Eleven homes with no mold problem.
Cladosporium, Penicillium and Alternaria showed up in most of the indoor air samples.
The study
Rachel Adams, Marzia Miletto, John Taylor and Thomas Bruns published “Dispersal in microbes: fungi in indoor air are dominated by outdoor air and show dispersal limitation at short distances” in The ISME Journal in 2013. The authors thank the Sloan Foundation for funding.
The site was a university family housing complex, chosen because it strips out most of the variables that usually differ between houses. A similar floor plan throughout. Fiber-cement siding, vinyl windows, painted sheetrock, vinyl tile in the kitchen and bathrooms and carpet everywhere else. Every building had forced-air heat and no air conditioning. The units were built between 1999 and 2008, families had lived in them about 2 years on average, and the whole complex occupies less than a square kilometer.
Critically, none of the units had a reported mold problem. The authors set out to catalog the fungi in the air of what they call “healthy” buildings.
The sampling method was deliberately plain. In each unit they suspended a sterile, empty, 9 cm petri dish 0.3 meters from the ceiling in four rooms: kitchen, living room, bathroom and bedroom. One more went outside on the patio or deck. Each dish collected whatever settled out of the air for 4 weeks. They ran the survey twice: 11 units in July 2011 and 8 of those same units in January 2012. Residents answered a questionnaire on how the home was used, and temperature and relative humidity were logged for the whole sampling period.
Then, instead of culturing what landed, they sequenced its DNA. That matters, because culture-based sampling only counts fungi willing to grow on a plate. As the authors note, the plant-associated fungi that dominate outdoor air are “essentially unknown from culture-based sampling of indoor air.”
What they found
986 distinct fungal types across the whole survey. The paper calls them operational taxonomic units, or OTUs: clusters of sequences grouped at 97% similarity, which is as close to “a species” as this method gets. After quality filtering and rarefaction, 84 samples went into the community analysis: 67 indoor and 17 outdoor.
The first result is a matter of quantity.
By the authors’ own estimate, outdoor samples have approximately 20-100-fold greater fungal biomass than indoor samples.
That difference was statistically significant within each season (summer P=0.03, winter P<0.01). Counted as types rather than mass, the mean observed richness was 179.7 per outdoor sample against 78.5 per indoor sample. Read that one carefully: the authors’ own estimated richness for indoor and outdoor air is statistically indistinguishable within a season, and their richness curves never leveled off. The biomass difference is the solid finding here, and nothing in this study says outdoor air is more diverse than indoor air.
The second result is a matter of composition. Sixteen of the 25 most abundant fungi were present in every single outdoor sample, and appeared in over half the indoor ones. The group with the most distinct types in the entire data set was Agaricomycetes, with 238 of them. Those are mushrooms and polypores, which decay plants outdoors. They were being counted in kitchens and bedrooms. Across the survey the authors also picked up plant pathogens, lichenized fungi, mushrooms and puffballs, none of which grow in a house.
What explained nothing
This is the part worth sitting with, because it is a set of null results and null results do not usually get written up in a mold brochure.
Room type never mattered. Bathrooms and kitchens, which have water, looked like bedrooms, which do not. Fungal biomass did not differ significantly across the four indoor rooms in either season (P=0.33 summer, P=0.71 winter).
Cleaning frequency looked significant, then evaporated. Considered on its own it explained a real share of the variation. Put into a model alongside the unit’s location, it dropped out entirely, along with floor level, number of bedrooms, number of residents, temperature and unit age. Relative humidity and number of bathrooms were never significant at all.
What survived was location. Geographic distance explained at least 33% of the variation in summer and 24% in winter in the multi-factor model, and it did so across a site less than 500 meters wide. Two apartments a few hundred meters apart, in the same complex and built to the same design, had measurably different fungal communities, and the difference grew with the distance between them.
The authors state the conclusion flatly:
“no taxa were indicative of the indoor environment or a particular unit or room.”
Simulations backed it up. When they compared the real indoor-versus-outdoor community distances against 999 randomized versions, the observed distances fell inside the randomized range. The fungi indoors behaved like a random subsample of the fungi outdoors, not like a filtered or resident population.
