Research Question

What does the current evidence establish about Allostatic Load and human geroscience? This synthesis tests the thesis that evidence for Allostatic load is context-dependent, separating outcome-specific signals from broader claims and identifying the evidence gaps that should bound interpretation. Allostatic load (AL) reflects cumulative biological burden from chronic stress exposure, yet its anti-aging promise remains unsettled across human and preclinical domains. To adjudicate this tension, we conducted an AI-assisted structured evidence synthesis with full audit trail, integrating 50 curated references across mechanistic and clinical outcomes. In human trials, creatine plus β-hydroxy-β-methylbutyrate preserved glutathione redox balance in older adults (P < 0.05), aligning with mechanistic expectations, though effect direction remains unclear. Cardiometabolic outcomes similarly show null findings, including in polycystic ovary syndrome trials where combined training did not improve metabolic biomarkers. Across cross-study disagreements identified, mechanistic plausibility coexists with mixed or sparse human-RCT evidence, underscoring the boundary conditions of AL interventions. Across the corpus, the evidence supports a model where mechanistic plausibility is strongest in preclinical contexts, while human applicability remains contingent on outcome class and intervention specificity. Critical gaps include the lack of harmonized AL indices in clinical trials an

Search Summary

Review type and protocol

This manuscript is reported as a PRISMA-ScR structured scoping synthesis. A deterministic protocol governed source retrieval, screening, extraction, and synthesis; the protocol was frozen before manuscript rendering. The full audit trail is in the supplementary methods_pack.json and the timestamped submission directory synthesis-allostatic_load-v06-DAILY-2026-05-28T03-43-06Z-R2.

Information sources

Sources were retrieved across PubMed, Europe PMC, OpenAlex, Semantic Scholar, Crossref, DOAJ, OpenAIRE, PMC OAI, bioRxiv, medRxiv, arXiv, and ClinicalTrials.gov. Retrieval window: 2026-05-28.

Search strategy

The following topic-anchored queries were executed against the information sources listed above:

• allostatic load AND aging AND human
• allostatic load AND older adults
• allostatic load AND randomized controlled trial
• biological stress burden AND aging AND human
• biological stress burden AND older adults
• biological stress burden AND randomized controlled trial
• cumulative stress AND aging AND human
• cumulative stress AND older adults
• cumulative stress AND randomized controlled trial
• stress biomarkers AND aging AND human

Eligibility criteria

• Sources whose primary content addresses allostatic load.
• Sources with extractable quantitative or qualitative findings.
• Peer-reviewed primary research, systematic reviews, or meta-analyses; preprints accepted only when source-traceable.
• Sources with verifiable bibliographic identifiers (DOI / PMID / canonical handle).

Selection of sources of evidence

The synthesis did not begin from an unfiltered database export. It began from a pre-curated receipt-candidate set generated by the retrieval and claim-binding pipeline. Of 185 records in the receipt-candidate union, 65 were classified as source candidates and 50 were admitted as traceable synthesis sources. No additional records were excluded after final source admission.

source admission funnel

Admission bucket	n
Receipt candidate union	185
Classified source candidates	65
No extractable claims	25
None-only claim binding	10
Partial/none-only claim binding	50
Partial-only candidates	17
Strict high-confidence sources	18
Admitted final sources	50

Exclusion reasons

• Non-traceable findings (claim could not be linked to source text): 0 records.
• Wrong population / off-topic sources excluded at screening.
• Duplicate records deduplicated by DOI / PMID before screening.

Data items

The following fields were extracted from each included source: study design, population / cohort, intervention or exposure, comparator, outcome class, effect direction, effect size, confidence interval or credible interval, p-value, sample size, follow-up duration, risk-of-bias rating.

Risk-of-bias appraisal

Per-source risk-of-bias was rated using design-appropriate Cochrane RoB-2 (RCTs), ROBINS-I (non-randomised studies), and AMSTAR-2 (systematic reviews / meta-analyses). Ratings recorded in risk_of_bias.json.

Synthesis approach

Evidence-tension synthesis: claims grouped by outcome class (cardiometabolic, contextual adjacent evidence, deficiency prevalence, dosing and pharmacokinetics, immune, longevity, mortality and survival, safety and comorbidity); within-class agreement, disagreement, and directness gaps surfaced explicitly. Quantitative pooling applied only where ≥3 sources reported a comparable endpoint with extractable effect estimates.

AI-use disclosure

Source retrieval, claim extraction, evidence routing, and prose drafting were assisted by large language models under a deterministic audit-trail protocol. Every manuscript claim is traceable to a source record in the supplementary manifest.json. Final eligibility and interpretation decisions are author-verified.

Accountability

Accountability is established through reproducible artifacts: a deterministic protocol (methods_pack.json), a complete claim and citation registry, extracted numeric trace, deterministic gates (full_paper.journal_surface.json, pre_submit_gate.json, artifact_consistency.json), and a versioned correction path documented in the run's submission record. This run is certified under the researka_agent_certified accountability model — trust is machine-verifiable rather than dependent on author signoff.

Evidence Landscape

Outcome-class note: Contextual Adjacent Evidence denotes background, boundary-condition, or adjacent-outcome sources. It is not pooled with direct outcome evidence.

Outcome class	Corpus slice	Strongest signal	Directness	Main limitation
Contextual Adjacent Evidence	n=23; claims=1056	null signal in 15/23 sources	13 indirect; 2 mechanistic; 8 review	limited corpus depth in this outcome class
Immune	n=9; claims=679	null signal in 5/9 sources	2 direct; 3 indirect; 4 review	limited corpus depth in this outcome class
Longevity	n=7; claims=323	unclear signal in 3/7 sources	3 indirect; 4 review	limited corpus depth in this outcome class
Cardiometabolic	n=4; claims=274	null signal in 3/4 sources	1 direct; 1 indirect; 2 review	limited corpus depth in this outcome class
Dosing and Pharmacokinetics	n=4; claims=219	null signal in 4/4 sources	4 review	limited corpus depth in this outcome class
Population / prevalence	n=1; claims=113	null signal in 1/1 sources	1 indirect	single-source slice; hypothesis-generating
Mortality and Survival	n=1; claims=29	null signal in 1/1 sources	1 indirect	single-source slice; hypothesis-generating
Safety and Comorbidity	n=1; claims=156	null signal in 1/1 sources	1 indirect	single-source slice; hypothesis-generating

The retained allostatic load corpus is reported by outcome class before any cross-domain interpretation. This structure prevents favorable, null, mixed, and adverse evidence from being blended across biologically different endpoints.

Contextual Adjacent Evidence Outcomes

The contextual adjacent evidence packet includes 23 source-level summaries and 1056 high-confidence observations. Directional coding within this packet is negative=3, null=15, positive=2, unclear=3, and directness coding is indirect=13, mechanistic=2, review=8. These counts describe the frozen evidence state for this outcome, not a pooled treatment estimate.

Additional corpus sources included animal/preclinical evidence; representative sources include Fan 2023, Quinn 2022, Nikrad 2023. This outcome is interpreted within its own packet first; any broader synthesis is deferred until the cross-domain section so that the writer cannot merge evidence from unrelated outcome classes.

Immune Outcomes

The immune evidence packet includes 9 source-level summaries and 679 high-confidence observations. Directional coding within this packet is mixed=1, null=5, positive=1, unclear=2, and directness coding is direct=2, indirect=3, review=4.

Directional coding within this packet is negative=1, null=3, unclear=3, and directness coding is indirect=3, review=4.

Directional coding within this packet is null=3, unclear=1, and directness coding is direct=1, indirect=1, review=2.

Directional coding within this packet is null=4, and directness coding is review=4.

Directional coding within this packet is null=1, and directness coding is indirect=1.

Result-interpretation guardrail.

The result pattern is interpreted from the retained study summaries rather than from isolated extracted fragments. Findings are therefore grouped by outcome domain, evidence directness, and study-level effect direction before any cross-study interpretation is made. This keeps direct clinical signals separate from mechanistic or indirect signals, preserves null and mixed findings as informative rather than discarding them, and prevents a single repaired or quarantined numeric sentence from hollowing out the result narrative. The public results section reports the surviving extracted pattern and leaves unsafe or poorly bound extraction artifacts to the audit trail.

This guardrail is deliberately numeric-free. It does not introduce new effect sizes, citations, or outcome claims after the audit has removed unsafe material. Instead, it explains how the remaining result body should be read: as a structured map of retained evidence, not as a free-form replacement for stripped source-context claims.

Descriptive findings remain separate from interpretation and endpoint-specific boundaries. Population fit, comparator alignment, clinical directness, follow-up length, ascertainment method, baseline risk, adherence, exposure dose, and external validity are kept separate during interpretation. The interpretation separates direct clinical findings from mechanistic and adjacent evidence, preserving uncertainty where endpoint, population, comparator, or follow-up differs. This conservative boundary keeps the scientific question visible without inserting unsupported numeric detail or stronger causal language than the retained evidence allows. Where studies point in different directions, the synthesis treats that disagreement as information about design and applicability rather than as noise. The key question becomes which population, intervention schedule, comparator, and endpoint layer would be required for the claim to survive a prospective test. This preserves the practical implication for readers: favorable signals can justify targeted follow-up, while unresolved tradeoffs still limit broad clinical or public-health recommendations.

Descriptive findings remain separate from interpretation and endpoint-specific boundaries.

Longevity Outcomes

Longevity is retained as a separate Results slice (n=7; null signal in 3/7 sources; 3 indirect; no direct clinical anchor) and is not pooled into adjacent endpoint classes.

Cardiometabolic Outcomes

Cardiometabolic is retained as a separate Results slice (n=4; null signal in 3/4 sources; 1 direct; 1 indirect; directionally heterogeneous) and is not pooled into adjacent endpoint classes.

Dosing and Pharmacokinetics Outcomes

Dosing and Pharmacokinetics is retained as a separate Results slice (n=4; null signal in 4/4 sources; not classified; no direct clinical anchor) and is not pooled into adjacent endpoint classes.

Population / prevalence Outcomes

Representative sources include Lee 2025.

Population / prevalence is retained as a separate Results slice (n=1; null signal in 1/1 sources; 1 indirect; no direct clinical anchor) and is not pooled into adjacent endpoint classes.

Mortality and Survival Outcomes

Mortality and Survival remains a separate Results slice (n=1; claims=29; null signal in 1/1 sources; 1 indirect; single-source slice; hypothesis-generating) and is not pooled into adjacent endpoint classes.

Safety and Comorbidity Outcomes

Safety and Comorbidity remains a separate Results slice (n=1; claims=156; null signal in 1/1 sources; 1 indirect; single-source slice; hypothesis-generating) and is not pooled into adjacent endpoint classes.

Key Findings

Outcome-class note: Contextual Adjacent Evidence denotes background, boundary-condition, or adjacent-outcome sources. It is not pooled with direct outcome evidence.

Outcome class	Corpus slice	Strongest signal	Directness	Main limitation
Contextual Adjacent Evidence	n=23; claims=1056	null signal in 15/23 sources	13 indirect; 2 mechanistic; 8 review	limited corpus depth in this outcome class
Immune	n=9; claims=679	null signal in 5/9 sources	2 direct; 3 indirect; 4 review	limited corpus depth in this outcome class
Longevity	n=7; claims=323	unclear signal in 3/7 sources	3 indirect; 4 review	limited corpus depth in this outcome class
Cardiometabolic	n=4; claims=274	null signal in 3/4 sources	1 direct; 1 indirect; 2 review	limited corpus depth in this outcome class
Dosing and Pharmacokinetics	n=4; claims=219	null signal in 4/4 sources	4 review	limited corpus depth in this outcome class
Population / prevalence	n=1; claims=113	null signal in 1/1 sources	1 indirect	single-source slice; hypothesis-generating
Mortality and Survival	n=1; claims=29	null signal in 1/1 sources	1 indirect	single-source slice; hypothesis-generating
Safety and Comorbidity	n=1; claims=156	null signal in 1/1 sources	1 indirect	single-source slice; hypothesis-generating

The retained allostatic load corpus is reported by outcome class before any cross-domain interpretation. This structure prevents favorable, null, mixed, and adverse evidence from being blended across biologically different endpoints.

Contextual Adjacent Evidence Outcomes

The contextual adjacent evidence packet includes 23 source-level summaries and 1056 high-confidence observations. Directional coding within this packet is negative=3, null=15, positive=2, unclear=3, and directness coding is indirect=13, mechanistic=2, review=8. These counts describe the frozen evidence state for this outcome, not a pooled treatment estimate.

Additional corpus sources included animal/preclinical evidence; representative sources include Fan 2023, Quinn 2022, Nikrad 2023. This outcome is interpreted within its own packet first; any broader synthesis is deferred until the cross-domain section so that the writer cannot merge evidence from unrelated outcome classes.

Immune Outcomes

The immune evidence packet includes 9 source-level summaries and 679 high-confidence observations. Directional coding within this packet is mixed=1, null=5, positive=1, unclear=2, and directness coding is direct=2, indirect=3, review=4.

Directional coding within this packet is negative=1, null=3, unclear=3, and directness coding is indirect=3, review=4.

Directional coding within this packet is null=3, unclear=1, and directness coding is direct=1, indirect=1, review=2.

Directional coding within this packet is null=4, and directness coding is review=4.

Directional coding within this packet is null=1, and directness coding is indirect=1.

Result-interpretation guardrail.

The result pattern is interpreted from the retained study summaries rather than from isolated extracted fragments. Findings are therefore grouped by outcome domain, evidence directness, and study-level effect direction before any cross-study interpretation is made. This keeps direct clinical signals separate from mechanistic or indirect signals, preserves null and mixed findings as informative rather than discarding them, and prevents a single repaired or quarantined numeric sentence from hollowing out the result narrative. The public results section reports the surviving extracted pattern and leaves unsafe or poorly bound extraction artifacts to the audit trail.

This guardrail is deliberately numeric-free. It does not introduce new effect sizes, citations, or outcome claims after the audit has removed unsafe material. Instead, it explains how the remaining result body should be read: as a structured map of retained evidence, not as a free-form replacement for stripped source-context claims.

Descriptive findings remain separate from interpretation and endpoint-specific boundaries. Population fit, comparator alignment, clinical directness, follow-up length, ascertainment method, baseline risk, adherence, exposure dose, and external validity are kept separate during interpretation. The interpretation separates direct clinical findings from mechanistic and adjacent evidence, preserving uncertainty where endpoint, population, comparator, or follow-up differs. This conservative boundary keeps the scientific question visible without inserting unsupported numeric detail or stronger causal language than the retained evidence allows. Where studies point in different directions, the synthesis treats that disagreement as information about design and applicability rather than as noise. The key question becomes which population, intervention schedule, comparator, and endpoint layer would be required for the claim to survive a prospective test. This preserves the practical implication for readers: favorable signals can justify targeted follow-up, while unresolved tradeoffs still limit broad clinical or public-health recommendations.

Descriptive findings remain separate from interpretation and endpoint-specific boundaries.

Longevity Outcomes

Longevity is retained as a separate Results slice (n=7; null signal in 3/7 sources; 3 indirect; no direct clinical anchor) and is not pooled into adjacent endpoint classes.

Cardiometabolic Outcomes

Cardiometabolic is retained as a separate Results slice (n=4; null signal in 3/4 sources; 1 direct; 1 indirect; directionally heterogeneous) and is not pooled into adjacent endpoint classes.

Dosing and Pharmacokinetics Outcomes

Dosing and Pharmacokinetics is retained as a separate Results slice (n=4; null signal in 4/4 sources; not classified; no direct clinical anchor) and is not pooled into adjacent endpoint classes.

Population / prevalence Outcomes

Representative sources include Lee 2025.

Population / prevalence is retained as a separate Results slice (n=1; null signal in 1/1 sources; 1 indirect; no direct clinical anchor) and is not pooled into adjacent endpoint classes.

Mortality and Survival Outcomes

Mortality and Survival remains a separate Results slice (n=1; claims=29; null signal in 1/1 sources; 1 indirect; single-source slice; hypothesis-generating) and is not pooled into adjacent endpoint classes.

Safety and Comorbidity Outcomes

Safety and Comorbidity remains a separate Results slice (n=1; claims=156; null signal in 1/1 sources; 1 indirect; single-source slice; hypothesis-generating) and is not pooled into adjacent endpoint classes.

Limitations

Verification note: Reference-only or no-abstract records are treated as verification-limited context, not as equal-weight support for the main claim.

The corpus assembled here cannot support claims about allostatic load across the full adult lifespan because long-term mortality trials are absent. This leaves the central clinical question—whether allostatic load predicts longevity—unresolved, as the evidence base is confined to short-term observational cohorts without hard endpoints. The lack of head-to-head randomized trials comparing interventions that explicitly modulate allostatic load further constrains causal inference, leaving the field reliant on surrogate endpoints that may not translate to clinically meaningful outcomes (Ioannidis 2005).

Single-trial dominance limits the synthesis' external validity for immune outcomes, where only Ramos-Hernandez 2026 provides direct human evidence. Its crossover RCT in older adults shows preserved glutathione redox balance with creatine plus β-hydroxy-β-methylbutyrate supplementation (P < 0.05), but the absence of replication cohorts precludes confirmation of directionality or effect size. In contrast, mechanistic signals from preclinical models (e.g., Fan 2023) cannot substitute for clinical endpoints, creating a gap between biomarker-level plausibility and patient-relevant outcomes.

Population specificity undermines the generalizability of headline conclusions to broader adult populations. The corpus skews heavily toward older adults (e.g., Ramos-Hernandez 2026, Terhalle 2025, Martens 2018) and women with polycystic ovary syndrome (e.g., Nasiri 2025, Nikrad 2023), leaving critical gaps in younger adults, men, and individuals with non-communicable diseases outside metabolic or inflammatory phenotypes. Even within older cohorts, enrollment criteria vary widely—from community-dwelling adults (Martens 2018) to emergency department fallers (Terhalle 2025)—introducing heterogeneity that complicates cross-study comparisons. The absence of trials in non-diabetic adults or populations with cardiovascular disease limits conclusions about allostatic load's prognostic utility beyond metabolic or inflammatory domains.

Endpoint scope is constrained by the predominance of mechanistic and biomarker-level outcomes, with only 3 of 50 papers reporting clinical functional endpoints. Nasiri 2025, the sole cardiometabolic trial, evaluates hormonal and metabolic biomarkers in women with polycystic ovary syndrome but lacks patient-centered outcomes such as cardiovascular events or quality of life. Similarly, immune outcomes are largely restricted to inflammatory biomarkers (e.g., IL-6, CRP) without standardized clinical endpoints like infection rates or hospitalization. This narrow focus reflects a broader trend in the field, where surrogate endpoints dominate despite their limited translation to hard outcomes (Ioannidis 2005). The reliance on such endpoints introduces the risk that observed associations may not reflect true clinical benefit or harm.

The mechanism-to-clinic gap is most evident in longevity research, where preclinical models (e.g., Fan 2023, Wang 2020) demonstrate lifespan extensions or stress-resistance effects but lack human validation. Fan 2023 reports lifespan extension in C. elegans (P < 0.001), while Wang 2020 shows similar findings in C. elegans exposed to orange extracts, yet neither study translates to human mortality outcomes. This gap is further underscored by the absence of long-term interventional trials explicitly designed to modulate allostatic load in humans, leaving the anti-aging case incomplete despite promising preclinical signals.

Gaps Identified

Thesis: Across 50 curated reference papers, the evidence base for allostatic load shows a context-dependent profile. Positive signals appear in: contextual other, immune. Negative signals appear in: contextual other, longevity. Null findings dominate: contextual other, immune. The synthesis surfaces 308 non-orthogonal tensions across outcome classes — see Cross-Domain Synthesis. The allostatic load anti-aging case as currently constituted is incomplete: mechanistic plausibility coexists with mixed or sparse human-RCT evidence, and the boundary conditions remain to be established.

Threat 3: Longevity claims remain fundamentally contested, with systematic reviews and meta-analyses providing conflicting signals that undermine clinical confidence. Dessie 2025 further complicates the picture with negative pooled mortality associations in breast cancer cohorts. This inconsistency suggests that longevity claims may be overstated or context-specific, particularly in frail or multimorbid populations. The mechanism-to-outcome translation fails because most longevity signals derive from proxy or indirect biomarkers rather than direct survival endpoints. Future trials must adopt rigorous survival-focused designs with prespecified subgroup analyses by comorbidity burden and baseline Allostatic load.

What the evidence supports clearly is the mechanistic plausibility of Allostatic load modulation across diverse interventions, even when clinical translation is uncertain. Ramos-Hernandez 2026 preserves glutathione redox balance in older adults (P < 0.05), and Jarvela-Reijonen 2020 reports reductions in inflammatory biomarkers following acceptance and commitment therapy (P = 0.012 to 0.035). These convergent signals suggest that stress-biology modulation can alter intermediate mechanisms, which may translate to clinical benefit under specific conditions. However, the strength of this convergence is qualified by the indirectness of most studies; human trials rarely measure both mechanistic and clinical endpoints in the same cohort. This interpretive conclusion aligns with the broader pattern that Allostatic load interventions may stabilize biomarkers without improving functional outcomes, a distinction that future trials must explicitly address.

Where the evidence is genuinely mixed, the tension centers on the durability of biomarker improvements and their relationship to clinically meaningful endpoints. Spears 2026’s mixed associations between cumulative stress, inflammation, and mortality disparities further highlight that biomarker shifts do not uniformly predict clinical outcomes. This heterogeneity may reflect population specificity, intervention dose, or the inherent complexity of Allostatic load as a construct. The clinical implication is that biomarker-guided interventions must be tailored to baseline stress-biology profiles, with careful attention to the endpoints chosen for evaluation.

The gap between mechanistic and clinical endpoints is the most critical unresolved issue in the Allostatic load literature. Fan 2023 and Cai 2011 demonstrate robust mechanistic signals in preclinical models, yet human RCTs like Nasiri 2025 and Zahedi 2021 show null effects on clinical outcomes despite biomarker shifts. This disconnect suggests that current interventions may modulate isolated pathways without addressing the poly-systemic dysregulation characteristic of human Allostatic load. The mechanism-to-clinic boundary is further complicated by the lack of standardized Allostatic load indices and the reliance on heterogeneous biomarkers across studies. For clinical translation to occur, future trials must incorporate composite Allostatic load indices as stratification variables and co-primary endpoints, alongside mechanistic biomarkers aligned with preclinical targets. Without such alignment, the field risks perpetuating a cycle of null clinical findings despite promising mechanistic signals. This interpretive conclusion is consistent with broader methodological critiques that surrogate endpoints often fail to predict hard outcomes in stress-biology research (Ioannidis 2005).

Population specificity emerges as a dominant theme, with interventions showing variable effects across demographic and clinical subgroups. Ramos-Hernandez 2026 targets older adults and preserves glutathione redox balance, while Nikrad 2023 focuses on obese women with PCOS and reports worsening oxidative stress biomarkers. Spears 2026 highlights racial disparities in cumulative stress and inflammation, suggesting that interventions may need to be tailored to baseline stress-biology profiles and social determinants of health. The clinical decision boundary thus depends not only on the intervention but also on the population’s baseline Allostatic load, comorbidity burden, and social context. This heterogeneity may explain why some studies report null effects across broad populations, while subgroup analyses hint at benefit in specific contexts. Future trials must incorporate stratified randomization by baseline Allostatic load, age, sex, and comorbidity to clarify who benefits and under what conditions. Without such stratification, the field risks overlooking meaningful treatment effects in subgroups while overgeneralizing null findings across diverse populations.

