Ann. occup. Hyg., Vol. 46, No. 1, pp. 113-118, 2002
© 2002 British Occupational Hygiene Society
Published by Oxford University Press
Article |
Effect of Sampling Time on the Culturability of Airborne Fungi and Bacteria Sampled by Filtration
1University of Washington, School of Public Health and Community Medicine, Department of Environmental Health; 2Harvard University, School of Public Health, Department of Environmental Health
Received 6 October 2000; in final form 20 March 2001.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Air sampling of bioaerosols by filtration may be preferable for many epidemiological studies because the methods can be used to collect personal samples for a full work-shift. There is some concern, however, that the viability of fungal spores and bacterial cells might be compromised by sampling for as long as a full shift. This study was designed to determine the effect of sampling up to 6 h on the viability (measured by culture) of airborne fungi and bacteria at composting facilities. Six side-by-side samples were collected in two locations at each of three composting facilities for 1 h at 2 l/m on polycarbonate filters. Two samples in each set were then capped while clean, HEPA-filtered air was drawn across two others for an additional 2 h and across the last two for an additional 5 h. Filters were washed and the samples were analyzed for culturable bacteria and fungi, and for total bacteria and fungi by microscopic counting. Concentrations ranged from 1.7 x 103to 6.2 x 107 c.f.u./m3 of culturable fungi and 1.17 x 104 to 1.0 x 106 c.f.u./m3 of culturable bacteria. In linear models that included duration of sampling, location, and the interaction of location and sample duration, neither sample duration nor the interaction term were significant predictors of the logs of the concentrations of culturable fungi or bacteria or of the ratio of the logs of the culturable concentrations to total concentrations for fungi or bacteria. This suggests that increased sampling time does not affect the viability of the organisms commonly found in the air at composting facilities.
Keywords: bioaerosols; sampling methods; composting; organic dusts; microbial viability; filter sampling
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Several different methods for the evaluation of bioaerosol exposures have been developed and used in a number of occupational and residential environments. However, no one standard method has been agreed upon to date. Historically, the Andersen sampler or glass impingers have been used to collect samples of bacteria and fungi for cultural analysis (Muilenberg, 1989; Burge, 1995; Macher et al., 1995; Lacey et al., 1996). In epidemiological studies of allergic illness, however, interest has also been shown in methods for collecting total bacterial cells and fungal spores irrespective of viability (Heikkila et al., 1988; Karlsson and Malmberg, 1989). In these cases, a combination of methods has often been used in order to obtain both culturable and total concentrations. For example, Hirst-type spore traps are often used in combination with Andersen samplers. Because many types of airborne cells are not culturable or readily identifiable in culture (e.g. mushroom spores, rusts and smuts) and environmental factors can impact cells of different taxa in different ways, the two types of measurements are not always predictive of one another. The two methods can be used to complement each other since the identification of different organisms is possible. Although culturable concentrations are not considered to be predictive of total concentrations (Eduard and Heederick, 1998), it has been assumed that the ratio of culturable to total organisms measured in a particular environment (i.e. with stable species composition) will remain constant over time (Palmgren et al., 1986).
In highly contaminated environments, however, culture plates quickly become overloaded such that sampling times must be reduced to as little as 15 or 30 s. Sampling onto a filter, which is then washed and cultured to obtain a measure of viable organisms, is used increasingly in epidemiological studies (Heikkila et al., 1988; Karlsson and Malmberg, 1989; Palmgren et al., 1986; Eduard and Heederick, 1998). Using this method, personal samples can be collected for entire work shifts rather than collecting area samples over a period of seconds to minutes, although Dillon et al. state that sampling times with filters are typically no longer than 30 min (Dillon et al., 1996). Personal sampling allows for better characterization of a worker’s average exposure during the work shift and was the chosen method for a study we conducted of exposures and respiratory symptoms in compost workers.
