Association between baseline HIV-1 DNA levels and clinical outcomes in people living with HIV: a meta-analysis of cohort studies

ABSTRACT

To assess the prognostic value of baseline HIV-1 DNA levels, a meta-analysis was conducted according to the PROSPERO protocol (CRD42024619050) based on data from studies published until March 4, 2025. Relevant studies were retrieved from the Web of Science, PubMed, Cochrane Library, Embase, and Scopus databases. Effect sizes (correlation coefficients, odds ratios [ORs], hazard ratios [HRs], and adjusted hazard ratios [aHRs]) were calculated using R software, with subgroup analyses and assessment of publication bias and sensitivity. Seventeen studies involving 4789 participants were included. The combined correlation coefficient between pre – and on-ART HIV-1 DNA levels was 0.71 (95% confidence Interval (CI): 0.63–0.78). Baseline DNA levels were significantly associated with viral rebound after viral suppression (combined OR = 1.74, 95% CI: 1.25–2.41; HR = 2.01, 95% CI: 1.58–2.56; aHR = 2.26, 95% CI: 1.75–2.92). For clinical progression, the combined HR and aHR for continuous baseline DNA were 3.66 (95% CI: 2.87-4.66) and 2.44 (95% CI: 1.87-3.20), respectively, with high baseline DNA levels associated with an increased risk of clinical progression (HR = 2.58, 95% CI: 1.96-3.39; aHR = 1.90, 95% CI: 1.41-2.55). For mortality, the HR and aHR were 3.22 (95% CI: 1.96-5.29) and 2.15 (95% CI: 1.21-3.84) respectively, with high baseline DNA levels associated with an increased risk of death (HR = 3.54, 95% CI: 1.39-9.00; aHR = 2.86, 95% CI: 1.01-8.08). Higher pre-ART HIV-1 DNA levels are associated with increased risks of viral rebound, clinical progression, and mortality. These results suggest that baseline HIV-1 DNA represents a potentially valuable supplementary biomarker for monitoring disease progression and treatment response.

Introduction

By the end of 2023, the global number of people living with HIV (PLWH) had reached 39.9 million. According to UNAIDS, progress towards the “95-95-95” targets was 86%, 89%, and 93%, respectively [1]. Approximately 35.51 million individuals are currently receiving antiretroviral therapy (ART). Despite the fact that the widespread implementation of ART has significantly improved the quality of life and increased the life expectancy of PLWH [2], HIV eradication is currently not achievable due to the persistence of the viral reservoir, thus patients require lifelong treatment [3]. HIV can rapidly reverse transcribe and integrate into the host genome after infection, whereas the normal function of infected cells remains unaffected by the latent integrated virus. This allows these cells to serve as long-term reservoirs that evade immune surveillance [4,5]. Previous studies have shown that even with long-term ART, the reduction in HIV DNA reservoirs is extremely slow, with an estimated median half-life of up to 44 months [6]. Moreover, more than 70 years are needed to completely eradicate these reservoirs even with intensified treatment [7], making this goal virtually unattainable for PLWH.

Till today, approximately 33 million people have achieved viral suppression worldwide [1]. One hundred and sixteen countries have incorporated dolutegravir (DTG), a drug from the integer strand transfer inhibitor (INSTI) class, as the backbone of first-line ART regimens [8] since the recommendation of the World Health Organization (WHO) in 2018. INSTIs offer several advantages, including good safety, potent antiviral activity, high resistance barriers, and good tolerability [9]. These benefits have alleviated previous limitations imposed by drug resistance mutations and side effects, improving patients’ experience and treatment outcomes. In recent years, the development of long-acting ARTs or long-acting ART combinations has further reduced the frequency of medication intake, significantly improving patient adherence [10] and ensuring the sustained efficacy of ART. Historically, the viral load (VL) and CD4⁺ T-cell count have served as the primary indicators for assessing treatment efficacy. However, given the extremely high rates of viral suppression in current ART-experienced patients, VL often remains undetectable during follow-up [1]. Additionally, CD4⁺ T-cell recovery tends to be slow, and a significant portion of patients can only maintain CD4⁺ T-cell counts at levels lower than those of healthy individuals even after prolonged ART, without fully returning to pre-infection levels [11]. Therefore, in the context of widespread ART treatment and stable VL and CD4⁺ T-cell counts, incorporating other markers to evaluate disease progression and treatment response is essential.

