Prevalence and determinants of unsuppressed HIV viral loads among children and adolescents living with HIV on antiretroviral therapy in Lubumbashi, Democratic Republic of the Congo: a retrospective cross-sectional study

Abstract

Background

Despite global improvements in antiretroviral therapy (ART) access for children and adolescents living with HIV (CALHIV), a significant proportion continue to experience unsuppressed viral load (USVL). Limited studies focus on the factors contributing to USVL among CALHIV in the Democratic Republic of the Congo (DRC), especially in the context of evolving treatment landscapes. Understanding these determinants is crucial for enhancing ART outcomes.

Objective

This study aimed to determine the prevalence of USVL and identify factors associated with USVL among CALHIV receiving ART in Lubumbashi, DRC.

Design

A multicentre retrospective cross-sectional study was conducted. Data were gathered using an observational checklist based on assessing patient file data and entered into Microsoft Excel. Analysis was performed using STATA V.16. Variables with a p value of 0.20 from the bivariable analysis were included in a multivariable logistic regression model, and significant variables (p<0.05) were retained in the final model.

Setting and participants

The study was conducted at 21 HIV care clinics in Lubumbashi from June to September 2024. It included 847 CALHIV aged 0–19 years who had been on ART for at least 6 months and had at least one available VL result.

Primary outcome measure

The rate of USVL among CALHIV, defined as achieving a VL below 1000 copies/mL, in those who had been on ART for at least 6 months.

Results

The prevalence of USVL among CALHIV was 24.68% (209/847; 95% CI: 21.89% to 27.69%). Multivariable logistic regression analysis revealed that CALHIV with married caregivers were more likely to have USVL (adjsuted OR, aOR=2.4; 95% CI: 1.2 to 5.0). Other factors associated with USVL included horizontal HIV transmission (aOR=2.3; 95% CI: 1.0 to 5.2), advanced WHO clinical stages (aOR=3.5; 95% CI: 1.0 to 13.7), poor/fair ART adherence (aOR=107.8; 95% CI: 50.3 to 231.1) and ART-induced side effects (aOR=3.8; 95% CI: 1.9 to 7.9).

Conclusions

The high rate of USVL among CALHIV in Lubumbashi highlights the need to strengthen ART adherence support, manage treatment side effects and improve early diagnosis and follow-up, particularly for those infected through horizontal transmission or presenting with advanced clinical stages. Special attention should also be given to caregiver-related factors, including marital status, which may influence treatment outcomes.

STRENGTHS AND LIMITATIONS OF THIS STUDY.

• Use of standardised questionnaires: The use of a validated questionnaire ensured the consistency and comparability of responses, reinforcing the reliability of the results obtained.

• Representative children and adolescents living with HIV (CALHIV) sample: The study focuses on a vulnerable group, CALHIV, providing valuable information on a subgroup often under-represented in research, particularly in sub-Saharan Africa.

• Cross-sectional design limitation: The study uses a cross-sectional design, which limits the ability to establish causality between the identified determinants and the outcome, unsuppressed viral loads.

• Geographical limitation: The study’s focus on Lubumbashi restricts the generalisability of the findings. Differences in healthcare systems, socioeconomic factors and cultural contexts across other regions of the Democratic Republic of the Congo or sub-Saharan Africa may yield different results.

• Lack of qualitative exploration: The study did not account for factors like mental health, psychosocial support or economic status, nor did it include a qualitative component to explore social, psychological and cultural barriers such as stigma and healthcare access, which could have impacted antiretroviral therapy adherence and viral load suppression.

Introduction

Over the past few decades, HIV/AIDS has emerged as one of the most severe global health challenges, particularly impacting sub-Saharan Africa (SSA), where the disease burden is the greatest.¹ The widespread introduction of antiretroviral therapy (ART) has dramatically shifted the outlook, transforming HIV from a fatal illness into a manageable chronic condition. The main goal of ART is to suppress viral load (VL) to undetectable levels, which not only enhances the health of people living with HIV (PLHIV) but also significantly reduces the potential for transmission.² However, despite substantial progress in expanding ART access, ensuring consistent VL suppression (VLS) remains a challenge, particularly among children and adolescents living with HIV (CALHIV) in SSA.^{3 4}

The UNAIDS 95-95-95 initiative aims to end the HIV/AIDS epidemic by ensuring that 95% of PLHIV know their status, 95% of those diagnosed are on ART, and 95% of those on treatment achieve VLS.⁴ Although many PLHIV can suppress their VL within 6 months of starting ART, CALHIV experience lower rates of VLS due to various obstacles in adhering to treatment.^{5 6} In SSA, where about 85% of all CALHIV live, the inability to meet the 95-95-95 goals will likely result in continued HIV-related deaths and new infections for many years.^{7 8}

