HIV treatment failure in tertiary care: clinical epidemiology and preliminary mechanistic characterisation

Abstract

Background

Despite antiretroviral therapy (ART) scale-up in sub-Saharan Africa, treatment failure remains a significant challenge. We characterised virological and immunological outcomes among people living with HIV (PLHIV) attending tertiary care facilities in Nigeria, with exploratory analysis of potential mechanistic factors.

Methods

This multi-centre cross-sectional study enrolled 517 HIV-positive adults from four Nigerian tertiary facilities between January 2019 and December 2021. Primary outcomes included viral load suppression (<1,000 copies/mL) and CD4 count. Exploratory mechanistic analyses examined drug resistance mutations (n = 50), immune activation markers (n = 40), and inflammatory biomarkers (n = 35) in pilot subsets.

Results

Among 412 participants with viral load data, only 111 (26.9%; 95% CI 22.7–31.5) achieved viral suppression, substantially below the UNAIDS 95% target. Of 387 with CD4 data, 149 (38.5%; 95% CI 33.6–43.6) had severe immunodeficiency (<200 cells/μL). Among 346 participants with complete data, discordant responses were common: 25.7% showed virological failure with preserved immunity, while 6.6% had immunological failure despite viral suppression. In pilot mechanistic subsets, 86% of viraemic participants harboured drug resistance mutations, with M184V (62%) and K103N (54%) predominating. CD8 T-cell activation (CD38+HLA-DR+) was significantly elevated in viraemic versus suppressed participants (median 28.6% vs. 12.4%; p < 0.001), correlating inversely with CD4 count (ρ = −0.46; p < 0.01).

Conclusion

HIV treatment outcomes at Nigerian tertiary facilities fall substantially short of global targets. The high prevalence of discordant immune-virological responses and preliminary evidence of drug resistance and immune activation suggest multiple interacting pathways to treatment failure. Larger mechanistic studies are warranted to inform targeted interventions.

Introduction

The global HIV response has achieved remarkable progress, with 39.9 million people living with HIV (PLHIV) worldwide and 30.7 million receiving antiretroviral therapy (ART) by 2023 (1, 2). Sub-Saharan Africa, home to approximately 26 million PLHIV, has made substantial strides toward the UNAIDS 95-95-95 targets, with several countries including Botswana, Eswatini, and Rwanda achieving epidemic control (3, 4). However, significant gaps persist, particularly regarding the third 95—viral load suppression among those on treatment.

Nigeria, with an estimated 1.9 million PLHIV, represents the second-largest HIV burden globally (5). Recent programmatic data suggest improving cascade performance, with national estimates approaching 87-98-95 toward the 95-95-95 targets (6). However, these aggregate figures may mask substantial heterogeneity across facilities and populations. Tertiary care facilities, which often receive complex cases and patients failing first-line therapy, may experience different treatment outcomes than primary care settings.

Treatment failure in HIV is multifactorial, potentially involving drug resistance, suboptimal adherence, pharmacokinetic factors, and host immune responses (7, 8). Chronic immune activation and inflammation persist even during suppressive ART and are associated with adverse clinical outcomes (9–14). Understanding the relative contributions of these factors in specific clinical contexts is essential for developing targeted interventions.

We conducted a multi-centre study to characterise virological and immunological outcomes among PLHIV attending Nigerian tertiary care facilities. Primary objectives were to determine rates of viral suppression and immunodeficiency, and to characterise discordant immune-virological responses. Secondary exploratory objectives examined potential mechanistic contributors to treatment failure, including drug resistance mutations, immune activation, and inflammatory biomarkers in pilot subsets.

Methods

Study design and setting

This multi-centre cross-sectional study was conducted at four Nigerian tertiary healthcare facilities: Abia State University Teaching Hospital (ABSUTH) Aba (Southern Nigeria), Federal Medical Centre (FMC) Makurdi and FMC Keffi (Central Nigeria), and Baru Diko Teaching Hospital Kaduna (Northern Nigeria). Participants were enrolled between January 2022 and December 2023. The study was approved by the institutional review boards of all participating facilities, and written informed consent was obtained from all participants.