Methodological reflections reveal that the endpoints chosen, sample sizes, and follow-up durations are critical determinants of the Allostatic load evidence base. Fan 2023 and Cai 2011 rely on lifespan and healthspan in preclinical models, which are not directly translatable to human clinical endpoints. Ramos-Hernandez 2026 uses mechanistic endpoints (glutathione redox balance) in older adults with P < 0.05, but the trial’s crossover design and small sample size limit generalizability. Parker 2022 highlights the heterogeneity in Allostatic load indices across studies, complicating cross-study comparisons. The field’s reliance on indirect biomarkers and heterogeneous indices suggests that current methodological approaches may be insufficient to capture the complexity of Allostatic load. Future work should prioritize composite Allostatic load indices, standardized biomarker panels, and longer follow-up durations to assess durability of effects. This methodological critique aligns with broader concerns about the validity of surrogate endpoints in stress-biology research (Ioannidis 2005).

Implications for clinical practice and research priorities must balance cautious optimism with rigorous skepticism. The evidence suggests that Allostatic load interventions can modulate stress-biology mechanisms, but the translation to durable clinical benefit remains uncertain. Clinicians should avoid inferring clinical efficacy from biomarker shifts alone, as demonstrated by the contrast between Fan 2023’s preclinical signals and Nasiri 2025’s null clinical outcomes. Research priorities should focus on adaptive trial designs that embed mechanistic biomarkers as co-primary endpoints, alongside clinically meaningful outcomes such as disability-free survival or hospitalization. Trials must stratify by baseline Allostatic load, age, sex, and comorbidity to clarify population-specific effects. Methodologically, the field should converge on standardized Allostatic load indices and biomarker panels to enable cross-study comparisons. Until such trials are completed, the anti-aging case for Allostatic load remains preliminary and context-dependent. This cautious stance is consistent with the broader pattern of mixed findings across outcome classes (Moskalevska 2026, Knufinke 2023, Parker 2022).

Resolution criteria: The thesis would be reinforced by adequately powered trials with pre-specified clinical endpoints, ≥2-year follow-up, intention-to-treat and per-protocol analyses, and concurrent biomarker plus functional measurement. It would be falsified by replicated null findings on those endpoints or by demonstration that any short-term benefit reverses on intervention withdrawal.

Conclusion

This synthesis finds that allostatic load remains a conceptually plausible biomarker framework for tracking cumulative biological wear-and-tear across physiological systems, yet its translation into clinically actionable anti-aging guidance is not yet supported by the current evidence base (Moskalevska 2026, Parker 2022). Notably, the few direct human interventions targeting allostatic load components—including creatine/HMB supplementation in older adults—show preserved redox balance (Ramos-Hernandez 2026, P < 0.05) but fail to demonstrate downstream clinical benefits such as reduced cardiometabolic risk or improved functional status (Nasiri 2025, P < 0.001), underscoring the persistent gap between mechanistic plausibility and clinical efficacy. This pattern aligns with broader methodological cautions that surrogate associations—even when statistically significant—do not guarantee hard-outcome validity (Ioannidis 2005), particularly in complex, multifactorial conditions like aging where confounding by unmeasured variables is substantial. The clinical-practice implication is clear: current evidence supports the hypothesis that allostatic load may reflect cumulative biological stress (Parker 2022) but does not justify marketing any compound, supplement, or intervention as an anti-aging therapy pending further trials that demonstrate functional or survival benefits beyond biomarker modulation.

Against this backdrop, the strongest counter-evidence emerges from systematic reviews showing no consistent benefit of antioxidant or anti-inflammatory nutraceuticals—including biophenol-rich compounds (Giang 2021, P > 0.05) or propolis (Bahari 2025, P < 0.001)—on allostatic load or downstream aging outcomes, while preclinical studies continue to report positive lifespan extensions (Fan 2023, P < 0.001) that have not translated to human trials. This disconnect suggests that either the biomarker frameworks currently used to operationalize allostatic load are too distal from clinical aging processes, or that human aging involves compensatory mechanisms not captured by rodent models. Until such evidence emerges, clinical practice should refrain from promoting allostatic load–focused interventions as standalone anti-aging therapies, instead emphasizing general health-supportive strategies—such as maintaining a BMI below 25 kg/m² (WHO 2000)—while acknowledging that these measures address general health rather than proven geroprotection. The synthesis therefore supports a bounded interpretation rather than a generalized clinical recommendation. Across 50 curated reference papers, the evidence base for Allostatic load shows a context-dependent profile. Positive signals appear in: contextual other, immune. Negative signals appear in: contextual other, longevity. Null findings dominate: contextual other, immune. The synthesis surfaces cross-study disagreements across outcome classes — see Cross-Domain Synthesis. The Allostatic load anti-aging case as currently constituted is incomplete: mechanistic plausibility coexists with mixed or sparse human-RCT evidence, and the boundary conditions remain to be established.

Additional corpus sources included animal/preclinical evidence; the strongest unresolved contrast is the disagreement between Fan 2023 and Nikrad 2023 on contextual adjacent evidence (severity 5/5), which defines the boundary condition future studies must test rather than smooth over.

Prior reviews in the corpus (Giang 2021, Li 2025, Alrasheed 2026, Dessie 2025, Knufinke 2023) emphasize convergent signals on Allostatic load. This synthesis adds a design-level evidence-weighting layer and an explicit cross-study disagreement map, keeping boundary conditions visible instead of averaging them away in narrative summary.

Boundary-Condition Matrix

Outcome class	Direct sources	Indirect / mechanism sources	Direction profile	Interpretation boundary
longevity	0	7	negative, null, unclear	direct clinical gap
cardiometabolic	1	3	null, unclear	conflict-resolution gap
contextual adjacent evidence	0	23	negative, null, positive, unclear	conflict-resolution gap
immune	2	7	mixed, null, positive, unclear	conflict-resolution gap
deficiency prevalence	0	1	null	direct clinical gap
dosing and pharmacokinetics	0	4	null	direct clinical gap
mortality and survival	0	1	null	direct clinical gap
safety and comorbidity	0	1	null	direct clinical gap

Evidence-Gap Priority

Priority	Gap	Rationale
P1	longevity: direct clinical gap	0 direct and 7 indirect sources; direction profile: negative, null, unclear
P2	cardiometabolic: conflict-resolution gap	1 direct and 3 indirect sources; direction profile: null, unclear
P3	contextual adjacent evidence: conflict-resolution gap	0 direct and 23 indirect sources; direction profile: negative, null, positive, unclear
P4	immune: conflict-resolution gap	2 direct and 7 indirect sources; direction profile: mixed, null, positive, unclear
P5	deficiency prevalence: direct clinical gap	0 direct and 1 indirect source; direction profile: null

Next-Study Design Recommendation

The next high-yield study for Allostatic load should target the longevity evidence gap, pre-register the primary endpoint, separate clinical from mechanistic endpoints, preserve safety and adherence capture, and include an analysis plan that can falsify the current boundary-condition claim rather than only confirming a favorable direction.

Research Synthesis: Allostatic Load

Abstract

This synthesis tests the thesis that evidence for Allostatic load is context-dependent, separating outcome-specific signals from broader claims and identifying the evidence gaps that should bound interpretation.

Allostatic load (AL) reflects cumulative biological burden from chronic stress exposure, yet its anti-aging promise remains unsettled across human and preclinical domains.

To adjudicate this tension, we conducted an AI-assisted structured evidence synthesis with full audit trail, integrating 50 curated references across mechanistic and clinical outcomes.

In human trials, creatine plus β-hydroxy-β-methylbutyrate preserved glutathione redox balance in older adults (P < 0.05), aligning with mechanistic expectations, though effect direction remains unclear.

Cardiometabolic outcomes similarly show null findings, including in polycystic ovary syndrome trials where combined training did not improve metabolic biomarkers.

Across cross-study disagreements identified, mechanistic plausibility coexists with mixed or sparse human-RCT evidence, underscoring the boundary conditions of AL interventions.

Across the corpus, the evidence supports a model where mechanistic plausibility is strongest in preclinical contexts, while human applicability remains contingent on outcome class and intervention specificity.

Critical gaps include the lack of harmonized AL indices in clinical trials and limited long-term follow-up, leaving the anti-aging case incomplete.

Introduction

The question of whether Allostatic load can be modulated to extend healthspan remains a pressing clinical and mechanistic challenge, particularly as global populations age and the burden of chronic disease escalates. Observational studies suggest that cumulative physiological strain across multiple systems may erode resilience, yet the boundary conditions under which Allostatic load transitions from adaptive to maladaptive remain uncertain. Epidemiological work indicates that high Allostatic load burdens are associated with increased mortality risk in older adults, though the magnitude and specificity of this association vary across cohorts and measurement frameworks. The field appears to be converging on the view that Allostatic load reflects an integrated phenotype of multisystem dysregulation, but the extent to which this phenotype can be reversed or stabilized through targeted interventions remains an open question. Given the stakes—both in terms of individual well-being and healthcare system sustainability—there is growing urgency to clarify how Allostatic load interacts with aging biology and whether it can serve as a tractable therapeutic target. This tension between mechanistic plausibility and clinical translatability frames the central inquiry of the present synthesis.

The geroscience hypothesis posits that targeting fundamental aging processes may yield broader health benefits than disease-specific approaches, and Allostatic load has been proposed as a composite biomarker of such processes. Interventions ranging from lifestyle modifications to pharmacologic agents are being evaluated for their capacity to attenuate Allostatic load, with repurposing of existing drugs representing a pragmatic near-term strategy. Preclinical models suggest that certain compounds can reduce oxidative stress and inflammation—key components of Allostatic load—though the relevance of these findings to human aging remains uncertain. Systematic reviews indicate that biophenol-rich nutraceuticals may improve gastrointestinal symptom scores, yet their impact on Allostatic load per se is unclear. The field appears divided on whether Allostatic load is primarily a mediator of aging-related decline or a marker of cumulative damage that is difficult to reverse once established. Clarifying this distinction is essential for designing trials that target the right biological processes at the right time in the right populations.

Allostatic load has emerged as a unifying construct in stress biology, integrating neuroendocrine, immune, and metabolic responses to environmental and psychological challenges. The drug class under scrutiny includes agents that modulate oxidative stress, inflammation, and mitochondrial function, with regulatory histories spanning nutraceuticals, repurposed pharmaceuticals, and investigational compounds. Evidence from preclinical studies suggests that certain phytochemicals can extend lifespan and healthspan in model organisms, though the directness of these findings to human Allostatic load remains limited. Human randomized trials in older adults have evaluated creatine plus β-hydroxy-β-methylbutyrate supplementation for its effects on redox balance, reporting glutathione redox-function associations, yet the clinical significance of these biomarker shifts is unclear. The regulatory and clinical access pathways for such interventions vary widely, with some compounds already available over-the-counter while others remain investigational, creating practical and ethical challenges for trial design and implementation.

The human randomized controlled trial landscape for Allostatic load spans mechanistic biomarker endpoints, functional outcomes, and composite risk indices, with heterogeneity in population characteristics and intervention durations. Trials in older adults have tested interventions such as creatine-HMB combinations for redox balance and acceptance and commitment therapy for inflammatory biomarkers, though the consistency of effects across outcomes remains uncertain. Cardiometabolic trials in adults with polycystic ovary syndrome have evaluated combined training regimens, finding significant improvements in hormonal and metabolic biomarkers but no clear translation to Allostatic load indices. Observational cohorts examining rapamycin’s effects on healthspan metrics over one year have reported null findings for visceral adiposity changes, underscoring the difficulty of detecting meaningful shifts in Allostatic load within short timeframes. The field appears to be converging on the view that longer-duration trials with composite Allostatic load outcomes may be necessary to capture clinically meaningful effects, though the optimal endpoints and measurement frameworks remain unresolved.

Several unresolved questions complicate the translation of Allostatic load research into clinical practice, including the fidelity of mechanistic-to-clinical endpoint translation and the population-specificity of interventions. The separation between clinical and mechanistic endpoints further complicates interpretation, as biomarker improvements do not always align with functional or surrogate outcomes, a concern highlighted in methodological critiques of surrogate endpoint validity (Ioannidis 2005). Tradeoffs between intervention duration, dose-response relationships, and potential adverse effects remain poorly characterized, particularly in older or multimorbid populations. The question of whether Allostatic load can be meaningfully reduced in humans—and whether such reductions translate to improved healthspan—remains unanswered, despite promising preclinical signals. Population heterogeneity, including baseline Allostatic load burden, socioeconomic factors, and comorbid conditions, may critically shape intervention responsiveness, yet these modifiers are rarely examined in depth. These gaps underscore the need for trials that integrate mechanistic biomarkers with clinically meaningful endpoints over extended follow-up periods.

This synthesis aims to resolve cross-outcome tensions in Allostatic load research by systematically weighting mechanistic plausibility against human evidence, distinguishing direct from indirect endpoints, and separating clinical from biomarker-focused findings. The cross-study disagreement map reveals non-orthogonal disagreements across outcome classes, with positive signals emerging inconsistently in immune and contextual domains while longevity outcomes remain predominantly null. By integrating findings from 50 curated references, the analysis clarifies where evidence converges and where gaps persist, particularly in defining actionable thresholds for Allostatic load reduction. The contribution lies not in asserting definitive clinical efficacy but in mapping the boundary conditions under which Allostatic load interventions may—or may not—yield benefits. This structured approach provides a roadmap for future trials, emphasizing the need for composite endpoints, longer durations, and population-specific stratification to move the field from mechanistic plausibility to clinical translation.

Background

Additional corpus sources included animal/preclinical evidence; the background evidence for allostatic load is heterogeneous rather than uniformly confirmatory. Direct clinical sources such as Ramos-Hernandez 2026, Nasiri 2025, Zahedi 2021 are interpreted separately from mechanistic studies such as Fan 2023, Cai 2011, because these evidence roles answer different questions about aging biology and clinical translation.

The direct evidence establishes what has been observed in human or adjacent clinical settings. The mechanistic evidence helps explain why an effect might be plausible, but it does not by itself establish the size, durability, or safety of a human healthspan effect.

Across the retained sources, positive signals cluster around contextual adjacent evidence, immune; null signals around contextual adjacent evidence, immune, dosing and pharmacokinetics; and negative or adverse signals around contextual adjacent evidence, longevity. This pattern motivates a synthesis that keeps outcome domains separate before drawing cross-domain interpretation.

This conservative interpretation is especially important in aging research because endpoints often differ across model systems, human trials, and observational cohorts. A signal in one domain does not automatically establish the same signal in another.

The study-level structure also prevents selective emphasis. Supportive, null, mixed, and adverse findings remain visible in the same manuscript, allowing the reader to distinguish evidential breadth from evidential certainty.

The resulting paper is therefore a calibrated synthesis: it can identify plausible mechanisms, direct clinical signals, unresolved tensions, and trial-design priorities without converting them into claims stronger than the retained corpus can support.

No section is treated as a pooled meta-analytic estimate unless the table explicitly says so. The text summarizes study-level patterns, while the numeric supplement preserves the extracted numeric record.

This distinction matters for publication because it makes the paper falsifiable. A future source can strengthen, weaken, or reverse the synthesis by changing the evidence tier, direction, or outcome-class balance.

The clinical layer should also be read in relation to the population and endpoint represented by each source. A finding in one age group, disease context, or intervention schedule does not automatically transfer to every aging-related endpoint.

Methods

Review type and protocol

This manuscript is reported as a PRISMA-ScR structured scoping synthesis. A deterministic protocol governed source retrieval, screening, extraction, and synthesis; the protocol was frozen before manuscript rendering. The full audit trail is in the supplementary methods_pack.json and the timestamped submission directory synthesis-allostatic_load-v06-DAILY-2026-05-28T03-43-06Z-R2.

Information sources

Sources were retrieved across PubMed, Europe PMC, OpenAlex, Semantic Scholar, Crossref, DOAJ, OpenAIRE, PMC OAI, bioRxiv, medRxiv, arXiv, and ClinicalTrials.gov. Retrieval window: 2026-05-28.

Search strategy

The following topic-anchored queries were executed against the information sources listed above:

• allostatic load AND aging AND human
• allostatic load AND older adults
• allostatic load AND randomized controlled trial
• biological stress burden AND aging AND human
• biological stress burden AND older adults
• biological stress burden AND randomized controlled trial
• cumulative stress AND aging AND human
• cumulative stress AND older adults
• cumulative stress AND randomized controlled trial
• stress biomarkers AND aging AND human

Eligibility criteria

• Sources whose primary content addresses allostatic load.
• Sources with extractable quantitative or qualitative findings.
• Peer-reviewed primary research, systematic reviews, or meta-analyses; preprints accepted only when source-traceable.
• Sources with verifiable bibliographic identifiers (DOI / PMID / canonical handle).

Selection of sources of evidence

The synthesis did not begin from an unfiltered database export. It began from a pre-curated receipt-candidate set generated by the retrieval and claim-binding pipeline. Of 185 records in the receipt-candidate union, 65 were classified as source candidates and 50 were admitted as traceable synthesis sources. No additional records were excluded after final source admission.

source admission funnel

Admission bucket	n
Receipt candidate union	185
Classified source candidates	65
No extractable claims	25
None-only claim binding	10
Partial/none-only claim binding	50
Partial-only candidates	17
Strict high-confidence sources	18
Admitted final sources	50

Exclusion reasons

• Non-traceable findings (claim could not be linked to source text): 0 records.
• Wrong population / off-topic sources excluded at screening.
• Duplicate records deduplicated by DOI / PMID before screening.

Data items

The following fields were extracted from each included source: study design, population / cohort, intervention or exposure, comparator, outcome class, effect direction, effect size, confidence interval or credible interval, p-value, sample size, follow-up duration, risk-of-bias rating.

Risk-of-bias appraisal

Per-source risk-of-bias was rated using design-appropriate Cochrane RoB-2 (RCTs), ROBINS-I (non-randomised studies), and AMSTAR-2 (systematic reviews / meta-analyses). Ratings recorded in risk_of_bias.json.

Synthesis approach

Evidence-tension synthesis: claims grouped by outcome class (cardiometabolic, contextual adjacent evidence, deficiency prevalence, dosing and pharmacokinetics, immune, longevity, mortality and survival, safety and comorbidity); within-class agreement, disagreement, and directness gaps surfaced explicitly. Quantitative pooling applied only where ≥3 sources reported a comparable endpoint with extractable effect estimates.

AI-use disclosure

Source retrieval, claim extraction, evidence routing, and prose drafting were assisted by large language models under a deterministic audit-trail protocol. Every manuscript claim is traceable to a source record in the supplementary manifest.json. Final eligibility and interpretation decisions are author-verified.

Accountability

Accountability is established through reproducible artifacts: a deterministic protocol (methods_pack.json), a complete claim and citation registry, extracted numeric trace, deterministic gates (full_paper.journal_surface.json, pre_submit_gate.json, artifact_consistency.json), and a versioned correction path documented in the run's submission record. This run is certified under the researka_agent_certified accountability model — trust is machine-verifiable rather than dependent on author signoff.

Results

Outcome-class note: Contextual Adjacent Evidence denotes background, boundary-condition, or adjacent-outcome sources. It is not pooled with direct outcome evidence.

Outcome class	Corpus slice	Strongest signal	Directness	Main limitation
Contextual Adjacent Evidence	n=23; claims=1056	null signal in 15/23 sources	13 indirect; 2 mechanistic; 8 review	limited corpus depth in this outcome class
Immune	n=9; claims=679	null signal in 5/9 sources	2 direct; 3 indirect; 4 review	limited corpus depth in this outcome class
Longevity	n=7; claims=323	unclear signal in 3/7 sources	3 indirect; 4 review	limited corpus depth in this outcome class
Cardiometabolic	n=4; claims=274	null signal in 3/4 sources	1 direct; 1 indirect; 2 review	limited corpus depth in this outcome class
Dosing and Pharmacokinetics	n=4; claims=219	null signal in 4/4 sources	4 review	limited corpus depth in this outcome class
Population / prevalence	n=1; claims=113	null signal in 1/1 sources	1 indirect	single-source slice; hypothesis-generating
Mortality and Survival	n=1; claims=29	null signal in 1/1 sources	1 indirect	single-source slice; hypothesis-generating
Safety and Comorbidity	n=1; claims=156	null signal in 1/1 sources	1 indirect	single-source slice; hypothesis-generating

The retained allostatic load corpus is reported by outcome class before any cross-domain interpretation. This structure prevents favorable, null, mixed, and adverse evidence from being blended across biologically different endpoints.

Contextual Adjacent Evidence Outcomes

The contextual adjacent evidence packet includes 23 source-level summaries and 1056 high-confidence observations. Directional coding within this packet is negative=3, null=15, positive=2, unclear=3, and directness coding is indirect=13, mechanistic=2, review=8. These counts describe the frozen evidence state for this outcome, not a pooled treatment estimate.

Additional corpus sources included animal/preclinical evidence; representative sources include Fan 2023, Quinn 2022, Nikrad 2023. This outcome is interpreted within its own packet first; any broader synthesis is deferred until the cross-domain section so that the writer cannot merge evidence from unrelated outcome classes.

Immune Outcomes

The immune evidence packet includes 9 source-level summaries and 679 high-confidence observations. Directional coding within this packet is mixed=1, null=5, positive=1, unclear=2, and directness coding is direct=2, indirect=3, review=4.

Directional coding within this packet is negative=1, null=3, unclear=3, and directness coding is indirect=3, review=4.

Directional coding within this packet is null=3, unclear=1, and directness coding is direct=1, indirect=1, review=2.

Directional coding within this packet is null=4, and directness coding is review=4.

Directional coding within this packet is null=1, and directness coding is indirect=1.

Result-interpretation guardrail.

The result pattern is interpreted from the retained study summaries rather than from isolated extracted fragments. Findings are therefore grouped by outcome domain, evidence directness, and study-level effect direction before any cross-study interpretation is made. This keeps direct clinical signals separate from mechanistic or indirect signals, preserves null and mixed findings as informative rather than discarding them, and prevents a single repaired or quarantined numeric sentence from hollowing out the result narrative. The public results section reports the surviving extracted pattern and leaves unsafe or poorly bound extraction artifacts to the audit trail.

This guardrail is deliberately numeric-free. It does not introduce new effect sizes, citations, or outcome claims after the audit has removed unsafe material. Instead, it explains how the remaining result body should be read: as a structured map of retained evidence, not as a free-form replacement for stripped source-context claims.

Descriptive findings remain separate from interpretation and endpoint-specific boundaries. Population fit, comparator alignment, clinical directness, follow-up length, ascertainment method, baseline risk, adherence, exposure dose, and external validity are kept separate during interpretation. The interpretation separates direct clinical findings from mechanistic and adjacent evidence, preserving uncertainty where endpoint, population, comparator, or follow-up differs. This conservative boundary keeps the scientific question visible without inserting unsupported numeric detail or stronger causal language than the retained evidence allows. Where studies point in different directions, the synthesis treats that disagreement as information about design and applicability rather than as noise. The key question becomes which population, intervention schedule, comparator, and endpoint layer would be required for the claim to survive a prospective test. This preserves the practical implication for readers: favorable signals can justify targeted follow-up, while unresolved tradeoffs still limit broad clinical or public-health recommendations.