Although the collection efficiency of the filter method has been compared with the collection efficiency of other methods (Eduard and Heederick, 1998; Blomquist et al., 1984), it has been assumed that cell damage during deposition or by desiccation during full shift sampling would reduce the viability of the organisms collected on filters. In fact, it has been shown that samples collected for 4 h at 30% relative humidity (RH) and 75% RH resulted in a lower percentage of germination of P. chrysogenum than samples collected at the same relative humidity for 15 min (Muilenberg and Burge, 1994). This study, however, used an experimental test atmosphere of spores that had been stored. The question of the effect of the increased sampling time on the viability of the organisms collected in the field has not yet been addressed in a scientific study. The purpose of this paper is to report the results of an experimental study designed to answer this question.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Participating facilities
Three composting facilities were chosen from among 10 facilities participating in a cross-sectional study of exposures and respiratory symptoms of compost workers. The facilities were chosen in order to ensure exposures high enough to enable measurements above the level of detection on 1 h samples and to provide a variety of environments in which samples could be collected. Facility A was a biosolids composting facility on the site of a municipal wastewater treatment plant in eastern Washington. Area 1 was next to the screening operation and area 2 was between some static piles at this facility. Composting of biosolids and wood chips was conducted entirely indoors using aerated piles. Facility B was a privately owned yard waste recycling center in northern Oregon. Composting at this facility was achieved using static piles and all operations were outdoors. Area 3 was next to the screening operation and area 4 was on top of a pile of fresh, actively composting yard waste. Facility C was a manure composting operation in southern Idaho. The turned windrows of cow manure in this operation were located outdoors on a dairy farm. Area 5 was on the turning equipment and area 6 was on the front-end loader used to build and straighten the windrows.
Sampling strategy
The two locations with the highest previously measured exposures were selected at each facility. Six side-by-side samples were collected within 12'' of each other at each location using constant flow pumps. Sample time was 1 h with a sampling rate of 2 l/min using closed-face, three-piece polystyrene cassettes loaded with 37 mm, 0.8 µm pore size polycarbonate membrane filters (Poretics, Inc.). After sampling, two of the six samples were recapped and stored. HEPA filters were attached upstream of the remaining four samples so that additional clean air could be drawn across the sample without any further collection of bioaerosols. Sampling was continued in this fashion at 2 l/min for an additional 2 h for two of the samples, and an additional 5 h for the other two samples. This yielded six samples expected to have identical concentrations of viable fungi and bacteria if there is no effect of extended sampling time on viability. The six filters, however, were subjected to varying durations of air flow. Two blank filters were included for each location. One blank was simply uncapped and immediately recapped without sampling (lab blank), and the second had HEPA-filtered air drawn over it for 5 h (field blank).
Sample analysis
All samples and blanks were sent overnight to the Environmental Microbiology Laboratory at the Harvard University School of Public Health where investigators were blinded as to which samples had been run for which time period. The filters were processed within 24 h of receipt using a variation of a previously described method (Dillon et al., 1996). Particles were washed from the filter by three times adding 2 ml 0.02% Tween 20 in distilled water to the intact cassette, agitating for 5 min, and combining all washes (6 ml).
A fraction of the resulting suspension was 10x serially diluted, and 0.1 ml plated in duplicate onto the following culture media: 2% malt extract agar for mesophilic fungi (room temperature incubation for 7–10 days) and thermophilic fungi (45°C incubation for 2 days), R2A for mesophilic bacteria (30°C incubation), and tryptic soy agar (TSA) for thermophilic bacteria (56°C for 2 days). After incubation, colonies on dilution plates with optimal colony density were analyzed. Bacteria were classified as Bacillus species, actinomycetes, Gram-positive cocci or rods, Gram-negative cocci or rods, or other. Fungi were classified to genus where possible and frequently recovered Aspergillus types were identified to subgenus.
