Nutritional depletion was strongly associated with stroke severity, but its role in modifying cold-related vulnerability remains uncertain, and the proposed 'ceiling-like' effect of malnutrition on cold-attributable stroke burden should be considered hypothesis-generating and requires prospective validation.
Key Findings
Results
Sodium-attributable stroke burden in China peaked at ages 70-79 years, while cold-attributable stroke burden increased in the oldest age groups.
Data were drawn from Global Burden of Disease (GBD) 2023 estimates for China spanning 1990-2021.
The divergent age patterns of sodium-attributable versus cold-attributable stroke burden informed the 'Collapsed Furnace' hypothesis about differential vulnerability in oldest adults.
Cold-attributable burden increasing in the oldest groups suggests age-related physiological vulnerability to cold exposure.
Results
In the GBD analysis, the malnutrition × low-temperature interaction on stroke burden was negative and statistically significant.
The interaction term was β = -12.20 (95% CI: -18.52 to -5.88; p < 0.001).
The negative interaction suggests that malnutrition may attenuate or 'ceiling' the additional stroke burden attributable to cold exposure.
This finding was derived from the GBD 2023 ecological-level analysis, not the clinical cohort.
Results
Nutritional risk, assessed via admission serum albumin, was strongly associated with higher admission NIHSS scores in the clinical stroke cohort.
The primary clinical analysis included patients aged ≥70 years; a sensitivity analysis retained the full cohort aged ≥67 years.
Nutritional risk was assessed using admission serum albumin as a nutritional-inflammatory marker.
The body mass index (BMI)-adjusted modified Geriatric Nutritional Risk Index (mGNRI) was used as a sensitivity measure, with directionally similar results.
CRP-based sensitivity analyses were also directionally consistent with the primary albumin-based findings.
Results
Cold exposure was associated with only a small additional NIHSS increase among nutritionally normal patients and no measurable increase among those with low albumin, but this interaction was not statistically significant.
Among patients with normal albumin, cold exposure was associated with a small additional increase in admission NIHSS score.
Among patients with low albumin, cold exposure was associated with no measurable increase in NIHSS score.
The interaction between cold exposure and nutritional status was not statistically significant in the clinical cohort.
Only 7 cold-exposed patients with low albumin aged ≥70 years were available, severely limiting statistical power.
Results
The China Health and Retirement Longitudinal Study (CHARLS) 2015 data suggested different seasonal blood-pressure patterns by physiological status, but interpretation was severely limited by sparse cold-season interviews.
CHARLS 2015 data were used to provide exploratory physiological context for the hypothesis.
Only 4 CHARLS participants aged ≥70 years with cold-season interviews were available.
The authors stated that 'interpretation was limited by sparse cold-season interviews.'
The seasonal blood-pressure pattern analysis was considered exploratory rather than confirmatory.
Discussion
The study identified multiple key data limitations that preclude definitive conclusions about the 'Collapsed Furnace' hypothesis.
Mortality data, pre-stroke albumin levels, individual cold exposure measures, and core temperature data were all unavailable.
The clinical cohort contained only 7 cold-exposed patients with low albumin in the primary age group (≥70 years).
The proposed ceiling-like effect of malnutrition on cold vulnerability is described as 'hypothesis-generating' requiring prospective validation.
Multivariable models adjusted for prespecified demographic, vascular, lifestyle, stroke-type, BMI, and CRP covariates.
What This Means
This research explores what the authors call the 'Collapsed Furnace' hypothesis — the idea that older adults who are malnourished or physically frail may respond differently to cold weather when it comes to stroke risk and severity. The study combined data from a large global disease burden database (covering China from 1990-2021), a national Chinese health survey, and a clinical group of stroke patients. The researchers found that poor nutritional status, measured by protein levels in the blood, was strongly linked to more severe strokes at the time of hospital admission. Interestingly, cold exposure appeared to add to stroke severity mainly in patients with normal nutritional status, but not in those who were already malnourished — suggesting that malnourished patients may already be at such high baseline severity that cold adds little additional measurable harm.
However, the researchers emphasize that these findings should be interpreted with great caution. The statistical interaction between cold exposure and malnutrition was not significant in the clinical data, and the number of patients in the critical comparison group was extremely small — only 7 patients who were both cold-exposed and malnourished in the primary analysis. Important data such as patients' temperatures, pre-existing nutritional status before the stroke, and death records were not available. The large global disease burden database did show a significant negative interaction between malnutrition and cold temperature on stroke burden, but this was an ecological analysis that cannot prove causation.
This research suggests that malnutrition is an important factor in stroke severity among older adults, and that the relationship between cold weather and stroke may differ depending on a patient's nutritional and physical condition. However, the specific 'Collapsed Furnace' idea — that malnutrition creates a kind of ceiling effect on cold-related stroke vulnerability — remains unproven and the authors explicitly describe it as hypothesis-generating. Larger, carefully designed prospective studies that measure cold exposure, nutritional status, and outcomes together would be needed to confirm or refute this idea.
Zhou Y, Huang Y, Li Q, Lv D, Cao Y, Tian H, et al.. (2026). Nutritional risk, cold exposure, and stroke severity in older adults: an exploratory analysis of the "collapsed furnace" hypothesis.. Frontiers in public health. https://doi.org/10.3389/fpubh.2026.1838018