{"publication_id":"45ba1bda-6940-4338-babd-4c360db9ad4d","screening":{"identified":46,"screened":46,"excluded":0,"included":46,"included_or_retained":46,"flow":["identified","screened","excluded_with_reasons","included"],"wording":"46 candidate receipts retained after source retrieval, deduplication, and topic filtering. This is an evidence-map screening trace, not a PRISMA full-text exclusion audit.","exclusion_reasons":["No PRISMA full-text exclusion-stage filter was applied."]},"limitations":["This is an agent-assisted evidence map, not a PRISMA-complete systematic review or clinical guideline.","It is not PROSPERO-registered and should not be read as medical advice.","Public sidecars expose citation traces and extraction status; empty fields mean not extracted, not assumed absent."],"contradictions":["Evidence-honesty note: 36/46 retained sources are indirect, review-level, adjacent, or mechanistic and are used only to bound interpretation. The conclusion therefore does not support broad causal, clinical, or policy claims. Probiotic supplementation has attracted intense interest as a low-cost adjunct to influence cardiometabolic, immune, musculoskeletal, and clinical-infection outcomes, yet the literature spans dozens of meta-analyses with non-overlapping conclusions and a long tail of single-site RCTs whose results have not always replicated (Liang 2022; Aparicio-Pascual 2025; Prokopidis 2022). We applied an AI-assisted structured evidence-synthesis framework that ingests curated reference sources, enforces direct-vs-indirect outcome stratification, and produces an audit trail linking every quantitative claim to its primary source, while keeping the analytic reasoning in clinical domain language. Pediatric and cognitively-oriented RCTs (Chen 2026, Kim 2020) and a positive bariatric/inflammation trial (Potrykus 2025, P < 0.001 for several biomarkers) round out a pattern in which short-chain-fatty-acid-linked, inflammation-modulating mechanisms are biologically coherent but clinically bounded. **Evidence-abstraction note.","We applied an AI-assisted structured evidence-synthesis framework that ingests curated reference sources, enforces direct-vs-indirect outcome stratification, and produces an audit trail linking every quantitative claim to its primary source, while keeping the analytic reasoning in clinical domain language.","The corpus contains 10 direct clinical sources, 36 adjacent, review, or context sources, and no sources classified primarily as mechanistic or model-system evidence. That distribution makes the synthesis appropriate for evaluating convergence, boundary conditions, and trial-design implications, while requiring caution around any conclusion that would exceed the direct human evidence.","Null findings have a specific role in this evidence model. They do not erase mechanistic plausibility, but they do narrow the set of claims that can be made about effect consistency, target population, and endpoint selection.","The evidence base also distinguishes breadth from certainty. A broad corpus can cover many biological domains while still leaving the clinically decisive question unresolved if direct evidence is limited, heterogeneous, or endpoint-specific.","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.","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.","Contextual Adjacent Evidence: n=14; claims=547; mixed signal in 8/14 sources | directness: 3 direct; 5 indirect; 6 review; main limitation: directionally heterogeneous.","Six curated reviews and meta-analyses comprise the cardiometabolic evidence base, spanning populations with type 2 diabetes, prediabetes, overweight/obesity, and polycystic ovary syndrome, as well as generally healthy adults and exercise-recovery cohorts. The dominant designs are systematic reviews with pooled analyses, with one observational synthesis, Shirkoohi 2025, contributing indirect data on body composition. Durations and dosing regimens were heterogeneous across the underlying trials, but endpoints were uniformly cardiometabolic: glycemic indices (HbA1c, fasting glucose, HOMA-IR, HOMA-β), lipid fractions, body weight, and islet β-cell function. The integrated thesis positions probiotic supplementation as context-dependent rather than uniformly favorable or unfavorable in this outcome class.","Quantitative signals within the cardiometabolic class are mixed and outcome-specific. Tabrizi 2022, in PCOS, reported several glycemic, lipid, and inflammatory contrasts reaching P = 0.01, P = 0.02, P < 0.001, P < 0.01, and P = 0.04. Across the class, the per-study p-value tuples are catalogued in the evidence synthesis.","Within-corpus tensions in the cardiometabolic class are pronounced. A plausible reconciling frame is that effect direction is conditional on baseline metabolic status, trial duration, and strain-specific formulations, but the corpus as currently constituted does not resolve which boundary condition drives the divergence. The same pattern recurs for body composition, where Shirkoohi 2025 reports predominantly null body-weight contrasts while Liang 2022 reports significant glycemic improvements, suggesting that probiotics decouple across cardiometabolic sub-domains within the same class."]}