Filter wash suspensions were acridine orange stained following the methods of Hobbie et al. (Hobbie et al., 1977). One milliliter of the remaining wash solution was fixed using 1% formaldehyde, and filtered through a 25 mm diameter, 0.2 µm pore polycarbonate filter in a 15 ml Poretics filter tower (Osmonics Laboratory, Livermore, CA). One milliliter of filter-sterilized solution of 0.1% acridine orange in distilled water was added to the tower and allowed to stand for 5 min, after which it was pulled through the filter, followed by three x 5 ml washes of filter-sterilized distilled water. The filter was placed on a glass microslide and allowed to dry, after which a drop of immersion oil was added then mounted with a cover glass. Particles were counted on the stained filter using an Olympus BH2 microscope with a 100 W mercury epifluorescence lamp, 490 nm excitation filter and 515 nm barrier filter. Fluorescently stained bacterial cells (particles of bacterial shape and size) and fungal spores in 10 random fields were counted at 1000x magnification.
Results of total fungi and bacteria from the acridine orange staining procedure were reported as number of spores or cells per filter respectively. Sample volumes from the 1 h of unfiltered sampling were used to calculate concentrations of total fungi (spores/m3), total bacteria (cells/m3), and total culturable fungi and bacteria (both c.f.u./m3). Several fungal samples had zero spores according to the acridine orange analysis. These were assigned a value of LOD/
2 = 4.3 x 104. Because the data were highly skewed, concentrations were log transformed (base 10) prior to analysis. Ratios of culturable to total fungi and bacteria were calculated after log transformation, e.g. ratio = log (c.f.u./m3)/log (spores/m3).
Data analysis
The hypothesis to be tested was whether a loss of viability was associated with longer sampling time. Loss of culturability was determined by evaluating the differences in the concentrations of culturable fungi and bacteria, as well as in the ratio of culturable to total fungi and bacteria for the three different time periods sampled. A general linear model including sample duration as a fixed factor, sample location as a random factor, and the interaction of sample duration and location was generated for each of four dependent variables of interest. The four independent variables evaluated were log of culturable fungi concentration, log of culturable bacteria concentration, ratio of log culturable fungi to log total fungi, and ratio of log culturable bacteria to log total bacteria.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The predominant fungi in virtually all samples were Aspergillus and Penicillium species with a smaller number of Cladosporium and very small numbers of other species. The predominant bacterial types were actinomycetes and Bacillus species in all locations, with some Gram-positive cocci and rods at facility B and some Gram-positive cocci at facility C. Results from the R2A agar were used to determine viable bacterial concentrations (all types) since this is the culture medium that is commonly used as a broad spectrum medium for ‘environmental source’ bacteria (bacteria that are not pathogenically or commensally dependent upon other plants or animals).
A summary of culturable and total concentrations of bacteria and fungi and their ratios by sample location is presented in Table 1. All concentrations have been log transformed to achieve a more normal distribution of the data. There are no results for total bacteria and fungi for location 2 due to a laboratory error. Geometric mean concentrations of culturable fungi from the different sampling locations ranged from 1.7 x 103 (log value 3.22) to 6.2 x 107 (log value 7.79) c.f.u./m3, and mean concentrations of total fungi ranged from 4.9 x 105 (log value 5.69) to 9.3 x 107 (log value 7.97) spores/m3. Geometric mean concentrations of culturable bacteria ranged from 1.17 x 104 (log value 4.07) to 1.0 x 106 (log value 6.0) c.f.u./m3, and concentrations of total bacteria ranged from 3.0 x 107 to 3.1 x 108 cells/m3. Ratios of log culturable fungi to log total fungi ranged from 0.53 to 1.01 and ratios of log culturable bacteria to log total bacteria ranged from 0.54 to 0.70. It should be noted that ratios >1 suggest that there were more culturable than total organisms, which is clearly not possible. It should also be noted that the ratios presented in the graphs are ratios of the logs of the concentrations and cannot be interpreted as percentages. The ratios of culturable to total bacteria concentrations (without log transformation) ranged from 0.0001 to 0.064 and ratios of culturable to total fungi concentrations ranged from 0.001 to 2.3.