Descriptive findings remain separate from interpretation and endpoint-specific boundaries.

Longevity Outcomes

Longevity is retained as a separate Results slice (n=7; null signal in 3/7 sources; 3 indirect; no direct clinical anchor) and is not pooled into adjacent endpoint classes.

Cardiometabolic Outcomes

Cardiometabolic is retained as a separate Results slice (n=4; null signal in 3/4 sources; 1 direct; 1 indirect; directionally heterogeneous) and is not pooled into adjacent endpoint classes.

Dosing and Pharmacokinetics Outcomes

Dosing and Pharmacokinetics is retained as a separate Results slice (n=4; null signal in 4/4 sources; not classified; no direct clinical anchor) and is not pooled into adjacent endpoint classes.

Population / prevalence Outcomes

Representative sources include Lee 2025.

Population / prevalence is retained as a separate Results slice (n=1; null signal in 1/1 sources; 1 indirect; no direct clinical anchor) and is not pooled into adjacent endpoint classes.

Mortality and Survival Outcomes

Mortality and Survival remains a separate Results slice (n=1; claims=29; null signal in 1/1 sources; 1 indirect; single-source slice; hypothesis-generating) and is not pooled into adjacent endpoint classes.

Safety and Comorbidity Outcomes

Safety and Comorbidity remains a separate Results slice (n=1; claims=156; null signal in 1/1 sources; 1 indirect; single-source slice; hypothesis-generating) and is not pooled into adjacent endpoint classes.

Cross-Domain Synthesis

A central tension in the Allostatic load literature arises from the discordance between mechanistic plausibility and human clinical translatability, where preclinical models consistently demonstrate positive outcomes in contextual other domains but human trials yield null or mixed results. Fan 2023 reports lifespan and healthspan extension in C. elegans through detoxification pathways modulated by a pregnane X receptor agonist, aligning with the broader expectation that oxidative-stress attenuation should confer biological benefits; however, Nikrad 2023 documents negative effects on oxidative stress biomarkers in obese women with polycystic ovary syndrome following calorie restriction with thylakoid membranes, underscoring that human metabolic contexts may not mirror the controlled simplicity of model organisms. The divergence is not merely quantitative but mechanistic: invertebrate detoxification capacity is highly conserved and experimentally tractable, whereas human oxidative biology is entangled with adiposity, insulin resistance, and endocrine feedback loops that can blunt or invert the intended therapeutic signal. Boundary conditions therefore likely include organismal complexity, baseline metabolic stress burden, and the presence of comorbid cardiometabolic dysregulation; resolution would require head-to-head translational studies that pair the same mechanistic readouts across species and human subgroups stratified by visceral adiposity and glycemic status, as suggested by the null visceral adiposity findings reported in Moel 2025.

A second unresolved tension pits immune biomarker modulation against cumulative clinical relevance, where human RCTs show inconsistent immune outcomes despite mechanistic rationale. Ramos-Hernandez 2026, a secondary analysis of a randomized crossover trial in older adults, reports preserved glutathione redox balance with creatine plus β-hydroxy-β-methylbutyrate supplementation, a finding that aligns with the expectation that redox homeostasis should translate into immune competence; yet Zahedi 2021, in a randomized double-blind placebo-controlled trial of curcuminoids in critically ill patients, found null effects on inflammatory cytokines despite statistically significant p-values, illustrating that biomarker preservation does not guarantee clinically meaningful immune modulation. The mismatch reflects differences in outcome directness and population vulnerability: older adults may retain sufficient reserve to benefit from redox support, whereas critically ill patients present with overwhelming inflammatory cascades that overwhelm targeted interventions. Boundary conditions therefore include baseline inflammatory burden and organ system reserve; evidence to resolve this tension would come from stratified RCTs that enroll both community-dwelling older adults and hospitalized patients, with immune endpoints standardized to functional assays (e.g., ex vivo pathogen challenge) rather than surrogate cytokines alone.

Another cross-domain tension emerges between longevity claims and direct mortality outcomes, where preclinical and review data suggest potential benefits but human cohorts fail to show survival gains. Wang 2020 reports dose-dependent lifespan extension in C. elegans exposed to orange extracts, and Moskalevska 2026 reviews interventions that extend healthspan and survival in model systems, yet Terhalle 2025 demonstrates that nonspecific stress biomarkers do not predict 30-day mortality in older emergency department patients presenting with falls, and Obeng-Gyasi 2022 finds no association between allostatic load and overall mortality in metastatic non–small cell lung cancer. The disconnect likely stems from species-level differences in stress-response architecture and the clinical heterogeneity of human cohorts, where comorbid conditions and competing risks dilute any modest survival signal. Boundary conditions include the absence of acute critical illness and low baseline comorbidity burden; resolution would require prospective human studies that track both allostatic load indices and hard mortality endpoints in populations free of advanced malignancy or acute decompensation, with sufficient follow-up to detect small effects.

Additional corpus sources included animal/preclinical evidence; another tension surfaces between cardiometabolic null findings and mechanistic plausibility, where direct clinical trials fail to show benefit despite oxidative-stress modulation. Nasiri 2025, a randomized controlled trial in women with polycystic ovary syndrome, reports null effects on hormonal and metabolic outcomes despite significant p-values in biomarker domains, while Cai 2011 demonstrates that icariin and its derivative extend healthspan via the insulin/IGF-1 pathway in C. elegans, a mechanism plausibly relevant to human metabolic regulation. The inconsistency suggests that human cardiometabolic disease is buffered by redundant pathways and lifestyle factors that blunt the impact of targeted oxidative interventions, or that the selected biomarkers do not capture the functional pathways driving clinical outcomes. Boundary conditions include the presence of severe insulin resistance and obesity, where compensatory mechanisms may mask intervention effects; evidence to adjudicate this tension would come from factorial trials that combine oxidative-stress modulators with established cardiometabolic therapies (e.g., metformin) and track composite functional endpoints such as insulin sensitivity and ovulatory frequency.

Another cross-domain tension contrasts mixed immune signals with consistent null findings in dosing and pharmacokinetics domains, where systematic reviews fail to identify robust biomarker responses. The divergence likely reflects heterogeneity in supplement composition, baseline inflammatory status, and outcome selection; CRP may be more sensitive to dietary polyphenols than to n-3 fatty acids, yet neither intervention demonstrates durable clinical translatability across populations. Boundary conditions include the absence of chronic inflammatory disease and standardized dosing protocols; resolution would require network meta-analyses that harmonize biomarker panels and dose-response gradients across diverse populations, with stratification by baseline inflammatory burden and comorbid conditions.

Boundary-condition synthesis

Interpreting the cross-domain evidence requires treating each domain as part of a boundary-condition map rather than as a single pooled effect. Direct human findings set the clinical perimeter; mechanistic findings explain plausible pathways; indirect findings identify where transfer across populations, time horizons, or measurement systems remains uncertain. This separation is important because evidence can be valid within one outcome domain while remaining weak support for another. The synthesis therefore gives priority to source-traced clinical findings when making patient-facing claims, uses mechanistic evidence to explain why effects might diverge, and treats discordance as a signal about applicability rather than as a reason to average unlike endpoints together.

Endpoint-Sensitivity Framework

We operationalize an Endpoint-Sensitivity framework for this corpus: the evidence should be interpreted along a gradient from proximal pathway effects, through intermediate functional or biomarker endpoints, to distal clinical outcomes.

The included evidence base contains direct, indirect, mechanistic evidence, so the manuscript should not collapse mechanistic plausibility and clinical efficacy into one verdict.

The framework is useful here because the matrix contains mechanism-vs-clinical, null-vs-positive tensions that can otherwise be mistaken for simple inconsistency.

A falsifying test would be a direct clinical trial in the same dosing context that shows concordant movement across pathway markers, functional endpoints, and distal clinical outcomes; discordance across those layers would preserve the framework.

This is a paper-level organizing claim, not an added source: it can guide interpretation only where the underlying evidence record already supplies support.

Discussion

Thesis: Across 50 curated reference papers, the evidence base for allostatic load shows a context-dependent profile. Positive signals appear in: contextual other, immune. Negative signals appear in: contextual other, longevity. Null findings dominate: contextual other, immune. The synthesis surfaces 308 non-orthogonal tensions across outcome classes — see Cross-Domain Synthesis. The allostatic load anti-aging case as currently constituted is incomplete: mechanistic plausibility coexists with mixed or sparse human-RCT evidence, and the boundary conditions remain to be established.

Threat 3: Longevity claims remain fundamentally contested, with systematic reviews and meta-analyses providing conflicting signals that undermine clinical confidence. Dessie 2025 further complicates the picture with negative pooled mortality associations in breast cancer cohorts. This inconsistency suggests that longevity claims may be overstated or context-specific, particularly in frail or multimorbid populations. The mechanism-to-outcome translation fails because most longevity signals derive from proxy or indirect biomarkers rather than direct survival endpoints. Future trials must adopt rigorous survival-focused designs with prespecified subgroup analyses by comorbidity burden and baseline Allostatic load.

What the evidence supports clearly is the mechanistic plausibility of Allostatic load modulation across diverse interventions, even when clinical translation is uncertain. Ramos-Hernandez 2026 preserves glutathione redox balance in older adults (P < 0.05), and Jarvela-Reijonen 2020 reports reductions in inflammatory biomarkers following acceptance and commitment therapy (P = 0.012 to 0.035). These convergent signals suggest that stress-biology modulation can alter intermediate mechanisms, which may translate to clinical benefit under specific conditions. However, the strength of this convergence is qualified by the indirectness of most studies; human trials rarely measure both mechanistic and clinical endpoints in the same cohort. This interpretive conclusion aligns with the broader pattern that Allostatic load interventions may stabilize biomarkers without improving functional outcomes, a distinction that future trials must explicitly address.

Where the evidence is genuinely mixed, the tension centers on the durability of biomarker improvements and their relationship to clinically meaningful endpoints. Spears 2026’s mixed associations between cumulative stress, inflammation, and mortality disparities further highlight that biomarker shifts do not uniformly predict clinical outcomes. This heterogeneity may reflect population specificity, intervention dose, or the inherent complexity of Allostatic load as a construct. The clinical implication is that biomarker-guided interventions must be tailored to baseline stress-biology profiles, with careful attention to the endpoints chosen for evaluation.

The gap between mechanistic and clinical endpoints is the most critical unresolved issue in the Allostatic load literature. Fan 2023 and Cai 2011 demonstrate robust mechanistic signals in preclinical models, yet human RCTs like Nasiri 2025 and Zahedi 2021 show null effects on clinical outcomes despite biomarker shifts. This disconnect suggests that current interventions may modulate isolated pathways without addressing the poly-systemic dysregulation characteristic of human Allostatic load. The mechanism-to-clinic boundary is further complicated by the lack of standardized Allostatic load indices and the reliance on heterogeneous biomarkers across studies. For clinical translation to occur, future trials must incorporate composite Allostatic load indices as stratification variables and co-primary endpoints, alongside mechanistic biomarkers aligned with preclinical targets. Without such alignment, the field risks perpetuating a cycle of null clinical findings despite promising mechanistic signals. This interpretive conclusion is consistent with broader methodological critiques that surrogate endpoints often fail to predict hard outcomes in stress-biology research (Ioannidis 2005).

Population specificity emerges as a dominant theme, with interventions showing variable effects across demographic and clinical subgroups. Ramos-Hernandez 2026 targets older adults and preserves glutathione redox balance, while Nikrad 2023 focuses on obese women with PCOS and reports worsening oxidative stress biomarkers. Spears 2026 highlights racial disparities in cumulative stress and inflammation, suggesting that interventions may need to be tailored to baseline stress-biology profiles and social determinants of health. The clinical decision boundary thus depends not only on the intervention but also on the population’s baseline Allostatic load, comorbidity burden, and social context. This heterogeneity may explain why some studies report null effects across broad populations, while subgroup analyses hint at benefit in specific contexts. Future trials must incorporate stratified randomization by baseline Allostatic load, age, sex, and comorbidity to clarify who benefits and under what conditions. Without such stratification, the field risks overlooking meaningful treatment effects in subgroups while overgeneralizing null findings across diverse populations.

Methodological reflections reveal that the endpoints chosen, sample sizes, and follow-up durations are critical determinants of the Allostatic load evidence base. Fan 2023 and Cai 2011 rely on lifespan and healthspan in preclinical models, which are not directly translatable to human clinical endpoints. Ramos-Hernandez 2026 uses mechanistic endpoints (glutathione redox balance) in older adults with P < 0.05, but the trial’s crossover design and small sample size limit generalizability. Parker 2022 highlights the heterogeneity in Allostatic load indices across studies, complicating cross-study comparisons. The field’s reliance on indirect biomarkers and heterogeneous indices suggests that current methodological approaches may be insufficient to capture the complexity of Allostatic load. Future work should prioritize composite Allostatic load indices, standardized biomarker panels, and longer follow-up durations to assess durability of effects. This methodological critique aligns with broader concerns about the validity of surrogate endpoints in stress-biology research (Ioannidis 2005).

Implications for clinical practice and research priorities must balance cautious optimism with rigorous skepticism. The evidence suggests that Allostatic load interventions can modulate stress-biology mechanisms, but the translation to durable clinical benefit remains uncertain. Clinicians should avoid inferring clinical efficacy from biomarker shifts alone, as demonstrated by the contrast between Fan 2023’s preclinical signals and Nasiri 2025’s null clinical outcomes. Research priorities should focus on adaptive trial designs that embed mechanistic biomarkers as co-primary endpoints, alongside clinically meaningful outcomes such as disability-free survival or hospitalization. Trials must stratify by baseline Allostatic load, age, sex, and comorbidity to clarify population-specific effects. Methodologically, the field should converge on standardized Allostatic load indices and biomarker panels to enable cross-study comparisons. Until such trials are completed, the anti-aging case for Allostatic load remains preliminary and context-dependent. This cautious stance is consistent with the broader pattern of mixed findings across outcome classes (Moskalevska 2026, Knufinke 2023, Parker 2022).

Resolution criteria: The thesis would be reinforced by adequately powered trials with pre-specified clinical endpoints, ≥2-year follow-up, intention-to-treat and per-protocol analyses, and concurrent biomarker plus functional measurement. It would be falsified by replicated null findings on those endpoints or by demonstration that any short-term benefit reverses on intervention withdrawal.

Limitations

Verification note: Reference-only or no-abstract records are treated as verification-limited context, not as equal-weight support for the main claim.

The corpus assembled here cannot support claims about allostatic load across the full adult lifespan because long-term mortality trials are absent. This leaves the central clinical question—whether allostatic load predicts longevity—unresolved, as the evidence base is confined to short-term observational cohorts without hard endpoints. The lack of head-to-head randomized trials comparing interventions that explicitly modulate allostatic load further constrains causal inference, leaving the field reliant on surrogate endpoints that may not translate to clinically meaningful outcomes (Ioannidis 2005).

Single-trial dominance limits the synthesis' external validity for immune outcomes, where only Ramos-Hernandez 2026 provides direct human evidence. Its crossover RCT in older adults shows preserved glutathione redox balance with creatine plus β-hydroxy-β-methylbutyrate supplementation (P < 0.05), but the absence of replication cohorts precludes confirmation of directionality or effect size. In contrast, mechanistic signals from preclinical models (e.g., Fan 2023) cannot substitute for clinical endpoints, creating a gap between biomarker-level plausibility and patient-relevant outcomes.

Population specificity undermines the generalizability of headline conclusions to broader adult populations. The corpus skews heavily toward older adults (e.g., Ramos-Hernandez 2026, Terhalle 2025, Martens 2018) and women with polycystic ovary syndrome (e.g., Nasiri 2025, Nikrad 2023), leaving critical gaps in younger adults, men, and individuals with non-communicable diseases outside metabolic or inflammatory phenotypes. Even within older cohorts, enrollment criteria vary widely—from community-dwelling adults (Martens 2018) to emergency department fallers (Terhalle 2025)—introducing heterogeneity that complicates cross-study comparisons. The absence of trials in non-diabetic adults or populations with cardiovascular disease limits conclusions about allostatic load's prognostic utility beyond metabolic or inflammatory domains.

Endpoint scope is constrained by the predominance of mechanistic and biomarker-level outcomes, with only 3 of 50 papers reporting clinical functional endpoints. Nasiri 2025, the sole cardiometabolic trial, evaluates hormonal and metabolic biomarkers in women with polycystic ovary syndrome but lacks patient-centered outcomes such as cardiovascular events or quality of life. Similarly, immune outcomes are largely restricted to inflammatory biomarkers (e.g., IL-6, CRP) without standardized clinical endpoints like infection rates or hospitalization. This narrow focus reflects a broader trend in the field, where surrogate endpoints dominate despite their limited translation to hard outcomes (Ioannidis 2005). The reliance on such endpoints introduces the risk that observed associations may not reflect true clinical benefit or harm.

The mechanism-to-clinic gap is most evident in longevity research, where preclinical models (e.g., Fan 2023, Wang 2020) demonstrate lifespan extensions or stress-resistance effects but lack human validation. Fan 2023 reports lifespan extension in C. elegans (P < 0.001), while Wang 2020 shows similar findings in C. elegans exposed to orange extracts, yet neither study translates to human mortality outcomes. This gap is further underscored by the absence of long-term interventional trials explicitly designed to modulate allostatic load in humans, leaving the anti-aging case incomplete despite promising preclinical signals.

Conclusion

This synthesis finds that allostatic load remains a conceptually plausible biomarker framework for tracking cumulative biological wear-and-tear across physiological systems, yet its translation into clinically actionable anti-aging guidance is not yet supported by the current evidence base (Moskalevska 2026, Parker 2022). Notably, the few direct human interventions targeting allostatic load components—including creatine/HMB supplementation in older adults—show preserved redox balance (Ramos-Hernandez 2026, P < 0.05) but fail to demonstrate downstream clinical benefits such as reduced cardiometabolic risk or improved functional status (Nasiri 2025, P < 0.001), underscoring the persistent gap between mechanistic plausibility and clinical efficacy. This pattern aligns with broader methodological cautions that surrogate associations—even when statistically significant—do not guarantee hard-outcome validity (Ioannidis 2005), particularly in complex, multifactorial conditions like aging where confounding by unmeasured variables is substantial. The clinical-practice implication is clear: current evidence supports the hypothesis that allostatic load may reflect cumulative biological stress (Parker 2022) but does not justify marketing any compound, supplement, or intervention as an anti-aging therapy pending further trials that demonstrate functional or survival benefits beyond biomarker modulation.

Against this backdrop, the strongest counter-evidence emerges from systematic reviews showing no consistent benefit of antioxidant or anti-inflammatory nutraceuticals—including biophenol-rich compounds (Giang 2021, P > 0.05) or propolis (Bahari 2025, P < 0.001)—on allostatic load or downstream aging outcomes, while preclinical studies continue to report positive lifespan extensions (Fan 2023, P < 0.001) that have not translated to human trials. This disconnect suggests that either the biomarker frameworks currently used to operationalize allostatic load are too distal from clinical aging processes, or that human aging involves compensatory mechanisms not captured by rodent models. Until such evidence emerges, clinical practice should refrain from promoting allostatic load–focused interventions as standalone anti-aging therapies, instead emphasizing general health-supportive strategies—such as maintaining a BMI below 25 kg/m² (WHO 2000)—while acknowledging that these measures address general health rather than proven geroprotection. The synthesis therefore supports a bounded interpretation rather than a generalized clinical recommendation. Across 50 curated reference papers, the evidence base for Allostatic load shows a context-dependent profile. Positive signals appear in: contextual other, immune. Negative signals appear in: contextual other, longevity. Null findings dominate: contextual other, immune. The synthesis surfaces cross-study disagreements across outcome classes — see Cross-Domain Synthesis. The Allostatic load anti-aging case as currently constituted is incomplete: mechanistic plausibility coexists with mixed or sparse human-RCT evidence, and the boundary conditions remain to be established.

Additional corpus sources included animal/preclinical evidence; the strongest unresolved contrast is the disagreement between Fan 2023 and Nikrad 2023 on contextual adjacent evidence (severity 5/5), which defines the boundary condition future studies must test rather than smooth over.

Prior reviews in the corpus (Giang 2021, Li 2025, Alrasheed 2026, Dessie 2025, Knufinke 2023) emphasize convergent signals on Allostatic load. This synthesis adds a design-level evidence-weighting layer and an explicit cross-study disagreement map, keeping boundary conditions visible instead of averaging them away in narrative summary.

Boundary-Condition Matrix

Outcome class	Direct sources	Indirect / mechanism sources	Direction profile	Interpretation boundary
longevity	0	7	negative, null, unclear	direct clinical gap
cardiometabolic	1	3	null, unclear	conflict-resolution gap
contextual adjacent evidence	0	23	negative, null, positive, unclear	conflict-resolution gap
immune	2	7	mixed, null, positive, unclear	conflict-resolution gap
deficiency prevalence	0	1	null	direct clinical gap
dosing and pharmacokinetics	0	4	null	direct clinical gap
mortality and survival	0	1	null	direct clinical gap
safety and comorbidity	0	1	null	direct clinical gap

Evidence-Gap Priority

Priority	Gap	Rationale
P1	longevity: direct clinical gap	0 direct and 7 indirect sources; direction profile: negative, null, unclear
P2	cardiometabolic: conflict-resolution gap	1 direct and 3 indirect sources; direction profile: null, unclear
P3	contextual adjacent evidence: conflict-resolution gap	0 direct and 23 indirect sources; direction profile: negative, null, positive, unclear
P4	immune: conflict-resolution gap	2 direct and 7 indirect sources; direction profile: mixed, null, positive, unclear
P5	deficiency prevalence: direct clinical gap	0 direct and 1 indirect source; direction profile: null

Next-Study Design Recommendation

The next high-yield study for Allostatic load should target the longevity evidence gap, pre-register the primary endpoint, separate clinical from mechanistic endpoints, preserve safety and adherence capture, and include an analysis plan that can falsify the current boundary-condition claim rather than only confirming a favorable direction.

Structured Evidence Tables

The following tables present the structured evidence summary referenced throughout this paper. Numbers live in the tables; prose references them. Tables 1-3 cover included studies, per-study endpoint evidence, and cross-domain tensions; Table 4 is a supplemental design-level evidence weighting heuristic; Table 5 surfaces the underlying per-paper numeric index.