In recent years, individualization and precision medicine have gradually become the focus of research on HIV treatment. The WHO first proposed individualized treatment strategies in the Comprehensive Guidelines for the Treatment of HIV Infection with Antiretroviral Drugs and Prevention of HIV Infection released in 2016 [12], emphasizing optimizing treatment based on virological and immunological indices. 2021 [13] update of the WHO Guidelines further clarified that dynamic assessment of immune function and treatment response is a key part of long-term management. China's 2024 Guidelines for the Diagnosis and Treatment of AIDS [14] provide the first management recommendations for patients with poor immune reconstitution, emphasizing the need to adjust treatment strategies based on immunological status. Meanwhile, the U.S. Department of Health and Human Services (HHS) [15] and the European AIDS Clinical Society (EACS) [16] have also consistently pointed out in their latest guidelines that optimizing long-term management through dynamic monitoring of virological and immunological indices is of great significance, especially for patients with incomplete immune reconstitution.

Accurate viral surveillance is essential to implementing the concept of precision medicine and reaching the goal of “95-95-95”. Considered an important molecular surveillance tool in precision medicine, HIV DNA testing identifies latent viruses in viral reservoirs, helping to estimate the number of infected cells and the size of latent viral reservoirs established soon after primary HIV infection. The persistent existence of these reservoirs provides further opportunities for HIV-1 DNA detection. Previous studies have shown that increased HIV-1 DNA levels in patients with treatment failure are associated with smaller increases in CD4⁺ T-cell counts, with the highest HIV-1 DNA levels typically observed in treatment failure cases [17]. In contrast, the lowest levels are usually observed in long-term nonprogressors and elite controllers [18,19]. Therefore, it is important to introduce HIV-1 DNA level testing on top of existing markers and to further investigate the association between HIV-1 DNA level and disease progression or treatment response. HIV DNA testing helps to assess the viral reservoir before treatment, providing a scientific basis for developing individualized treatment strategies, while after treatment it can compensate for the shortcomings of the traditional methods by providing information on the latent viral reservoir, thus supporting the adjustment and optimization of long-term management. With advances in technology and continued policy impetus, HIV DNA testing is expected to be an important complement to traditional surveillance, particularly for patients with poor immune reconstitution. It can support the optimizations of treatment strategies, which is also critical to improving HIV prevention and control strategies.

A previous study summarized the relationship between baseline HIV-1 DNA levels and severe clinical outcomes [20]; however, this study was published a long time ago and may have had some methodological issues. In recent years, numerous studies have focused on similar topics, providing more information on various forms of adverse clinical outcomes and treatment-related drug response cycles, reservoir clearance, and immune reconstitution. However, comprehensive, systematic reviews of these related studies are lacking. Therefore, this research aimed to aggregate and analyse data from cohort studies that investigated the relationships between baseline HIV-1 DNA levels and various clinical outcomes and indicators. By transforming data and combining similar outcomes, we explored how variations in HIV-1 DNA levels correlate with clinical outcomes and indicators. We hope that this study will provide valuable reference information for understanding the prognostic value of baseline HIV-1 DNA levels, thereby offering direction and evidence for identifying supplementary evaluation markers beyond the viral load and immunological indicators in the post-AIDS era.

Methods and materials

Search strategy and selection criteria

This study was registered on PROSPERO with the registration ID CRD42024619050. The data for the study were sourced from multiple databases, including Web of Science, PubMed, the Cochrane Library, Embase, and Scopus. The search utilized terms including HIV, AIDS, human immunodeficiency virus, acquired immune deficiency syndrome, DNA, provirus, reservoir, quantification, and level. Before conducting the formal search, all terms were verified against the MeSH Word List, and additional commonly used expressions were incorporated to enhance comprehensiveness. A manual retrieval strategy was also employed to identify and supplement relevant literature cited within existing records. The search spanned from the earliest available records in the databases up to March 4, 2025, with no restrictions on article type.

The inclusion criteria were as follows: (1) publications in English released before March 4, 2025; (2) cohort study design; (3) participants who were diagnosed with HIV infection or AIDS; (4) participants aged 18 years and above; (5) studies examining the relationship between baseline HIV-1 DNA levels and subsequent clinical outcomes or indicators; and (6) studies presenting results on associations between variables (correlation coefficients, odds ratios (ORs), hazard ratios (HRs)) or providing raw data sufficient for calculating these parameters.

