Prediction of long-term outcomes of HIV-infected patients developing non-AIDS events using a multistate approach

Persistent poor clinical outcomes of people living with HIV presenting with AIDS and late HIV diagnosis - results from the ICONA cohort in Italy, 2009-2022

OBJECTIVES

Limited data are available on the long-term outcomes in recent years for late HIV diagnosis (LD).

METHODS

All subjects with HIV enrolled in the ICONA cohort in 2009-2022 who started antiretroviral treatment (ART) within 4 months from diagnosis were included and divided into: (i) pre-ART CD4 count ≥350/mm3 without AIDS (non-LD), (ii) pre-ART CD4 count <350/mm3 without AIDS (LD asymptomatic), and (iii) with AIDS events pre-ART (LD-AIDS). The estimated probability and independent risk for mortality (all-cause and cause-specific) and treatment failure were evaluated.

RESULTS

Of 6813 participants (2448 non-LD, 3198 LD asymptomatic, and 1167 LD-AIDS), 161 (2.4%) died after ART initiation. At survival analysis, a higher probability of all-cause mortality has been identified for LD than non-LD (P <0.001) and within the former, for LD-AIDS over LD asymptomatic (P <0.001). After adjusting for confounders, LD showed a higher risk of all-cause mortality (vs non-LD adjusted hazard ratio (aHR) 5.51, P <0.001) and, in particular, being an AIDS presenter predicted a greater risk of all-cause (aHR = 4.42, P <0.001), AIDS-related (adjusted subhazard ratio [aSHR] = 16.86, P <0.001), and non-AIDS-related mortality (aSHR = 1.74, P = 0.022) than the rest of the late presenters. Among the short-term survivors in the LD-AIDS group, the long-term mortality was mediated by the lack of immune recovery at 2 years. Finally, LD compared with non-LD and, particularly, among the former, LD-AIDS over LD asymptomatic showed a greater risk of treatment failure.

CONCLUSIONS

In recent years, LD subjects, particularly, AIDS presenters, remained at a higher risk of poorer outcomes. Public health strategies for early HIV diagnosis are urgently needed to constrain the mortality gap.

Non-AIDS-Defining Events in Human Immunodeficiency Virus Controllers Versus Antiretroviral Therapy-Controlled Patients: A Cohort Collaboration From the French National Agency for Research on AIDS CO21 (CODEX) and CO06 (PRIMO) Cohorts

Background

Low-grade chronic inflammation may persist in spontaneous human immunodeficiency virus controllers (HICs), leading to non-AIDS-defining events (nADEs).

Methods

Two hundred twenty-seven antiretroviral therapy (ART)-naive HICs (known human immunodeficiency virus type 1 [HIV-1] infection ≥5 years and at least 5 consecutive viral loads [VLs] <400 HIV RNA copies/mL) were compared with 328 patients who initiated ART ≤1 month after primary HIV infection diagnosis and had undetectable VL within 12 months following ART initiation for at least 5 years. Incidence rates of first nADEs were compared between HICs and ART-treated patients. Determinants of nADEs were assessed by using Cox regression models.

Results

All-cause nADEs incidence rates were 7.8 (95% confidence interval [CI], 5.9-9.6) and 5.2 (95% CI, 3.9-6.4) per 100 person-months among HICs and ART patients, respectively (incidence rate ratio [IRR], 1.5 [95% CI, 1.1-2.2]; adjusted IRR, 1.93 [95% CI, 1.16-3.20]). After adjustment for the cohort, demographic, and immunological characteristics, the only other factor associated with all-cause nADE occurrence was age ≥43 (vs <43) years at the beginning of viral control (IRR, 1.69 [95% CI, 1.11-2.56]). The most frequent events observed in the 2 cohorts were non-AIDS-related benign infections (54.6% and 32.9% of all nADEs, respectively, for HICs and ART patients). No differences in cardiovascular or psychiatric events were observed.

Conclusions

HICs experienced 2 times more nADEs than virologically suppressed patients on ART, mainly non-AIDS-related benign infections. Older age was associated with nADE occurrence, independent of immune or virologic parameters. These results do not argue in favor of expanding the ART indication for HICs but rather a case-by-case approach considering clinical outcomes such as nADEs besides immune activation.

