Int. J. Environ. Res. Public Health 2013, 10, 541-555; doi:10.3390/ijerph10020541 OPEN ACCESS International Journal of Environmental Research and Public Health ISSN 1660-4601 www.mdpi.com/journal/ijerph Article Assessment of the Levels of Airborne Bacteria, Gram-Negative Bacteria, and Fungi in Hospital Lobbies Dong-Uk Park 1,*, Jeong-Kwan Yeom 2,3, Won Jae Lee 3 and Kyeong-Min Lee 1,4 1 Department of Environmental Health, Korea National Open University, Seoul, 110-791, Korea 2 Hanbul Energy Manufacturing, Sampyeong-Dong, Bundang-Gu, Gyeonggido, 463-413, Korea 3 Department of Global Healthcare Management, Gachon University, Gyeonggido, 461-701, Korea 4 Institute for Occupational Health & Graduate School of Public Health, Yonsei University, Seoul, 120-752, Korea * Author to whom correspondence should be addressed; E-Mail: [email protected]; Tel.: +82-2-3668-4707; Fax: +82-2-741-4701. Received: 31 October 2012; in revised form: 4 January 2013 / Accepted: 16 January 2013 / Published: 31 January 2013 Abstract: Aims: We assessed the levels of airborne bacteria, Gram-negative bacteria (GNB), and fungi in six hospital lobbies, and investigated the environmental and hospital characteristics that affected the airborne microorganism levels. Methods: An Andersen single-stage sampler equipped with appropriate nutrition plate agar was used to collect the samples. The three types of microorganisms were repeatedly collected at a fixed location in each hospital (assumed to be representative of the entire hospital lobby) from 08:00 through 24:00, with a sampling time of less than 5 min. Temperature and relative humidity were simultaneously monitored. Results: Multiple regression analysis was used to identify the major factors affecting microorganism levels. The average levels of bacteria (7.2 × 102 CFU/m3), GNB (1.7 × 10 CFU/m3), and fungi (7.7 × 10 CFU/m3) indicated that all hospital lobbies were generally contaminated. Season was the only factor that significantly affected the levels of all microorganisms (p < 0.0001), where contamination was the highest during the summer, significantly higher than during the winter. Other significant factors varied by microorganism, as follows: airborne bacteria (number of people in the lobby, sampling time), GNB (scale of hospital), and fungi (humidity and air temperature). Conclusions: Hospital lobby air was generally contaminated with microorganisms, including bacteria, GNB, and fungi. Environmental factors that may significantly influence the airborne concentrations of these agents should be managed to minimize airborne levels. Int. J. Environ. Res. Public Health 2013, 10 542 Keywords: hospital lobby; bacteria; fungi; Gram-negative bacteria (GNB) 1. Introduction Exposure to biological agents (including microorganisms) is associated with a wide range of major public health issues, such as infectious diseases, acute toxic effects and allergies [1]. Although airborne microorganisms encountered in hospital lobbies are apparently harmless to healthy people, they can cause adverse health effects in immune-compromised people. Many of those who pass through hospital lobbies belong to the vulnerable group of weak, elderly, and infirm people, and thus may be very sensitive to biological hazards. In particular, hospitalized patients could be at significantly increased risk of bio-aerosol exposure [2]. Airborne microorganisms originate not only from people (including patients), but are also spawned by various indoor hospital characteristics and outdoor environmental sources. There are a number of such factors that may be related to the generation of bio-aerosols in the lobby of a hospital. Airborne microorganisms and other sources of contamination in hospitals must be reduced to a minimum, because many of the people passing through hospital lobbies could be very sensitive to these hazardous agents. Thus, to maintain the lowest possible airborne microorganism levels in hospital lobbies, it is crucial to identify the factors influencing these levels. Hospital buildings may be regarded as dynamic environments affected by season, weather conditions, indoor ventilation system design and operation, intrusion of moisture, the outdoor microbial load and the number of occupants and visitors, and human activities. These factors can be related to a condition for microbial