From:  Janus-faced role of anti-infective drugs: a revisit through the lens of vascular ageing

 Risk-of-bias assessment of human studies (clinical trials and observational cohorts) examining anti-infective drugs in relation to vascular ageing.

Ref #Study typeDrug/ClassModel/PopulationTools usedKey bias domains (summary)Overall riskAgeing effects (biomarkers)
[11]ClinicalAntibiotics (mixed)Patients with vascular disease (meta-analysis)RoB-2Moderate heterogeneity; unclear blinding; limited reporting.Some concernsMixed outcomes; no consistent biomarker tracking; some ↓ CRP and vascular inflammation.
[136]ClinicalAzithromycin/RoxithromycinSenescent fibroblasts (clinical translation)RoB-2In vitro-clinical extrapolation; not randomized.High↓ SA-β-gal+ cells; ↓ p16/p21 expression in fibroblast biopsies.
[65]ClinicalAzithromycinCystic fibrosis patientsRoB-2Observational; small cohort; confounding not adjusted.High↑ Senescence markers in airway cells (p16, p21, SA-β-gal); systemic spillover likely.
[66]ClinicalAzithromycin (COVID-19)COVID-19 patientsRoB-2Not randomized; strong confounding; small sample.HighTheoretical ↓ SASP load proposed; not biomarker validated.
[67]ClinicalAzithromycinWomen with endometriosisRoB-2Randomization not reported; allocation concealment unclear.Some concerns↓ Lesion progression; signals of ↓ senescence markers (p16, SA-β-gal) in tissue.
[61]ClinicalVDA (Vit C, Doxycycline, Azithromycin)Human adipose-derived MSCsRoB-2Ex vivo/MSC work; randomization unclear.High↓ Senescence gene expression (p16, p21, SASP mediators) in MSCs.
[62]ClinicalDoxycyclineEndothelial cells from ethanol-exposed humansRoB-2Small cohorts; observational; incomplete adjustment.Some concerns↓ Inflammageing markers (IL-6, TNF-α); partial ↓ p16/p21 signals in endothelial cells.
[73]ClinicalART (Efavirenz/Zidovudine)HIV+ patients under ARTROBINS-IObservational; confounding by indication; attrition bias.Some concerns↑ ROS, mitochondrial dysfunction, endothelial dysfunction; no direct senescence markers.
[109]ClinicalProtease inhibitorsHIV+ patientsRoB-2Observational; moderate confounding; incomplete reporting.Some concerns↑ Senescence phenotype (reversible); mitochondrial dysfunction.
[111]ClinicalProtease inhibitorsHIV+ patients (vascular biopsies)RoB-2Non-randomized; mechanistic endpoints limited.High↑ Prelamin A accumulation; ↑ p16/p21; premature vascular senescence.
[112]ClinicalViral therapy contextHIV/Viral infections (review translational)NarrativeNot interventional; observational extrapolation.Some concerns↑ Senescence in the viral infection context; drug exposure exacerbates markers.
[114]ClinicalHAARTHIV+ patient endothelial cellsRoB-2Observational; incomplete adjustment; confounding.High↑ Oxidative stress; ↑ endothelial senescence; ↑ monocyte adhesion.
[113]ClinicalAntiviralsEndothelial cells from patientsRoB-2High-content imaging assays; small cohorts.Some concernsDistinct ↑ senescence signatures across drugs; drug-specific biomarker profiles.
[135]ClinicalNiclosamideAgeing/Frailty patients (clinical models)RoB-2Early-stage; randomization not clear; limited duration.Some concerns↓ mTORC1 hyperactivation; ↓ frailty index; no direct p16/p21/SA-β-gal measured.
[105]Human genomic analysis/observational mechanistic studyGanciclovir (antiviral nucleoside analog)Hematopoietic stem & progenitor cells (HSPCs) from HSCT recipients (n = 12 recipients; 15 expanded clones). Donors = matched controlsROBINS-ISelection bias: moderate (recipients selected based on post-transplant samples).
Confounding: moderate (immunosuppression, CMV infection, other antivirals).
Detection bias: low (whole-genome sequencing, validated pipelines).
Reporting bias: low (complete genomic datasets shared).
Attrition bias: low (all sequenced samples included).
ModerateSignificant increase in somatic mutation load in recipients exposed to ganciclovir.
Mutation burden increased 3–6-fold vs. non-exposed donor HSPCs.
Distinct antiviral-associated mutational signature (“Signature GX”) identified.
Mutations enriched in T>A and C>A substitutions, consistent with nucleoside-analog misincorporation.
Same signature observed in therapy-related cancers in transplant recipients.
It indicates genomic instability, a hallmark of aging.
[79]ClinicalMicrobiome-ageing/antibiotic exposureHuman gut microbiome cohortsROBINS-IObservational; confounding not excluded; no specific antibiotics mentioned.Some concerns↑ Inflammageing: altered microbiome linked to ↑ systemic senescence load.
[50]ClinicalMicrobiome-sarcopenia/nonspecific antibioticsAgeing frail patientsROBINS-IObservational; small cohorts/antibiotics were used, but no specific class was mentioned.Some concerns↑ Frailty; ↓ muscle mass; inflammageing.
[58]ClinicalAntibiotic exposureDysbiosis in the human gutROBINS-IObservational; no blinding.High↑ Dysbiosis; ↑ pro-inflammatory tone; indirect vascular ageing.
[59]ClinicalAntibiotics/Antiviral drugs (nonspecific)Elderly patients with bacterial infectionsROBINS-IObservational; moderate confounding.Some concerns↑ Susceptibility to infection; ↑ SASP cytokines; ↑ vascular inflammation.
[49]ClinicalGut microbiome dysbiosisCKD patientsROBINS-IObservational; no blinding; dietary confounding.Some concerns↑ TMAO; ↑ oxidative stress; ↑ arterial stiffness.