The exclusion criteria were as follows: (1) nonoriginal research (reviews, meta-analyses), case reports, conference abstracts, letters, or clinical trial registration protocols; (2) studies in which all patients have comorbidities; (3) Studies that solely assess the efficacy of DNA quantification methods; and (4) studies with incomplete or missing relevant data.

Literature screening, data extraction, and quality assessment

The references retrieved from the database were managed using EndNote software (version X6). Two authors independently screened the deduplicated references, and the final inclusion of studies was determined on the basis of the predefined inclusion and exclusion criteria. In cases of discrepancies during the assessment of references, a discussion was held with a third author to reach a consensus. The following information was extracted from the included studies: title, first author, publication year, journal name, sample size, sample type, outcome variables, and effect sizes (such as correlation coefficients, ORs, or HRs). This study included only cohort studies; therefore, the Newcastle‒Ottawa Scale (NOS) was used for quality assessment of the included studies. Adjustments were made on the basis of the study objectives and the characteristics of the included studies, as detailed in Supplementary Material 1.

Data analysis

The statistical analysis for this study was performed using R software (version 4.4.2). The Meta package calculates the combined value for correlation coefficients, ORs, and HRs. The correlation coefficients were initially transformed using Fisher’s Z before being combined. Different DNA level variables (HR and adjusted hazard ratios [aHR]) were analysed separately through meta-analysis and presented in one figure by subgroup analyses for the same outcome. Heterogeneity was assessed using the I² statistic [21], publication bias was tested using Egger’s test, and a P value ≤ 0.05 indicated significant publication bias. Sensitivity analysis was performed to assess the stability of the results.

Results

Literature Screening Process

Figure 1 [22] illustrates the article selection process for this study. A total of 3,865 records were obtained from the five databases. After the removal of duplications, the titles and abstracts of the remaining 1,853 records were reviewed, and 214 full-text articles were further assessed. In the full-text review process, 199 records failed to meet the inclusion criteria, in which 68 studies were solely focused on the assessment of the efficacy of DNA quantitative testing methods without providing any association between DNA quantitative results and clinical outcomes or indicators. Consequently, a total of 15 studies were included in the analysis. Additionally, during the full-text review process, 26 potentially relevant articles that were not initially retrieved were identified. After these full texts were downloaded and reviewed, two studies met the inclusion criteria. Therefore, a total of 17 studies [23–39] were included in this research.

Figure 1.

PRISMA 2020 study selection flow diagram for the meta-analysis of the associations between pre-ART HIV-1 DNA levels and clinical outcomes.

Characteristics of studies on the associations between pre-ART HIV-1 DNA levels and clinical outcomes

The 17 included studies involved 4,789 HIV-infected individuals from 20 countries across five continents. Thirteen studies used peripheral blood mononuclear cell (PBMC) samples, three used whole blood samples, and one used both PBMC and whole blood samples. All studies were rated as high quality. Detailed information is provided in Table 1 and Supplementary Material 2.

Table 1.

Characteristics of 17 included studies.

ID	Sample				Score
	Source	Time	Size	Type
2020. N'Takpe, J.B	West Africa	2008.3-2012.7	2019	PBMC	9
2018. Francesca, C.S	Italy	After 2000	433	PBMC/Whole Blood	8
2017. Lombardi, F	Italy	2011.7-2014.6	201	Whole blood	8
2017. Gandhi, R.T	United States	-	101	PBMC	9
2014. Williams, J.P	Eight countries	2003.8-2007.7	51	PBMC	7
2014. Besson, G.J	United States	2000.1-2007.6	30	PBMC	8
2011. Sidonie, L.N	Franch	2010.6-2011.2	193	Whole blood	9
2010. Carmen, R.S	Spain	1997–2001	115	PBMC	8
2010. Avettand-Fenoel, V	Five countries	2003.10-2005.2	72	Whole blood	7
2008. Minga, A.K	Abidjan, Co^te d’Ivoire	1997.6-2006.2	200	PBMC	9
2008. Avettand-Fènoël, V	Amsterdam/Franch	-	422	PBMC	7
2007. Hoen, B	Ten countries	1998.2-1999.10	78	PBMC	8
2006. Goujard, C	Franch	1996.11-2004.10	163	PBMC	9
2005. Rouzioux, C	Franch	1988–1996	383	PBMC	9
2003. Tierney, C	United States/Puerto Rico	1991.12-1994.11	111	PBMC	7
2002. Kostrikis, L.G	Greece	-	127	PBMC	9
2002. Katzenstein, T.L	Copenhagen	1984–1988	90	PBMC	8

The correlation between pre-ART HIV-1 DNA levels and on-ART HIV-1 DNA levels

Three studies explored the relationship between pre – and on-ART HIV-1 DNA levels, providing 8 data points. One studies (2 data points) reported Pearson correlation coefficients (r), whereas the other two studies (6 data points) provided Spearman correlation coefficients (ρ). After Fisher’s Z transformation of the 8 data points, the combined effect size was 0.71 (95% CI: 0.63–0.78) (Figure 2).