Participants

Adults (≥18 years) attending HIV care services at participating facilities were eligible for enrolment. HIV status was confirmed using the national HIV testing algorithm. Participants with confirmed HIV infection were included regardless of ART status or duration. Exclusion criteria included inability to provide informed consent and acute intercurrent illness at the time of enrolment.

Clinical and laboratory assessments

Demographic and clinical data were collected using standardised case report forms. Plasma HIV-1 RNA viral load was quantified using the Roche COBAS AmpliPrep/COBAS TaqMan HIV-1 Test (lower limit of detection 40 copies/mL). CD4+ T-lymphocyte counts were determined by flow cytometry (BD FACSCount or FACSCalibur). Viral suppression was defined as <1,000 copies/mL per WHO guidelines. Severe immunodeficiency was defined as CD4 <200 cells/μL.

Exploratory mechanistic analyses (pilot subsets)

Drug resistance genotyping was performed in a convenience subset of 50 viraemic participants using population-based Sanger sequencing of the HIV-1 pol gene. Sequences were analysed using the Stanford HIV Drug Resistance Database. Immune activation was assessed in 40 participants by flow cytometry, measuring CD38 and HLA-DR co-expression on CD4+ and CD8+ T cells. Inflammatory biomarkers (IL-6, TNF-α, CRP) were quantified by ELISA in 35 participants. These analyses were exploratory and hypothesis-generating given the small sample sizes.

Statistical analysis

Categorical variables were summarised as frequencies and percentages with 95% confidence intervals calculated using the Wilson method. Continuous variables were expressed as medians with interquartile ranges (IQR). Between-group comparisons used Mann–Whitney U tests for continuous variables and chi-square or Fisher’s exact tests for categorical variables. Correlations were assessed using Spearman’s rank coefficient. Multivariable logistic regression examined factors associated with viral non-suppression, adjusting for age, sex, duration on ART, pregnancy status, and region, with site included as a fixed effect. Analyses were performed using Stata 17.0 (StataCorp, College Station, TX). A two-sided p < 0.05 was considered statistically significant.

Results

Study population

Of 1,580 participants enrolled, 1,523 (96.4%) completed HIV testing, yielding 517 HIV-positive individuals (33.9% prevalence). The cohort was predominantly female (362/517; 70.0%), with a median age of 38 years (IQR 31–46). Among women, 86 (23.8%) were pregnant. Participants were distributed across Northern (242; 46.8%), Central (207; 40.0%), and Southern (68; 13.2%) Nigeria. Biospecimen collection was successful in 492 participants (95.2%). Viral load data were available for 412 participants (79.7%), CD4 data for 387 (74.9%), and complete immunovirological data for 346 (66.9%) (Table 1; Figure 1).

Table 1.

Study population characteristics (N = 517).

Characteristic	n	%	95% CI
Sex
Female	362	70.0	65.9–73.9
Male	155	30.0	26.1–34.1
Age, years (median, IQR)	38	—	31–46
Pregnancy status (among women)
Pregnant	86	23.8	19.5–28.5
Non-pregnant	276	76.2	71.5–80.5
Region
Southern Nigeria	68	13.2	10.4–16.4
Central Nigeria	207	40.0	35.8–44.4
Northern Nigeria	242	46.8	42.5–51.2
Testing completeness
Viral load available	412	79.7	75.9–83.1
CD4 count available	387	74.9	70.9–78.5
Both VL and CD4 available	346	66.9	62.7–70.9

IQR, interquartile range; VL, viral load.

Figure 1.

Study flow diagram. STROBE-compliant flow diagram depicting participant recruitment and data availability across four Nigerian tertiary care facilities. Of 1,580 participants enrolled, 1,523 (96.4%) completed HIV testing, yielding 517 HIV-positive individuals (33.9% prevalence). Viral load data were available for 412 participants (79.7%) and CD4 count data for 387 participants (74.9%). Complete immunovirological data enabling treatment response classification were available for 346 participants (66.9%). Exploratory mechanistic analyses were performed in pilot subsets: drug resistance genotyping (n = 50), immune activation (n = 40), and inflammatory biomarkers (n = 35).