|
Changes in culturable and total concentrations and in the ratio of culturable to total concentrations over time by area are shown in Figs 1 and 2. It is clear from these plots that there was little change over sampling time in either the concentration or the ratios of culturable to total bacteria and fungi for any of the areas sampled. In some cases, however, the variability of measurements seemed to increase with sample duration.
|
Results of general linear models fit for each of the four dependent variables of interest are presented in Table 2. It is clear that the location of the sample was highly significant in predicting all four dependent variables. This is expected since the different locations chosen would be expected to have different bioaerosol concentrations. Neither sampling time nor the interaction of sampling time and location were significant contributors to the model. For the ratio of culturable to total bacteria, the estimated effect of time did have a P-value <0.05. In Fig. 2, however, the trend appears to be such that the ratio of culturable to total bacteria increased with sample duration, which is the opposite of what would be expected if the viability of the cells was being compromised by the additional sampling time. This appearance of a trend was probably a chance observation due to the variability in the analytical method.
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The results of this study suggest that sampling times up to 6 h at 2 l/min do not affect the viability (as measured by culture) of either bacteria or fungi in environments such as composting facilities in which selected microbial species predominate. Although there may be some species of bacteria and fungi that are more susceptible to desiccation, the species found in the composting facilities, which were predominantly Aspergillus and Penicillium species, maintained their viability. This finding is encouraging since full-shift personal samples are preferable to short-term area samples for exposure assessment in epidemiological studies, and greater confidence can now be placed on longer-term samples collected on filter media.
It should be noted, however, that the recovery of bacteria from filter samples has been shown to be significantly lower than recovery from samples collected using either the Andersen sampler or an impinger (Jensen et al., 1992). Without additional studies comparing sampling instruments in composting environments, we are unable to determine if bacterial recoveries may be reduced using the filter methods.
In addition, although both the concentration of culturable organisms and the ratio of the logs of culturable to total organisms remained constant over different sampling times for side-by-side samples, the ratios of culturable to total organisms varied widely in different locations. In Table 1, the range of ratios for the log of the concentration of fungi is 0.53–1.01, and for bacteria it is 0.54–0.70. Even when samples were collected at two different locations at the same facility, the ratio can be drastically different, as is seen in areas 3 and 4 where ratios of logs of culturable to total fungi are 0.53 and 0.95 respectively. Given that these are ratios of logs, the variability in ratios of concentrations is enormous (the ratios of the average concentrations for areas 3 and 4 are 0.001 and 0.43 respectively). Therefore, this study supports the previously held notion that culturable counts do not accurately predict total counts and that culturable counts should not be relied upon solely in epidemiological studies of allergic illness.
Interestingly, for fungi the variability in the total organisms measured was higher than for culturable cells, while for bacteria the opposite was true (Table 1). For bacteria, the species that grew in culture depended to a great extent on the chosen agar. For example, when dilutions from these samples were grown on MacConkey’s agar rather than R2A, Gram-negative and Gram-positive cocci predominated and the actinomycetes and bacilli virtually disappeared. The numbers of recovered bacteria for any given culture medium was therefore highly dependent on the species present and if this was highly variable, so was the result. Fungi are much more robust in nature and perhaps not as dependent on culture media, so that variability is not quite as high as it is for culturable bacteria. The counts of total fungi, however, were extremely variable. There were several samples on which there were no spores observed causing the ratio of viable to total organisms to exceed 1. There are a number of possible explanations for this. First, the total counts were derived from a smaller portion of the sample than were cultured samples, possibly resulting in increased variability. About 0.16–1.6% of the sample (depending on the dilution used) was analyzed by culture. With acridine orange, only 10 fields (representing 0.01% of the entire filter and 0.0017% of the sample) were analyzed. Second, colonies are much easier to see and count than are microscopic, fluorescently stained particles. Third, fungal spores stain variably with acridine orange, with some staining very poorly, possibly due to wall pigments (Muilenberg and Burge, 1994).
