Oxidative stress markers show selective links to cancer risk in people with HIV

People living with HIV who receive antiretroviral therapy remain at elevated risk of cancers that are not directly caused by advanced immunodeficiency. A case-cohort study from the Spanish HIV Research Network has now tested whether common indicators of oxidative stress in blood can help identify that risk. The analysis, involving 332 people with HIV, found no clear overall association between baseline oxidative stress biomarkers and later non-AIDS-defining cancers, but it detected possible signals linked to sex and age that could guide future validation studies.

The researchers used a case-cohort design within the CoRIS cohort, drawing on participants who were receiving antiretroviral therapy. In this design, cancer cases are assembled from the wider cohort, while a comparison group is selected through stratified sampling. Statistical weights are then used so that the sample adequately represents the source population. The investigators analyzed incident non-AIDS-defining cancers, meaning malignancies other than the cancer categories classically associated with AIDS, and estimated cause-specific hazards with Borgan II-weighted Cox proportional hazards models. This approach allows the timing of a cancer event to be related to baseline biomarker measurements while accounting for the sampling structure of the case-cohort study.

Oxidative stress describes an imbalance in which reactive oxygen species and related products exceed the body’s capacity to neutralize them. These molecules can damage lipids, proteins and DNA, creating conditions that may favor abnormal cell growth. The study examined several plasma markers that capture different aspects of this process: 4-hydroxynonenal, commonly abbreviated 4-HNE; oxidized low-density lipoprotein, or ox-LDL; carbonyl proteins, which indicate oxidation of proteins; and 8-hydroxy-2’-deoxyguanosine, or 8-OHdG, a marker associated with oxidative DNA damage. By measuring several indicators together, the researchers could test whether information about lipid peroxidation, protein oxidation and genomic injury was related to subsequent cancer incidence.

The main result was cautious. Across the full study population, the oxidative stress biomarkers were not associated with non-AIDS-defining cancer risk in the overall analysis. That finding does not mean oxidative stress is irrelevant to cancer biology. Rather, it indicates that a single baseline plasma measurement did not provide a broadly reproducible signal across all participants. Oxidative stress is a dynamic process, and blood concentrations at one point may not capture cumulative exposure, tissue-specific damage or the effects of antiretroviral treatment, smoking, inflammation and other factors that influence both biomarker levels and cancer risk.

More pronounced patterns emerged when the researchers examined whether sex and age modified the relationships between biomarkers and cancer. Among females, higher baseline 4-HNE was associated with a higher hazard of developing a non-AIDS-defining cancer, with an adjusted hazard ratio of 2.79 and a 95% confidence interval of 1.14 to 6.80. The same association was not observed among males. This apparent sex-specific pattern is biologically plausible in principle because oxidative-stress responses can differ between women and men, but the researchers emphasized that it was exploratory. Only 15.4% of the study sample was female, so the estimate is vulnerable to imprecision and instability and requires confirmation in a larger cohort with more participants.

Age also appeared to modify the association between oxidative stress and cancer. Statistical interaction tests for 4-HNE and ox-LDL produced p-values below 0.050, indicating that the strength or direction of their relationships with cancer risk could differ across age groups. However, estimates restricted to adults aged 50 or older had wide confidence intervals and were not statistically significant. The interaction results therefore identify a hypothesis for further study rather than a definitive age-specific risk signal. The contrast between a statistically detectable interaction and imprecise estimates within the older subgroup illustrates why subgroup findings in a modest-sized cohort should not be treated as clinical evidence on their own.

The investigators also tested whether adding oxidative stress biomarkers could improve a clinical model for predicting non-AIDS-defining cancer among older adults. In the study sample, the model’s discrimination improved from a C-index of 0.863 to 0.877, a change associated with a p-value of 0.003. The C-index measures how well a model separates participants who experience the event of interest from those who do not, with higher values indicating better ranking. Internal fivefold cross-validation produced a smaller but directionally consistent gain: the out-of-fold C-index increased from 0.586 to 0.632. Cross-validation is important because a model evaluated only on the data used to develop it can appear more accurate than it will be in new populations.

Even so, the clinical implications remain limited. The results come from one Spanish cohort, the sample was relatively small, and only a minority of participants were women. The model evaluation was internal rather than external, meaning the researchers did not test performance in an independent cohort. Cancer risk also depends on factors that these biomarkers cannot replace, including age, smoking, viral suppression, treatment history, viral co-infections such as hepatitis B or C, and established clinical risk factors. The results are best interpreted as evidence that lipid-peroxidation markers may add information for selected groups, not as justification for routinely measuring 4-HNE or ox-LDL to guide cancer screening.

The study therefore offers a more nuanced picture of oxidative stress as a potential cancer marker in people with HIV. The broad hypothesis that baseline biomarkers could identify elevated non-AIDS-defining cancer risk was not supported at the population level, but exploratory analyses identified a stronger association between 4-HNE and cancer in females and evidence that age might alter the relationships of 4-HNE and ox-LDL with risk. The gains in model discrimination, including during cross-validation, suggest that this information could eventually complement established predictors. Larger, independently validated cohorts, with balanced representation by sex and sufficient numbers of older adults, are now needed to determine whether these signals are reproducible and clinically useful.

Subject of Research: Association of baseline plasma oxidative stress biomarkers with incident non-AIDS-defining cancer risk in people with HIV

Article Title: Association of baseline plasma oxidative stress biomarkers with incident non-AIDS-defining cancer risk in people with HIV: a case-cohort study

Article References: Coca-LaCalle, D., Pita-Martínez, C., García-Lumbreras, L., Rava, M., Sánchez-Villegas, J., Sánchez-Molina, A., Muga Bustamante, R., Alemán Valls, M. D. L. R., Morano Amado, L., Ayerbe, C. G., Jiménez-Sousa, M. Á., Resino, S., Martín-Escolano, R., & CoRIS cohort (2026). Association of baseline plasma oxidative stress biomarkers with incident non-AIDS-defining cancer risk in people with HIV: a case-cohort study. GeroScience. https://doi.org/10.1007/s11357-026-02613-0

Image Credits: AI Generated

DOI: 10.1007/s11357-026-02613-0

Keywords: HIV, antiretroviral therapy, non-AIDS-defining cancers, oxidative stress, 4-hydroxynonenal, oxidized LDL, biomarkers, cancer risk, sex differences, aging, Spain, CoRIS