Unsuppressed VL (USVL) among CALHIV in SSA remains a major public health concern. In 2023, an estimated 76 000 (53 000–110 000) children died from HIV-related causes globally, with the highest burden among those under 10 years of age—often due to insufficient access to HIV prevention, testing and ART.⁸,⁹ Death rates are highest during infancy, largely associated with challenges in early infant diagnosis and late HIV diagnosis.¹⁰ Attaining VLS in CALHIV is crucial not only for individual health outcomes but also for curbing HIV transmission and advancing towards the UNAIDS targets. The WHO has emphasised the necessity of providing comprehensive counselling on ART adherence for individuals with USVL, highlighting the ongoing need for adherence support, particularly during adolescence.³

Studies indicate that CALHIV face higher rates of virological failure and poorer treatment outcomes compared with adults, largely due to developmental changes, stigma and discrimination.¹¹ Factors such as the burden of daily pill-taking, fear of stigma, lack of psychosocial support and systemic barriers, including limited access to healthcare services, contribute to these challenges globally, with these issues often being more pronounced in resource-limited settings. For example, research from Tanzania has shown that perceived stigma and fear of unintentional disclosure of HIV status significantly hinder achieving suppressed VL among CALHIV.^{12 13} Other factors, including orphan status, male gender, poor nutritional status, poor adherence to treatment, advanced WHO stage, low CD4+T cell counts, presence of opportunistic infections, nevirapine-based treatment and drug substitution, were also associated with USVL.613,15

Globally, VLS among CALHIV remains suboptimal. According to UNAIDS,⁸ only 84% of children aged 0–14 years on ART achieve VLS, falling short of the final 95% target of the global 95–95–95 goals. Additionally, a large cohort study found that 24% (1354/5641) of CALHIV, and 10% (6517/64487) of adults in follow-up with available VL data, had USVLs 3 years after initiating ART.¹¹ In SSA, where the majority of CALHIV reside, these gaps are particularly pronounced. In Lubumbashi, Democratic Republic of the Congo (DRC), data on VLS among CALHIV are scarce, but studies from SSA report USVL rates ranging from 15.89% in Tanzania⁵ to 51.28% in Nigeria,¹⁶ highlighting ongoing challenges in achieving treatment success. In Lubumbashi, achieving VLS among CALHIV is further hindered by structural barriers such as inadequate healthcare infrastructure, stigma, and inconsistent ART adherence.¹⁷ While existing studies have explored factors contributing to USVL in SSA, there is a critical gap in localised data specific to the DRC. This study aims to fill this gap by estimating the prevalence of USVL and identifying factors associated with VLS among CALHIV receiving ART in Lubumbashi. The findings will inform the development of targeted strategies to improve HIV care, treatment outcomes and reduce HIV-related morbidity and mortality in this vulnerable population.

Materials and methods

Study design, period and setting

A hospital-based retrospective cross-sectional study was conducted at 21 HIV care clinics in the eleven health zones in Lubumbashi city (figure 1), specifically focusing on HIV care clinics for CALHIV in the area. The study period covered 5 years, from January 2018 to December 2023, and data were collected from June to September 2024.

Figure 1. Maps of Africa (A), the Democratic Republic of the Congo (B), and Lubumbashi city (C). The star represents Lubumbashi city.

The DRC’s second-largest city, Lubumbashi, is situated in the southeastern part of the country. With an estimated population of approximately 2 933 962 as of 2024, spread across 747 km², the city has a population density exceeding 3927 individuals per km².¹⁸

Lubumbashi City is divided into 11 health zones, with all health zones except one having a General Referral Hospital. Each health zone has an average of 15 health centres, resulting in over 350 healthcare facilities (hospitals, polyclinics and health centres) in the city, with 70% of them situated in the densely populated urban core. The private sector comprises more than 60% of the healthcare facilities. Among these facilities, approximately 50 healthcare facilities are offering HIV care in Lubumbashi.

In Lubumbashi city, HIV service activities in the HIV care clinics are primarily funded by the US President’s Emergency Plan for AIDS Relief (PEPFAR), with coordination overseen by the National HIV/AIDS Programme. However, recent shifts in PEPFAR’s strategic priorities—including an increased focus on local ownership, health system strengthening and sustainability—may influence the future delivery and organisation of these services in the region.¹⁹ This restricted definition excludes several vulnerable groups—including children and adolescents—from benefiting from expanded prevention and support packages, such as access to pre-exposure prophylaxis, comprehensive sexual and reproductive health services, mental health and psychosocial support, and targeted interventions for key populations.¹⁹

Population and eligibility criteria

The study included CALHIV aged 0–19 years who were receiving ART at specific HIV care clinics, with a minimum follow-up period of 6 months and at least one available VL result at the clinic (figure 2).

Figure 2. Flow chart showing a selection of HIV-positive children and adolescents on ART in 21 HIV care clinics in Lubumbashi, DRC. ART, antiretroviral therapy; DRC, Democratic Republic of the Congo.

The following participants were excluded from the study: those whose records were inaccessible or had incomplete data; those who were lost to follow-up; those not receiving ART; those with a follow-up period of less than 6 months; and those with missing VL results during the study period.