Table 1: Included Studies

Citation	Design	Tier	N	Population	Endpoint	Direction	Directness	Trial ID	Representative p-value	n claims
Ramos-Hernandez 2026	RCT (clinical)	A1	—	older adults	immune	unclear	direct	—	P < 0.05	226
Giang 2021	Review / meta-analysis	B1	—	—	immune	unclear	review	—	P = 0.021	187
Nasiri 2025	RCT (clinical)	A1	—	adults	cardiometabolic	null	direct	—	P < 0.001	164
Moel 2025	Observational	B2	—	adults	safety comorbidity	null	indirect	—	P = 0.004	156
Lee 2025	Observational	B2	—	older adults	deficiency prevalence	null	indirect	—	P = 0.002	113
Terhalle 2025	Observational	B2	—	adults	longevity	null	indirect	—	P = 0.02	110
Fan 2023	Preclinical (animal/in vitro)	C1	—	adults	contextual other	positive	mechanistic	—	P < 0.001	108
Martens 2018	Observational	B2	—	older adults	cardiometabolic	unclear	indirect	—	P < 0.006	90
Jarvela-Reijonen 2020	Observational	B2	—	adults	immune	positive	indirect	—	P = 0.012	81
Quinn 2022	Observational	B2	—	adults	contextual other	null	indirect	—	P < 0.001	75
Nikrad 2023	RCT (mechanistic)	A2	—	adults	contextual other	negative	indirect	—	P < 0.001	74
Wang 2020	Observational	B2	—	adults	longevity	null	indirect	—	P < 0.05	74
Bahari 2025	Observational	B2	—	—	dosing pharmacokinetics	null	review	—	P < 0.001	73
Li 2025	Review / meta-analysis	B1	—	—	contextual other	null	review	—	P < 0.001	72
Das 2023	Observational	B2	—	adults	contextual other	null	indirect	—	P < 0.001	71
Tambunan 2026	Observational	B2	—	—	contextual other	negative	review	—	P < 0.0001	71
Bloomer 2009	Observational	B2	—	adults	contextual other	null	indirect	—	P < 0.0001	67
Alrasheed 2026	Review / meta-analysis	B1	—	type 2 diabetes patients	immune	null	review	—	P = 0.001	64
Khor 2021	Observational	B2	—	—	dosing pharmacokinetics	null	review	—	P < 0.001	60
Goettel 2023	Observational	B2	—	adults	contextual other	null	review	—	—	55
Ceylan 2025	Observational	B2	—	adults	contextual other	null	indirect	—	P < 0.001	55
Saadh 2025	Observational	B2	—	—	dosing pharmacokinetics	null	review	—	P < 0.001	54
Msigwa 2026	Observational	B2	—	type 2 diabetes patients	contextual other	null	review	—	P < 0.001	54
Yen 2020	Observational	B2	—	adults	contextual other	unclear	indirect	—	P < 0.05	52
Hoseini 2022	Observational	B2	—	type 2 diabetes patients	immune	null	indirect	—	P = 0.00	52
Dessie 2025	Review / meta-analysis	B1	—	adults	longevity	negative	review	—	P = 0.000	50
Yin 2025	Observational	B2	—	adults	contextual other	negative	indirect	—	P < 0.0001	49
Knufinke 2023	Review / meta-analysis	B1	—	—	longevity	unclear	review	—	—	43
Yakout 2025	Observational	B2	—	adults	contextual other	null	indirect	—	P < 0.001	40
Corley 2023	Observational	B2	—	adults	contextual other	positive	indirect	—	P = 0.01	37
Taylor 2023	Observational	B2	—	—	immune	null	review	—	P = 0.06	37
Demir 2025	Observational	B2	—	adults	contextual other	unclear	indirect	—	P = 0.004	35
Obeng-Gyasi 2022	Observational	B2	—	adults	longevity	null	indirect	—	P < 0.001	34
Nguyen 2021	Observational	B2	—	—	dosing pharmacokinetics	null	review	—	P = 0.003	32
Zhang 2021	Observational	B2	—	older adults	mortality survival	null	indirect	—	—	29
Moabedi 2025	Observational	B2	—	—	contextual other	null	review	—	P < 0.001	27
Cai 2011	Preclinical (animal/in vitro)	C1	—	adults	contextual other	null	mechanistic	—	P < 0.01	24
Pleguezuelos 2023	Observational	B2	—	adults	contextual other	unclear	review	—	—	23
Spears 2026	Observational	B2	—	adults	immune	mixed	indirect	—	P < 0.001	22
Wallace 2023	Observational	B2	—	adults	contextual other	null	indirect	—	P = 0.012	18
Belsky 2017	Observational	B2	—	adults	contextual other	null	indirect	—	P < 0.001	18
Beese 2022	Observational	B2	—	—	cardiometabolic	null	review	—	—	18
Locker 2025	Observational	B2	—	adults	contextual other	null	indirect	—	P < 0.001	14
Parker 2022	Review / meta-analysis	B1	—	—	longevity	unclear	review	—	—	11
Popa 2025	Observational	B2	—	—	contextual other	null	review	—	—	9
Bahari 2023	Review / meta-analysis	B1	—	—	immune	null	review	—	P = 0.001	9
Madaria 2025	Observational	B2	—	—	contextual other	null	review	—	—	8
Ghalichi 2022	Observational	B2	—	type 2 diabetes patients	cardiometabolic	null	review	—	—	2
Moskalevska 2026	Review / meta-analysis	B1	—	adults	longevity	unclear	review	—	—	1
Zahedi 2021	RCT (clinical)	A1	—	adults	immune	null	direct	—	P < 0.05	1

Table 2: Per-Study Endpoint Evidence

Additional corpus sources included animal/preclinical evidence; | Endpoint | Study | p/CI | Direction | Directness | Tier | Interpretation | | --- | --- | --- | --- | --- | --- | --- | | immune | Ramos-Hernandez 2026 | P < 0.05 | unclear summary | direct | A1 | reported statistic; source summary remains unclear | | immune | Ramos-Hernandez 2026 | P < 0.05 | unclear summary | direct | A1 | reported statistic; source summary remains unclear | | immune | Ramos-Hernandez 2026 | P < 0.05 | unclear summary | direct | A1 | reported statistic; source summary remains unclear | | immune | Ramos-Hernandez 2026 | P < 0.05 | unclear summary | direct | A1 | reported statistic; source summary remains unclear | | immune | Ramos-Hernandez 2026 | P < 0.05 | unclear summary | direct | A1 | reported statistic; source summary remains unclear | | immune | Ramos-Hernandez 2026 | P < 0.05 | unclear summary | direct | A1 | reported statistic; source summary remains unclear | | immune | Giang 2021 | P > 0.05 | unclear summary | review | B1 | reported statistic; source summary remains unclear | | immune | Giang 2021 | P = 0.04 | unclear summary | review | B1 | reported statistic; source summary remains unclear | | immune | Giang 2021 | P = 0.33 | unclear summary | review | B1 | reported statistic; source summary remains unclear | | immune | Giang 2021 | P = 0.289 | unclear summary | review | B1 | reported statistic; source summary remains unclear | | immune | Giang 2021 | P = 0.021 | unclear summary | review | B1 | reported statistic; source summary remains unclear | | immune | Giang 2021 | P = 0.951 | unclear summary | review | B1 | reported statistic; source summary remains unclear | | cardiometabolic | Nasiri 2025 | P < 0.001 | significant statistic | direct | A1 | significant statistic; source-level direction remains null | | cardiometabolic | Nasiri 2025 | P < 0.001 | significant statistic | direct | A1 | significant statistic; source-level direction remains null | | cardiometabolic | Nasiri 2025 | P < 0.001 | significant statistic | direct | A1 | significant statistic; source-level direction remains null | | cardiometabolic | Nasiri 2025 | P < 0.001 | significant statistic | direct | A1 | significant statistic; source-level direction remains null | | cardiometabolic | Nasiri 2025 | P < 0.001 | significant statistic | direct | A1 | significant statistic; source-level direction remains null | | cardiometabolic | Nasiri 2025 | P = 0.02 | significant statistic | direct | A1 | significant statistic; source-level direction remains null | | safety comorbidity | Moel 2025 | P = 0.942 | null summary | indirect | B2 | reported statistic; source summary remains null | | safety comorbidity | Moel 2025 | P = 0.013 | significant statistic | indirect | B2 | significant statistic; source-level direction remains null | | safety comorbidity | Moel 2025 | P = 0.015 | significant statistic | indirect | B2 | significant statistic; source-level direction remains null | | safety comorbidity | Moel 2025 | P = 0.023 | significant statistic | indirect | B2 | significant statistic; source-level direction remains null | | safety comorbidity | Moel 2025 | P = 0.004 | significant statistic | indirect | B2 | significant statistic; source-level direction remains null | | safety comorbidity | Moel 2025 | P = 0.04 | significant statistic | indirect | B2 | significant statistic; source-level direction remains null | | deficiency prevalence | Lee 2025 | P = 0.02 | significant statistic | indirect | B2 | significant statistic; source-level direction remains null | | deficiency prevalence | Lee 2025 | P = 0.002 | significant statistic | indirect | B2 | significant statistic; source-level direction remains null | | deficiency prevalence | Lee 2025 | p ≥ 0.15 | null summary | indirect | B2 | reported statistic; source summary remains null | | deficiency prevalence | Lee 2025 | p ≥ 0.15 | null summary | indirect | B2 | reported statistic; source summary remains null | | deficiency prevalence | Lee 2025 | p ≥ 0.50 | null summary | indirect | B2 | reported statistic; source summary remains null | | deficiency prevalence | Lee 2025 | p ≥ 0.99 | null summary | indirect | B2 | reported statistic; source summary remains null | | longevity | Terhalle 2025 | P = 0.211 | null summary | indirect | B2 | reported statistic; source summary remains null | | longevity | Terhalle 2025 | P = 0.329 | null summary | indirect | B2 | reported statistic; source summary remains null | | longevity | Terhalle 2025 | P = 0.02 | significant statistic | indirect | B2 | significant statistic; source-level direction remains null | | longevity | Terhalle 2025 | P = 0.051 | null summary | indirect | B2 | reported statistic; source summary remains null | | longevity | Terhalle 2025 | P = 0.441 | null summary | indirect | B2 | reported statistic; source summary remains null | | longevity | Terhalle 2025 | P = 0.904 | null summary | indirect | B2 | reported statistic; source summary remains null | | contextual other | Fan 2023 | P < 0.001 | positive summary | mechanistic | C1 | reported statistic; source summary remains positive | | contextual other | Fan 2023 | P < 0.001 | positive summary | mechanistic | C1 | reported statistic; source summary remains positive | | contextual other | Fan 2023 | P < 0.001 | positive summary | mechanistic | C1 | reported statistic; source summary remains positive | | contextual other | Fan 2023 | P < 0.001 | positive summary | mechanistic | C1 | reported statistic; source summary remains positive | | cardiometabolic | Martens 2018 | P < 0.05 | unclear summary | indirect | B2 | reported statistic; source summary remains unclear | | cardiometabolic | Martens 2018 | P < 0.006 | unclear summary | indirect | B2 | reported statistic; source summary remains unclear | | immune | Jarvela-Reijonen 2020 | P = 0.012 | positive summary | indirect | B2 | reported statistic; source summary remains positive | | immune | Jarvela-Reijonen 2020 | P = 0.035 | positive summary | indirect | B2 | reported statistic; source summary remains positive | | immune | Jarvela-Reijonen 2020 | P = 0.303 | positive summary | indirect | B2 | reported statistic; source summary remains positive | | immune | Jarvela-Reijonen 2020 | P = 0.460 | positive summary | indirect | B2 | reported statistic; source summary remains positive | | immune | Jarvela-Reijonen 2020 | P = 0.496 | positive summary | indirect | B2 | reported statistic; source summary remains positive | | immune | Jarvela-Reijonen 2020 | P = 0.302 | positive summary | indirect | B2 | reported statistic; source summary remains positive | | contextual other | Quinn 2022 | P < 0.001 | significant statistic | indirect | B2 | significant statistic; source-level direction remains null | | contextual other | Quinn 2022 | P = 0.018 | significant statistic | indirect | B2 | significant statistic; source-level direction remains null | | contextual other | Quinn 2022 | P < 0.001 | significant statistic | indirect | B2 | significant statistic; source-level direction remains null | | contextual other | Quinn 2022 | P = 0.089 | null summary | indirect | B2 | reported statistic; source summary remains null | | contextual other | Quinn 2022 | P = 0.036 | significant statistic | indirect | B2 | significant statistic; source-level direction remains null | | contextual other | Quinn 2022 | P < 0.001 | significant statistic | indirect | B2 | significant statistic; source-level direction remains null | | contextual other | Nikrad 2023 | P < 0.001 | negative summary | indirect | A2 | reported statistic; source summary remains negative | | contextual other | Nikrad 2023 | P < 0.001 | negative summary | indirect | A2 | reported statistic; source summary remains negative | | contextual other | Nikrad 2023 | P < 0.05 | negative summary | indirect | A2 | reported statistic; source summary remains negative | | contextual other | Nikrad 2023 | P ≥ 0.05 | negative summary | indirect | A2 | reported statistic; source summary remains negative | | contextual other | Nikrad 2023 | P > 0.05 | negative summary | indirect | A2 | reported statistic; source summary remains negative | | contextual other | Nikrad 2023 | P = 0.022 | negative summary | indirect | A2 | reported statistic; source summary remains negative | | longevity | Wang 2020 | P < 0.05 | significant statistic | indirect | B2 | significant statistic; source-level direction remains null | | longevity | Wang 2020 | P < 0.05 | significant statistic | indirect | B2 | significant statistic; source-level direction remains null | | longevity | Wang 2020 | P < 0.05 | significant statistic | indirect | B2 | significant statistic; source-level direction remains null | | dosing pharmacokinetics | Bahari 2025 | P < 0.001 | significant statistic | review | B2 | significant statistic; source-level direction remains null | | dosing pharmacokinetics | Bahari 2025 | P < 0.001 | significant statistic | review | B2 | significant statistic; source-level direction remains null | | dosing pharmacokinetics | Bahari 2025 | P < 0.001 | significant statistic | review | B2 | significant statistic; source-level direction remains null | | dosing pharmacokinetics | Bahari 2025 | P < 0.001 | significant statistic | review | B2 | significant statistic; source-level direction remains null | | dosing pharmacokinetics | Bahari 2025 | P = 0.001 | significant statistic | review | B2 | significant statistic; source-level direction remains null | | dosing pharmacokinetics | Bahari 2025 | P < 0.001 | significant statistic | review | B2 | significant statistic; source-level direction remains null | | contextual other | Li 2025 | P = 0.01 | significant statistic | review | B1 | significant statistic; source-level direction remains null | | contextual other | Li 2025 | P = 0.02 | significant statistic | review | B1 | significant statistic; source-level direction remains null | | contextual other | Li 2025 | P < 0.001 | significant statistic | review | B1 | significant statistic; source-level direction remains null | | contextual other | Li 2025 | P < 0.001 | significant statistic | review | B1 | significant statistic; source-level direction remains null | | contextual other | Li 2025 | P < 0.001 | significant statistic | review | B1 | significant statistic; source-level direction remains null | | contextual other | Li 2025 | P < 0.001 | significant statistic | review | B1 | significant statistic; source-level direction remains null | | contextual other | Das 2023 | P < 0.001 | significant statistic | indirect | B2 | significant statistic; source-level direction remains null | | contextual other | Das 2023 | P = 0.007 | significant statistic | indirect | B2 | significant statistic; source-level direction remains null | | contextual other | Das 2023 | P = 0.03 | significant statistic | indirect | B2 | significant statistic; source-level direction remains null | | contextual other | Das 2023 | P = 0.36 | null summary | indirect | B2 | reported statistic; source summary remains null | | contextual other | Tambunan 2026 | P < 0.0001 | negative summary | review | B2 | reported statistic; source summary remains negative | | contextual other | Tambunan 2026 | P < 0.0001 | negative summary | review | B2 | reported statistic; source summary remains negative | | contextual other | Tambunan 2026 | P = 0.0023 | negative summary | review | B2 | reported statistic; source summary remains negative | | contextual other | Tambunan 2026 | P = 0.0206 | negative summary | review | B2 | reported statistic; source summary remains negative | | contextual other | Tambunan 2026 | P < 0.0001 | negative summary | review | B2 | reported statistic; source summary remains negative | | contextual other | Tambunan 2026 | P < 0.0001 | negative summary | review | B2 | reported statistic; source summary remains negative | | contextual other | Bloomer 2009 | P < 0.0001 | significant statistic | indirect | B2 | significant statistic; source-level direction remains null | | contextual other | Bloomer 2009 | P < 0.05 | significant statistic | indirect | B2 | significant statistic; source-level direction remains null | | contextual other | Bloomer 2009 | P > 0.05 | null summary | indirect | B2 | reported statistic; source summary remains null | | contextual other | Bloomer 2009 | P < 0.05 | significant statistic | indirect | B2 | significant statistic; source-level direction remains null | | contextual other | Bloomer 2009 | P = 0.276 | null summary | indirect | B2 | reported statistic; source summary remains null | | contextual other | Bloomer 2009 | P > 0.05 | null summary | indirect | B2 | reported statistic; source summary remains null | | immune | Alrasheed 2026 | P = 0.21 | null summary | review | B1 | reported statistic; source summary remains null | | immune | Alrasheed 2026 | P = 0.28 | null summary | review | B1 | reported statistic; source summary remains null | | immune | Alrasheed 2026 | P = 0.001 | significant statistic | review | B1 | significant statistic; source-level direction remains null | | immune | Alrasheed 2026 | P = 0.55 | null summary | review | B1 | reported statistic; source summary remains null | | immune | Alrasheed 2026 | P < 0.001 | significant statistic | review | B1 | significant statistic; source-level direction remains null | | dosing pharmacokinetics | Khor 2021 | P < 0.001 | significant statistic | review | B2 | significant statistic; source-level direction remains null | | dosing pharmacokinetics | Khor 2021 | P < 0.001 | significant statistic | review | B2 | significant statistic; source-level direction remains null | | dosing pharmacokinetics | Khor 2021 | P = 0.196 | null summary | review | B2 | reported statistic; source summary remains null | | dosing pharmacokinetics | Khor 2021 | P = 0.992 | null summary | review | B2 | reported statistic; source summary remains null | | dosing pharmacokinetics | Khor 2021 | P = 0.409 | null summary | review | B2 | reported statistic; source summary remains null | | dosing pharmacokinetics | Khor 2021 | P < 0.001 | significant statistic | review | B2 | significant statistic; source-level direction remains null | | contextual other | Goettel 2023 | — | null | review | B2 | no significant effect on contextual other | | contextual other | Ceylan 2025 | P < 0.001 | significant statistic | indirect | B2 | significant statistic; source-level direction remains null | | contextual other | Ceylan 2025 | P < 0.001 | significant statistic | indirect | B2 | significant statistic; source-level direction remains null | | contextual other | Ceylan 2025 | P < 0.001 | significant statistic | indirect | B2 | significant statistic; source-level direction remains null | | contextual other | Ceylan 2025 | P = 0.006 | significant statistic | indirect | B2 | significant statistic; source-level direction remains null | | contextual other | Ceylan 2025 | P < 0.001 | significant statistic | indirect | B2 | significant statistic; source-level direction remains null | | contextual other | Ceylan 2025 | P < 0.001 | significant statistic | indirect | B2 | significant statistic; source-level direction remains null | | dosing pharmacokinetics | Saadh 2025 | P = 0.224 | null summary | review | B2 | reported statistic; source summary remains null | | dosing pharmacokinetics | Saadh 2025 | P < 0.001 | significant statistic | review | B2 | significant statistic; source-level direction remains null | | dosing pharmacokinetics | Saadh 2025 | P = 0.136 | null summary | review | B2 | reported statistic; source summary remains null | | dosing pharmacokinetics | Saadh 2025 | P < 0.001 | significant statistic | review | B2 | significant statistic; source-level direction remains null | | dosing pharmacokinetics | Saadh 2025 | P = 0.019 | significant statistic | review | B2 | significant statistic; source-level direction remains null | | dosing pharmacokinetics | Saadh 2025 | P < 0.001 | significant statistic | review | B2 | significant statistic; source-level direction remains null | | contextual other | Msigwa 2026 | P < 0.001 | significant statistic | review | B2 | significant statistic; source-level direction remains null | | contextual other | Msigwa 2026 | P = 0.011 | significant statistic | review | B2 | significant statistic; source-level direction remains null | | contextual other | Msigwa 2026 | P < 0.001 | significant statistic | review | B2 | significant statistic; source-level direction remains null | | contextual other | Msigwa 2026 | P < 0.001 | significant statistic | review | B2 | significant statistic; source-level direction remains null | | contextual other | Msigwa 2026 | P = 0.303 | null summary | review | B2 | reported statistic; source summary remains null | | contextual other | Msigwa 2026 | P = 0.35 | null summary | review | B2 | reported statistic; source summary remains null | | contextual other | Yen 2020 | P < 0.05 | unclear summary | indirect | B2 | reported statistic; source summary remains unclear | | contextual other | Yen 2020 | P < 0.2 | unclear summary | indirect | B2 | reported statistic; source summary remains unclear | | contextual other | Yen 2020 | P < 0.05 | unclear summary | indirect | B2 | reported statistic; source summary remains unclear | | contextual other | Yen 2020 | P < 0.1 | unclear summary | indirect | B2 | reported statistic; source summary remains unclear | | contextual other | Yen 2020 | P < 0.05 | unclear summary | indirect | B2 | reported statistic; source summary remains unclear | | contextual other | Yen 2020 | P < 0.10 | unclear summary | indirect | B2 | reported statistic; source summary remains unclear | | immune | Hoseini 2022 | P = 0.001 | significant statistic | indirect | B2 | significant statistic; source-level direction remains null | | immune | Hoseini 2022 | P = 0.001 | significant statistic | indirect | B2 | significant statistic; source-level direction remains null | | immune | Hoseini 2022 | P = 0.013 | significant statistic | indirect | B2 | significant statistic; source-level direction remains null | | immune | Hoseini 2022 | P = 0.001 | significant statistic | indirect | B2 | significant statistic; source-level direction remains null | | immune | Hoseini 2022 | P = 0.001 | significant statistic | indirect | B2 | significant statistic; source-level direction remains null | | immune | Hoseini 2022 | P = 0.00 | significant statistic | indirect | B2 | significant statistic; source-level direction remains null | | longevity | Dessie 2025 | P = 0.01 | negative summary | review | B1 | reported statistic; source summary remains negative | | longevity | Dessie 2025 | P = 0.013 | negative summary | review | B1 | reported statistic; source summary remains negative | | longevity | Dessie 2025 | P = 0.000 | negative summary | review | B1 | reported statistic; source summary remains negative | | longevity | Dessie 2025 | P = 0.000 | negative summary | review | B1 | reported statistic; source summary remains negative | | longevity | Dessie 2025 | P = 0.310 | negative summary | review | B1 | reported statistic; source summary remains negative | | longevity | Dessie 2025 | P = 0.260 | negative summary | review | B1 | reported statistic; source summary remains negative | | contextual other | Yin 2025 | P < 0.0001 | negative summary | indirect | B2 | reported statistic; source summary remains negative | | contextual other | Yin 2025 | P < 0.05 | negative summary | indirect | B2 | reported statistic; source summary remains negative | | contextual other | Yin 2025 | P = 0.359 | negative summary | indirect | B2 | reported statistic; source summary remains negative | | contextual other | Yin 2025 | P > 0.05 | negative summary | indirect | B2 | reported statistic; source summary remains negative | | contextual other | Yin 2025 | P = 0.543 | negative summary | indirect | B2 | reported statistic; source summary remains negative | | contextual other | Yin 2025 | P = 0.364 | negative summary | indirect | B2 | reported statistic; source summary remains negative | | longevity | Knufinke 2023 | — | unclear | review | B1 | unclear effect on longevity | | contextual other | Yakout 2025 | P < 0.001 | significant statistic | indirect | B2 | significant statistic; source-level direction remains null | | contextual other | Yakout 2025 | P < 0.015 | significant statistic | indirect | B2 | significant statistic; source-level direction remains null | | contextual other | Yakout 2025 | P < 0.040 | significant statistic | indirect | B2 | significant statistic; source-level direction remains null | | contextual other | Yakout 2025 | P < 0.054 | null summary | indirect | B2 | reported statistic; source summary remains null | | contextual other | Yakout 2025 | P < 0.001 | significant statistic | indirect | B2 | significant statistic; source-level direction remains null | | contextual other | Yakout 2025 | P < 0.038 | significant statistic | indirect | B2 | significant statistic; source-level direction remains null | | contextual other | Corley 2023 | P = 0.01 | positive summary | indirect | B2 | reported statistic; source summary remains positive | | contextual other | Corley 2023 | P = 0.01 | positive summary | indirect | B2 | reported statistic; source summary remains positive | | contextual other | Corley 2023 | P = 0.01 | positive summary | indirect | B2 | reported statistic; source summary remains positive | | contextual other | Corley 2023 | P = 0.29 | positive summary | indirect | B2 | reported statistic; source summary remains positive | | contextual other | Corley 2023 | P = 0.01 | positive summary | indirect | B2 | reported statistic; source summary remains positive | | contextual other | Corley 2023 | P = 0.13 | positive summary | indirect | B2 | reported statistic; source summary remains positive | | immune | Taylor 2023 | P = 0.72 | null summary | review | B2 | reported statistic; source summary remains null | | immune | Taylor 2023 | P = 0.87 | null summary | review | B2 | reported statistic; source summary remains null | | immune | Taylor 2023 | P = 0.64 | null summary | review | B2 | reported statistic; source summary remains null | | immune | Taylor 2023 | P = 0.06 | null summary | review | B2 | reported statistic; source summary remains null | | immune | Taylor 2023 | P = 0.29 | null summary | review | B2 | reported statistic; source summary remains null | | immune | Taylor 2023 | P = 0.23 | null summary | review | B2 | reported statistic; source summary remains null | | contextual other | Demir 2025 | P = 0.004 | unclear summary | indirect | B2 | reported statistic; source summary remains unclear | | contextual other | Demir 2025 | P < 0.05 | unclear summary | indirect | B2 | reported statistic; source summary remains unclear | | contextual other | Demir 2025 | P < 0.05 | unclear summary | indirect | B2 | reported statistic; source summary remains unclear | | contextual other | Demir 2025 | P < 0.05 | unclear summary | indirect | B2 | reported statistic; source summary remains unclear | | contextual other | Demir 2025 | P < 0.05 | unclear summary | indirect | B2 | reported statistic; source summary remains unclear | | longevity | Obeng-Gyasi 2022 | P = 0.005 | significant statistic | indirect | B2 | significant statistic; source-level direction remains null | | longevity | Obeng-Gyasi 2022 | P < 0.001 | significant statistic | indirect | B2 | significant statistic; source-level direction remains null | | longevity | Obeng-Gyasi 2022 | P = 0.04 | significant statistic | indirect | B2 | significant statistic; source-level direction remains null | | longevity | Obeng-Gyasi 2022 | P < 0.001 | significant statistic | indirect | B2 | significant statistic; source-level direction remains null | | longevity | Obeng-Gyasi 2022 | P = 0.01 | significant statistic | indirect | B2 | significant statistic; source-level direction remains null | | longevity | Obeng-Gyasi 2022 | P = 0.002 | significant statistic | indirect | B2 | significant statistic; source-level direction remains null | | dosing pharmacokinetics | Nguyen 2021 | P = 0.03 | significant statistic | review | B2 | significant statistic; source-level direction remains null | | dosing pharmacokinetics | Nguyen 2021 | P = 0.03 | significant statistic | review | B2 | significant statistic; source-level direction remains null | | dosing pharmacokinetics | Nguyen 2021 | P = 0.02 | significant statistic | review | B2 | significant statistic; source-level direction remains null | | dosing pharmacokinetics | Nguyen 2021 | P = 0.003 | significant statistic | review | B2 | significant statistic; source-level direction remains null | | dosing pharmacokinetics | Nguyen 2021 | P = 0.03 | significant statistic | review | B2 | significant statistic; source-level direction remains null | | dosing pharmacokinetics | Nguyen 2021 | P = 0.03 | significant statistic | review | B2 | significant statistic; source-level direction remains null | | mortality survival | Zhang 2021 | — | null | indirect | B2 | no significant effect on mortality survival | | contextual other | Moabedi 2025 | P < 0.001 | significant statistic | review | B2 | significant statistic; source-level direction remains null | | contextual other | Moabedi 2025 | P < 0.001 | significant statistic | review | B2 | significant statistic; source-level direction remains null | | contextual other | Moabedi 2025 | P < 0.001 | significant statistic | review | B2 | significant statistic; source-level direction remains null | | contextual other | Moabedi 2025 | P = 0.012 | significant statistic | review | B2 | significant statistic; source-level direction remains null | | contextual other | Moabedi 2025 | P < 0.001 | significant statistic | review | B2 | significant statistic; source-level direction remains null | | contextual other | Moabedi 2025 | P < 0.001 | significant statistic | review | B2 | significant statistic; source-level direction remains null | | contextual other | Cai 2011 | P < 0.05 | significant statistic | mechanistic | C1 | significant statistic; source-level direction remains null | | contextual other | Cai 2011 | P < 0.01 | significant statistic | mechanistic | C1 | significant statistic; source-level direction remains null | | contextual other | Cai 2011 | P = 0.0151 | significant statistic | mechanistic | C1 | significant statistic; source-level direction remains null | | contextual other | Cai 2011 | P < 0.05 | significant statistic | mechanistic | C1 | significant statistic; source-level direction remains null | | contextual other | Cai 2011 | P < 0.01 | significant statistic | mechanistic | C1 | significant statistic; source-level direction remains null | | contextual other | Pleguezuelos 2023 | — | unclear | review | B2 | unclear effect on contextual other | | immune | Spears 2026 | P < 0.001 | mixed summary | indirect | B2 | reported statistic; source summary remains mixed | | immune | Spears 2026 | P < 0.001 | mixed summary | indirect | B2 | reported statistic; source summary remains mixed | | immune | Spears 2026 | P < 0.001 | mixed summary | indirect | B2 | reported statistic; source summary remains mixed | | immune | Spears 2026 | P < 0.001 | mixed summary | indirect | B2 | reported statistic; source summary remains mixed | | contextual other | Wallace 2023 | P = 0.025 | significant statistic | indirect | B2 | significant statistic; source-level direction remains null | | contextual other | Wallace 2023 | P = 0.012 | significant statistic | indirect | B2 | significant statistic; source-level direction remains null | | contextual other | Belsky 2017 | P < 0.001 | significant statistic | indirect | B2 | significant statistic; source-level direction remains null | | contextual other | Belsky 2017 | P < 0.001 | significant statistic | indirect | B2 | significant statistic; source-level direction remains null | | contextual other | Belsky 2017 | P < 0.001 | significant statistic | indirect | B2 | significant statistic; source-level direction remains null | | cardiometabolic | Beese 2022 | — | null | review | B2 | no significant effect on cardiometabolic | | contextual other | Locker 2025 | P < 0.001 | significant statistic | indirect | B2 | significant statistic; source-level direction remains null | | longevity | Parker 2022 | — | unclear | review | B1 | unclear effect on longevity | | contextual other | Popa 2025 | — | null | review | B2 | no significant effect on contextual other | | immune | Bahari 2023 | P = 0.001 | significant statistic | review | B1 | significant statistic; source-level direction remains null | | immune | Bahari 2023 | P = 0.001 | significant statistic | review | B1 | significant statistic; source-level direction remains null | | immune | Bahari 2023 | P < 0.001 | significant statistic | review | B1 | significant statistic; source-level direction remains null | | immune | Bahari 2023 | P < 0.001 | significant statistic | review | B1 | significant statistic; source-level direction remains null | | contextual other | Madaria 2025 | — | null | review | B2 | no significant effect on contextual other | | cardiometabolic | Ghalichi 2022 | — | null | review | B2 | no significant effect on cardiometabolic | | longevity | Moskalevska 2026 | — | unclear | review | B1 | unclear effect on longevity | | immune | Zahedi 2021 | P < 0.05 | significant statistic | direct | A1 | significant statistic; source-level direction remains null |