Figure 2.

Forest plots of the meta-analysis for the correlation between pre – and on-ART HIV-1 DNA levels.

Association between pre-ART HIV-1 DNA levels and viral suppression under ART

Two studies used Cox regression models to explore the association between baseline DNA levels and viral load suppression. Although both studies revealed that for each unit increase in baseline DNA, the likelihood of achieving viral suppression decreased, the meta-analysis using a random effects model yielded combined HRs and aHRs of 0.49 (95% CI: 0.19–1.25) and 0.52 (95% CI: 0.19–1.42), respectively. Both P values were greater than 0.05 (Figure 3), suggesting that the association between baseline DNA levels and viral suppression did not reach statistical significance.

Figure 3.

Forest plots of the meta-analysis for the association between pre-ART HIV-1 DNA levels and viral suppression under ART.

Association between pre-ART HIV-1 DNA levels and viral rebound (VR) after viral suppression under ART

Two studies used logistic regression models to explore the association between baseline DNA levels and posttreatment viral load rebound, providing three unadjusted OR data points. Meta-analysis using a random model was used to calculate a combined OR (Figure 4). The results suggested that for each 1 log10 copy/10⁶ PBMC increase in baseline DNA level, the risk of viral load rebound after treatment was 1.74 (95% CI: 1.25-2.41) times greater.

Figure 4.

Forest plots of the meta-analysis for the association between pre-ART HIV-1 DNA levels and viral rebound according to logistic regression.

Three studies used Cox regression models to explore the association between baseline DNA levels (as a continuous variable) and posttreatment viral load rebound. One study reported the HR and aHR (3 data points each), whereas the other two studies reported one HR and one aHR. Figure 5 shows that the combined HR and aHR were 2.01 (95% CI: 1.58–2.56) and 2.26 (95% CI: 1.75–2.92), respectively.

Figure 5.

Forest plots of the meta-analysis for the association between pre-ART HIV-1 DNA levels and viral rebound according to Cox regression.

Association between pre-ART HIV-1 DNA levels and clinical progression

Seven studies used Cox regression models to explore the associations between baseline HIV-1 DNA levels, both as continuous and categorical variables, and clinical progression in patients (Figure 6). When baseline DNA levels were treated as a continuous variable, seven studies provided 14 HR data points, with six contributing ten aHR data points. The combined HR and aHR were 3.66 (95% CI: 2.87-4.66) and 2.44 (95% CI: 1.87-3.20), respectively. When baseline DNA levels were treated as a categorical variable, two studies provided five HR and five aHR data points. Compared with patients in the lower baseline DNA group, patients in the higher baseline DNA group had a 2.58-fold (95% CI: 1.96-3.39) and 1.90-fold (95% CI: 1.41-2.55) increased risk of clinical disease progression for HR and aHR, respectively.

Figure 6.

Forest plots of the meta-analysis for the associations between pre-ART HIV-1 DNA levels and clinical progression.

The association between pre-ART HIV-1 DNA levels and death

Six studies used Cox regression models to explore the associations between baseline HIV-1 DNA levels, both as continuous and categorical variables, and mortality outcomes in patients (Figure 7). When baseline DNA levels were treated as a continuous variable, four studies provided HR data points, and three provided aHR data points. The combined HR and aHR were 3.22 (95% CI: 1.96-5.29) and 2.15 (95% CI: 1.21-3.84), respectively. When baseline DNA levels were treated as a categorical variable, patients in the higher baseline DNA group had 3.54-fold (95% CI: 1.39-9.00) and 2.86-fold (95% CI: 1.01-8.08) increased mortality risk, respectively, compared with patients in the lower baseline DNA group.

Figure 7.

Forest plots of the meta-analysis for the association between pre-ART HIV-1 DNA levels and death.