The study involved a review of all medical records of the study population. We used a simple random sampling technique to select participants. The sample size was calculated using the following formula:

$${\displaystyle n=\frac{z^2.\mathrm{p}\left(1-\mathrm{p}\right)}{e^2}}$$

Where: n represents the minimum sample size, z is the CI (1.96), p was set at 0.5 (due to unknown prevalence of USVL in CALHIV in the DRC) and e represents the margin of error (0.05). Using this calculation, we determined a minimum sample size of 382 participants. To account for potential missing data in medical records, an additional 10% was added, resulting in a total study sample size of 420 participants. The sample size calculation based on the determinants of USVL is detailed in online supplemental table S1.

The study involved a review of all medical records of CALHIV who met the inclusion criteria (using the census method), which were sourced from the VL registration books of the HIV care clinics. To ensure the presence of all required data for the study, all medical records were thoroughly examined.

In our study, a total of 847 participants met the inclusion criteria during the recruitment process, as detailed in figure 2. This sample is considered representative of the broader population, as it includes patients from 21 different health facilities, ensuring a diverse and comprehensive representation.

Data collection and quality control

All HIV care clinics across the 11 health zones in Lubumbashi used standardised clinical files and registers. Data were retrospectively collected using a Microsoft Excel datasheet specifically developed for this study. 10 qualified ART providers, trained during a 3-day session, gathered data from patient intake forms, ART monitoring books, follow-up charts and clinic registers. The training covered the study’s objectives, proper use of institutional recording tools and detailed instructions on completing the data extraction form.

Data collection was guided by a pretested checklist and focused on key indicators related to VL testing and routine HIV care. The principal investigator (OM) supervised the entire process through regular clinic visits and daily oversight, ensuring data completeness, accuracy and consistency. Any errors identified during the review were promptly corrected. Before entering the data into Excel, all forms were checked for missing values, duplicate entries and overall quality. To maintain confidentiality, each form was anonymised and identified solely by a unique study code.

Study variables and operational definitions

The final dataset comprised deidentified data for each participant, including information on VL tests and results, sociodemographic and clinical characteristics, and treatment outcomes.

The primary outcome variable was VLS, categorised as either non-suppressed or suppressed status.

The independent variables were as follows:

• Sociodemographic characteristics included variables such as sex, age at the last VL test, age at ART initiation, education level, caregiver relations and status, vital status of mother and father, HIV status of caregivers, disclosure status (disclosed HIV status to the child or other relatives), orphan status, caregiver’s age, caregiver’s sex, caregiver’s occupation, caregiver’s HIV status, caregiver alcohol use and caregiver’s marital status.

• Clinical characteristics, such as HIV infection route, nutritional status, WHO clinical stage and history of tuberculosis in the last 6 months.

• Medication-related characteristics included the current ART regimen type, history of regimen change, duration of ART, side effects of medications, history of ART interruption since ART initiation, ART-induced side effects, cotrimoxazole prophylactic treatment at the start of ART, isoniazid prophylaxis use and level of adherence.

VLS was defined according to the guidelines of the DRC National HIV/AIDS Programme²⁰ and the WHO,^{21 22} wherein patients achieving a VL of less than 1000 copies/mL were considered virally suppressed. Following the guidelines in the DRC, the first VL test is recommended to be conducted 6 months after the initiation of ART.

Adherence to ART was assessed based on the percentage of pills reportedly taken over the preceding month compared with the prescribed amount, following WHO criteria, as recorded in the patients’ medical files. All adherence information was extracted retrospectively from clinical records. For adolescents, these records were primarily based on self-reported adherence documented during clinical visits, whereas for younger children, adherence was reported by caregivers and noted by the healthcare provider. CALHIV were classified as poor/fair adherents if they reportedly took less than 95% of their prescribed doses, and as good adherents if their reported adherence was above 95%.²³

Regimen change was defined as any modification to the initial ART drugs prescribed to the patient.²²

Statistical analysis

The data were cleaned, coded and reviewed for completeness after the completion of data collection. Subsequently, we entered data into Microsoft Excel version 2019 and then uploaded them to STATA software V.16 for analysis.

To describe the independent variables of the study population, descriptive analysis was used. The proportion of USVL was defined as the percentage of the total number of individuals with USVL among the total number of VL tested. Further assessment of the prevalence of USVL in CALHIV was based on sociodemographic, clinical and therapeutic factors.