Virological outcomes

Among 412 participants with viral load data, only 111 (26.9%; 95% CI 22.7–31.5) achieved viral suppression (<1,000 copies/mL), with 43 (10.4%) having undetectable viral load (<40 copies/mL). The remaining 301 participants (73.1%; 95% CI 68.5–77.3) had persistent viraemia, including 127 (30.8%) with very high viral loads (≥100,000 copies/mL). Median viral load among non-suppressed participants was 45,230 copies/mL (IQR 8,450–185,000) (Table 2). These suppression rates fall substantially below the UNAIDS target of 95% and recent Nigerian programmatic estimates (6).

Table 2.

Virological and immunological outcomes.

Outcome	n	%	95% CI
Virological outcomes (n = 412)
Viral suppression (<1,000 copies/mL)	111	26.9	22.7–31.5
Undetectable (<40 copies/mL)	43	10.4	7.7–13.9
Persistent viraemia (≥1,000 copies/mL)	301	73.1	68.5–77.3
High-level (≥100,000 copies/mL)	127	30.8	26.4–35.6
Median VL, non-suppressed (copies/mL)	45,230	IQR:	8,450–185,000
Immunological outcomes (n = 387)
CD4 ≥500 cells/μL	90	23.3	19.1–28.0
CD4 350–499 cells/μL	70	18.1	14.4–22.3
CD4 200–349 cells/μL	78	20.2	16.3–24.6
CD4 < 200 cells/μL (severe immunodeficiency)	149	38.5	33.6–43.6
Median CD4 count (cells/μL)	285	IQR:	145–456

VL, viral load; IQR, interquartile range.

Immunological outcomes

Among 387 participants with CD4 data, 149 (38.5%; 95% CI 33.6–43.6) had severe immunodeficiency (CD4 <200 cells/μL), including 53 (13.7%) with profound immunodeficiency (<100 cells/μL). Only 90 participants (23.3%) had CD4 counts ≥500 cells/μL indicating good immune function. Median CD4 count was 285 cells/μL (IQR 145–456). These findings indicate a high burden of advanced HIV disease in this population, consistent with recent estimates suggesting 1.8–1.9 million people in sub-Saharan Africa live with advanced HIV disease (4, 7).

Discordant immune-virological responses

Among 346 participants with complete viral load and CD4 data, treatment response patterns revealed substantial discordance (Table 3). Only 62 participants (17.9%; 95% CI 14.0–22.4) achieved concordant success (viral suppression with CD4 ≥350 cells/μL), while 172 (49.7%; 95% CI 44.3–55.1) had concordant failure. Notably, 89 participants (25.7%; 95% CI 21.2–30.7) demonstrated discordant virological failure—persistent viraemia despite preserved CD4 counts (≥350 cells/μL). Conversely, 23 participants (6.6%; 95% CI 4.3–9.9) showed discordant immunological failure—inadequate CD4 recovery despite viral suppression. Among virologically suppressed participants (n = 85), 27.1% had incomplete immune recovery. Among viraemic participants (n = 261), 34.1% maintained preserved immunity (Figure 2).

Table 3.

Treatment response patterns (n = 346 with complete data).

Response category	Virological status	Immunological status	n (%)	95% CI
Concordant success	Suppressed	CD4 ≥350 cells/μL	62 (17.9)	14.0–22.4
Concordant failure	Viraemic	CD4 <350 cells/μL	172 (49.7)	44.3–55.1
Discordant immunological failure	Suppressed	CD4 <350 cells/μL	23 (6.6)	4.3–9.9
Discordant virological failure	Viraemic	CD4 ≥350 cells/μL	89 (25.7)	21.2–30.7

Among virologically suppressed (n = 85): 27.1% had incomplete immune recovery. Among viraemic (n = 261): 34.1% maintained preserved immunity.

Figure 2.

Treatment response patterns. Matrix representation of concordant and discordant immune-virological responses among 346 participants with complete data. Concordant success: viral suppression with CD4 ≥350 cells/μL. Concordant failure: viraemia with CD4 <350 cells/μL. Discordant immunological failure: viral suppression with incomplete immune recovery. Discordant virological failure: viraemia with preserved CD4 counts. Notably, 32.3% of participants exhibited discordant responses.