Publication bias and sensitivity analysis

In the 14 meta-analyses conducted, Egger's test results indicated no significant publication bias except for the four analyses where pre-ART DNA was treated as a continuous variable with outcomes of death and clinical progression when the combined HR and aHR were calculated. The sensitivity analysis suggested stable results after the studies were removed one by one.

Discussion

This meta-analysis of cohort studies quantifying pre-ART HIV-1 DNA levels revealed that, with the exception of the lack of a statistically significant association with posttreatment viral suppression, pre-ART HIV-1 DNA levels were positively correlated with long-term DNA levels following treatment. Additionally, increased pre-ART HIV-1 DNA levels were associated with an increased risk of VR after virological suppression, clinical disease progression in any form postinfection, and even mortality risk in patients.

This study revealed a significant positive correlation between pre – and on-ART HIV-1 DNA levels. Specifically, the findings indicate that the higher the initial reservoir level formed after HIV infection is, the higher the residual reservoir level after treatment. Previous cross-sectional studies failed to confirm a correlation between patient inflammatory factors and HIV DNA [40]. However, more recent studies, mainly longitudinal studies on acute-phase patients with different ART initiation times, revealed that even starting ART during acute HIV infection could not completely prevent the low-level inflammation caused by myeloid immune system activation. Some soluble mediators produced during acute HIV infection exhibit a long-term association with the HIV reservoir during viral suppression, highlighting the importance of the inflammatory environment during the earliest stages of HIV infection [41]. Although it remains unclear whether early initiation of treatment influences the decay of the reservoir by modulating inflammatory factors, some studies have indicated that the timing of ART initiation significantly affects the slope of the initial decline in DNA levels. Specifically, the earlier ART starts, the faster the cell associated HIV-DNA levels decline within the first 8 months post infection, with significant differences in the degree of decline observed among groups with different initiation times after long-term ART [42]. Therefore, shortening the time interval between diagnosis and ART initiation to prevent the accumulation of the initial reservoir as early as possible while maintaining lower DNA levels after treatment not only benefits individual patients’ outcomes but also contributes to HIV prevention and control in the broader population.

The results of this study indicate that the association between pre-ART HIV-1 DNA levels and viral suppression did not reach statistical significance. However, the original research data revealed a negative correlation between baseline DNA levels and viral suppression. This phenomenon may be attributed to the high heterogeneity among studies and the small sample size. Nonetheless, an increase in baseline DNA levels was associated with a greater risk of VR after achieving viral suppression, and this relationship was statistically significant according to both logistic regression and Cox regression models. We previously summarized the factors influencing VR [43]. However, the studies included at that time focused primarily on patients with low-level viremia (LLV) under different criteria and did not address treatment failure cases. Additionally, the role of baseline DNA levels in LLV was not discussed in that study. The associations between DNA levels and various forms of VR found in this study provide indirect evidence for guiding the selection of first-line treatment regimens for newly diagnosed patients. Compared with most traditional recommended regimens, ART regimens based on integrase strand transfer inhibitors (INSTIs) have been shown to reduce the HIV-1 viral load more rapidly [44] and significantly lower total HIV-1 DNA levels in patients who achieve virological suppression [45]. On the basis of the findings of this study, it may be more appropriate to recommend an INSTI-based regimen for patients with higher baseline HIV-1 DNA levels, as this approach may more effectively reduce HIV-1 DNA levels in patients with reasonable viral control, potentially offering long-term benefits. This recommendation is also consistent with the current guidelines from the WHO and various national health authorities [14,46].

This study revealed that regardless of whether baseline DNA levels were treated as a continuous or categorical variable and whether other factors were adjusted for, pre-ART HIV-1 DNA levels were associated with clinical disease progression and adverse outcomes, including death. It is worth noting that, we converted the groupings in each study into higher and lower DNA level groups during our analysis, rather than using a specific threshold. The results can only indicate that the higher group has a greater risk of adverse events compared to the lower group. Although the viral load and CD4⁺ T-cell count have been demonstrated to be reliable and consistent biomarkers for assessing disease status and treatment response and remain the primary indicators for evaluating treatment effectiveness in clinical practice, the associations identified in this study suggest the prognostic importance of the number of HIV-infected cells during acute infection. Furthermore, some studies included in this research had already adjusted for HIV-1 RNA and CD4⁺ T-cell count, and the combined aHR still indicated that higher DNA levels in various forms increased the risk of adverse events. This suggests that results of the same clinical outcome show consistency across formats. Such consistency provides multiple perspectives on our understanding of the association between DNA levels and these clinical outcomes, and is difficult to achieve in a single study. The integration of multicentre data integration also revealed a key phenomenon: while DNA levels varied across studies influenced by study design or assay methodology, the relative risk ratios between DNA levels and each clinical outcome showed consistency across geographic regions.