The VL results were classified into two categories: <1000 copies/mL was considered ‘suppressed’, and ≥1000 copies/mL was considered ‘unsuppressed’. Bivariable analysis was used to explore the association between the independent variables and the outcome variable. To identify factors independently associated with virological non-suppression, crude ORs and 95% CIs were computed. Variables with a p<0.20 in the bivariable analysis were included in the multivariable logistic regression model, following the commonly recommended approach to avoid prematurely excluding potentially relevant variables. Adjusted ORs (aORs) and 95% CIs were then calculated, and variables with a p<0.05 were considered to have a statistically significant association with the outcome variable. Multicollinearity among independent variables was assessed using the variance inflation factor (VIF). A VIF threshold of >5 was considered indicative of potential collinearity. For model performance and assumption checks, we used several diagnostic tools. Sensitivity, specificity and predictive values were evaluated to assess the model’s performance in classifying observed outcomes. Discrimination capacity was evaluated via the receiver Operating Characteristic (ROC) curve and its associated area under the curve (AUC). The Hosmer-Lemeshow test was applied to examine model calibration, and the link test assessed model specification. Overall model fit was evaluated using the pseudo-R², log likelihood and likelihood ratio χ² statistics.

Patient and public involvement

There was no direct interaction with patients in this study and no direct patient involvement in the design or conduct of this study.

Results

Figure 2 shows the selection process of CALHIV on ART across 21 HIV care clinics in Lubumbashi, DRC. Out of 9789 PLHIV enrolled in these clinics, 1648 were under the age of 20. Among them, 847 were included in the study, while 801 were excluded for the following reasons: 86 had died, 462 were lost to follow-up, 161 were transferred to other facilities, 28 had incomplete data and 64 had been on ART for less than 6 months.

Tables1 2 present the descriptive statistics, bivariable and multivariable analyses of the sociodemographic, clinical and therapeutic characteristics of CALHIV on ART in Lubumbashi.

Table 1. Sociodemographic characteristics of children and adolescents living with HIV on antiretroviral therapy and their caregivers attending HIV care clinics in Lubumbashi, by viral load suppression status.

Variable	Total (N=847), n (%)	HIV viral load		cOR (95% CI)	aOR (95% CI)	P value
		Unsuppressed (n=209), n (%)	Suppressed (n=638), n (%)
CALHIV’s age*
<5 years	74 (8.7)	6 (8.1)	68 (91.9)	Reference	Reference
5–9 years	244 (28.8)	41 (16.8)	203 (83.2)	2.3 (0.9 to 5.6)	0.6 (0.1 to 4.1)	0.614
10–14 years	232 (27.4)	59 (25.4)	173 (74.6)	3.9 (1.6 to 9.4)	0.8 (0.1 to 6.2)	0.862
15–19 years	297 (35.1)	103 (34.7)	194 (65.3)	6.0 (2.5 to 14.3)	0.7 (0.1 to 6.1)	0.783
CALHIV’s sex
Male	341 (40.3)	83 (24.3)	258 (75.7)	Reference
Female	506 (59.7)	126 (24.9)	380 (75.0)	1.0 (0.7 to 1.4)
CALHIV’s educational level*
Never been to school	127 (15.0)	17 (13.4)	110 (86.6)	Reference	Reference
Primary	396 (46.7)	90 (22.7)	306 (77.3)	1.9 (1.1 to 3.3)	4.0 (0.9 to 14.4)	0.055
Secondary	286 (33.8)	88 (30.8)	198 (69.2)	2.9 (1.6 to 5.1)	2.6 (0.7 to 9.7)	0.162
Higher/university	38 (4.5)	14 (36.8)	24 (63.2)	3.8 (1.6 to 8.7)	1.0 (0.1 to 7.9)	0.971
Currently living with/in*
At least one biological parent	592 (69.9)	128 (21.6)	464 (78.4)	Reference	Reference
A non-parent family member	191 (22.6)	62 (32.5)	129 (67.5)	1.7 (1.2 to 2.5)	2.1 (0.7 to 6.3)	0.187
Adoptive parents	50 (5.9)	12 (24.0)	38 (76.0)	1.1 (0.6 to 2.3)	0.5 (0.1 to 3.3)	0.473
Group housing	14 (1.6)	7 (50.0)	7 (50.0)	3.6 (1.2 to 10.5)	1.7 (0.1 to 31.4)	0.735
Orphan status*
Non-orphan	390 (46.0)	81 (20.8)	309 (79.2)	Reference	Reference
At least one biological parent died	330 (39.0)	85 (25.8)	245 (74.2)	1.3 (0.9 to 1.9)	0.7 (0.3 to 1.5)	0.376
Double orphan	127 (15.0)	43 (33.9)	84 (66.1)	2.0 (1.3 to 3.0)	0.6 (0.1 to 2.4)	0.449
Caregiver’s age
<40 years	454 (53.6)	95 (20.9)	359 (79.1)	Reference	Reference
≥40 years	393 (46.4)	114 (29.0)	279 (71.0)	1.5 (1.1 to 2.1)	0.7 (0.3 to 1.5)	0.327
Caregiver’s sex
Male	158 (18.6)	53 (33.5)	105 (66.5)	1.7 (1.2 to 2.5)	2.0 (0.8 to 4.9)	0.132
Female	689 (81.4)	156 (22.6)	533 (77.4)	Reference	Reference
Caregiver’s occupation
Unemployed	435 (51.4)	96 (22.1)	339 (77.9)	Reference
Employed	412 (48.6)	113 (27.4)	299 (72.6)	1.3 (0.9 to 1.8)
Caregiver’s HIV status*
Positive	497 (58.7)	102 (20.5)	395 (79.5)	Reference	Reference
Negative	209 (24.7)	68 (32.5)	141 (67.5)	1.9 (1.3 to 2.7)	0.6 (0.2 to 1.9)	0.403
Unknown	141 (16.7)	39 (27.7)	102 (72.3)	1.5 (0.9 to 2.3)	0.9 (0.3 to 3.0)	0.900
Caregiver alcohol use
No	406 (47.9)	95 (23.4)	311 (76.6)	Reference
Yes	441 (52.1)	114 (25.9)	327 (74.1)	1.1 (0.8 to 1.6)
Caregiver’s marital status
Married	514 (60.7)	139 (27.0)	375 (73.0)	1.4 (1.0 to 1.9)	2.4 (1.2 to 5.0)	0.019
Single/divorced/widowed	333 (39.3)	70 (21.0)	263 (79.0)	Reference	Reference