growth. Nutrient availability, indoor temperature, and shelter generally cannot be managed to limit indoor microbial survival and growth. Several studies have been conducted to determine the factors influencing microbial growth in hospital environments. Scaltriti et al. reported that risk factors for microbial contamination of air in operating theatres were the frequency of door-opening, taken as an index of staff and visitor movement, was a positive predictor of raised bacterial counts [3]. Obbard et al. identified the occupant density as the key factor influencing the level of airborne bacteria. Humidity was also found to be important factor depending on the particular location within the hospital [4]. To date, airborne microorganism levels have been reported in hospital clean rooms, such as operating rooms, hospital rooms, intensive care units, surgical units, hematological wards, and maternity wards [5–7]. To the best of our knowledge, no previous studies have investigated how hospital characteristics are associated with airborne levels of microorganisms in lobbies. Accordingly, we examined the levels of airborne bacteria, gram negative bacteria (GNB), and fungi with respect to environmental and other characteristics in the lobbies of six hospitals. 2. Methods 2.1. Monitoring Strategy Microorganism levels, relative humidity, and temperature were measured in six Korean hospitals with more than 500 beds, between July (summer) and January (winter). General characteristics for general hospital lobby were shown in Table 1. We are unaware of any hospital that has installed carpet Int. J. Environ. Res. Public Health 2013, 10 543 on the entire hospital floor (including the lobby) in Korea. No hospitals have a wooden construction or use other materials that would provide a good environment for microbial growth. A different hospital was examined every month. Air samples were repeatedly taken from the start of lobby service (08:00 AM) through the end of lobby service (18:00 PM), continuing until midnight (24:00). The air was monitored after lobby service hours for comparison with samples taken during lobby service hours. Sampling instruments were placed at a height of around 1.2–1.5 m, in a fixed location assumed to be representative of the entire lobby. Table 1. General characteristics for the hospitals investigated Hospital No of bed Building age, year Area of lobby (m2) A 710 23 2,764 B 1,000 10 2,717 C 1,200 3 2,413 D 530 26 1,674 E 867 19 2,712 F 800 13 1,710 Our microbial sampling strategy was designed to best represent the level of microorganisms generated in hospital lobby environment. An Anderson single-stage cascade sampler was used for all collections, with a flow rate of 28.3 L/min and a sampling time of less than 5 min to avoid the collection of unaccountable microorganisms. The airborne microorganisms were targeted one after another, using a 20 mL nutrient plate (tryptic soy agar, TSA) for bacteria, MacConkey agar (MAC) for GNB, and sabouraud dextroseagar with chloramphenicol (SDAC) for fungi) coupled inside the stage sampler. When changing the collection plates, the stage hole was sterilized with a 70% ethanol solution to prevent cross-contamination. To evaluate the representative level of airborne microorganisms during the morning (08:00–13:00), afternoon (13:00–18:00) and after lobby service (18:00–24:00), three or four samples by type of microorganism were those taken in each respective period. The sampling interval between types of microorganisms was the duration required to both change the collection plates and take samples for the other two types of microorganisms. After sample collection, the agar plates were transported to the laboratory and incubated at 37 ± 1 °C for 2 days for bacteria and GNB and at 25 ± 1 °C for 4 days for fungi [8]. Specific statistical corrections were made using correction factors obtained for each growing medium from Conversion Tables provided by single stage operating manual [9]. For all the media the minimum detectable number of colony forming units (CFU) was 30 per sample. Data below minimum detectable number was treated as missing. The microbial counts were expressed in terms of colony-forming units (CFU) per unit volume of air (m3). Total culturable bacterial and fungal counts were compared to ambient environmental levels (background) to