[75]ClinicalMicrobiome/CKDCKD patientsROBINS-IObservational; dietary confounding.Some concerns↑ TMAO; ↑ endothelial dysfunction.
[69]ClinicalTMAO/Nonspecific antibioticHuman cohortsROBINS-IObservational; dietary confounding and lack of mention of the employed antibiotic.Some concerns↑ TMAO associated with ↑ arterial stiffness, inflammaging.
[76]ClinicalBroad spectrum antibiotics/microbiota metabolitesHuman cohortsROBINS-IObservational; no blinding.Some concerns↑ SASP signaling; microbiota-derived metabolites drive vascular inflammation.
[104]ClinicalmiRNA mimics/inhibitorsCOVID-19 patientsRoB-2Small trials; confounding; not blinded.High↓ Viral-induced senescence signaling; ↓ cytokine storm; indirect vascular benefit.
[103]ClinicalLncRNA antifungalPatients with candidiasis (biopsy models)RoB-2Translational; not blinded.Some concerns↓ Fungal burden; ↓ inflammatory senescence response.
[108]ClinicalMicrobiome histone modification/antibiotic treatmentHuman hepatology patientsROBINS-IObservational; confounding not excluded/exact antibiotic not mentioned.Some concernsIrreversible histone acetylation; independent of SCFA; links to epigenetic ageing.
[82]ClinicalAntiviralsViral infection patientsRoB-2Non-randomized; confounding.High↑ Stress proteins; mitochondrial strain; senescence pathway engagement.
[143]ClinicalAntiviralsHIV+ patients (old regimens)ROBINS-IObservational; incomplete adjustment.Some concerns↑ Mitochondrial toxicity; ↑ oxidative stress; indirect senescence induction.
[127]ClinicalItraconazoleEndothelial cells from patientsRoB-2Mechanistic study; no blinding.Some concerns↓ mTORC1 activation; ↓ angiogenesis; senescence-modulating pathway engaged.
[128]ClinicalItraconazoleEndothelial cells (clinical setting)RoB-2Small cohorts; limited blinding.Some concerns↓ VEGFR2 glycosylation; ↓ angiogenesis; possible senescence-modulating pathway.
[129]ClinicalItraconazoleEndothelial angiogenesis (patients)RoB-2Observational, translational; confounding.Some concerns↓ Angiogenesis; modulation of senescence pathways.
[131]ClinicalAmphotericin BHuman endothelial cellsRoB-2Mechanistic; small sample.Some concerns↑ ROS; mitochondrial dysfunction; senescence induction.
[20]Retrospective cohort (observational)Multiple systemic antibiotics (all classes aggregated)2,159,864 adults, age 40–79, Korea; 10-year follow-upROBINS-IConfounding: Infection severity & comorbidities may influence antibiotic use; selection bias: possible immortal-time bias; exposure misclassification: antibiotic days aggregated, no dose/type specificity; outcome bias: relies on hospital records.Moderate, seriousLong-term antibiotic exposure (≥ 365 days) increased CVD risk: aHR = 1.10 (95% CI 1.07–1.13). No direct vascular aging biomarkers measured; evidence suggests epidemiological association, not mechanistic aging effects.
[21]Prospective cohort (children)Systemic antibiotics (various classes)5-year-old children from the WHISTLER birth cohort (Netherlands)ROBINS-IConfounding (infections vs. antibiotic use)
Exposure misclassification (prescription ≠ actual ingestion; no class/dose stratification).
Outcomes limited to subclinical vascular markers (CIMT, distensibility).
Loss to follow-up/selection bias (children with vascular measurements ≠ full cohort).
Moderate↑ CIMT: +18.1 µm (3-mo), +10.7 µm (6-mo).
↓ Carotid distensibility: –8.3 mPa–1 (6-mo).
[145]Human tissue observational + multiomics analysisMaraviroc (CCR5 antagonist)Human skeletal muscle from donors across the age spectrumROBINS-ITissue: skeletal muscle, not vascular tissue
• Senescence mapping in muscle may not reflect vascular cell senescence.
• Drug tested for senotherapeutic potential, but not in a vascular context.
• Cross-sectional design (aging donors), observational, not longitudinal.
Moderate, high risk/indirect relevanceIdentification of senescent cell populations; heterogenous SASP profiles; epigenomic alterations in aged muscle; evidence that blocking SASP-receptor interactions (via Maraviroc) can modulate senescence-associated pathology.
[100]Clinical/ObservationalAminoglycosides (geneticin/G418, hygromycin B, streptomycin); rifamycin (rifampicin); β-lactams (amoxicillin, ampicillin, cefepime, cefuroxime)Human-derived glioma cell lines and fibroblast Narrative In-vitro model; cancer cell lines; short exposure duration; no organism-level ageing outcomes; no randomisation/blinding applicable.Moderate↑ Alpha satellite DNA transcription (1.5–3-fold); ↓ H3K9me3; ↑ H3K18ac → heterochromatin erosion, epigenetic drift, genome instability (aging hallmarks).

RoB-2 was applied for interventional trials, and ROBINS-I for observational studies. Ageing biomarkers reported included SA-β-gal activity, p16/p21 expression, mitochondrial dysfunction, oxidative stress, prelamin A accumulation, TMAO levels, and frailty indices (n = 36). ↑: increased/upregulated/elevated, ↓: decreased/downregulated/reduced. ART: antiretroviral therapy; CIMT: carotid intima-media thickness; HAART: highly active ART; ROS: reactive oxygen species; SASP: senescence-associated secretory phenotype; SCFA: short-chain fatty acid; TMAO: trimethylamine N-oxide; mTORC1: mechanistic target of rapamycin complex 1.