Variables with an overall p<0.2 in the bivariable χ2 analysis were included in the multivariable logistic regression model. Based on this criterion, CALHIV’s sex, cotrimoxazole prophylaxis, caregiver’s occupation and caregiver alcohol use were excluded from the multivariable logistic regression model.

^*Overall p<0.2 for categorical variables with more than two modalities.

%, percentage; aOR, adjusted OR; ART, antiretroviral therapy; CALHIV, children and adolescents living with HIV; cOR, crude OR; n, number.

Table 2. Clinical and treatment features of children and adolescents living with HIV on antiretroviral therapy attending HIV care clinics in Lubumbashi, by viral load suppression status.

Variable	Total (N=847), n (%)	HIV viral load		cOR (95% CI)	aOR (95% CI)	P value
		Unsuppressed (n=209), n (%)	Suppressed (n=638), n (%)
HIV infection route
Vertical (perinatal)	589 (69.5)	120 (20.4)	469 (79.6)	Reference	Reference
Horizontal (transfusion, sexual, unknown)	258 (30.5)	89 (34.5)	169 (65.5)	2.1 (1.5 to 2.9)	2.3 (1.0 to 5.2)	0.042
HIV status disclosure to CALHIV
Yes	322 (38.0)	101 (31.4)	221 (68.6)	1.8 (1.3 to 2.4)	1.4 (0.6 to 3.4)	0.446
No	525 (62.0)	108 (20.6)	417 (79.4)	Reference	Reference
HIV status disclosure to others
Yes	642 (75.8)	151 (23.5)	491 (76.5)	Reference	Reference
No	205 (24.2)	58 (28.3)	147 (71.7)	1.3 (0.9 to 1.8)	1.3 (0.6 to 2.8)	0.466
WHO clinical stage
1–2	738 (87.1)	121 (16.4)	617 (83.6)	Reference	Reference
3–4	109 (12.9)	88 (80.7)	21 (19.3)	21.4 (12.8 to 35.7)	3.5 (1.0 to 13.7)	0.048
Nutritional status
Normal	761 (89.9)	143 (18.8)	618 (81.2)	Reference	Reference
Moderate/severe malnutrition	86 (10.2)	66 (76.7)	20 (23.3)	14.3 (8.4 to 24.3)	2.0 (0.7 to 6.1)	0.202
History of TB in the last 6 months
Yes	129 (15.2)	98 (76.0)	31 (24.0)	17.3 (11.0 to 27.1)	0.9 (0.3 to 2.8)	0.819
No	718 (84.8)	111 (15.5)	607 (84.5)	Reference	Reference
Cotrimoxazole prophylaxis
Yes	842 (99.4)	207 (24.6)	635 (75.4)	Reference
No	5 (0.6)	2 (40.0)	3 (60.0)	2.0 (0.2 to 18.0)
Isoniazid prophylaxis
Yes	787 (92.9)	175 (22.2)	612 (77.8)	Reference	Reference
No	60 (7.1)	34 (56.7)	26 (43.3)	4.6 (2.7 to 7.8)	1.8 (0.6 to 5.8)	0.290
History of ART interruption since ART initiation
Yes	136 (16.1)	55 (40.4)	81 (59.6)	2.5 (1.7 to 3.6)	1.3 (0.6 to 3.0)	0.551
No	711 (83.9)	154 (21.7)	557 (78.3)	Reference
ART-induced side effects
Yes	228 (26.9)	147 (64.5)	81 (35.5)	16.3 (11.2 to 23.8)	3.8 (1.9 to 7.9)	<0.001
No	619 (73.1)	62 (10.0)	557 (90.0)	Reference	Reference
ART adherence
Good	613 (72.4)	19 (3.1)	594 (96.9)	Reference	Reference
Poor/Fair	234 (27.6)	190 (81.2)	44 (18.8)	135.0 (76.9 to 236.9)	107.8 (50.3 to 231.1)	<0.001
Current ART regimen*
Dolutegravir-based regimen	795 (93.9)	184 (23.1)	611 (76.9)	Reference	Reference
Nevirapine-based regimen	21 (2.5)	12 (57.1)	9 (42.9)	4.4 (1.8 to 10.7)	1.2 (0.2 to 5.9)	0.833
Efavirenz-based regimen	31 (3.7)	13 (41.9)	18 (58.1)	2.4 (1.2 to 5.0)	0.4 (0.1 to 1.3)	0.116
ART duration*
6–24 months	176 (20.8)	50 (28.4)	126 (71.6)	1.4 (0.9 to 2.1)	2.0 (0.8 to 4.8)	0.132
25–48 months	161 (19.0)	48 (29.8)	113 (70.2)	1.5 (1.0 to 2.3)	2.3 (0.9 to 5.5)	0.059
>48 months	510 (60.2)	111 (21.8)	399 (78.2)	Reference	Reference