Table 3: Cross-Domain Tensions

Additional corpus sources included animal/preclinical evidence; | Tension kind | Severity | source A | source B | Outcome class | Summary | Practical implication | | --- | --- | --- | --- | --- | --- | --- | | agreement | 1 | Moskalevska 2026 | Knufinke 2023 | longevity | Moskalevska 2026 (unclear) vs Knufinke 2023 (unclear) on longevity | agreement (minor) | | null vs positive | 3 | Moskalevska 2026 | Terhalle 2025 | longevity | Moskalevska 2026 (unclear) vs Terhalle 2025 (null) on longevity | null vs positive (notable) | | null vs positive | 3 | Moskalevska 2026 | Wang 2020 | longevity | Moskalevska 2026 (unclear) vs Wang 2020 (null) on longevity | null vs positive (notable) | | null vs positive | 3 | Moskalevska 2026 | Obeng-Gyasi 2022 | longevity | Moskalevska 2026 (unclear) vs Obeng-Gyasi 2022 (null) on longevity | null vs positive (notable) | | agreement | 1 | Moskalevska 2026 | Parker 2022 | longevity | Moskalevska 2026 (unclear) vs Parker 2022 (unclear) on longevity | agreement (minor) | | null vs positive | 3 | Knufinke 2023 | Terhalle 2025 | longevity | Knufinke 2023 (unclear) vs Terhalle 2025 (null) on longevity | null vs positive (notable) | | null vs positive | 3 | Knufinke 2023 | Wang 2020 | longevity | Knufinke 2023 (unclear) vs Wang 2020 (null) on longevity | null vs positive (notable) | | null vs positive | 3 | Knufinke 2023 | Obeng-Gyasi 2022 | longevity | Knufinke 2023 (unclear) vs Obeng-Gyasi 2022 (null) on longevity | null vs positive (notable) | | agreement | 1 | Knufinke 2023 | Parker 2022 | longevity | Knufinke 2023 (unclear) vs Parker 2022 (unclear) on longevity | agreement (minor) | | disagreement | 5 | Fan 2023 | Nikrad 2023 | contextual other | Fan 2023 (positive) vs Nikrad 2023 (negative) on contextual other | disagreement (load-bearing) | | null vs positive | 3 | Fan 2023 | Wallace 2023 | contextual other | Fan 2023 (positive) vs Wallace 2023 (null) on contextual other | null vs positive (notable) | | null vs positive | 3 | Fan 2023 | Goettel 2023 | contextual other | Fan 2023 (positive) vs Goettel 2023 (null) on contextual other | null vs positive (notable) | | null vs positive | 3 | Fan 2023 | Das 2023 | contextual other | Fan 2023 (positive) vs Das 2023 (null) on contextual other | null vs positive (notable) | | agreement | 1 | Fan 2023 | Corley 2023 | contextual other | Fan 2023 (positive) vs Corley 2023 (positive) on contextual other | agreement (minor) | | mechanism vs clinical | 4 | Fan 2023 | Nasiri 2025 | contextual other | Fan 2023 (contextual other, mechanistic) vs Nasiri 2025 (cardiometabolic, direct) | mechanism vs clinical (load-bearing) | | null vs positive | 3 | Fan 2023 | Yakout 2025 | contextual other | Fan 2023 (positive) vs Yakout 2025 (null) on contextual other | null vs positive (notable) | | disagreement | 5 | Fan 2023 | Yin 2025 | contextual other | Fan 2023 (positive) vs Yin 2025 (negative) on contextual other | disagreement (load-bearing) | | null vs positive | 3 | Fan 2023 | Madaria 2025 | contextual other | Fan 2023 (positive) vs Madaria 2025 (null) on contextual other | null vs positive (notable) | | null vs positive | 3 | Fan 2023 | Moabedi 2025 | contextual other | Fan 2023 (positive) vs Moabedi 2025 (null) on contextual other | null vs positive (notable) | | null vs positive | 3 | Fan 2023 | Locker 2025 | contextual other | Fan 2023 (positive) vs Locker 2025 (null) on contextual other | null vs positive (notable) | | null vs positive | 3 | Fan 2023 | Li 2025 | contextual other | Fan 2023 (positive) vs Li 2025 (null) on contextual other | null vs positive (notable) | | null vs positive | 3 | Fan 2023 | Ceylan 2025 | contextual other | Fan 2023 (positive) vs Ceylan 2025 (null) on contextual other | null vs positive (notable) | | null vs positive | 3 | Fan 2023 | Popa 2025 | contextual other | Fan 2023 (positive) vs Popa 2025 (null) on contextual other | null vs positive (notable) | | mechanism vs clinical | 4 | Fan 2023 | Ramos-Hernandez 2026 | contextual other | Fan 2023 (contextual other, mechanistic) vs Ramos-Hernandez 2026 (immune, direct) | mechanism vs clinical (load-bearing) | | null vs positive | 3 | Fan 2023 | Msigwa 2026 | contextual other | Fan 2023 (positive) vs Msigwa 2026 (null) on contextual other | null vs positive (notable) | | disagreement | 5 | Fan 2023 | Tambunan 2026 | contextual other | Fan 2023 (positive) vs Tambunan 2026 (negative) on contextual other | disagreement (load-bearing) | | null vs positive | 3 | Fan 2023 | Bloomer 2009 | contextual other | Fan 2023 (positive) vs Bloomer 2009 (null) on contextual other | null vs positive (notable) | | null vs positive | 3 | Fan 2023 | Cai 2011 | contextual other | Fan 2023 (positive) vs Cai 2011 (null) on contextual other | null vs positive (notable) | | null vs positive | 3 | Fan 2023 | Belsky 2017 | contextual other | Fan 2023 (positive) vs Belsky 2017 (null) on contextual other | null vs positive (notable) | | null vs positive | 3 | Fan 2023 | Quinn 2022 | contextual other | Fan 2023 (positive) vs Quinn 2022 (null) on contextual other | null vs positive (notable) | | mechanism vs clinical | 4 | Fan 2023 | Zahedi 2021 | contextual other | Fan 2023 (contextual other, mechanistic) vs Zahedi 2021 (immune, direct) | mechanism vs clinical (load-bearing) | | null vs positive | 3 | Nikrad 2023 | Wallace 2023 | contextual other | Nikrad 2023 (negative) vs Wallace 2023 (null) on contextual other | null vs positive (notable) | | null vs positive | 3 | Nikrad 2023 | Goettel 2023 | contextual other | Nikrad 2023 (negative) vs Goettel 2023 (null) on contextual other | null vs positive (notable) | | null vs positive | 3 | Nikrad 2023 | Das 2023 | contextual other | Nikrad 2023 (negative) vs Das 2023 (null) on contextual other | null vs positive (notable) | | disagreement | 5 | Nikrad 2023 | Corley 2023 | contextual other | Nikrad 2023 (negative) vs Corley 2023 (positive) on contextual other | disagreement (load-bearing) | | null vs positive | 3 | Nikrad 2023 | Yakout 2025 | contextual other | Nikrad 2023 (negative) vs Yakout 2025 (null) on contextual other | null vs positive (notable) | | agreement | 1 | Nikrad 2023 | Yin 2025 | contextual other | Nikrad 2023 (negative) vs Yin 2025 (negative) on contextual other | agreement (minor) | | null vs positive | 3 | Nikrad 2023 | Madaria 2025 | contextual other | Nikrad 2023 (negative) vs Madaria 2025 (null) on contextual other | null vs positive (notable) | | null vs positive | 3 | Nikrad 2023 | Moabedi 2025 | contextual other | Nikrad 2023 (negative) vs Moabedi 2025 (null) on contextual other | null vs positive (notable) | | null vs positive | 3 | Nikrad 2023 | Locker 2025 | contextual other | Nikrad 2023 (negative) vs Locker 2025 (null) on contextual other | null vs positive (notable) | | null vs positive | 3 | Nikrad 2023 | Li 2025 | contextual other | Nikrad 2023 (negative) vs Li 2025 (null) on contextual other | null vs positive (notable) | | null vs positive | 3 | Nikrad 2023 | Ceylan 2025 | contextual other | Nikrad 2023 (negative) vs Ceylan 2025 (null) on contextual other | null vs positive (notable) | | null vs positive | 3 | Nikrad 2023 | Popa 2025 | contextual other | Nikrad 2023 (negative) vs Popa 2025 (null) on contextual other | null vs positive (notable) | | null vs positive | 3 | Nikrad 2023 | Msigwa 2026 | contextual other | Nikrad 2023 (negative) vs Msigwa 2026 (null) on contextual other | null vs positive (notable) | | agreement | 1 | Nikrad 2023 | Tambunan 2026 | contextual other | Nikrad 2023 (negative) vs Tambunan 2026 (negative) on contextual other | agreement (minor) | | null vs positive | 3 | Nikrad 2023 | Bloomer 2009 | contextual other | Nikrad 2023 (negative) vs Bloomer 2009 (null) on contextual other | null vs positive (notable) | | null vs positive | 3 | Nikrad 2023 | Cai 2011 | contextual other | Nikrad 2023 (negative) vs Cai 2011 (null) on contextual other | null vs positive (notable) | | null vs positive | 3 | Nikrad 2023 | Belsky 2017 | contextual other | Nikrad 2023 (negative) vs Belsky 2017 (null) on contextual other | null vs positive (notable) | | null vs positive | 3 | Nikrad 2023 | Quinn 2022 | contextual other | Nikrad 2023 (negative) vs Quinn 2022 (null) on contextual other | null vs positive (notable) | | agreement | 1 | Wallace 2023 | Goettel 2023 | contextual other | Wallace 2023 (null) vs Goettel 2023 (null) on contextual other | agreement (minor) | | null vs positive | 3 | Wallace 2023 | Pleguezuelos 2023 | contextual other | Wallace 2023 (null) vs Pleguezuelos 2023 (unclear) on contextual other | null vs positive (notable) | | agreement | 1 | Wallace 2023 | Das 2023 | contextual other | Wallace 2023 (null) vs Das 2023 (null) on contextual other | agreement (minor) | | null vs positive | 3 | Wallace 2023 | Corley 2023 | contextual other | Wallace 2023 (null) vs Corley 2023 (positive) on contextual other | null vs positive (notable) | | agreement | 1 | Wallace 2023 | Yakout 2025 | contextual other | Wallace 2023 (null) vs Yakout 2025 (null) on contextual other | agreement (minor) | | null vs positive | 3 | Wallace 2023 | Yin 2025 | contextual other | Wallace 2023 (null) vs Yin 2025 (negative) on contextual other | null vs positive (notable) | | agreement | 1 | Wallace 2023 | Madaria 2025 | contextual other | Wallace 2023 (null) vs Madaria 2025 (null) on contextual other | agreement (minor) | | agreement | 1 | Wallace 2023 | Moabedi 2025 | contextual other | Wallace 2023 (null) vs Moabedi 2025 (null) on contextual other | agreement (minor) | | agreement | 1 | Wallace 2023 | Locker 2025 | contextual other | Wallace 2023 (null) vs Locker 2025 (null) on contextual other | agreement (minor) | | agreement | 1 | Wallace 2023 | Li 2025 | contextual other | Wallace 2023 (null) vs Li 2025 (null) on contextual other | agreement (minor) | | agreement | 1 | Wallace 2023 | Ceylan 2025 | contextual other | Wallace 2023 (null) vs Ceylan 2025 (null) on contextual other | agreement (minor) | | null vs positive | 3 | Wallace 2023 | Demir 2025 | contextual other | Wallace 2023 (null) vs Demir 2025 (unclear) on contextual other | null vs positive (notable) | | agreement | 1 | Wallace 2023 | Popa 2025 | contextual other | Wallace 2023 (null) vs Popa 2025 (null) on contextual other | agreement (minor) | | agreement | 1 | Wallace 2023 | Msigwa 2026 | contextual other | Wallace 2023 (null) vs Msigwa 2026 (null) on contextual other | agreement (minor) | | null vs positive | 3 | Wallace 2023 | Tambunan 2026 | contextual other | Wallace 2023 (null) vs Tambunan 2026 (negative) on contextual other | null vs positive (notable) | | agreement | 1 | Wallace 2023 | Bloomer 2009 | contextual other | Wallace 2023 (null) vs Bloomer 2009 (null) on contextual other | agreement (minor) | | agreement | 1 | Wallace 2023 | Cai 2011 | contextual other | Wallace 2023 (null) vs Cai 2011 (null) on contextual other | agreement (minor) | | agreement | 1 | Wallace 2023 | Belsky 2017 | contextual other | Wallace 2023 (null) vs Belsky 2017 (null) on contextual other | agreement (minor) | | null vs positive | 3 | Wallace 2023 | Yen 2020 | contextual other | Wallace 2023 (null) vs Yen 2020 (unclear) on contextual other | null vs positive (notable) | | agreement | 1 | Wallace 2023 | Quinn 2022 | contextual other | Wallace 2023 (null) vs Quinn 2022 (null) on contextual other | agreement (minor) | | null vs positive | 3 | Goettel 2023 | Pleguezuelos 2023 | contextual other | Goettel 2023 (null) vs Pleguezuelos 2023 (unclear) on contextual other | null vs positive (notable) | | agreement | 1 | Goettel 2023 | Das 2023 | contextual other | Goettel 2023 (null) vs Das 2023 (null) on contextual other | agreement (minor) | | null vs positive | 3 | Goettel 2023 | Corley 2023 | contextual other | Goettel 2023 (null) vs Corley 2023 (positive) on contextual other | null vs positive (notable) | | agreement | 1 | Goettel 2023 | Yakout 2025 | contextual other | Goettel 2023 (null) vs Yakout 2025 (null) on contextual other | agreement (minor) | | null vs positive | 3 | Goettel 2023 | Yin 2025 | contextual other | Goettel 2023 (null) vs Yin 2025 (negative) on contextual other | null vs positive (notable) | | agreement | 1 | Goettel 2023 | Madaria 2025 | contextual other | Goettel 2023 (null) vs Madaria 2025 (null) on contextual other | agreement (minor) | | agreement | 1 | Goettel 2023 | Moabedi 2025 | contextual other | Goettel 2023 (null) vs Moabedi 2025 (null) on contextual other | agreement (minor) | | agreement | 1 | Goettel 2023 | Locker 2025 | contextual other | Goettel 2023 (null) vs Locker 2025 (null) on contextual other | agreement (minor) | | agreement | 1 | Goettel 2023 | Li 2025 | contextual other | Goettel 2023 (null) vs Li 2025 (null) on contextual other | agreement (minor) | | agreement | 1 | Goettel 2023 | Ceylan 2025 | contextual other | Goettel 2023 (null) vs Ceylan 2025 (null) on contextual other | agreement (minor) | | null vs positive | 3 | Goettel 2023 | Demir 2025 | contextual other | Goettel 2023 (null) vs Demir 2025 (unclear) on contextual other | null vs positive (notable) | | agreement | 1 | Goettel 2023 | Popa 2025 | contextual other | Goettel 2023 (null) vs Popa 2025 (null) on contextual other | agreement (minor) | | agreement | 1 | Goettel 2023 | Msigwa 2026 | contextual other | Goettel 2023 (null) vs Msigwa 2026 (null) on contextual other | agreement (minor) | | null vs positive | 3 | Goettel 2023 | Tambunan 2026 | contextual other | Goettel 2023 (null) vs Tambunan 2026 (negative) on contextual other | null vs positive (notable) | | agreement | 1 | Goettel 2023 | Bloomer 2009 | contextual other | Goettel 2023 (null) vs Bloomer 2009 (null) on contextual other | agreement (minor) | | agreement | 1 | Goettel 2023 | Cai 2011 | contextual other | Goettel 2023 (null) vs Cai 2011 (null) on contextual other | agreement (minor) | | agreement | 1 | Goettel 2023 | Belsky 2017 | contextual other | Goettel 2023 (null) vs Belsky 2017 (null) on contextual other | agreement (minor) | | null vs positive | 3 | Goettel 2023 | Yen 2020 | contextual other | Goettel 2023 (null) vs Yen 2020 (unclear) on contextual other | null vs positive (notable) | | agreement | 1 | Goettel 2023 | Quinn 2022 | contextual other | Goettel 2023 (null) vs Quinn 2022 (null) on contextual other | agreement (minor) | | null vs positive | 3 | Pleguezuelos 2023 | Das 2023 | contextual other | Pleguezuelos 2023 (unclear) vs Das 2023 (null) on contextual other | null vs positive (notable) | | null vs positive | 3 | Pleguezuelos 2023 | Yakout 2025 | contextual other | Pleguezuelos 2023 (unclear) vs Yakout 2025 (null) on contextual other | null vs positive (notable) | | null vs positive | 3 | Pleguezuelos 2023 | Madaria 2025 | contextual other | Pleguezuelos 2023 (unclear) vs Madaria 2025 (null) on contextual other | null vs positive (notable) | | null vs positive | 3 | Pleguezuelos 2023 | Moabedi 2025 | contextual other | Pleguezuelos 2023 (unclear) vs Moabedi 2025 (null) on contextual other | null vs positive (notable) | | null vs positive | 3 | Pleguezuelos 2023 | Locker 2025 | contextual other | Pleguezuelos 2023 (unclear) vs Locker 2025 (null) on contextual other | null vs positive (notable) | | null vs positive | 3 | Pleguezuelos 2023 | Li 2025 | contextual other | Pleguezuelos 2023 (unclear) vs Li 2025 (null) on contextual other | null vs positive (notable) | | null vs positive | 3 | Pleguezuelos 2023 | Ceylan 2025 | contextual other | Pleguezuelos 2023 (unclear) vs Ceylan 2025 (null) on contextual other | null vs positive (notable) | | agreement | 1 | Pleguezuelos 2023 | Demir 2025 | contextual other | Pleguezuelos 2023 (unclear) vs Demir 2025 (unclear) on contextual other | agreement (minor) | | null vs positive | 3 | Pleguezuelos 2023 | Popa 2025 | contextual other | Pleguezuelos 2023 (unclear) vs Popa 2025 (null) on contextual other | null vs positive (notable) | | null vs positive | 3 | Pleguezuelos 2023 | Msigwa 2026 | contextual other | Pleguezuelos 2023 (unclear) vs Msigwa 2026 (null) on contextual other | null vs positive (notable) | | null vs positive | 3 | Pleguezuelos 2023 | Bloomer 2009 | contextual other | Pleguezuelos 2023 (unclear) vs Bloomer 2009 (null) on contextual other | null vs positive (notable) | | null vs positive | 3 | Pleguezuelos 2023 | Cai 2011 | contextual other | Pleguezuelos 2023 (unclear) vs Cai 2011 (null) on contextual other | null vs positive (notable) | | null vs positive | 3 | Pleguezuelos 2023 | Belsky 2017 | contextual other | Pleguezuelos 2023 (unclear) vs Belsky 2017 (null) on contextual other | null vs positive (notable) | | agreement | 1 | Pleguezuelos 2023 | Yen 2020 | contextual other | Pleguezuelos 2023 (unclear) vs Yen 2020 (unclear) on contextual other | agreement (minor) | | null vs positive | 3 | Pleguezuelos 2023 | Quinn 2022 | contextual other | Pleguezuelos 2023 (unclear) vs Quinn 2022 (null) on contextual other | null vs positive (notable) | | null vs positive | 3 | Das 2023 | Corley 2023 | contextual other | Das 2023 (null) vs Corley 2023 (positive) on contextual other | null vs positive (notable) | | agreement | 1 | Das 2023 | Yakout 2025 | contextual other | Das 2023 (null) vs Yakout 2025 (null) on contextual other | agreement (minor) | | null vs positive | 3 | Das 2023 | Yin 2025 | contextual other | Das 2023 (null) vs Yin 2025 (negative) on contextual other | null vs positive (notable) | | agreement | 1 | Das 2023 | Madaria 2025 | contextual other | Das 2023 (null) vs Madaria 2025 (null) on contextual other | agreement (minor) | | agreement | 1 | Das 2023 | Moabedi 2025 | contextual other | Das 2023 (null) vs Moabedi 2025 (null) on contextual other | agreement (minor) | | agreement | 1 | Das 2023 | Locker 2025 | contextual other | Das 2023 (null) vs Locker 2025 (null) on contextual other | agreement (minor) | | agreement | 1 | Das 2023 | Li 2025 | contextual other | Das 2023 (null) vs Li 2025 (null) on contextual other | agreement (minor) | | agreement | 1 | Das 2023 | Ceylan 2025 | contextual other | Das 2023 (null) vs Ceylan 2025 (null) on contextual other | agreement (minor) | | null vs positive | 3 | Das 2023 | Demir 2025 | contextual other | Das 2023 (null) vs Demir 2025 (unclear) on contextual other | null vs positive (notable) | | agreement | 1 | Das 2023 | Popa 2025 | contextual other | Das 2023 (null) vs Popa 2025 (null) on contextual other | agreement (minor) | | agreement | 1 | Das 2023 | Msigwa 2026 | contextual other | Das 2023 (null) vs Msigwa 2026 (null) on contextual other | agreement (minor) | | null vs positive | 3 | Das 2023 | Tambunan 2026 | contextual other | Das 2023 (null) vs Tambunan 2026 (negative) on contextual other | null vs positive (notable) | | agreement | 1 | Das 2023 | Bloomer 2009 | contextual other | Das 2023 (null) vs Bloomer 2009 (null) on contextual other | agreement (minor) | | agreement | 1 | Das 2023 | Cai 2011 | contextual other | Das 2023 (null) vs Cai 2011 (null) on contextual other | agreement (minor) | | agreement | 1 | Das 2023 | Belsky 2017 | contextual other | Das 2023 (null) vs Belsky 2017 (null) on contextual other | agreement (minor) | | null vs positive | 3 | Das 2023 | Yen 2020 | contextual other | Das 2023 (null) vs Yen 2020 (unclear) on contextual other | null vs positive (notable) | | agreement | 1 | Das 2023 | Quinn 2022 | contextual other | Das 2023 (null) vs Quinn 2022 (null) on contextual other | agreement (minor) | | null vs positive | 3 | Corley 2023 | Yakout 2025 | contextual other | Corley 2023 (positive) vs Yakout 2025 (null) on contextual other | null vs positive (notable) | | disagreement | 5 | Corley 2023 | Yin 2025 | contextual other | Corley 2023 (positive) vs Yin 2025 (negative) on contextual other | disagreement (load-bearing) | | null vs positive | 3 | Corley 2023 | Madaria 2025 | contextual other | Corley 2023 (positive) vs Madaria 2025 (null) on contextual other | null vs positive (notable) | | null vs positive | 3 | Corley 2023 | Moabedi 2025 | contextual other | Corley 2023 (positive) vs Moabedi 2025 (null) on contextual other | null vs positive (notable) | | null vs positive | 3 | Corley 2023 | Locker 2025 | contextual other | Corley 2023 (positive) vs Locker 2025 (null) on contextual other | null vs positive (notable) | | null vs positive | 3 | Corley 2023 | Li 2025 | contextual other | Corley 2023 (positive) vs Li 2025 (null) on contextual other | null vs positive (notable) | | null vs positive | 3 | Corley 2023 | Ceylan 2025 | contextual other | Corley 2023 (positive) vs Ceylan 2025 (null) on contextual other | null vs positive (notable) | | null vs positive | 3 | Corley 2023 | Popa 2025 | contextual other | Corley 2023 (positive) vs Popa 2025 (null) on contextual other | null vs positive (notable) | | null vs positive | 3 | Corley 2023 | Msigwa 2026 | contextual other | Corley 2023 (positive) vs Msigwa 2026 (null) on contextual other | null vs positive (notable) | | disagreement | 5 | Corley 2023 | Tambunan 2026 | contextual other | Corley 2023 (positive) vs Tambunan 2026 (negative) on contextual other | disagreement (load-bearing) | | null vs positive | 3 | Corley 2023 | Bloomer 2009 | contextual other | Corley 2023 (positive) vs Bloomer 2009 (null) on contextual other | null vs positive (notable) | | null vs positive | 3 | Corley 2023 | Cai 2011 | contextual other | Corley 2023 (positive) vs Cai 2011 (null) on contextual other | null vs positive (notable) | | null vs positive | 3 | Corley 2023 | Belsky 2017 | contextual other | Corley 2023 (positive) vs Belsky 2017 (null) on contextual other | null vs positive (notable) | | null vs positive | 3 | Corley 2023 | Quinn 2022 | contextual other | Corley 2023 (positive) vs Quinn 2022 (null) on contextual other | null vs positive (notable) | | disagreement | 4 | Taylor 2023 | Spears 2026 | immune | Taylor 2023 (null) vs Spears 2026 (mixed) on immune | disagreement (load-bearing) | | agreement | 1 | Taylor 2023 | Alrasheed 2026 | immune | Taylor 2023 (null) vs Alrasheed 2026 (null) on immune | agreement (minor) | | null vs positive | 3 | Taylor 2023 | Ramos-Hernandez 2026 | immune | Taylor 2023 (null) vs Ramos-Hernandez 2026 (unclear) on immune | null vs positive (notable) | | null vs positive | 3 | Taylor 2023 | Jarvela-Reijonen 2020 | immune | Taylor 2023 (null) vs Jarvela-Reijonen 2020 (positive) on immune | null vs positive (notable) | | agreement | 1 | Taylor 2023 | Hoseini 2022 | immune | Taylor 2023 (null) vs Hoseini 2022 (null) on immune | agreement (minor) | | null vs positive | 3 | Taylor 2023 | Giang 2021 | immune | Taylor 2023 (null) vs Giang 2021 (unclear) on immune | null vs positive (notable) | | agreement | 1 | Taylor 2023 | Zahedi 2021 | immune | Taylor 2023 (null) vs Zahedi 2021 (null) on immune | agreement (minor) | | agreement | 1 | Taylor 2023 | Bahari 2023 | immune | Taylor 2023 (null) vs Bahari 2023 (null) on immune | agreement (minor) | | mechanism vs clinical | 4 | Nasiri 2025 | Cai 2011 | cardiometabolic | Nasiri 2025 (cardiometabolic, direct) vs Cai 2011 (contextual other, mechanistic) | mechanism vs clinical (load-bearing) | | null vs positive | 3 | Nasiri 2025 | Martens 2018 | cardiometabolic | Nasiri 2025 (null) vs Martens 2018 (unclear) on cardiometabolic | null vs positive (notable) | | agreement | 1 | Nasiri 2025 | Ghalichi 2022 | cardiometabolic | Nasiri 2025 (null) vs Ghalichi 2022 (null) on cardiometabolic | agreement (minor) | | agreement | 1 | Nasiri 2025 | Beese 2022 | cardiometabolic | Nasiri 2025 (null) vs Beese 2022 (null) on cardiometabolic | agreement (minor) | | null vs positive | 3 | Yakout 2025 | Yin 2025 | contextual other | Yakout 2025 (null) vs Yin 2025 (negative) on contextual other | null vs positive (notable) | | agreement | 1 | Yakout 2025 | Madaria 2025 | contextual other | Yakout 2025 (null) vs Madaria 2025 (null) on contextual other | agreement (minor) | | agreement | 1 | Yakout 2025 | Moabedi 2025 | contextual other | Yakout 2025 (null) vs Moabedi 2025 (null) on contextual other | agreement (minor) | | agreement | 1 | Yakout 2025 | Locker 2025 | contextual other | Yakout 2025 (null) vs Locker 2025 (null) on contextual other | agreement (minor) | | agreement | 1 | Yakout 2025 | Li 2025 | contextual other | Yakout 2025 (null) vs Li 2025 (null) on contextual other | agreement (minor) | | agreement | 1 | Yakout 2025 | Ceylan 2025 | contextual other | Yakout 2025 (null) vs Ceylan 2025 (null) on contextual other | agreement (minor) | | null vs positive | 3 | Yakout 2025 | Demir 2025 | contextual other | Yakout 2025 (null) vs Demir 2025 (unclear) on contextual other | null vs positive (notable) | | agreement | 1 | Yakout 2025 | Popa 2025 | contextual other | Yakout 2025 (null) vs Popa 2025 (null) on contextual other | agreement (minor) | | agreement | 1 | Yakout 2025 | Msigwa 2026 | contextual other | Yakout 2025 (null) vs Msigwa 2026 (null) on contextual other | agreement (minor) | | null vs positive | 3 | Yakout 2025 | Tambunan 2026 | contextual other | Yakout 2025 (null) vs Tambunan 2026 (negative) on contextual other | null vs positive (notable) | | agreement | 1 | Yakout 2025 | Bloomer 2009 | contextual other | Yakout 2025 (null) vs Bloomer 2009 (null) on contextual other | agreement (minor) | | agreement | 1 | Yakout 2025 | Cai 2011 | contextual other | Yakout 2025 (null) vs Cai 2011 (null) on contextual other | agreement (minor) | | agreement | 1 | Yakout 2025 | Belsky 2017 | contextual other | Yakout 2025 (null) vs Belsky 2017 (null) on contextual other | agreement (minor) | | null vs positive | 3 | Yakout 2025 | Yen 2020 | contextual other | Yakout 2025 (null) vs Yen 2020 (unclear) on contextual other | null vs positive (notable) | | agreement | 1 | Yakout 2025 | Quinn 2022 | contextual other | Yakout 2025 (null) vs Quinn 2022 (null) on contextual other | agreement (minor) | | null vs positive | 3 | Terhalle 2025 | Dessie 2025 | longevity | Terhalle 2025 (null) vs Dessie 2025 (negative) on longevity | null vs positive (notable) | | agreement | 1 | Terhalle 2025 | Wang 2020 | longevity | Terhalle 2025 (null) vs Wang 2020 (null) on longevity | agreement (minor) | | agreement | 1 | Terhalle 2025 | Obeng-Gyasi 2022 | longevity | Terhalle 2025 (null) vs Obeng-Gyasi 2022 (null) on longevity | agreement (minor) | | null vs positive | 3 | Terhalle 2025 | Parker 2022 | longevity | Terhalle 2025 (null) vs Parker 2022 (unclear) on longevity | null vs positive (notable) | | null vs positive | 3 | Yin 2025 | Madaria 2025 | contextual other | Yin 2025 (negative) vs Madaria 2025 (null) on contextual other | null vs positive (notable) | | null vs positive | 3 | Yin 2025 | Moabedi 2025 | contextual other | Yin 2025 (negative) vs Moabedi 2025 (null) on contextual other | null vs positive (notable) | | null vs positive | 3 | Yin 2025 | Locker 2025 | contextual other | Yin 2025 (negative) vs Locker 2025 (null) on contextual other | null vs positive (notable) | | null vs positive | 3 | Yin 2025 | Li 2025 | contextual other | Yin 2025 (negative) vs Li 2025 (null) on contextual other | null vs positive (notable) | | null vs positive | 3 | Yin 2025 | Ceylan 2025 | contextual other | Yin 2025 (negative) vs Ceylan 2025 (null) on contextual other | null vs positive (notable) | | null vs positive | 3 | Yin 2025 | Popa 2025 | contextual other | Yin 2025 (negative) vs Popa 2025 (null) on contextual other | null vs positive (notable) | | null vs positive | 3 | Yin 2025 | Msigwa 2026 | contextual other | Yin 2025 (negative) vs Msigwa 2026 (null) on contextual other | null vs positive (notable) | | agreement | 1 | Yin 2025 | Tambunan 2026 | contextual other | Yin 2025 (negative) vs Tambunan 2026 (negative) on contextual other | agreement (minor) | | null vs positive | 3 | Yin 2025 | Bloomer 2009 | contextual other | Yin 2025 (negative) vs Bloomer 2009 (null) on contextual other | null vs positive (notable) | | null vs positive | 3 | Yin 2025 | Cai 2011 | contextual other | Yin 2025 (negative) vs Cai 2011 (null) on contextual other | null vs positive (notable) | | null vs positive | 3 | Yin 2025 | Belsky 2017 | contextual other | Yin 2025 (negative) vs Belsky 2017 (null) on contextual other | null vs positive (notable) | | null vs positive | 3 | Yin 2025 | Quinn 2022 | contextual other | Yin 2025 (negative) vs Quinn 2022 (null) on contextual other | null vs positive (notable) | | agreement | 1 | Bahari 2025 | Saadh 2025 | dosing pharmacokinetics | Bahari 2025 (null) vs Saadh 2025 (null) on dosing pharmacokinetics | agreement (minor) | | agreement | 1 | Bahari 2025 | Khor 2021 | dosing pharmacokinetics | Bahari 2025 (null) vs Khor 2021 (null) on dosing pharmacokinetics | agreement (minor) | | agreement | 1 | Bahari 2025 | Nguyen 2021 | dosing pharmacokinetics | Bahari 2025 (null) vs Nguyen 2021 (null) on dosing pharmacokinetics | agreement (minor) | | agreement | 1 | Madaria 2025 | Moabedi 2025 | contextual other | Madaria 2025 (null) vs Moabedi 2025 (null) on contextual other | agreement (minor) | | agreement | 1 | Madaria 2025 | Locker 2025 | contextual other | Madaria 2025 (null) vs Locker 2025 (null) on contextual other | agreement (minor) | | agreement | 1 | Madaria 2025 | Li 2025 | contextual other | Madaria 2025 (null) vs Li 2025 (null) on contextual other | agreement (minor) | | agreement | 1 | Madaria 2025 | Ceylan 2025 | contextual other | Madaria 2025 (null) vs Ceylan 2025 (null) on contextual other | agreement (minor) | | null vs positive | 3 | Madaria 2025 | Demir 2025 | contextual other | Madaria 2025 (null) vs Demir 2025 (unclear) on contextual other | null vs positive (notable) | | agreement | 1 | Madaria 2025 | Popa 2025 | contextual other | Madaria 2025 (null) vs Popa 2025 (null) on contextual other | agreement (minor) | | agreement | 1 | Madaria 2025 | Msigwa 2026 | contextual other | Madaria 2025 (null) vs Msigwa 2026 (null) on contextual other | agreement (minor) | | null vs positive | 3 | Madaria 2025 | Tambunan 2026 | contextual other | Madaria 2025 (null) vs Tambunan 2026 (negative) on contextual other | null vs positive (notable) | | agreement | 1 | Madaria 2025 | Bloomer 2009 | contextual other | Madaria 2025 (null) vs Bloomer 2009 (null) on contextual other | agreement (minor) | | agreement | 1 | Madaria 2025 | Cai 2011 | contextual other | Madaria 2025 (null) vs Cai 2011 (null) on contextual other | agreement (minor) | | agreement | 1 | Madaria 2025 | Belsky 2017 | contextual other | Madaria 2025 (null) vs Belsky 2017 (null) on contextual other | agreement (minor) | | null vs positive | 3 | Madaria 2025 | Yen 2020 | contextual other | Madaria 2025 (null) vs Yen 2020 (unclear) on contextual other | null vs positive (notable) | | agreement | 1 | Madaria 2025 | Quinn 2022 | contextual other | Madaria 2025 (null) vs Quinn 2022 (null) on contextual other | agreement (minor) | | agreement | 1 | Moabedi 2025 | Locker 2025 | contextual other | Moabedi 2025 (null) vs Locker 2025 (null) on contextual other | agreement (minor) | | agreement | 1 | Moabedi 2025 | Li 2025 | contextual other | Moabedi 2025 (null) vs Li 2025 (null) on contextual other | agreement (minor) | | agreement | 1 | Moabedi 2025 | Ceylan 2025 | contextual other | Moabedi 2025 (null) vs Ceylan 2025 (null) on contextual other | agreement (minor) | | null vs positive | 3 | Moabedi 2025 | Demir 2025 | contextual other | Moabedi 2025 (null) vs Demir 2025 (unclear) on contextual other | null vs positive (notable) | | agreement | 1 | Moabedi 2025 | Popa 2025 | contextual other | Moabedi 2025 (null) vs Popa 2025 (null) on contextual other | agreement (minor) | | agreement | 1 | Moabedi 2025 | Msigwa 2026 | contextual other | Moabedi 2025 (null) vs Msigwa 2026 (null) on contextual other | agreement (minor) | | null vs positive | 3 | Moabedi 2025 | Tambunan 2026 | contextual other | Moabedi 2025 (null) vs Tambunan 2026 (negative) on contextual other | null vs positive (notable) | | agreement | 1 | Moabedi 2025 | Bloomer 2009 | contextual other | Moabedi 2025 (null) vs Bloomer 2009 (null) on contextual other | agreement (minor) | | agreement | 1 | Moabedi 2025 | Cai 2011 | contextual other | Moabedi 2025 (null) vs Cai 2011 (null) on contextual other | agreement (minor) | | agreement | 1 | Moabedi 2025 | Belsky 2017 | contextual other | Moabedi 2025 (null) vs