These findings offer a new perspective on the role of HIV-1 DNA levels in disease monitoring and long-term prognosis evaluation. Although the direct correlation between baseline DNA levels and viral suppression did not reach statistical significance, the predictive value of baseline DNA levels for long-term clinical outcomes is consistent with the previous studies. This discovery aligns with the mechanism of early treatment intervention's impact on HIV DNA dynamics, emphasizing the importance of early initiation of ART. Multiple studies have demonstrated that initiating ART during the acute phase can significantly reduce baseline HIV DNA levels and hinder the formation of viral reservoirs. For instance, Hongzhou Lu’s team found that the HIV DNA level in the acute treatment group was approximately five times lower than that in the late treatment group [47]. Simultaneously, Barbehenn et al. confirmed through mathematical modelling that early treatment can lead to a biphasic rapid decay of the HIV DNA reservoir, with a decay rate five times faster than that of chronic phase treatment. This intervention effect explains the association between baseline DNA levels and long-term prognosis observed in this study. Early treatment reduces the initial viral integration and clonal expansion, lowering the baseline DNA load and indirectly delaying viral rebound and improving immune recovery [48]. Furthermore, the positive impact of early treatment on clinical outcomes has been validated in several cohort studies. For example, the START trial demonstrated that early ART can significantly reduce the size of the viral reservoir and delay disease progression. This suggests that uninterrupted viral replication after infection may accelerate the accumulation of the viral reservoir, leading to increased baseline DNA levels and ultimately increasing the risk of rebound and mortality after treatment. Based on this, early initiation of ART may improve clinical outcomes by limiting the initial expansion of the reservoir [49]. This study further supports the integration of HIV DNA testing in early intervention strategies. Quantitative detection of baseline HIV-1 DNA levels can help identify individuals at high risk of progression, facilitating the advancement of individualized treatment windows [50]. Meanwhile, the combination of DNA testing and conventional detection indicators not only suppresses viral replication and reduces the heterogeneity of the viral reservoir, but also improves long-term survival by reducing the level of immune activation. This aligns with the “treatment at diagnosis” policy advocated by various guidelines, and more PLWH will benefit from it. Notably, in addition to baseline HIV-1 DNA levels, some studies have measured HIV-1 DNA in patients who interrupted ART after achieving viral suppression and reported that patients with HIV-DNA levels ≥ 150 copies/10⁶ PBMCs had a significantly shorter time to lose viral control [51]. This further underscores the potential value of monitoring HIV-DNA levels during treatment. However, current ART regimens have limited effectiveness in clearing reservoirs, and routine measurement of HIV-1 DNA during treatment has limited significance in clinical practice.

Owing to differences in the methods used for HIV DNA quantification across studies, the parameters selected to explore the relationship between DNA levels and clinical outcomes also varied. Although most studies treated DNA levels as continuous variables and reported changes in outcome risk for each unit change in DNA level (OR or HR), some studies presented different results. For example, in one study using Cox regression, the calculated HR represented the change in outcome risk for every 0.5-unit increase in the DNA level, and another study provided HR values for every tenfold increase in the DNA level. Previous meta-analyses may have experienced issues with combining raw data without standardizing units and parameters, which could lead to a misestimation of the combined effect size. To address this, we converted and standardized the parameter values from different units to minimize their impact on the final combined effect size in this study to address such issues. Detailed information can be found in Supplementary Material 2.

Compared to traditional indicators such as viral load and CD4⁺ T-cell count, baseline HIV-1 DNA levels not only directly quantify proviral DNA integrated into the host genome, reflecting the core size of the viral reservoir, but also enable the assessment of the residual status of the viral reservoir, offering potential benefits in real-world applications [23,36]. Furthermore, patients with higher levels of HIV-1 DNA typically experience a significantly accelerated viral rebound upon treatment interruption. The decline in CD4⁺ T-cell count often lags behind HIV-1 DNA levels, and thus may serve as an earlier warning signal [27,52,53]. Although the CD4⁺ T-cell count remains the cornerstone of immune status assessment, and the <200 cell count in particular is the strongest predictor of AIDS-related events (e.g. opportunistic infections and malignancies), its prediction of viral reservoir dynamics and long-term prognosis is susceptible to interference by age, comorbidities, individual differences in immune recovery and medications, and does not directly reflect the effects of viral reservoir clearance. For instance, a higher CD4⁺ T-cell count does not necessarily indicate a reduction in the body's viral reservoir, whereas HIV-1 DNA levels more accurately reflect the viral burden in the body [54,55].

