHD. A further strength of our study is that our definition of awareness of HIV status incorporated self-report and the results of ARV detection assays.

Limitations include exclusion of people aged <15 years and >60 years because these age groups were not uniformly included in all country surveys. As with all household surveys, all who were hospitalized at the time of the survey were excluded, and this may have led to an underestimation of AHD prevalence. We were unable to account for treatment failure or mortality data over time, which likely contributed to the reduction in AHD, and this is an important area to explore for future analyses. We did not perform WHO clinical staging to further classify those with AHD, as this clinical evaluation was not included in PHIA survey procedures. Finally, a small number of observations may have been misclassified by awareness and treatment status because of nondisclosure of HIV infection status or if the ARVs taken were not detected with available assays.

CONCLUSION

Overall, these findings indicate tremendous success in the implementation of test-and-treat strategies and other interventions in sub-Saharan Africa countries with a high prevalence of HIV, as evidenced in the reduction of AHD prevalence and burden over time. Challenges remain in preventing and addressing AHD through early diagnosis, linkage to care, screening for AHD and opportunistic infections, and delivery of other components of the AHD package of care, which reduces risk of mortality and improves the quality of life for PWH. The majority of AHD cases in our survey were among those who were aware and on treatment, highlighting the need for continued efforts to understand treatment interruption, barriers for reengagement in care, and other causes of treatment failure. This analysis supports current guidance in use of CD4 monitoring for those who present for the first time or return to care to screen for AHD in clinical settings. In addition, accurate estimation of mortality and AHD population indicators is recommended as part of national monitoring and surveillance frameworks to better plan for services and address the issues that lead to AHD. Public health efforts aimed at addressing AHD are crucial to reducing mortality and morbidity in PWH.

Supplementary Data

Supplementary materials are available at Open Forum Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.