Variables with an overall p<0.2 in the bivariable χ2 analysis were included in the multivariable logistic regression model. Based on this criterion, CALHIV’s sex, cotrimoxazole prophylaxis, caregiver’s occupation and caregiver alcohol use were excluded from the multivariable logistic regression model.

^*Overall p<0.2 for categorical variables with more than two modalities.

%, percentage; aOR, adjusted OR; ART, antiretroviral therapy; CALHIV, children and adolescents living with HIV; cOR, crude OR; n, number; TB, tuberculosis.

A total of 847 CALHIV included, with a mean age of 11.6±5.0 years (range: 2 and 19 years); most were adolescents aged 15–19 years (35.1%) and female (59.7%). Nearly half (46.7%) had primary education, and most lived with a biological parent (69.9%). About 15% were double orphans. Most caregivers were female (81.4%), with a mean age of 39.0±9.9 years (range: 20 and 75 years). 51.4% were unemployed, 58.7% were living with HIV, and 52.1% reported alcohol use. Most CALHIV acquired HIV perinatally (69.5%) and had not been informed of their status (62.0%), while 24.2% had disclosed it to others. WHO clinical stage 1 was predominant (62.0%), and 89.9% had normal nutritional status. Tuberculosis history in the last 6 months was reported by 15.2%. Most received cotrimoxazole (99.4%) and isoniazid prophylaxis (92.9%). ART side effects affected 26.9%, and 27.6% had poor/fair adherence. Most were on a dolutegravir-based regimen (93.9%) and had been on ART for >48 months (60.2%).

The prevalence of USVL was 24.68% (209 out of 847 CALHIV; 95% CI: 21.89% to 27.69%).

In bivariable analysis, older age (10–14 and 15–19 years), primary/secondary/higher educational levels, orphanhood, not living with at least one biological parent, having a male, not living with HIV or married caregiver were significantly associated with USVL. Clinical factors linked to USVL included horizontal transmission, early HIV status disclosure, advanced WHO clinical stage (3–4), malnutrition, TB history, absence of isoniazid prophylaxis, ART interruptions, side effects, poor or fair adherence, use of Nevirapine/Efavirenz-based regimens and shorter ART duration (<48 months). In contrast, CALHIV’s sex, caregiver’s occupation, caregiver alcohol use and cotrimoxazole prophylaxis were not significantly associated with USVL. Variables with a p<0.2 in the bivariable analysis were included in the multivariable logistic regression model, except for the four non-significant ones above.

The multiple logistic regression analysis identified several significant factors associated with USVL among CALHIV. Specifically, CALHIV whose caregivers are married are almost three times more likely to have USVL compared with those whose caregivers are single, divorced or widowed (aOR=2.4; 95% CI: 1.2 to 5.0; p=0.019). CALHIV infected through horizontal transmission (eg, transfusion, sexual contact or unknown route) are more than twice as likely to have USVL compared with those infected perinatally (aOR=2.3; 95% CI: 1.0 to 5.2; p=0.042). Those in WHO clinical stages 3 or 4 are 3.5 times more likely to have USVL compared with those in stages 1 or 2 (aOR=3.5; 95% CI: 1.0 to 13.7; p=0.048). CALHIV who experience ART-induced side effects are nearly four times more likely to have USVL compared with those not reporting such side effects (aOR=3.8; 95% CI: 1.9 to 7.9; p<0.001). Finally, poor or fair ART adherence is the strongest predictor of USVL. CALHIV with poor/fair adherence are more than 100 times more likely to have USVL compared with those with good adherence (aOR=107.8; 95% CI: 50.3 to 231.1; p<0.001).