The limits, stated plainly
Nobody was tested. No symptoms, no lung function, no allergy panel, no health outcome of any kind. This is a study of air.
Eleven units, dropping to eight in winter, in one complex, in one mild climate. The authors say so: “the mild climate of our sampling locality increases the exposure of building interiors to the outdoors in both winter and summer in this study system.” They describe it elsewhere as a Mediterranean climate, and none of the buildings had air conditioning. A sealed, air-conditioned house in a different climate could behave differently.
Every home was healthy by design. The study never sampled a water-damaged building. It describes the baseline, and says nothing about what a moisture problem looks like against it.
DNA is not viability. The authors flag their own weakness: PCR-based studies have “an inability to separate living from dead tissue and cell fragments, or the nature of the starting material in general (spore versus fragment).”
They did not catch everything. Their coverage estimate ranged from 20.7% to 58.1% of the fungal types present, median 36.8%, and the richness curves never leveled off. A single sample missed most of what was in the room.
The sampler itself is unvalidated. A settled-dust dish at a fixed height may bias toward particular particle sizes, though the authors did detect both 3 micrometer and 9 micrometer spored taxa, and they write that “further work is required to determine the individual bias introduced with this particular sampling device.”
What gets lost in the mold-testing marketing
Cladosporium, Penicillium, Alternaria and Aspergillus are the four names that make a mold report look frightening. Here is how often a single sequence type from each turned up in the indoor air of these ordinary, problem-free homes, out of 67 indoor samples:
- Cladosporium (one sequence type): 91.0%, 61 of 67
- Penicillium (one sequence type): 89.6%, 60 of 67
- Alternaria (one sequence type): 71.6%, 48 of 67
- Aspergillus (one sequence type): 47.8%, 32 of 67
Finding those names is the baseline. A report that lists them without an outdoor comparison has described outdoor air.
The EPA is direct about the regulatory vacuum underneath all of this. From its mold testing page:
“Since no EPA or other federal limits have been set for mold or mold spores, sampling cannot be used to check a building’s compliance with federal mold standards.”
And on when sampling is worth doing at all:
“In most cases, if visible mold growth is present, sampling is unnecessary.”
The EPA does allow that surface sampling “may be useful to determine if an area has been adequately cleaned or remediated,” and says sampling “should be conducted by professionals who have specific experience in designing mold sampling protocols, sampling methods and interpreting results.”
None of this makes air sampling useless. It makes the comparison, not the raw count, the part that carries the information.
What the data supports
Insist on a same-day outdoor sample from the same property. An indoor number read on its own has no baseline. This study is a measurement of how large that baseline is and how much it moves.
Do not compare across seasons. Summer and winter communities separated cleanly in this data. A January result and a July result from the same house are not the same measurement.
Do not treat one sample as a census. The median sample detected roughly a third of the fungal types present.
Look for water first. Moisture is what creates an indoor source. This study shows what the air looks like when there is not one, which is exactly the picture you need before you can say a house has a problem.
Treat a single hot room in a dry house with caution. In these homes, biomass did not differ by room and no fungus was tied to a room. In a house with no moisture damage, a high reading in one room is more likely sampling variation than a source.
Bottom line
In houses without a moisture problem, the fungi in the air tracked the outdoors and the address. Room type and cleaning frequency did not survive the analysis.
Next time you see a mold air report, check whether an outdoor sample was taken the same day. If it was not, the indoor number has nothing to compare itself to.
illumenair tests indoor air in Portland and Vancouver homes.
971-363-5626 · info@illumenair.com · www.illumenair.com
References
Adams RI, Miletto M, Taylor JW, Bruns TD. Dispersal in microbes: fungi in indoor air are dominated by outdoor air and show dispersal limitation at short distances. ISME J 2013;7(7):1262-1273. doi:10.1038/ismej.2013.28. (Erratum: ISME J 2013;7(7):1460, correcting a sequence accession number only.)
US Environmental Protection Agency. Mold Testing or Sampling. https://www.epa.gov/mold/mold-testing-or-sampling