However, this study has certain limitations. First, there was significant heterogeneity between studies regarding the relationships between baseline DNA levels and specific outcomes. Owing to the limited number of included studies and the limited information that could be extracted from each original study, we could not explore the sources of heterogeneity, such as patient demographic characteristics, infecting viral strain subtypes, and DNA measurement methods. Second, although this study revealed associations between baseline HIV-DNA levels and various clinical progressions and adverse outcomes, we were unable to further analyse clinically meaningful HIV-DNA threshold levels given a lack of original data. Future research should continue to track relevant studies to gather more information on these aspects. Third, it is worth noting that although adherence indeed has an impact on treatment outcomes, only four of the included studies mentioned ART adherence. Due to the lack of sufficient data, we were unable to conduct a more comprehensive supplementary analysis on adherence. Future research should place greater emphasis on the collection and analysis of adherence data. Finally, some analyses in this study were limited by the small number of included studies. To enhance the robustness of the estimation, we effectively addressed this issue by standardizing the correlation coefficients using Fisher's Z transformation, as detailed in Supplementary Material 1.

In addition to commonly used indicators such as the viral load and CD4⁺ T-cell count, baseline HIV-1 DNA levels not only reflect the size of the initial reservoir in PLWH but are also associated with long-term DNA levels after treatment. Higher pre-ART HIV-1 DNA levels are associated with an increased risk of viral rebound following viral suppression, clinical progression, and even death, suggesting that baseline HIV-1 DNA could serve as a potential supplementary biomarker for evaluating disease progression and treatment response. Therefore, it is recommended that HIV-1 DNA levels be monitored in combination with CD4⁺ T-cell counts, complementing the current HIV RNA-centred monitoring programme. Especially for patients with high viral reservoirs or low CD4⁺ T-cell counts, stratified management and regular follow-up should be implemented to optimize clinical risk assessment and intervention strategies. Meanwhile, HIV-1 DNA testing can significantly reduce the time window from diagnosis to treatment, fully implementing the “rapid initiation of ART” strategy and improving patient survival rates.

Funding Statement

This project is financially supported by the National Key R&D Program of China [grant numbers 2023YFC2308300, 2023YFC2308302, 2023YFE0116000, 2022YFC2305200, 2022YFC2305202, 2021YFC2301900, 2021YFC2301905], the High-Level Public Health Specialized Talents Project of Beijing Municipal Health Commission [grant number 2022-2-018], the Beijing Research Ward Excellence Program (BRWEP), [grant number 2024W042180108], the Beijing Key Laboratory for HIV/AIDS Research [grant number BZ0089], the Postgraduate Research＆Practice Innovation Program of Jiangsu Province [grant number KYCX24_0484], and the SEU Innovation Capability Enhancement Plan for Doctoral Students [grant number CXJH_SEU 24042]. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Author contributions

Conceptualization, F.Z., DF.Y., B.W. and B.S.; methodology, F.Z., DF.Y. and YY.L.; software, F.Z., DF.Y., SS.L., YY.L., L.L., HX.Y. and LF.L.; validation, B.W., B.S. and C.M.; formal analysis, F.Z., DF.Y., SS.L., YY.L., L.L., HX.Y. and LF.L.; resources, C.M., B.W. and B.S.; data curation, F.Z., DF.Y., L.L. and HX.Y.; writing – original draft preparation, F.Z., DF.Y. and SS.L.; writing – review and editing, C.M., B.W., and B.S.; visualization, DF.Y., F.Z., YY.L., SS.L., HX.Y. and LF.L.; supervision, C.M., B.W. and B.S.; project administration, C.M., B.W., and B.S.; funding acquisition, B.W., and B.S. All authors had full access to all data in the study and had final responsibility to submit for publication. All authors read and approved the final version of the manuscript.

Data availability statement

All data generated or analyzed during this study are included in this published article.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Supplemental Material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/22221751.2025.2508759.