To assess the robustness and validity of our multivariable logistic regression model, we conducted a series of diagnostic evaluations. The model demonstrated excellent predictive performance, with a sensitivity of 86.12%, specificity of 95.61%, positive predictive value of 86.54%, and negative predictive value of 95.46%, resulting in an overall correct classification rate of 93.27% (online supplemental tables S2, S3 and figure S1). The ROC curve analysis yielded an AUC of 0.9699, indicating excellent discriminative ability (online supplemental figure S2). The Hosmer-Lemeshow goodness-of-fit test showed a non-significant result (χ²=3.01, df=8, p=0.9340), suggesting good model calibration. Additionally, the link test confirmed proper model specification, with a significant _hat term (p<0.001) and a non-significant _hatsq term (p=0.391), indicating no evidence of misspecification (online supplemental table S4). The model also exhibited a strong overall fit, with a pseudo-R² of 0.6745 and a highly significant likelihood ratio χ² (χ²=638.39, df=2, p<0.001), explaining a substantial proportion of the variation in VLS among CALHIV. Multicollinearity diagnostics showed no concerning levels of collinearity among the independent variables, with a mean VIF of 2.69—well below the commonly accepted threshold of 10 (online supplemental table S5).

Discussion

This study sheds light on the factors driving USVL among CALHIV in Lubumbashi, DRC. Despite significant progress in expanding ART coverage and the global push to meet the UNAIDS 95-95-95 goals, our findings reveal that a substantial portion of this vulnerable population continues to struggle with USVL, underscoring critical gaps in treatment effectiveness.

In our study, the prevalence of USVL among CALHIV was 24.68%, indicating that nearly one in four CALHIV were not virally suppressed at the time of data collection. This mirrors trends observed in other regions of SSA, where achieving sustained VLS remains a significant hurdle due to socioeconomic and structural barriers to optimal ART adherence. When compared with other studies across SSA, the prevalence rates of USVL vary. In Tanzania, Khamadi et al⁵ reported the lowest prevalence of 15.84% in the southern region, followed by Mchomvu et al¹⁵ with a prevalence of 32.80% in the Tabora region, and Bitwale et al¹³ who observed a slightly higher rate of 34% in Dodoma Municipality. In Ethiopia, Abera et al²⁴ found a prevalence of 17.28% in East Shewa hospitals. In Kenya, Tsikhutsu et al²⁵ reported a prevalence of 19.79% in the South Rift Valley and Kisumu regions. More concerning figures were observed in Cameroon, where Dobseu Soudebto et al²⁶ found a prevalence of 45.59% in the Central Region, and in Nigeria, where Isaac et al¹⁶ reported a staggering 51.28%. These findings highlight an urgent need for targeted interventions to improve ART adherence and VLS among CALHIV. Bridging the gap between ART availability and treatment success requires a nuanced understanding of the socioeconomic, cultural and health system barriers that hinder progress. Without addressing these root causes, achieving sustainable VLS will remain elusive for many CALHIV across the region.

Multivariable logistic regression analysis indicated that five factors associated with USVL were identified, including caregiver marital status, non-perinatal (horizontal) HIV transmission routes, advanced HIV disease stage, poor ART adherence and ART-related side effects. Adherence to ART is key for VLS and long-term treatment success in CALHIV. While caregivers play a central role in early childhood adherence, their ongoing support remains vital throughout adolescence due to persistent developmental and psychosocial challenges.²⁷ Our study shows that CALHIV with married caregivers had higher odds of USVL compared with those with single, divorced or widowed caregivers. This may be due to divided responsibilities in two-caregiver households, leading to miscommunication and inconsistent adherence. Marital stress, financial strain and managing multiple children may also reduce adherence oversight.²⁸ These findings contrast with previous studies that reported poorer outcomes among CALHIV with single or separated caregivers.^{29 30} This highlights the complexity of family dynamics. In some settings, single caregivers may develop structured routines that enhance adherence, while shared responsibility in dual-caregiver households may cause lapses.³¹ Moreover, social support—regardless of marital status—can influence adherence. Extended family, community networks, and health services strengthen caregivers’ ability to manage ART.³² Interventions should, therefore, target caregiver support through counselling, mental healthcare and community-based programmes to improve ART outcomes.³²

The significantly higher odds of USVL in CALHIV infected through horizontal routes—defined in our study as acquisition via blood transfusion, sexual contact or unknown means—compared with those with vertical infection, can be explained by several factors. Horizontal infections are often diagnosed later, leading to delays in the initiation of ART and prolonged periods of uncontrolled viraemia, which hinder VLS efforts.³³ This increased pill burden could negatively impact adherence to ART. Regarding psychosocial aspects, while the sexual acquisition of HIV in adolescents is widely recognised as highly stigmatising—due to the intersection of HIV-related and sexual stigma—transfusion-acquired HIV may also carry stigma in certain contexts. In our setting, repeated transfusions are often associated with chronic illnesses such as sickle cell disease or HIV, which can lead to perceptions of the individual as weak, frequently ill or socially ‘unfit’. These perceptions may foster exclusion, internalised stigma and reduced self-efficacy in managing ART. Such stigma can discourage adherence to ART and reduce access to consistent healthcare, further hindering VLS.³⁴ Additionally, CALHIV infected through these horizontal routes may have more complex health histories, potentially including other comorbidities or complications related to their mode of infection. These added health challenges can interfere with effective ART adherence and treatment outcomes. Lastly, unlike vertical infections, which benefit from early diagnosis and structured follow-up through maternal and child health programmes,³⁵ horizontal cases often lack early intervention, resulting in poorer long-term VLS.³³ Further research is needed to better understand these nuanced stigma dynamics among CALHIV infected through horizontal routes. These results underscore the urgency of differentiated service delivery models that address adherence barriers. From a public health perspective, reinforcing HIV counselling, strengthening community-based support and promoting timely disclosure strategies could improve treatment outcomes. Further research, especially longitudinal and qualitative, is essential to design and evaluate such interventions in resource-limited settings like Lubumbashi.