Outcomes of persistent low-level viremia among HIV patients on antiretroviral therapy: A prospective cohort study

INTRODUCTION

We aimed to investigate the relationship between low-level viremia (LLV) and virological failure (VF), death, and non-AIDS events (NAEs).

METHODS

A prospective cohort study of people living with HIV (PLHIV) on antiretroviral therapy (ART) was conducted from 2011-2018 at an HIV clinic in Shenyang, China. The incidence of VF and the mortality and NAEs due to LLV were assessed. Cox proportional hazards regression was performed to investigate risk factors for VF, mortality, and NAEs.

RESULTS

In total, 1288 patients, contributing 3915 person-years of follow-up (median follow-up, 2.5 years [interquartile range: 2-4 years]), were enrolled. Thirty-one patients (2.4%) experienced VF, 5 (0.4%) died, and 38 (3.0%) experienced NAEs. The risk of VF was significantly increased among patients with a viral load (VL) of 200-499 copies/mL (adjusted hazard ratio [aHR]: 14.92, 95% confidence interval [CI]: 5.92-37.60) or 500-999 copies/mL (aHR: 13.68, 95% CI: 3.61-51.87), but not among patients with a VL of 50-199 copies/mL (aHR: 3.10, 95% CI: 0.86-11.09). The risk of NAEs was significantly increased among patients with LLV (aHR: 7.33, 95% CI: 3.73-14.42). Compared to no LLV, a VL of 50-199 copies/mL (aHR: 4.11, 95% CI: 1.73-9.74), 200-499 copies/mL (aHR: 18.31, 95% CI: 6.66-50.33), and 500-999 copies/mL (aHR: 21.34, 95% CI: 5.69-80.01) showed higher risk of NAEs.

CONCLUSION

Low-level viremia was associated with VF and NAEs. Patients with LLV, especially those with a VL ≥200 copies/mL, may need more frequent VL testing and NAE screening.

Late presentation for HIV remains a major health issue in Spain: Results from a multicenter cohort study, 2004-2018

OBJECTIVES

With the purpose of reducing the well-known negative impact of late presentation (LP) on people living with HIV (PLWH), guidelines on early HIV diagnosis were published in 2014 in Spain, but since then no data on LP prevalence have been published. To estimate prevalence and risk factors of LP and to evaluate their impact on the development of clinical outcomes in the Cohort of the Spanish HIV/AIDS Research Network (CoRIS) during 2004-2018.

METHODS

CoRIS is an open prospective multicenter cohort of PLWH, adults, naive to ART at entry. LP was defined as HIV diagnosis with CD4 count ≤350 cells/μL or an AIDS defining event (ADE). Multivariable Poisson regression models were used to estimate both prevalence ratios (PR) for the association of potential risk factors with LP and Incidence rate ratios (IRRs) for its impact on the development of the composite endpoint (first ADE, first serious non-AIDS event [SNAE] or overall mortality).

RESULTS

14,876 individuals were included. Overall, LP prevalence in 2004-2018 was 44.6%. Risk factors for LP included older age, having been infected through injection drug use or heterosexual intercourse, low educational level and originating from non-European countries. LP was associated with an increased risk of the composite endpoint (IRR: 1.34; 95%CI 1.20, 1.50), ADE (1.39; 1.18, 1.64), SNAE (1.22; 1.01, 1.47) and mortality (1.71; 1.41, 2.08).

CONCLUSIONS

LP remains a health problem in Spain, mainly among certain populations, and is associated with greater morbidity and mortality. Public policies should be implemented to expand screening and early diagnosis of HIV infection, for a focus on those at greatest risk of LP.

Methods of competing risks flexible parametric modeling for estimation of the risk of the first disease among HIV infected men

Background

Patients infected with the Human Immunodeficiency Virus (HIV) are susceptible to many diseases. In these patients, the occurrence of one disease alters the chance of contracting another. Under such circumstances, methods for competing risks are required. Recently, competing risks analyses in the scope of flexible parametric models have risen to address this requirement. These lesser-known analyses have considerable advantages over conventional methods.

Methods

Using data from Multi Centre AIDS Cohort Study (MACS), this paper reviews and applies methods of competing risks flexible parametric models to analyze the risk of the first disease (AIDS or non-AIDS) among HIV-infected patients. We compared two alternative subdistribution hazard flexible parametric models (SDH_FPM1 and SDH_FPM2) with the Fine & Gray model. To make a complete inference, we performed cause-specific hazard flexible parametric models for each event separately as well.