Exploratory mechanistic findings

Note: The following analyses were performed in small pilot subsets and should be interpreted as hypothesis-generating.

Drug resistance genotyping in 50 viraemic participants revealed that 43 (86%) harboured at least one resistance-associated mutation. NRTI mutations were common, with M184V/I detected in 31 (62%) and K65R in 14 (28%). NNRTI mutations included K103N in 27 (54%) and Y181C in 11 (22%). Multi-class resistance affecting ≥2 drug classes was present in 19 participants (38%). These mutation patterns are consistent with regional surveillance data from East and Southern Africa (15–18).

Immune activation analysis in 40 participants showed significantly elevated CD8+ T-cell activation (CD38+HLA-DR+ co-expression) in viraemic compared to suppressed participants (median 28.6% vs. 12.4%; p < 0.001). CD8 activation correlated inversely with CD4 count (Spearman ρ = −0.46; p < 0.01), consistent with the established role of chronic immune activation in CD4 depletion (9, 13, 14, 19).

Inflammatory biomarker analysis in 35 participants demonstrated elevated IL-6 (median 4.2 vs. 1.9 pg/mL; p = 0.002) and TNF-α (median 12.3 vs. 6.7 pg/mL; p = 0.01) in viraemic versus suppressed participants (Table 4). These findings align with evidence that persistent inflammation contributes to HIV-associated morbidity (10–12).

Table 4.

Exploratory mechanistic findings (pilot subsets).

Parameter	Suppressed	Viraemic	p-value
Drug resistance (n = 50 viraemic)
Any resistance mutation	—	43/50 (86%)	—
M184V/I (NRTI)	—	31/50 (62%)	—
K103N (NNRTI)	—	27/50 (54%)	—
Multi-class resistance	—	19/50 (38%)	—
Immune activation (n = 40)	n = 15	n = 25
CD8+ CD38+ HLA-DR+ %, median (IQR)	12.4 (8.7–17.2)	28.6 (20.1–35.4)	<0.001
Correlation with CD4 (Spearman ρ)		−0.46	<0.01
Inflammatory biomarkers (n = 35)	n = 12	n = 23
IL-6 (pg/mL), median (IQR)	1.9 (1.2–2.8)	4.2 (2.9–6.8)	0.002
TNF-α (pg/mL), median (IQR)	6.7 (4.2–9.1)	12.3 (8.4–18.6)	0.01

NRTI, nucleoside reverse transcriptase inhibitor; NNRTI, non-nucleoside reverse transcriptase inhibitor; IQR, interquartile range. Exploratory analyses in small pilot subsets; findings require validation.

Factors associated with viral non-suppression

In multivariable analysis (n = 346), lower CD4 count was strongly associated with viral non-suppression (adjusted OR 0.53 per 100 cells/μL increase; 95% CI 0.43–0.66; p < 0.001). Age, sex, duration on ART, pregnancy status, and region were not significantly associated with non-suppression after adjustment (Table 5).

Table 5.

Multivariable logistic regression for viral non-suppression (n = 346).

Variable	Adjusted OR	95% CI	p-value
CD4 count (per 100 cells/μL increase)	0.53	0.43–0.66	<0.001
Age (per 10 years)	1.12	0.89–1.41	0.34
Sex (male vs. female)	0.87	0.52–1.46	0.60
Duration on ART (per year)	1.03	0.96–1.11	0.42
Pregnancy (yes vs. no)	1.28	0.68–2.41	0.44
Region (ref: Southern)
Central	1.34	0.71–2.53	0.37
Northern	1.52	0.79–2.92	0.21

Model includes site fixed effects; AIC = 389.2. OR, odds ratio; ART, antiretroviral therapy.

Discussion

This multi-centre study reveals substantial gaps in HIV treatment outcomes at Nigerian tertiary care facilities, with viral suppression rates (26.9%) falling far below the UNAIDS 95% target and recent national programmatic estimates (6). The high prevalence of severe immunodeficiency (38.5%), discordant immune-virological responses (32.3%), and preliminary evidence of widespread drug resistance and immune activation suggest multiple interacting pathways to treatment failure in this population (25–29).