As in previous studies,^{6 24 36} this study shows that advanced WHO clinical stages are strongly associated with USVL. The WHO staging refers to the highest ever recorded stage, typically assessed at diagnosis. This aligns with the literature indicating poorer health outcomes in individuals with advanced disease. If current staging were used, a bidirectional effect could arise, where USVL contributes to disease progression, complicating VLS. As the disease progresses, patients are more likely to face opportunistic infections, increased pill burden and side effects, all of which hinder ART adherence.³⁷ Early ART initiation is crucial to prevent progression,^{38 39} but for those at advanced stages, intensified medical support and adherence counselling are vital to improving VLS outcomes.

Poor or fair adherence to ART emerged as the most significant predictor of USVL. This finding is consistent with studies from SSA^{6 12 13 15 24 36} and highlights the central role of adherence in achieving VLS and ART success. While this relationship is well established across all age groups and settings, it is important to note the increasing use of dolutegravir-based regimens, which are more ‘forgiving’ of occasional missed doses. These regimens may sustain VLS even with suboptimal adherence, although consistent adherence remains essential for optimal outcomes.^{40 41} Various factors contribute to poor adherence, including stigma, forgetfulness, pill fatigue and lack of social support. Given the high odds of USVL associated with poor adherence in this study, interventions aimed at enhancing adherence—such as peer support groups, mobile health reminders and community-based adherence support—are crucial.⁴¹ These strategies could address the multifaceted barriers that CALHIV face, ultimately improving their long-term health outcomes.

ART-induced side effects were significantly associated with USVL in CALHIV. Consistent with findings from a study conducted in Uganda and Kenya,^{35 42} where ART-related side effects were linked to USVL, our results similarly highlight side effects as a key factor in USVL. Side effects are a well-established predictor of ART non-adherence, with symptoms such as nausea and fatigue often prompting patients to skip or discontinue treatment.⁴³ The recent shift to dolutegravir-based regimens—which are generally better tolerated and associated with fewer side effects than earlier regimens—may help mitigate this barrier and improve adherence, particularly among adolescents. Nevertheless, comprehensive side effect management remains crucial. This includes offering tolerable alternatives, when necessary, symptom relief and targeted counselling for CALHIV and caregivers to encourage open communication with healthcare providers. Equipping healthcare staff with the skills to identify and manage side effects can further support adherence, reduce USVL and improve long-term treatment outcomes.

To our knowledge, this is the first study to investigate the factors associated with USVL among CALHIV in the DRC. However, several limitations should be acknowledged. First, its cross-sectional design precludes any causal inference between the identified factors and USVL. Longitudinal studies are needed to better understand temporal relationships and causal pathways. Second, the study was conducted in urban Lubumbashi, limiting the generalisability of the findings to rural areas with different socioeconomic and healthcare contexts. Third, some associations—such as those involving ART adherence—showed high ORs with wide CIs, possibly due to sparse data or misclassification bias. These results should be interpreted with caution despite their biological plausibility. Finally, reliance on routinely collected medical records may have introduced information bias, as data may be incomplete, outdated or inaccurately recorded, potentially affecting the accuracy of some variables.

To address high USVL rates among CALHIV in Lubumbashi, future research should explore barriers at the individual, interpersonal, health system and community levels. We recommend qualitative studies to identify these barriers and guide targeted, context-specific interventions.

Conclusions

This study sheds light on the prevalence and key determinants of USVL among CALHIV in Lubumbashi, DRC. Despite improvements in ART, nearly a quarter of CALHIV in the study population failed to achieve VLS. Factors such as caregiver marital status, horizontal HIV transmission, advanced HIV clinical stages, poor ART adherence and ART-induced side effects were identified as significant contributors to USVL. To address these challenges, targeted interventions are necessary. These include improving ART adherence through counselling, managing side effects and providing comprehensive support for both adolescents and caregivers. Strengthening healthcare systems for better follow-up and early intervention is also essential. The findings emphasise the need for a holistic approach that integrates medical, psychological and social support to enhance health outcomes for CALHIV and reduce the high rates of USVL in this vulnerable population.

Data availability statement

Data are available on reasonable request.