determine the degree of incremental levels generated during household waste handling. Our microbial sampling strategy mentioned above was determined to best represent the level of microorganisms generated in hospital lobby. Direct-reading instruments for monitoring the temperature and humidity (IAQ, Model; 8762, TSI, Shoreview, MN, USA) were placed next to the cascade sampler to simultaneously monitor air temperature and humidity during each sampling interval. Int. J. Environ. Res. Public Health 2013, 10 544 2.2. Data Analysis The data-logging interval was set at 30 s, averaged with a time-weighted average. Direct-reading temperature, humidity, and CO results were matched with airborne microorganism levels measured 2 during the sampling period of microorganisms. We selected a number of environmental variables and hospital characteristics that may influence airborne microorganism levels in hospital lobbies. The quantitative variables were categorized based on subjective classification, as follows: the scale of a hospital expressed by the number of beds (<1,000 vs. ≥1,000), lobby age (<10 years, 10–19 years, and ≥20 years), the input location of outdoor air in the HVAC (Heating, Ventilating and air conditioning) system (underground vs. ≥1st floor), the number of filters in the HVAC system (pre and medium vs. only medium), and lobby volume (<10,000 m3 vs. ≥10,000 m3). Both lobby service time and cleaning time were dichotomized into “on (08:00–18:00)” and “off (18:00–24:00)”. Correlations between the selected independent variables were evaluated to identify potential problems prior to offering to the multiple regression model. Correlations involving at least one continuous variable were tested using Spearman’s rho, and between two dichotomous variables using Kendall’s tau. The absolute values of all correlation coefficients were less than 0.5, so all of the above variables were considered appropriate for input to a single model. Simple linear regression analysis was used to examine the relationships among quantitative variables such as air temperature, humidity, CO level, and the number of people in the lobby during sampling, 2 and microbial concentrations. Analysis of variance (ANOVA) was employed to compare mean microorganism levels between categories of qualitative variables. Only variables with p-values less than 0.25 were ultimately included in the multiple regression analysis [10]. Stepwise analysis was used to identify the final factors that significantly affected the airborne concentrations of microorganisms. Descriptive statistics were obtained and data analysis was conducted using the STATA (Version 11.0) software (Stata Corp. LP, College Station, TX, USA). The concentrations of all microorganisms were found to be distributed log-normally. To improve the statistical models used in this study, the microorganism data employed as dependent variables were all log-transformed. One of the categories served as a reference group for comparison with the other groups. Each descriptive factor was given a value of zero when used as a reference variable. 3. Results Temperature, relative humidity and carbon dioxide level (CO ) was shown by hospital in Table 2. 2 The mean level of airborne bacteria (7.2 × 102 CFU/m3) was lower than the indoor air quality standard (8.0 × 102 CFU/m3) enforced by the Korean Environmental Protection Agency [11], while 30% of the 76 samples exceeded the standard. The average concentration of GNB was 1.7 × 10 CFU/m3, although 12 samples were below the limit of detection. The airborne fungi level ranged from 1.1 to 2.2 × 102 CFU/m3. Significant differences were found among the six hospitals (p < 0.0001; Table 3). Int. J. Environ. Res. Public Health 2013, 10 545 Table 2. Temperature, relative humidity and carbon dioxide level (CO ) by hospital. 