Results

Both SDH_FPM1 and SDH_FPM2 provided consistent results regarding the magnitude of coefficients and risk estimations compared with estimations obtained from the Fine & Gray model, However, competing risks flexible parametric models provided more efficient and smoother estimations for the baseline risks of the first disease. We found that age at HIV diagnosis indirectly affected the risk of AIDS as the first event by increasing the number of patients who experience a non-AIDS disease prior to AIDS among > 40 years. Other significant covariates had direct effects on the risks of AIDS and non-AIDS.

Discussion

The choice of an appropriate model depends on the research goals and computational challenges. The SDH_FPM1 models each event separately and requires calculating censoring weights which is time-consuming. In contrast, SDH_FPM2 models all events simultaneously and is more appropriate for large datasets, however, when the focus is on one particular event SDH_FPM1 is more preferable.

Study on Prediction Model of HIV Incidence Based on GRU Neural Network Optimized by MHPSO

Acquired Immune Deficiency Syndrome (AIDS) is still one of the most life-threatening diseases in the world. Moreover, new infections are still potentially increasing. This difficult problem must be solved. Early warning is the most effective way to solve this problem. Here, we aim to determine the best performing model to track the epidemic of AIDS, which will provide a methodological basis for testing the time characteristics of the disease. From January 2004 to January 2018, we built four computing methods based on AIDS dataset: BPNN model, RNN model, LSTM model and MHPSO-GRU model. Compare the final estimated performance to determine the preferred method. Result. Considering the root mean square error (RMSE), mean absolute error (MAE), mean error rate (MER) and mean absolute percentage error (MAPE) in the simulation and prediction subsets, the MHPSO-GRU model is determined as the best performance technology. Estimates for the period from May 2018 to December 2020 suggest that the event appears to continue to increase and remain high.

Evolving understanding of cardiovascular, cerebrovascular and peripheral arterial disease in people living with HIV and role of novel biomarkers. A study of the Spanish CoRIS cohort, 2004-2015

OBJECTIVES

To analyze the incidence rates (IR) and spectrum of vascular events in people living with HIV (PLWH) in Spain from 2004 to 2015. Serial measurements of different plasma cardiovascular biomarkers were assessed in relation to disease development.

METHODS

Longitudinal study in a nationwide contemporary multicenter cohort of PLWH. A nested case-control study was performed to evaluate the predictive value of cardiovascular biomarkers. Additive generalized and Cox mixed models were used for the analyses.

RESULTS

9,712 PLWH and 48,341 person-years of follow-up were analysed. During 2004-2015, 147 persons developed 154 vascular events; 80 (54.42%) coronary-related; 65 (44.22%) cerebrovascular-related, and 9 (6.12%) peripheral arterial disease. The 2004-2015 IR (95% confidence interval) of vascular events was 3.17 (2.69-3.71) x1,000 person-years; 1.64 (1.30-2.05) for coronary events; 1.34 (1.03-1.70) for cerebrovascular events; and 0.19 (0.09-0.35) for peripheral arterial disease (p<0.001). IR of vascular events gradually increased from 0.37 (0.12-0.85) x1,000 patient-years in the stratum 25-34-years to 19.65 (6.38-45.85) x1,000 patient-years in the stratum 75-84-years. Compared to the general population, there was a higher incidence of acute myocardial infarction (AMI) in men (sIR ratio 1.29 [95% CI 1.16-1.42]), of cerebrovascular events in women (sIR ratio 2.44 [95% CI 1.68-3.19]), and of both types of events specifically among the younger age-strata. CD4 count (hazard ratio 0.80, [95% CI, 0.79-0.81]), age (1.86 [1.47-2.34] for 45-65 years and 3.44 [2.37-4.97] for >65 years) and vascular event (1.81 [1.12-2.94]) were associated with total mortality. Adjusted levels of intercellular-adhesion-molecule (sICAM), pro-b-type-natriuretic-peptide (pro-BNP) and marginally sCD14, were higher among patients who subsequently developed vascular events.

CONCLUSION

Vascular events in PLWH do preferentially occur in the older age-strata, they are associated with increased mortality and, compared to the general population, the excess risk occurs at younger ages. Peripheral arterial disease is unusual. Vascular events are preceded by increased levels of sICAM, pro-BNP and, marginally, sCD14.