At a single Texas lung clinic, 54 of 231 patients with hypersensitivity pneumonitis had mold in their own home named as the trigger. The striking part is not the count. It is what the water looked like.
The mold that put these 54 people in a pulmonary clinic was not dramatic. In 49 of the 54, it was a leak that kept coming back.
Traci Adams and Craig Glazer at the University of Texas Southwestern Medical Center in Dallas, with Carrie Redlich and Mridu Gulati at the Yale School of Medicine, went back through their interstitial lung disease database and pulled every patient given a diagnosis of hypersensitivity pneumonitis between 2011 and 2019. The paper was published on 8 May 2025 in PLOS ONE.
What hypersensitivity pneumonitis is
Hypersensitivity pneumonitis, or HP, is not asthma and it is not allergy in the everyday sense. It is an immune reaction in the lung tissue itself to something inhaled repeatedly over time, and in its fibrotic form it leaves scarring. It is the family of disease that gave us the names “farmer’s lung” and “hot tub lung.” In Japan, a summer-type HP attributed to indoor mold contamination, principally by Trichosporon species, is widely reported. Outside that region, the authors write, the evidence has been thinner, and clinicians have tended to overlook a patient’s own house as the source.
What they did
231 patients in the database carried a moderate, high or definite confidence diagnosis of HP under current American Thoracic Society criteria, decided in multidisciplinary discussion using HRCT pattern, bronchoalveolar lavage lymphocyte percentage, transbronchial biopsy and surgical lung biopsy. An occupational lung disease specialist then reviewed every exposure history to confirm that the antigen was visible, present in enough quantity to induce an immunologic response, persistent, and there before the disease developed.
Be clear about what that review was. The confirmation step was a chart review, not a site visit: the presence of residential mold exposure was “confirmed by medical chart review by a pulmonologist trained in exposure assessment.”
What they found
Home mold was the culprit antigen in 54 of the 231 patients, or 23.4% of the entire HP cohort. 48 of those 54 had the fibrotic form. Mean age at diagnosis was 62.9 years. Mean FVC was 70.9% of predicted and mean DLCO 50.9%, and 22 patients (40.7%) were on supplemental oxygen. Median transplant-free survival was 97 months. Seven patients died and five were transplanted.
Where the mold was growing, among the three locations the authors call out:
- 22.2% — central air conditioning unit
- 18.5% — bedroom
- 16.7% — bathroom
18 patients (33.3%) had mold in more than one area of the home, and 8 patients had mold in more than two. The majority (81.5%) lived in a house; 8 patients (14.8%) lived in an apartment.
How the water got in:
- 33.3% — pipe leak or burst
- 14.1% — roof leak
- 11.1% — leak under a pier and beam foundation
Nine patients (16.7%) had more than one source of water intrusion.
One caution on that middle figure. The roof-leak share is printed in the paper as 14.1%, but with 54 patients no whole number of people produces 14.1% (7 would be 13.0%, 8 would be 14.8%). Its neighbors are exact: 33.3% is 18 of 54 and 11.1% is 6 of 54. We reproduce the published figure rather than correct it, and flag it rather than lean on it.
The part that surprises people
Five patients (9.3%) reported mold developing after a single episode of a pipe burst that was never remediated. For the other 49 (90.7%), the leak was due to a recurrent or chronic issue: a chronically leaking pipe, a basement that flooded again and again. The lung disease people associate with catastrophe arrived, in this cohort, by way of water that was never fully stopped. A chronically leaking pipe is often invisible.
What happened when the exposure was removed
41 patients (75.9%) removed the exposure. 20 (37.0%) had a professional mold remediation and 21 (38.9%) moved out of the affected residence. The median time from diagnosis to remediation or moving was 11.6 months.
On pulmonary function tests performed three to four months after removal, 5 of the 41 (12.2%) had more than a 10% improvement in FVC percent predicted, 24 (58.5%) had no significant change, none had more than a 10% decline, and 12 (29.3%) had insufficient data recorded to judge. All five who improved still held that gain a year after removal.
Then the follow-up thins out, and this part matters. Three of those five were lost to follow-up within 18 months. Of the two who were still being followed past two years, both declined again by 10% predicted between three and five years out, and both had fibrotic HP. “Two of five declined” reads like a minority. Among the patients anyone kept watching, it was all of them.