Our suppression rates contrast with recent Nigerian data showing national suppression approaching 95% (6). This discrepancy likely reflects the tertiary care setting, which concentrates patients with treatment challenges, suspected failure, and complex clinical presentations. Similar facility-level heterogeneity has been documented elsewhere in sub-Saharan Africa (20–22).

The finding that 25.7% of participants had viraemia despite preserved CD4 counts is clinically significant. This discordant pattern may represent early treatment failure before immunological decline, host factors conferring immune resilience, or measurement timing effects. Regardless of mechanism, these individuals remain at risk of disease progression and onward transmission, highlighting the importance of viral load monitoring over CD4-only approaches (20, 23).

The 86% prevalence of drug resistance mutations in our pilot subset, while requiring confirmation in larger studies, aligns with regional data showing high acquired resistance in failing patients (15–18). The predominance of M184V and K103N reflects selection by commonly used NRTI/NNRTI-based regimens and underscores the importance of resistance testing to guide regimen switches (16, 23, 24).

Elevated immune activation (CD38+HLA-DR+CD8+ T cells) and inflammatory biomarkers (IL-6, TNF-α) in viraemic participants are consistent with established pathophysiological mechanisms linking chronic immune activation to CD4 depletion and disease progression (9, 13, 14, 19). The inverse correlation between CD8 activation and CD4 count supports a potential mechanistic link, though causality cannot be established from cross-sectional data (30–35).

Limitations

Several limitations warrant consideration. First, the cross-sectional design precludes causal inference and assessment of temporal relationships. Second, the tertiary care setting limits generalisability to primary care populations where treatment outcomes may differ. Third, mechanistic analyses were performed in small convenience subsets (n = 35–50), limiting statistical power and generalisability; these findings require validation in larger, systematically sampled populations. Fourth, adherence was not directly measured, precluding assessment of its contribution to treatment failure. Fifth, missing viral load (20.3%) and CD4 (25.1%) data may introduce selection bias. Sixth, single time-point sampling cannot distinguish persistent from transient viraemia. Despite these limitations, our findings provide important insights into treatment outcomes and potential failure mechanisms in a high-burden setting.

Conclusion

HIV treatment outcomes at Nigerian tertiary care facilities fall substantially short of global targets, with high rates of viral non-suppression, severe immunodeficiency, and discordant immune-virological responses. Preliminary mechanistic data suggest contributions from drug resistance, immune activation, and inflammation. These findings highlight the need for enhanced viral load monitoring, resistance testing, and research into interventions targeting immune activation. Larger, longitudinal studies with systematic sampling are warranted to validate these mechanistic associations and inform targeted interventions to improve treatment outcomes in this high-burden setting.

Funding Statement

The author(s) declared that financial support was not received for this work and/or its publication.

Data availability statement

The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author.

Ethics statement

The studies involving humans were approved by Ethical approval was obtained from Universitätsklinikum Freiburg Ethics Committee (NO/140/19). The studies were conducted in accordance with the local legislation and institutional requirements. Written informed consent for participation in this study was provided by the participants’ legal guardians/next of kin.

Author contributions

PM: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Project administration, Resources, Software, Supervision, Visualization, Writing – original draft, Writing – review & editing. AK: Conceptualization, Investigation, Methodology, Project administration, Resources, Supervision, Validation, Writing – original draft, Writing – review & editing. SO: Data curation, Investigation, Methodology, Project administration, Supervision, Writing – original draft, Writing – review & editing. PA: Conceptualization, Data curation, Investigation, Methodology, Resources, Writing – original draft, Writing – review & editing. AO: Data curation, Investigation, Methodology, Resources, Writing – original draft, Writing – review & editing. DB: Data curation, Formal analysis, Methodology, Project administration, Validation, Visualization, Writing – original draft, Writing – review & editing. CA: Data curation, Investigation, Methodology, Project administration, Resources, Writing – original draft, Writing – review & editing.

Conflict of interest

The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Correction note

This article has been corrected with minor changes. These changes do not impact the scientific content of the article.

Generative AI statement

The author(s) declared that Generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.