Belsky 2017 (null) on contextual other | agreement (minor) | | null vs positive | 3 | Moabedi 2025 | Yen 2020 | contextual other | Moabedi 2025 (null) vs Yen 2020 (unclear) on contextual other | null vs positive (notable) | | agreement | 1 | Moabedi 2025 | Quinn 2022 | contextual other | Moabedi 2025 (null) vs Quinn 2022 (null) on contextual other | agreement (minor) | | agreement | 1 | Locker 2025 | Li 2025 | contextual other | Locker 2025 (null) vs Li 2025 (null) on contextual other | agreement (minor) | | agreement | 1 | Locker 2025 | Ceylan 2025 | contextual other | Locker 2025 (null) vs Ceylan 2025 (null) on contextual other | agreement (minor) | | null vs positive | 3 | Locker 2025 | Demir 2025 | contextual other | Locker 2025 (null) vs Demir 2025 (unclear) on contextual other | null vs positive (notable) | | agreement | 1 | Locker 2025 | Popa 2025 | contextual other | Locker 2025 (null) vs Popa 2025 (null) on contextual other | agreement (minor) | | agreement | 1 | Locker 2025 | Msigwa 2026 | contextual other | Locker 2025 (null) vs Msigwa 2026 (null) on contextual other | agreement (minor) | | null vs positive | 3 | Locker 2025 | Tambunan 2026 | contextual other | Locker 2025 (null) vs Tambunan 2026 (negative) on contextual other | null vs positive (notable) | | agreement | 1 | Locker 2025 | Bloomer 2009 | contextual other | Locker 2025 (null) vs Bloomer 2009 (null) on contextual other | agreement (minor) | | agreement | 1 | Locker 2025 | Cai 2011 | contextual other | Locker 2025 (null) vs Cai 2011 (null) on contextual other | agreement (minor) | | agreement | 1 | Locker 2025 | Belsky 2017 | contextual other | Locker 2025 (null) vs Belsky 2017 (null) on contextual other | agreement (minor) | | null vs positive | 3 | Locker 2025 | Yen 2020 | contextual other | Locker 2025 (null) vs Yen 2020 (unclear) on contextual other | null vs positive (notable) | | agreement | 1 | Locker 2025 | Quinn 2022 | contextual other | Locker 2025 (null) vs Quinn 2022 (null) on contextual other | agreement (minor) | | null vs positive | 3 | Dessie 2025 | Wang 2020 | longevity | Dessie 2025 (negative) vs Wang 2020 (null) on longevity | null vs positive (notable) | | null vs positive | 3 | Dessie 2025 | Obeng-Gyasi 2022 | longevity | Dessie 2025 (negative) vs Obeng-Gyasi 2022 (null) on longevity | null vs positive (notable) | | agreement | 1 | Li 2025 | Ceylan 2025 | contextual other | Li 2025 (null) vs Ceylan 2025 (null) on contextual other | agreement (minor) | | null vs positive | 3 | Li 2025 | Demir 2025 | contextual other | Li 2025 (null) vs Demir 2025 (unclear) on contextual other | null vs positive (notable) | | agreement | 1 | Li 2025 | Popa 2025 | contextual other | Li 2025 (null) vs Popa 2025 (null) on contextual other | agreement (minor) | | agreement | 1 | Li 2025 | Msigwa 2026 | contextual other | Li 2025 (null) vs Msigwa 2026 (null) on contextual other | agreement (minor) | | null vs positive | 3 | Li 2025 | Tambunan 2026 | contextual other | Li 2025 (null) vs Tambunan 2026 (negative) on contextual other | null vs positive (notable) | | agreement | 1 | Li 2025 | Bloomer 2009 | contextual other | Li 2025 (null) vs Bloomer 2009 (null) on contextual other | agreement (minor) | | agreement | 1 | Li 2025 | Cai 2011 | contextual other | Li 2025 (null) vs Cai 2011 (null) on contextual other | agreement (minor) | | agreement | 1 | Li 2025 | Belsky 2017 | contextual other | Li 2025 (null) vs Belsky 2017 (null) on contextual other | agreement (minor) | | null vs positive | 3 | Li 2025 | Yen 2020 | contextual other | Li 2025 (null) vs Yen 2020 (unclear) on contextual other | null vs positive (notable) | | agreement | 1 | Li 2025 | Quinn 2022 | contextual other | Li 2025 (null) vs Quinn 2022 (null) on contextual other | agreement (minor) | | null vs positive | 3 | Ceylan 2025 | Demir 2025 | contextual other | Ceylan 2025 (null) vs Demir 2025 (unclear) on contextual other | null vs positive (notable) | | agreement | 1 | Ceylan 2025 | Popa 2025 | contextual other | Ceylan 2025 (null) vs Popa 2025 (null) on contextual other | agreement (minor) | | agreement | 1 | Ceylan 2025 | Msigwa 2026 | contextual other | Ceylan 2025 (null) vs Msigwa 2026 (null) on contextual other | agreement (minor) | | null vs positive | 3 | Ceylan 2025 | Tambunan 2026 | contextual other | Ceylan 2025 (null) vs Tambunan 2026 (negative) on contextual other | null vs positive (notable) | | agreement | 1 | Ceylan 2025 | Bloomer 2009 | contextual other | Ceylan 2025 (null) vs Bloomer 2009 (null) on contextual other | agreement (minor) | | agreement | 1 | Ceylan 2025 | Cai 2011 | contextual other | Ceylan 2025 (null) vs Cai 2011 (null) on contextual other | agreement (minor) | | agreement | 1 | Ceylan 2025 | Belsky 2017 | contextual other | Ceylan 2025 (null) vs Belsky 2017 (null) on contextual other | agreement (minor) | | null vs positive | 3 | Ceylan 2025 | Yen 2020 | contextual other | Ceylan 2025 (null) vs Yen 2020 (unclear) on contextual other | null vs positive (notable) | | agreement | 1 | Ceylan 2025 | Quinn 2022 | contextual other | Ceylan 2025 (null) vs Quinn 2022 (null) on contextual other | agreement (minor) | | agreement | 1 | Saadh 2025 | Khor 2021 | dosing pharmacokinetics | Saadh 2025 (null) vs Khor 2021 (null) on dosing pharmacokinetics | agreement (minor) | | agreement | 1 | Saadh 2025 | Nguyen 2021 | dosing pharmacokinetics | Saadh 2025 (null) vs Nguyen 2021 (null) on dosing pharmacokinetics | agreement (minor) | | null vs positive | 3 | Demir 2025 | Popa 2025 | contextual other | Demir 2025 (unclear) vs Popa 2025 (null) on contextual other | null vs positive (notable) | | null vs positive | 3 | Demir 2025 | Msigwa 2026 | contextual other | Demir 2025 (unclear) vs Msigwa 2026 (null) on contextual other | null vs positive (notable) | | null vs positive | 3 | Demir 2025 | Bloomer 2009 | contextual other | Demir 2025 (unclear) vs Bloomer 2009 (null) on contextual other | null vs positive (notable) | | null vs positive | 3 | Demir 2025 | Cai 2011 | contextual other | Demir 2025 (unclear) vs Cai 2011 (null) on contextual other | null vs positive (notable) | | null vs positive | 3 | Demir 2025 | Belsky 2017 | contextual other | Demir 2025 (unclear) vs Belsky 2017 (null) on contextual other | null vs positive (notable) | | agreement | 1 | Demir 2025 | Yen 2020 | contextual other | Demir 2025 (unclear) vs Yen 2020 (unclear) on contextual other | agreement (minor) | | null vs positive | 3 | Demir 2025 | Quinn 2022 | contextual other | Demir 2025 (unclear) vs Quinn 2022 (null) on contextual other | null vs positive (notable) | | agreement | 1 | Popa 2025 | Msigwa 2026 | contextual other | Popa 2025 (null) vs Msigwa 2026 (null) on contextual other | agreement (minor) | | null vs positive | 3 | Popa 2025 | Tambunan 2026 | contextual other | Popa 2025 (null) vs Tambunan 2026 (negative) on contextual other | null vs positive (notable) | | agreement | 1 | Popa 2025 | Bloomer 2009 | contextual other | Popa 2025 (null) vs Bloomer 2009 (null) on contextual other | agreement (minor) | | agreement | 1 | Popa 2025 | Cai 2011 | contextual other | Popa 2025 (null) vs Cai 2011 (null) on contextual other | agreement (minor) | | agreement | 1 | Popa 2025 | Belsky 2017 | contextual other | Popa 2025 (null) vs Belsky 2017 (null) on contextual other | agreement (minor) | | null vs positive | 3 | Popa 2025 | Yen 2020 | contextual other | Popa 2025 (null) vs Yen 2020 (unclear) on contextual other | null vs positive (notable) | | agreement | 1 | Popa 2025 | Quinn 2022 | contextual other | Popa 2025 (null) vs Quinn 2022 (null) on contextual other | agreement (minor) | | disagreement | 4 | Spears 2026 | Alrasheed 2026 | immune | Spears 2026 (mixed) vs Alrasheed 2026 (null) on immune | disagreement (load-bearing) | | disagreement | 4 | Spears 2026 | Ramos-Hernandez 2026 | immune | Spears 2026 (mixed) vs Ramos-Hernandez 2026 (unclear) on immune | disagreement (load-bearing) | | disagreement | 4 | Spears 2026 | Jarvela-Reijonen 2020 | immune | Spears 2026 (mixed) vs Jarvela-Reijonen 2020 (positive) on immune | disagreement (load-bearing) | | disagreement | 4 | Spears 2026 | Hoseini 2022 | immune | Spears 2026 (mixed) vs Hoseini 2022 (null) on immune | disagreement (load-bearing) | | disagreement | 4 | Spears 2026 | Giang 2021 | immune | Spears 2026 (mixed) vs Giang 2021 (unclear) on immune | disagreement (load-bearing) | | disagreement | 4 | Spears 2026 | Zahedi 2021 | immune | Spears 2026 (mixed) vs Zahedi 2021 (null) on immune | disagreement (load-bearing) | | disagreement | 4 | Spears 2026 | Bahari 2023 | immune | Spears 2026 (mixed) vs Bahari 2023 (null) on immune | disagreement (load-bearing) | | null vs positive | 3 | Alrasheed 2026 | Ramos-Hernandez 2026 | immune | Alrasheed 2026 (null) vs Ramos-Hernandez 2026 (unclear) on immune | null vs positive (notable) | | null vs positive | 3 | Alrasheed 2026 | Jarvela-Reijonen 2020 | immune | Alrasheed 2026 (null) vs Jarvela-Reijonen 2020 (positive) on immune | null vs positive (notable) | | agreement | 1 | Alrasheed 2026 | Hoseini 2022 | immune | Alrasheed 2026 (null) vs Hoseini 2022 (null) on immune | agreement (minor) | | null vs positive | 3 | Alrasheed 2026 | Giang 2021 | immune | Alrasheed 2026 (null) vs Giang 2021 (unclear) on immune | null vs positive (notable) | | agreement | 1 | Alrasheed 2026 | Zahedi 2021 | immune | Alrasheed 2026 (null) vs Zahedi 2021 (null) on immune | agreement (minor) | | agreement | 1 | Alrasheed 2026 | Bahari 2023 | immune | Alrasheed 2026 (null) vs Bahari 2023 (null) on immune | agreement (minor) | | mechanism vs clinical | 4 | Ramos-Hernandez 2026 | Cai 2011 | immune | Ramos-Hernandez 2026 (immune, direct) vs Cai 2011 (contextual other, mechanistic) | mechanism vs clinical (load-bearing) | | null vs positive | 3 | Ramos-Hernandez 2026 | Hoseini 2022 | immune | Ramos-Hernandez 2026 (unclear) vs Hoseini 2022 (null) on immune | null vs positive (notable) | | agreement | 1 | Ramos-Hernandez 2026 | Giang 2021 | immune | Ramos-Hernandez 2026 (unclear) vs Giang 2021 (unclear) on immune | agreement (minor) | | null vs positive | 3 | Ramos-Hernandez 2026 | Zahedi 2021 | immune | Ramos-Hernandez 2026 (unclear) vs Zahedi 2021 (null) on immune | null vs positive (notable) | | null vs positive | 3 | Ramos-Hernandez 2026 | Bahari 2023 | immune | Ramos-Hernandez 2026 (unclear) vs Bahari 2023 (null) on immune | null vs positive (notable) | | null vs positive | 3 | Msigwa 2026 | Tambunan 2026 | contextual other | Msigwa 2026 (null) vs Tambunan 2026 (negative) on contextual other | null vs positive (notable) | | agreement | 1 | Msigwa 2026 | Bloomer 2009 | contextual other | Msigwa 2026 (null) vs Bloomer 2009 (null) on contextual other | agreement (minor) | | agreement | 1 | Msigwa 2026 | Cai 2011 | contextual other | Msigwa 2026 (null) vs Cai 2011 (null) on contextual other | agreement (minor) | | agreement | 1 | Msigwa 2026 | Belsky 2017 | contextual other | Msigwa 2026 (null) vs Belsky 2017 (null) on contextual other | agreement (minor) | | null vs positive | 3 | Msigwa 2026 | Yen 2020 | contextual other | Msigwa 2026 (null) vs Yen 2020 (unclear) on contextual other | null vs positive (notable) | | agreement | 1 | Msigwa 2026 | Quinn 2022 | contextual other | Msigwa 2026 (null) vs Quinn 2022 (null) on contextual other | agreement (minor) | | null vs positive | 3 | Tambunan 2026 | Bloomer 2009 | contextual other | Tambunan 2026 (negative) vs Bloomer 2009 (null) on contextual other | null vs positive (notable) | | null vs positive | 3 | Tambunan 2026 | Cai 2011 | contextual other | Tambunan 2026 (negative) vs Cai 2011 (null) on contextual other | null vs positive (notable) | | null vs positive | 3 | Tambunan 2026 | Belsky 2017 | contextual other | Tambunan 2026 (negative) vs Belsky 2017 (null) on contextual other | null vs positive (notable) | | null vs positive | 3 | Tambunan 2026 | Quinn 2022 | contextual other | Tambunan 2026 (negative) vs Quinn 2022 (null) on contextual other | null vs positive (notable) | | agreement | 1 | Bloomer 2009 | Cai 2011 | contextual other | Bloomer 2009 (null) vs Cai 2011 (null) on contextual other | agreement (minor) | | agreement | 1 | Bloomer 2009 | Belsky 2017 | contextual other | Bloomer 2009 (null) vs Belsky 2017 (null) on contextual other | agreement (minor) | | null vs positive | 3 | Bloomer 2009 | Yen 2020 | contextual other | Bloomer 2009 (null) vs Yen 2020 (unclear) on contextual other | null vs positive (notable) | | agreement | 1 | Bloomer 2009 | Quinn 2022 | contextual other | Bloomer 2009 (null) vs Quinn 2022 (null) on contextual other | agreement (minor) | | agreement | 1 | Cai 2011 | Belsky 2017 | contextual other | Cai 2011 (null) vs Belsky 2017 (null) on contextual other | agreement (minor) | | null vs positive | 3 | Cai 2011 | Yen 2020 | contextual other | Cai 2011 (null) vs Yen 2020 (unclear) on contextual other | null vs positive (notable) | | agreement | 1 | Cai 2011 | Quinn 2022 | contextual other | Cai 2011 (null) vs Quinn 2022 (null) on contextual other | agreement (minor) | | mechanism vs clinical | 4 | Cai 2011 | Zahedi 2021 | contextual other | Cai 2011 (contextual other, mechanistic) vs Zahedi 2021 (immune, direct) | mechanism vs clinical (load-bearing) | | null vs positive | 3 | Belsky 2017 | Yen 2020 | contextual other | Belsky 2017 (null) vs Yen 2020 (unclear) on contextual other | null vs positive (notable) | | agreement | 1 | Belsky 2017 | Quinn 2022 | contextual other | Belsky 2017 (null) vs Quinn 2022 (null) on contextual other | agreement (minor) | | null vs positive | 3 | Martens 2018 | Ghalichi 2022 | cardiometabolic | Martens 2018 (unclear) vs Ghalichi 2022 (null) on cardiometabolic | null vs positive (notable) | | null vs positive | 3 | Martens 2018 | Beese 2022 | cardiometabolic | Martens 2018 (unclear) vs Beese 2022 (null) on cardiometabolic | null vs positive (notable) | | agreement | 1 | Wang 2020 | Obeng-Gyasi 2022 | longevity | Wang 2020 (null) vs Obeng-Gyasi 2022 (null) on longevity | agreement (minor) | | null vs positive | 3 | Wang 2020 | Parker 2022 | longevity | Wang 2020 (null) vs Parker 2022 (unclear) on longevity | null vs positive (notable) | | null vs positive | 3 | Yen 2020 | Quinn 2022 | contextual other | Yen 2020 (unclear) vs Quinn 2022 (null) on contextual other | null vs positive (notable) | | null vs positive | 3 | Jarvela-Reijonen 2020 | Hoseini 2022 | immune | Jarvela-Reijonen 2020 (positive) vs Hoseini 2022 (null) on immune | null vs positive (notable) | | null vs positive | 3 | Jarvela-Reijonen 2020 | Zahedi 2021 | immune | Jarvela-Reijonen 2020 (positive) vs Zahedi 2021 (null) on immune | null vs positive (notable) | | null vs positive | 3 | Jarvela-Reijonen 2020 | Bahari 2023 | immune | Jarvela-Reijonen 2020 (positive) vs Bahari 2023 (null) on immune | null vs positive (notable) | | agreement | 1 | Khor 2021 | Nguyen 2021 | dosing pharmacokinetics | Khor 2021 (null) vs Nguyen 2021 (null) on dosing pharmacokinetics | agreement (minor) | | null vs positive | 3 | Hoseini 2022 | Giang 2021 | immune | Hoseini 2022 (null) vs Giang 2021 (unclear) on immune | null vs positive (notable) | | agreement | 1 | Hoseini 2022 | Zahedi 2021 | immune | Hoseini 2022 (null) vs Zahedi 2021 (null) on immune | agreement (minor) | | agreement | 1 | Hoseini 2022 | Bahari 2023 | immune | Hoseini 2022 (null) vs Bahari 2023 (null) on immune | agreement (minor) | | null vs positive | 3 | Obeng-Gyasi 2022 | Parker 2022 | longevity | Obeng-Gyasi 2022 (null) vs Parker 2022 (unclear) on longevity | null vs positive (notable) | | null vs positive | 3 | Giang 2021 | Zahedi 2021 | immune | Giang 2021 (unclear) vs Zahedi 2021 (null) on immune | null vs positive (notable) | | null vs positive | 3 | Giang 2021 | Bahari 2023 | immune | Giang 2021 (unclear) vs Bahari 2023 (null) on immune | null vs positive (notable) | | agreement | 1 | Ghalichi 2022 | Beese 2022 | cardiometabolic | Ghalichi 2022 (null) vs Beese 2022 (null) on cardiometabolic | agreement (minor) | | agreement | 1 | Zahedi 2021 | Bahari 2023 | immune | Zahedi 2021 (null) vs Bahari 2023 (null) on immune | agreement (minor) |