In conclusion, full-shift sampling on polycarbonate filters is not only the most practical method for collecting bioaerosols in some situations, but a method that does not significantly compromise the viability of at least certain types of fungi or bacteria, such as those seen at the composting facilities in this study. In addition, since viable counts are not predictive of total counts, this is an extremely useful method in that it allows for analysis of either total organisms, culturable organisms, or both from the same sample.
Acknowledgements—This project was supported by cooperative agreement U60/CCU014539 from the National Institute for Occupational Safety and Health of the US Centers for Disease Control and Prevention.
|
FOOTNOTES |
|---|
* Author to whom correspondence and reprint requests should be addressed: University of Washington, Department of Environmental Health, PO Box 357234, Seattle, WA 98195, USA.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Blomquist G, Palmgren U, Strom G. (1984) Improved techniques for sampling airborne fungal particles in highly contaminated environments. Scand J Work Environ Health; 10: 253–8.[Medline]
Burge HA. (1995) Bioaerosols. Boca Raton, FL: Lewis, pp. 16–23.
Dillon HK, Heinsohn PA, Miller JD. (1996) Field guide for the determination of biological contaminants in environmental samples. Fairfax, VA: American Industrial Hygiene Association.
Eduard W, Heederick D. (1998) Methods for quantitative assessment of airborne levels of noninfectious microorganisms in highly contaminated work environments. Am Ind Hyg Assoc J; 59: 113–27.[Web of Science][Medline]
Heikkila P, Kotimma M, Tuomi T, Salmi T, Louhelainen K. (1988) Identification and counting of fungal spores by scanning electron microscope. Ann Occ Hyg; 32: 241–8.
Hobbie JE, Daley RJ, Jasper S. (1977) Use of Nuclepore filters for counting bacteria by fluorescence microscopy. Appl Environ Microbiol; 33: 1225–8.
Jensen PA, Todd WF, Davis GN, Scarpino PV. (1992) Evaluation of eight bioaerosol samplers challenged with aerosols of free bacteria. Am Ind Hyg Assoc J; 53: 660–7.[Medline]
Karlsson K, Malmberg, P. (1989) Characterization of exposure to molds and actinomycetes in agricultural dusts by scanning electron microscopy, fluorescence microscopy and the culture method. Scand J Work Environ Health; 15: 353–9.[Web of Science][Medline]
Lacey J, Williamson PAM, Crook B. (1996) Microbial emissions from composts and associated risks—trials and tribulations of an occupational aerobiologist. In Muilenberg M, Burge H, editors. Aerobiology. Boca Raton, FL: Lewis, pp. 1–15.
Macher JM, Chatigny MA, Burge HA. (1995) Sampling airborne microorganisms and allergens. In Cohen BS, Hering SV, editors. Air sampling instruments. Cincinnati, OH: American Conference of Governmental Hygienists, pp. 589–608.
Muilenberg ML. (1989) Aeroallergen assessment by microscopy and culture. Immun All Clin N Am; 9: 245–68.
Muilenberg ML, Burge HA. (1994) Filter cassette sampling for bacterial and fungal aerosols. In Samson RA, Flannigan B, Flannigan ME, Verhoeff AP, Adan OCG, Hoekstra ES, editors. Health implications of fungi in indoor environments. Amsterdam: Elsevier, pp. 75–89.
Palmgren U, Strom G, Blomquist G, Malmberg P. (1986) Collection of airborne micro-organisms on nuclepore filters, estimation and analysis—CAMNEA method. J Appl Bacteriol; 61: 401–6.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  K. Fallschissel, P. Kampfer, and U. Jackel Direct Detection of Salmonella Cells in the Air of Livestock Stables by Real-Time PCR Ann. Hyg., November 1, 2009; 53(8): 859 - 868. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. M. Madsen and A. K. Sharma Sampling of High Amounts of Bioaerosols Using a High-Volume Electrostatic Field Sampler Ann. Hyg., April 1, 2008; 52(3): 167 - 176. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||