Read plainly: for most of these patients, removing the exposure did not reverse the disease. It stopped the fall.
The limits, stated plainly
This is one center, one region, and a retrospective case series. There is no unexposed comparison group, so it cannot tell you the odds that a damp house causes HP in anyone. (The paper does compare home-mold HP with the 50 patients whose mold exposure was outside the home, and separately with the 97 who had avian exposure, and found no difference in transplant-free survival. The avian comparison is reported without the data.) The authors quote the objection themselves: home mold exposure is common and HP is rare, and there is no accurate antibody test to confirm sensitization.
Only 20 of the 54 patients (37%) had their home professionally inspected by an industrial hygienist. In the rest, visible mold was confirmed by the patient, a family member, or a contractor, and only the clinician’s brief summary of any formal report was in the chart. No mold species and no concentrations were recorded anywhere in the study.
Serum precipitins were not available for the cohort, and the reason the paper gives is specific: they “are not routinely used at our institution for patients with a clear exposure due to poor negative predictive value in that setting.” Separately, in its discussion of testing generally, the paper notes that commercial IgG and precipitin panels have “at best modest sensitivity and specificity due to the lack of standardization of methodology and quality.” Follow-up was limited for some patients, and follow-up imaging was missing for some of those whose lung function improved.
What survives all of it: 23% of this HP cohort traced to mold in the patient’s own house, with what the authors describe as more specific detail on the source, extent and duration of the water damage than earlier studies provide.
What gets lost in the mold-testing marketing
This paper is unusually blunt about testing, and it does not flatter the industry we work in.
There are no reference ranges for mold. Standardized sampling methods and reference ranges have not been established, which is why a number on a lab report cannot be read as pass or fail the way a lead or radon result can.
A surface sample is not an air measurement. Surface samples may identify specific species but do not quantify airborne levels, and culture media may not allow growth of certain suspected organisms.
The authors’ own conclusion: “visual inspection is often superior to quantitative air sampling, and a history of water intrusion provides strong supporting evidence of the presence of conditions ideal mold overgrowth.”
A spore count is not a diagnosis. The water history is strong supporting evidence, and note what the authors say it supports: conditions favorable to mold growth. Not a diagnosis either.
What the data supports
Treat a leak that keeps returning as the finding, not the puddle. For 49 of these 54 patients, the water was recurrent or chronic rather than a single event.
Count the rooms. A third of these patients had mold in more than one area of the home, and some in more than two. Finding one spot does not close the question.
Do not skip the mechanical equipment. Of the three locations the authors call out, the central air conditioning unit carried the highest share. In the Pacific Northwest the nearest equivalent is the furnace or air handler cabinet and the duct run beside it. That translation is our inference, not the paper’s finding.
Bring the water history to the pulmonologist. If someone in the house has breathlessness or a dry cough that has dragged on with no clear cause, the exposure history matters clinically. Identifying an inciting antigen raises HP diagnostic confidence, which the authors note may reduce the number of diagnostic procedures a patient needs.
Translate the geography. In Portland and Vancouver the water usually arrives slowly: chronic roof and flashing leaks, wet crawl spaces under the floor, and eight months of rain. Different cause, same category of exposure: water that stays.
None of this is medical advice, and none of it says that a particular damp home will cause lung disease. Hypersensitivity pneumonitis is uncommon and home mold is not. What this study establishes is narrower and still worth knowing: at one major lung center, nearly a quarter of HP cases were traced to mold in the patient’s own residence, and the water that caused it was overwhelmingly a slow, recurring leak rather than a flood. illumenair tests indoor air quality and mold. We do not diagnose illness and we do not perform remediation.
Wondering what your house is actually holding?
illumenair tests indoor air quality and mold in Portland and Vancouver homes, room by room. We measure and report, and we do not remediate, so the result is not a sales pitch for the repair.
971-363-5626 · info@illumenair.com · www.illumenair.com
Sources: Adams TN, Redlich CA, Glazer CS, Gulati M. Hypersensitivity pneumonitis associated with home mold exposure: A retrospective cohort analysis. PLOS ONE, 2025;20(5):e0323093, doi:10.1371/journal.pone.0323093.