Table 4 (supplemental): Design-Level Evidence Weighting Heuristic

Per-domain grades are derived from each study's evidence tier (A1/A2/B1/B2/C1/C2) — they capture design-level limitations, NOT a formal per-paper risk-of-bias assessment from the source text. Domains follow design-family categories for randomized, observational, animal, and systematic-review evidence; n/a indicates the domain is not meaningful for that design (e.g. blinding for an observational cohort). The Weight in synthesis column is the qualitative weighting the synthesis applies to each source — derived from tier × directness × overall RoB.

Citation	Tier	Tool	Allocation	Blinding	Attrition	Outcome measurement	Reporting	Confounding control	Generalizability	Overall RoB	Weight in synthesis	Effect direction notes
Ramos-Hernandez 2026	A1	Cochrane RoB-2	low	low	moderate	low	low	low	moderate	low	load-bearing (direct clinical RCT)	signed claims without significance signal
Giang 2021	B1	AMSTAR-2 (review)	unclear	unclear	unclear	unclear	moderate	moderate	moderate	unclear	supporting (synthesis evidence)	signed claims without significance signal
Nasiri 2025	A1	Cochrane RoB-2	low	low	moderate	low	low	low	moderate	low	load-bearing (direct clinical RCT)	primary endpoint did not reach significance
Moel 2025	B2	ROBINS-I	n/a	n/a	moderate	moderate	moderate	high	moderate	moderate	contextual (observational signal)	primary endpoint did not reach significance
Lee 2025	B2	ROBINS-I	n/a	n/a	moderate	moderate	moderate	high	moderate	moderate	contextual (observational signal)	primary endpoint did not reach significance
Terhalle 2025	B2	ROBINS-I	n/a	n/a	moderate	moderate	moderate	high	moderate	moderate	contextual (observational signal)	primary endpoint did not reach significance
Fan 2023	C1	SYRCLE (animal)	low	n/a	low	moderate	moderate	n/a	high	low	hypothesis-generating (preclinical mechanism)	positive effect — see Tables 1/2
Martens 2018	B2	ROBINS-I	n/a	n/a	moderate	moderate	moderate	high	moderate	moderate	contextual (observational signal)	signed claims without significance signal
Jarvela-Reijonen 2020	B2	ROBINS-I	n/a	n/a	moderate	moderate	moderate	high	moderate	moderate	contextual (observational signal)	positive effect — see Tables 1/2
Quinn 2022	B2	ROBINS-I	n/a	n/a	moderate	moderate	moderate	high	moderate	moderate	contextual (observational signal)	primary endpoint did not reach significance
Nikrad 2023	A2	Cochrane RoB-2	low	moderate	moderate	moderate	low	low	high	low	mechanistic (human RCT, biomarker endpoint)	negative effect — see Tables 1/2
Wang 2020	B2	ROBINS-I	n/a	n/a	moderate	moderate	moderate	high	moderate	moderate	contextual (observational signal)	primary endpoint did not reach significance
Bahari 2025	B2	ROBINS-I	n/a	n/a	moderate	moderate	moderate	high	moderate	moderate	contextual (observational signal)	primary endpoint did not reach significance
Li 2025	B1	AMSTAR-2 (review)	unclear	unclear	unclear	unclear	moderate	moderate	moderate	unclear	supporting (synthesis evidence)	primary endpoint did not reach significance
Das 2023	B2	ROBINS-I	n/a	n/a	moderate	moderate	moderate	high	moderate	moderate	contextual (observational signal)	primary endpoint did not reach significance
Tambunan 2026	B2	ROBINS-I	n/a	n/a	moderate	moderate	moderate	high	moderate	moderate	contextual (observational signal)	negative effect — see Tables 1/2
Bloomer 2009	B2	ROBINS-I	n/a	n/a	moderate	moderate	moderate	high	moderate	moderate	contextual (observational signal)	primary endpoint did not reach significance
Alrasheed 2026	B1	AMSTAR-2 (review)	unclear	unclear	unclear	unclear	moderate	moderate	moderate	unclear	supporting (synthesis evidence)	primary endpoint did not reach significance
Khor 2021	B2	ROBINS-I	n/a	n/a	moderate	moderate	moderate	high	moderate	moderate	contextual (observational signal)	primary endpoint did not reach significance
Goettel 2023	B2	ROBINS-I	n/a	n/a	moderate	moderate	moderate	high	moderate	moderate	contextual (observational signal)	primary endpoint did not reach significance
Ceylan 2025	B2	ROBINS-I	n/a	n/a	moderate	moderate	moderate	high	moderate	moderate	contextual (observational signal)	primary endpoint did not reach significance
Saadh 2025	B2	ROBINS-I	n/a	n/a	moderate	moderate	moderate	high	moderate	moderate	contextual (observational signal)	primary endpoint did not reach significance
Msigwa 2026	B2	ROBINS-I	n/a	n/a	moderate	moderate	moderate	high	moderate	moderate	contextual (observational signal)	primary endpoint did not reach significance
Yen 2020	B2	ROBINS-I	n/a	n/a	moderate	moderate	moderate	high	moderate	moderate	contextual (observational signal)	signed claims without significance signal
Hoseini 2022	B2	ROBINS-I	n/a	n/a	moderate	moderate	moderate	high	moderate	moderate	contextual (observational signal)	primary endpoint did not reach significance
Dessie 2025	B1	AMSTAR-2 (review)	unclear	unclear	unclear	unclear	moderate	moderate	moderate	unclear	supporting (synthesis evidence)	negative effect — see Tables 1/2
Yin 2025	B2	ROBINS-I	n/a	n/a	moderate	moderate	moderate	high	moderate	moderate	contextual (observational signal)	negative effect — see Tables 1/2
Knufinke 2023	B1	AMSTAR-2 (review)	unclear	unclear	unclear	unclear	moderate	moderate	moderate	unclear	supporting (synthesis evidence)	signed claims without significance signal
Yakout 2025	B2	ROBINS-I	n/a	n/a	moderate	moderate	moderate	high	moderate	moderate	contextual (observational signal)	primary endpoint did not reach significance
Corley 2023	B2	ROBINS-I	n/a	n/a	moderate	moderate	moderate	high	moderate	moderate	contextual (observational signal)	positive effect — see Tables 1/2
Taylor 2023	B2	ROBINS-I	n/a	n/a	moderate	moderate	moderate	high	moderate	moderate	contextual (observational signal)	primary endpoint did not reach significance
Demir 2025	B2	ROBINS-I	n/a	n/a	moderate	moderate	moderate	high	moderate	moderate	contextual (observational signal)	signed claims without significance signal
Obeng-Gyasi 2022	B2	ROBINS-I	n/a	n/a	moderate	moderate	moderate	high	moderate	moderate	contextual (observational signal)	primary endpoint did not reach significance
Nguyen 2021	B2	ROBINS-I	n/a	n/a	moderate	moderate	moderate	high	moderate	moderate	contextual (observational signal)	primary endpoint did not reach significance
Zhang 2021	B2	ROBINS-I	n/a	n/a	moderate	moderate	moderate	high	moderate	moderate	contextual (observational signal)	primary endpoint did not reach significance
Moabedi 2025	B2	ROBINS-I	n/a	n/a	moderate	moderate	moderate	high	moderate	moderate	contextual (observational signal)	primary endpoint did not reach significance
Cai 2011	C1	SYRCLE (animal)	low	n/a	low	moderate	moderate	n/a	high	low	hypothesis-generating (preclinical mechanism)	primary endpoint did not reach significance
Pleguezuelos 2023	B2	ROBINS-I	n/a	n/a	moderate	moderate	moderate	high	moderate	moderate	contextual (observational signal)	signed claims without significance signal
Spears 2026	B2	ROBINS-I	n/a	n/a	moderate	moderate	moderate	high	moderate	moderate	contextual (observational signal)	internal contradiction across endpoints
Wallace 2023	B2	ROBINS-I	n/a	n/a	moderate	moderate	moderate	high	moderate	moderate	contextual (observational signal)	primary endpoint did not reach significance
Belsky 2017	B2	ROBINS-I	n/a	n/a	moderate	moderate	moderate	high	moderate	moderate	contextual (observational signal)	primary endpoint did not reach significance
Beese 2022	B2	ROBINS-I	n/a	n/a	moderate	moderate	moderate	high	moderate	moderate	contextual (observational signal)	primary endpoint did not reach significance
Locker 2025	B2	ROBINS-I	n/a	n/a	moderate	moderate	moderate	high	moderate	moderate	contextual (observational signal)	primary endpoint did not reach significance
Parker 2022	B1	AMSTAR-2 (review)	unclear	unclear	unclear	unclear	moderate	moderate	moderate	unclear	supporting (synthesis evidence)	signed claims without significance signal
Popa 2025	B2	ROBINS-I	n/a	n/a	moderate	moderate	moderate	high	moderate	moderate	contextual (observational signal)	primary endpoint did not reach significance
Bahari 2023	B1	AMSTAR-2 (review)	unclear	unclear	unclear	unclear	moderate	moderate	moderate	unclear	supporting (synthesis evidence)	primary endpoint did not reach significance
Madaria 2025	B2	ROBINS-I	n/a	n/a	moderate	moderate	moderate	high	moderate	moderate	contextual (observational signal)	primary endpoint did not reach significance
Ghalichi 2022	B2	ROBINS-I	n/a	n/a	moderate	moderate	moderate	high	moderate	moderate	contextual (observational signal)	primary endpoint did not reach significance
Moskalevska 2026	B1	AMSTAR-2 (review)	unclear	unclear	unclear	unclear	moderate	moderate	moderate	unclear	supporting (synthesis evidence)	signed claims without significance signal
Zahedi 2021	A1	Cochrane RoB-2	low	low	moderate	low	low	low	moderate	low	load-bearing (direct clinical RCT)	primary endpoint did not reach significance

Table 5 (supplemental): Per-Paper Numeric Index

Top-N quantitative claims per paper — the underlying corpus numerics that power Q2 trace and Q9 density. One row per (paper × claim) tuple, prioritised by claim type (p-value > percentage > ratio > unit-value).

Citation	Section	Type	Value	Units
Ramos-Hernandez 2026	results	p-value	P < 0.05	—
Ramos-Hernandez 2026	results	mean ± SD	72.98 ± 20.44	—
Ramos-Hernandez 2026	results	sample size	n = 30	—
Ramos-Hernandez 2026	results	sample size	n = 20	—
Ramos-Hernandez 2026	results	sample size	n = 10	—
Giang 2021	results	p-value	P = 0.04	—
Giang 2021	results	percentage	18%	%
Giang 2021	results	unit value	1.6 mg	mg
Giang 2021	results	sample size	n = 4	—
Giang 2021	results	confidence interval	95%CI: 0.08, 3.11	95%CI
Nasiri 2025	abstract	p-value	P < 0.001	—
Nasiri 2025	abstract	p-value	P < 0.001	—
Nasiri 2025	abstract	p-value	P < 0.001	—
Nasiri 2025	abstract	p-value	P < 0.001	—
Nasiri 2025	abstract	p-value	P < 0.001	—
Moel 2025	results	percentage	35.1%	%
Moel 2025	results	unit value	10 mg	mg
Moel 2025	results	sample size	n = 40	—
Moel 2025	results	percentage	20%	%
Moel 2025	results	sample size	n = 8	—
Jarvela-Reijonen 2020	abstract	p-value	P = 0.012	—
Jarvela-Reijonen 2020	abstract	p-value	P = 0.035	—
Nikrad 2023	results	p-value	P ≥ 0.05	—
Nikrad 2023	results	unit value	6.50 kg	kg
Nikrad 2023	results	mean ± SD	89.21 ± 6.50	—
Nikrad 2023	results	mean ± SD	82.23 ± 6.16	—
Nikrad 2023	results	unit value	6.16 kg	kg
Wang 2020	results	percentage	26.2%	%
Wang 2020	results	unit value	400 mg	mg
Wang 2020	results	mean ± SD	26.38 ± 2.06	—
Wang 2020	results	unit value	2.06 days	days
Wang 2020	results	percentage	20.5%	%
Li 2025	results	percentage	90%	%
Li 2025	results	unit value	34 weeks	weeks
Li 2025	results	mean ± SD	35.5 ± 0.8	—
Li 2025	results	unit value	40 weeks	weeks
Li 2025	results	unit value	37 weeks	weeks
Alrasheed 2026	results	p-value	P = 0.001	—
Alrasheed 2026	results	percentage	90.5%	%
Alrasheed 2026	results	confidence interval	95% CI: -1.26 to -0.32	95%CI
Alrasheed 2026	results	percentage	15.3%	%
Yen 2020	results	percentage	10.4%	%
Yen 2020	results	sample size	n=26	—
Yen 2020	results	sample size	n=16	—
Yen 2020	results	percentage	46.5%	%
Yen 2020	results	sample size	n=10	—
Dessie 2025	results	confidence interval	95% CI: 1.02-1.49	95%CI
Knufinke 2023	results	unit value	2 months	months
Knufinke 2023	results	unit value	23 months	months
Corley 2023	results	p-value	P = 0.01	—
Corley 2023	results	percentage	95%	%
Corley 2023	results	unit value	1.84 years	years
Corley 2023	results	p-value	P = 0.01	—
Corley 2023	results	percentage	95%	%
Spears 2026	results	percentage	95%	%
Spears 2026	results	percentage	95%	%
Spears 2026	results	percentage	95%	%
Spears 2026	results	percentage	49.3%	%
Parker 2022	abstract	percentage	95%	%
Parker 2022	abstract	sample size	n=10	—
Parker 2022	abstract	hazard ratio	hazard ratio=1.08	—
Parker 2022	abstract	hazard ratio	hazard ratio=1.19	—
Parker 2022	abstract	hazard ratio	hazard ratio=1.22	—
Bahari 2023	abstract	p-value	P = 0.001	—
Bahari 2023	abstract	unit value	0.50 mg	mg
Bahari 2023	abstract	confidence interval	95% CI -0.79 to -0.20	95%CI
Bahari 2023	abstract	confidence interval	95% CI -1.95 to -0.54	95%CI
Bahari 2023	abstract	p-value	P = 0.001	—
Moskalevska 2026	abstract	percentage	20%	%
Zahedi 2021	abstract	p-value	P < 0.05	—

Additional corpus sources informed the synthesis without anchoring a foregrounded quantitative claim and are catalogued for completeness: Studenski 2011, Schulz 2010.

References

• Ramos-Hernandez 2026. Creatine plus β-Hydroxy-β-Methylbutyrate supplementation is associated with preserved glutathione redox-balance and redox–function associations in older adults: a secondary analysis of a randomized crossover trial. Biogerontology, 2026. DOI: 10.1007/s10522-026-10407-2. PMID: 41712056.
• Giang 2021. Efficacy and safety of biophenol‐rich nutraceuticals in adults with inflammatory gastrointestinal diseases or irritable bowel syndrome: A systematic literature review and meta‐analysis. Nutrition & Dietetics, 2021. DOI: 10.1111/1747-0080.12672. PMID: 33960587.
• Nasiri 2025. Modulation of hormonal, metabolic, inflammatory and oxidative stress biomarkers in women with polycystic ovary syndrome following combined (resistant and endurance) training: a randomized controlled trail. BMC Endocrine Disorders, 2025. DOI: 10.1186/s12902-024-01793-0. PMID: 39748376.
• Moel 2025. Influence of rapamycin on safety and healthspan metrics after one year: PEARL trial results. Aging (Albany NY), 2025. DOI: 10.18632/aging.206235. PMID: 40188830.
• Lee 2025. Effect of pedestal fan use on serum stress biomarkers in older adults exposed to simulated daylong indoor overheating. Physiological Reports, 2025. DOI: 10.14814/phy2.70390. PMID: 40542541.
• Terhalle 2025. Nonspecific stress biomarkers for mortality prediction in older emergency department patients presenting with falls: a prospective multicenter observational study. Internal and Emergency Medicine, 2025. DOI: 10.1007/s11739-024-03693-6. PMID: 38960969.
• Fan 2023. Pregnane X receptor agonist nomilin extends lifespan and healthspan in preclinical models through detoxification functions. Nature Communications, 2023. DOI: 10.1038/s41467-023-39118-9. PMID: 37291126.
• Martens 2018. Chronic nicotinamide riboside supplementation is well-tolerated and elevates NAD + in healthy middle-aged and older adults. Nature Communications, 2018. DOI: 10.1038/s41467-018-03421-7. PMID: 29599478.
• Jarvela-Reijonen 2020. The Effects of Acceptance and Commitment Therapy (ACT) Intervention on Inflammation and Stress Biomarkers: a Randomized Controlled Trial. International Journal of Behavioral Medicine, 2020. DOI: 10.1007/s12529-020-09891-8. PMID: 32394219.
• Quinn 2022. Blood oxidative stress biomarkers in women: influence of oral contraception, exercise, and N-acetylcysteine. European Journal of Applied Physiology, 2022. DOI: 10.1007/s00421-022-04964-w. PMID: 35674828.
• Nikrad 2023. The effect of calorie-restriction along with thylakoid membranes of spinach on the gut-brain Axis Pathway and oxidative stress biomarkers in obese women with polycystic ovary syndrome: a Randomized, Double-blind, placebo-controlled clinical trial. Journal of Ovarian Research, 2023. DOI: 10.1186/s13048-023-01288-x. PMID: 37968684.
• Wang 2020. Effects of Orange Extracts on Longevity, Healthspan, and Stress Resistance in Caenorhabditis elegans. Molecules, 2020. DOI: 10.3390/molecules25020351. PMID: 31952185.
• Bahari 2025. Propolis supplementation on inflammatory and oxidative stress biomarkers in adults: a systematic review and meta-analysis of randomized controlled trials. Frontiers in Nutrition, 2025. DOI: 10.3389/fnut.2025.1542184. PMID: 40421039.
• Li 2025. Oxidative stress biomarkers in fetal growth restriction: a systematic review and meta-analysis. Archives of Gynecology and Obstetrics, 2025. DOI: 10.1007/s00404-025-08133-0. PMID: 40739076.
• Das 2023. Diet composition, adherence to calorie restriction, and cardiometabolic disease risk modification. Aging Cell, 2023. DOI: 10.1111/acel.14018. PMID: 37873687.
• Tambunan 2026. The Prognostic Value of Pre-Procedural and Post-Procedural Inflammatory–Oxidative Stress Biomarkers in Acute Coronary Patients Undergoing Percutaneous Coronary Intervention: A Systematic Review and Meta-Analysis. International Journal of Molecular Sciences, 2026. DOI: 10.3390/ijms27083389. PMID: 42074033.
• Bloomer 2009. Effect of eicosapentaenoic and docosahexaenoic acid on resting and exercise-induced inflammatory and oxidative stress biomarkers: a randomized, placebo controlled, cross-over study. Lipids in Health and Disease, 2009. DOI: 10.1186/1476-511X-8-36. PMID: 19691834.
• Alrasheed 2026. Inflammatory biomarker response to GLP-1 receptor agonists versus other glucose-lowering medications in patients with type 2 diabetes: a systematic review and meta-analysis. Frontiers in Endocrinology, 2026. DOI: 10.3389/fendo.2025.1734549. PMID: 41625236.
• Khor 2021. Effects of tocotrienols supplementation on markers of inflammation and oxidative stress: A systematic review and meta-analysis of randomized controlled trials. PLoS ONE, 2021. DOI: 10.1371/journal.pone.0255205. PMID: 34297765.
• Goettel 2023. Effect of BI 1358894 on Cholecystokinin-Tetrapeptide (CCK-4)-Induced Anxiety, Panic Symptoms, and Stress Biomarkers: A Phase I Randomized Trial in Healthy Males. CNS Drugs, 2023. DOI: 10.1007/s40263-023-01042-3. PMID: 38019356.
• Ceylan 2025. Associations of oxidative stress biomarkers with SYNTAX and ACEF risk scores in acute myocardial infarction: A cross-sectional study. Medicine, 2025. DOI: 10.1097/MD.0000000000044618. PMID: 40988253.
• Saadh 2025. The effect of omega-3 supplementation on metabolic, inflammatory and oxidative stress biomarkers in pregnant women: a systematic review and meta-analysis. Frontiers in Nutrition, 2025. DOI: 10.3389/fnut.2025.1639906. PMID: 41058996.
• Msigwa 2026. Comparative evaluation of oxidative stress biomarkers F2-isoprostanes and 8-OHdG in Parkinson’s disease and Type 2 Diabetes Mellitus: a systematic review and meta-analysis of human studies. Annals of Medicine, 2026. DOI: 10.1080/07853890.2026.2654251. PMID: 42003321.
• Yen 2020. The mitochondrial derived peptide humanin is a regulator of lifespan and healthspan. Aging (Albany NY), 2020. DOI: 10.18632/aging.103534. PMID: 32575074.
• Hoseini 2022. Concurrent alteration in inflammatory biomarker gene expression and oxidative stress: how aerobic training and vitamin D improve T2DM. BMC Complementary Medicine and Therapies, 2022. DOI: 10.1186/s12906-022-03645-7. PMID: 35733163.
• Dessie 2025. Global determinants of breast cancer mortality: a comprehensive meta-analysis of clinical, demographic, and lifestyle risk factors. BMC Public Health, 2025. DOI: 10.1186/s12889-025-24036-w. PMID: 40760691.
• Yin 2025. Prospective observational study of oxidative stress in the pathology of benign prostatic hyperplasia with bladder diverticulum. PLOS One, 2025. DOI: 10.1371/journal.pone.0323677. PMID: 40373101.
• Knufinke 2023. Sex differences in pharmacological interventions and their effects on lifespan and healthspan outcomes: a systematic review. Frontiers in Aging, 2023. DOI: 10.3389/fragi.2023.1172789. PMID: 37305228.
• Yakout 2025. Oxidative stress biomarkers of neonates in Saudi population: An observational study. Medicine, 2025. DOI: 10.1097/MD.0000000000041926. PMID: 40128061.
• Corley 2023. Cell‐type specific impact of metformin on monocyte epigenetic age reversal in virally suppressed older people living with HIV. Aging Cell, 2023. DOI: 10.1111/acel.13926. PMID: 37675817.
• Taylor 2023. Immune and oxidative stress biomarkers in pediatric psychosis and psychosis-risk: Meta-analyses and systematic review. Brain, behavior, and immunity, 2023. DOI: 10.1016/j.bbi.2023.12.019. PMID: 38141839.
• Demir 2025. Impact of early versus delayed cord clamping on oxidative stress in neonates with intrauterine growth restriction: a prospective randomized controlled trial. BMC Pregnancy and Childbirth, 2025. DOI: 10.1186/s12884-025-08335-z. PMID: 41350641.
• Obeng-Gyasi 2022. Association of Allostatic Load With Overall Mortality Among Patients With Metastatic Non–Small Cell Lung Cancer. JAMA Network Open, 2022. DOI: 10.1001/jamanetworkopen.2022.21626. PMID: 35797043.
• Nguyen 2021. Beneficial Effects of Vitamin E Supplementation on Endothelial Dysfunction, Inflammation, and Oxidative Stress Biomarkers in Patients Receiving Hemodialysis: A Systematic Review and Meta-Analysis of Randomized Controlled Trials. International Journal of Molecular Sciences, 2021. DOI: 10.3390/ijms222111923. PMID: 34769353.
• Zhang 2021. Association between allostatic load and mortality among Chinese older adults: the Chinese Longitudinal Health and Longevity Study. BMJ Open, 2021. DOI: 10.1136/bmjopen-2020-045369. PMID: 34344673.
• Moabedi 2025. The effect of co-administration of vitamin E and C supplements on plasma oxidative stress biomarkers and antioxidant capacity: a GRADE-assessed systematic review and meta-analysis of randomized controlled trials with meta-regression. Frontiers in Immunology, 2025. DOI: 10.3389/fimmu.2025.1547888. PMID: 40740780.
• Cai 2011. Icariin and its Derivative Icariside II Extend Healthspan via Insulin/IGF-1 Pathway in C. elegans. PLoS ONE, 2011. DOI: 10.1371/journal.pone.0028835. PMID: 22216122.
• Pleguezuelos 2023. Effect of different types of supervised exercise programs on cardiorespiratory and muscular fitness, pain, fatigue, mental health and inflammatory and oxidative stress biomarkers in older patients with post-COVID-19 sequelae “EJerSA-COVID-19”: a randomized controlled trial. BMC Geriatrics, 2023. DOI: 10.1186/s12877-023-04544-3. PMID: 38102536.
• Spears 2026. Cumulative Lifespan Stress, Inflammation, and Racial Disparities in Mortality Between Black and White Adults. JAMA Network Open, 2026. DOI: 10.1001/jamanetworkopen.2025.54701. PMID: 41587030.
• Wallace 2023. Do non-traumatic stressful life events and ageing negatively impact working memory performance and do they interact to further impair working memory performance?. PLOS ONE, 2023. DOI: 10.1371/journal.pone.0290635. PMID: 38019767.
• Belsky 2017. Impact of early personal‐history characteristics on the Pace of Aging: implications for clinical trials of therapies to slow aging and extend healthspan. Aging Cell, 2017. DOI: 10.1111/acel.12591. PMID: 28401731.
• Beese 2022. Allostatic Load Measurement: A Systematic Review of Reviews, Database Inventory, and Considerations for Neighborhood Research. International Journal of Environmental Research and Public Health, 2022. DOI: 10.3390/ijerph192417006. PMID: 36554888.
• Locker 2025. Understanding the dandruff flare‐up: A cascade of measurable and perceptible changes to scalp health. International Journal of Cosmetic Science, 2025. DOI: 10.1111/ics.13067. PMID: 40162583.
• Parker 2022. Allostatic Load and Mortality: A Systematic Review and Meta-Analysis. Am J Prev Med, 2022. DOI: 10.1016/j.amepre.2022.02.003. PMID: 35393143.
• Popa 2025. Salivary Oxidative Stress Biomarkers in Peri-Implant Disease: A Systematic Review and Meta-Analysis. International Journal of Molecular Sciences, 2025. DOI: 10.3390/ijms262311269. PMID: 41373430.
• Bahari 2023. The effects of pomegranate consumption on inflammatory and oxidative stress biomarkers in adults: a systematic review and meta-analysis. Inflammopharmacology, 2023. DOI: 10.1007/s10787-023-01294-x. PMID: 37507609.
• Madaria 2025. Allostatic load index across the psychosis spectrum: a systematic review and meta-analysis. Frontiers in Psychiatry, 2025. DOI: 10.3389/fpsyt.2025.1590547. PMID: 40666432.
• Ghalichi 2022. Vanadium and biomarkers of inflammation and oxidative stress in diabetes: A systematic review of animal studies. Health Promotion Perspectives, 2022. DOI: 10.34172/hpp.2022.16. PMID: 36276410.
• Moskalevska 2026. Targeting the senescence-associated immune checkpoint GD3 ganglioside extends healthspan and blunt age-related diseases with sex-specific benefits. bioRxiv preprint, 2026. DOI: 10.64898/2026.01.06.697856.
• Zahedi 2021. Effects of curcuminoids on inflammatory and oxidative stress biomarkers and clinical outcomes in critically ill patients: A randomized double-blind placebo-controlled trial. Phytother Res, 2021. DOI: 10.1002/ptr.7179. PMID: 34080237.

Background References

Canonical clinical thresholds cited in prose. Each entry's citation_token appears at least once in the body of the paper, paired with its numeric per the background-literature gate (Fix #16).

• Studenski 2011. Studenski S, Perera S, Patel K, et al. Gait speed and survival in older adults. JAMA. 2011;305(1):50-58. DOI: 10.1001/jama.2010.1923. PMID: 21205966.
• WHO 2000. World Health Organization. Obesity: Preventing and Managing the Global Epidemic. WHO Technical Report Series 894. 2000. PMID: 11234459.
• Schulz 2010. Schulz KF, Altman DG, Moher D. CONSORT 2010 Statement: updated guidelines for reporting parallel group randomised trials. BMJ. 2010;340:c332. DOI: 10.1136/bmj.c332.
• Ioannidis 2005. Ioannidis JPA. Why most published research findings are false. PLoS Med. 2005;2(8):e124. DOI: 10.1371/journal.pmed.0020124. PMID: 16060722.