# Consensus checked against original sources: Sauna and Cardiovascular Prevention

Cutoff: 13 October 2026. Consensus supports discovery and checks of specific source aspects; it is not a primary study, a full-paper audit or a causal probability meter.

## Finnish cohort hard endpoints and causal wording
Consensus accurately reports the2315-men cohort,190 sudden deaths,407 fatal cardiovascular events and the adjusted frequent-use SCD association. These are prospective observed events, not randomized sauna prevention. Reference groups are predominantly different sauna frequencies, rather than a representative no-sauna population; repeated papers cannot multiply independent exposure allocation. [Original PMID25705824](https://pubmed.ncbi.nlm.nih.gov/25705824/). Access: abstract_metadata.

## Fitness adjustment and study independence
Consensus correctly highlights measured fitness and cohort overlap. The original joint-analysis associations strengthen a confounder-specific check but reuse the KIHD male cohort family. Baseline gas-analyzed fitness and statistical adjustment do not make sauna exposure randomized or eliminate every health/social selection process. [Original PMID28972808](https://pubmed.ncbi.nlm.nih.gov/28972808/). Access: abstract_metadata.

## WAON-CHF endpoint hierarchy
Consensus summarizes review-mediated BNP decreases and function gains without visibly retaining the requested original trial primary-null distinction. The original149-person randomized report designates pre/postBNP ratio as primary; its between-group BNP change did not reach significance while selectedNYHA/6MWD/cardiothoracic secondary outcomes improved. Primary and secondary results remain separately reported, and no broad event prevention follows. [Original PMID27001189](https://pubmed.ncbi.nlm.nih.gov/27001189/). Access: abstract_metadata.

## Added sauna effect and fitness measurement
Consensus correctly identifies an eight-week added-sauna systolic-pressure and fitness estimate over exercise alone. Original Methods show48 enrolled/47 analyzed, most participants women, and estimatedVO2max from a submaximal heart-rate test. Thus the real positive is retained, with estimated rather than directly gas-measured fitness and no hard-event endpoint. [Original PMID35785965](https://pubmed.ncbi.nlm.nih.gov/35785965/). Access: full_text_partial.

## CAD acclimation versus vascular outcomes
Consensus accurately notes the41-person CAD randomized null for FMD, stiffness andBP despite acclimation. The result constrains vascular mechanism generalization; it does not directly demonstrate zero mortality benefit. A separate review-mediated93percent ischemia claim was not verified against an original and is excluded from our numeric evidence cards. [Original PMID37650138](https://pubmed.ncbi.nlm.nih.gov/37650138/). Access: abstract_metadata.

## Sickness absence versus illness days, control and correction
Consensus describes reduced absence despite zero change in actual illness days. Original Methods establish unblinded self-reported outcomes and unverified compliance; Results report no significant overall illness-day group effect, rather than exact zero or objective infection surveillance. Sickness absenceIRR0.71 is retained. The examined original correction adds omitted underlyingTable2 data, not a revised outcome or retraction. [Original PMID27631616](https://pubmed.ncbi.nlm.nih.gov/27631616/). Access: full_text_partial.

## Acute recovery versus chronic training adaptation
Consensus notes attenuated resistance adaptation after cold recovery. The original reports a21-man twelve-week training experiment and a separate nine-man acute biopsy experiment. Smaller strength/mass gains with repeated cold are a genuine training-context boundary; a separate acute submaximal recovery positive25121612 is not discarded or converted into durable health. [Original PMID26174323](https://pubmed.ncbi.nlm.nih.gov/26174323/). Access: abstract_metadata.

## Auricular consumer-style devices and genuine controlled evidence
Consensus categorically contrasts validated clinical cervical/implanted paths with consumer auricular devices lacking comparable sham-controlled evidence. OriginalTREAT-AF Methods name a home Parasym device, stimulated tragus versus stimulated earlobe, blinded outcome assessors and ECG-patch AF-burden primary measurement. Its53-person positive contradicts a blanket no-evidence assertion for every auricular/home device, while not establishing all-brand or implanted-equivalence claims. [Original PMID32192678](https://pubmed.ncbi.nlm.nih.gov/32192678/). Access: full_text_partial.

## Disease-specific null and secondary report overlap
Consensus coverage cells report no vagal literature in several domains; such retrieval gaps are not literature absence. The original multicenter cervical-diabetes trial had no active-sham gastrointestinal symptom superiority, and cytokine-null39251831 is a secondaryNCT04143269 analysis. Defined negative and positive trials should coexist rather than be generalized by device label. [Original PMID38546822](https://pubmed.ncbi.nlm.nih.gov/38546822/). Access: abstract_metadata.

## Five-year Waon hard events and absolute versus relative effect
Consensus cites a review-mediated 38% reduction in cardiac events over five years. The verified original reports HF/cardiac-death composite rates 68.7% in 65 matched traditional-care controls versus 31.3% in 64 Waon recipients: a 37.4 percentage-point difference, not an unambiguously stated 38% relative reduction. The study reports matching, without stated random allocation, in NYHA III/IV heart failure; deaths were twelve versus eight. We retain this clinically important event association without treating it as a mortality-prevention RCT or ordinary-sauna primary prevention. [Original PMID19304125](https://pubmed.ncbi.nlm.nih.gov/19304125/). Access: abstract_metadata.

## Methods
Consensus was used to identify literature and cross-check specific claims about design, endpoints, scope and research gaps. Those aspects were checked against original primary abstracts/status records and selected original Methods/Results passages. Consensus output was not treated as original full text. Provider search omissions are not evidence that a study does not exist. Source access and examination receipts are retained. Scite use or monthly unavailability is recorded per case; provider citation counts do not change causal grades.