2 Temperature, °C Relative Humidity, % CO , ppm 2 No. of Hospital Range Range Range Samples * Mean ± SD Mean ± SD Mean ± SD (min–max) (min–max) (min–max) A 13 25.1 ± 0.4 24.5–25.9 66.8 ± 3.9 62.2–72.6 801 ± 221 458–1,061 B 12 26.5 ± 0.1 26.3–26.6 70.6 ± 1.7 67.6–72.8 432 ± 44 377–509 C 12 25.9 ± 0.4 25.1–66.3 65.7 ± 0.1 65.6–65.8 503 ± 87 393–625 D 12 25.8 ± 0.9 24.0–26.9 65.3 ± 0.7 63.5–65.8 481 ± 109 342–651 E 13 22.9 ± 0.8 21.7–24.1 65.8 ± 0.1 65.6–66.0 558 ± 111 366–692 F 14 20.7 ± 1.2 17.4–21.8 65.0 ± 0.4 64.4–65.6 409 ± 50 345–486 Total 76 24.4 ± 2.2 17.4–26.9 66.5 ± 2.5 62.2–72.8 531 ± 176 342–1,061 No of sample: number of time-weighted average (TWA) of data logged every 30 s during the sampling period of microorganisms. Table 3. Airborne microorganism levels by hospital. Bacteria, CFU/m3 GNB, CFU/m3 Fungi, CFU/m3 No of Hospital Range Range Range sample GM GSD GM GSD GM GSD (min-max) (min-max) (min-max) A 13 8.7 × 102 1.6 3.8 × 102 – 1.8 × 103 10 2.7 4 – 1.1 × 102 53 1.9 21 – 2.1 × 102 B 12 6.2 × 102 3.2 50 – 2.3 × 103 21 2.1 5 – 5.6 × 10 79 1.8 29 – 2.0 × 102 C 12 7.6 × 102 1.8 2.8 × 102 – 1.8 × 103 18 1.9 7 – 5.2 × 10 66 2.1 18 – 1.5 × 102 D 12 6.8 × 102 1.7 3.3 × 102 – 1.8 × 103 11 2.3 4 – 4.7 × 10 83 2.1 18 – 2.2 × 102 E 13 5.4 × 102 1.6 2.7 × 102 – 1.7 × 103 16 2.1 LOD – 4.0 × 10 59 1.9 25 – 2.1 × 102 F 14 2.0 × 102 1.7 80 – 4.5 × 102 5 1.5 LOD – 1.1 × 10 36 2.7 11 – 1.5 × 102 Total 76 5.5 × 102 2.3 5.0 × 10 – 2.3 × 103 13 2.4 LOD – 1.1 × 102 59 2.2 11 – 2.2 × 102 GM = geometric mean; GSD = geometric standard deviation, unitless; GNB = Gram-negative bacteria; LOD = Limit of detection (30 CFU/sample). Int. J. Environ. Res. Public Health 2013, 10 546 Our results indicate that the lobby air in all six hospitals was generally contaminated to some extent. We observed a clear pattern in microorganism levels over time, from the start of lobby service (08:00) to the end of lobby service, and after hours until midnight (Figure 1). Figure 1. Daily variation in airborne microorganism concentrations (mean and range measured at six hospitals during summer, fall, and winter). Int. J. Environ. Res. Public Health 2013, 10 547 The airborne bacteria level was higher during lobby service hours (between 08:00 and 18:00), when crowds of people were present and during summer (Figure 2). Figure 2. Variation in airborne microorganism concentrations by season. Int. J. Environ. Res. Public Health 2013, 10 548 Table 4. Relationship between the levels of log-transformed airborne microorganism and quantitative variables. Airborne microorganism Relative humidity, % Temperature, °C Number of people present CO , ppm 2 level, CFU/m3 β SE Model p R2 β SE Model p R2 β SE Model p R2 β SE Model p R2 Bacteria NS NS NS 0.01 0.18 0.038 0.0001 0.23 0.01 0.001 <0.0001 0.37 0.01 0.001 <0.0001 0.37 GNB 0.05 0.04 0.2197 0.02 0.14 0.05 0.0068 0.1 NS NS NS NS NS NS NS NS Fungi 0.09 0.034 0.0127 0.08 NS NS NS NS 0.002 0.001 0.056 0.05 0.002 0.001 0.056 0.05 GNB: Gram-negative bacteria; NS = not statistically significant at p = 0.25. Table 5. The levels of log-transformed bacteria, gram-negative bacteria, and fungi among the categories of occupational and environmental variables. Bacteria, CFU/m3 GNB, CFU/m3 Fungi, CFU/m3 Independent No of ANOVA ANOVA variables sample Mean SD Mean SD Mean SD ANOVA model p model p model p Sampling time Service hours 44 9.3 × 102 5.7 × 102 P ≤ 0.0001 1.5 × 10 1.3 × 10 P = 0.0422 5.6 × 10 3.3 × 10 p < 0.0001 (08:00–18:00) After service hours 32 4.4 × 102 3.0 × 102 2.0 × 10 2.2 × 10 1.1 × 102 6.4 × 10 (18:00–24:00) Season Summer 25 9.7 × 102 6.2 × 102 p = 0.0001 2.2 × 10 2.4 × 10 p = 0.0005 7.8 × 10 5.4 × 10 p = 0.0206 Fall 37 7.5 × 102 4.3 × 102 1.8 × 10 1.3 × 10 8.6 × 10 5.6 × 10 Winter 14 2.3 × 102 1.1 × 102 0.4 × 10 0.3 × 10 5.3 × 10 4.7 × 10 Scale of hospital (number of beds) <1,000 52 6.3 × 102 4.4 × 102 p = 0.0192 1.3 × 10 1.7 × 10 p = 0.0016 7.3 × 10 5.6 × 10 p = 0.1365 ≥1,000 24 9.3 × 102 6.4 × 102 2.4 × 10 1.6 × 10 8.7 × 10 5.1 × 10 Lobby age (years) <10 24 9.3 × 102 6.4 × 102 p = 0.0001 2.4 × 10 1.6 × 10 p = 0.0059 8.7 × 10 5.1 × 10 p = 0.0788 10–19 27 4.1 × 102 3.2 × 102 1.0 × 10 1.2 × 10 6.3 × 10 5.1 × 10 ≥20 25 8.7 × 102 4.4 × 102 1.7 × 10 2.2 × 10 8.4 × 10 6.0 × 10 Int. J. Environ. Res. Public Health 2013, 10 549 Table 5. Cont. Cleaning time Service hours 37 8.7 × 102 4.4 × 102 p = 0.0001 1.8 × 10 1.9 × 10 p = 0.4623 8.3 × 10 5.7 × 10 p = 0.2351 (08:00–18:00) After service hours 26 5.7 × 102 6.6 × 102 1.4 × 10 1.7 × 10 7.1 × 10 5.3 × 10 (18:00–24:00) Location of fresh air inlet underground 52 6.8 × 102 5.6 × 102 p = 0.0446 1.6 × 10 1.9 × 10 p = 0.3524 7.0 × 10 5.2 × 10 p = 0.0902 & 1st floor higher than 24 8.3 × 102 4.5 × 102 1.8 × 10 1.3 × 10 9.3 × 10 5.7 × 10 2nd floor Number of filters in HVAC medium 14 2.3 × 102 1.1 × 102 p ≤ 0.0001 0.4 × 10 0.3 × 10 p = 0.0001 5.3 × 10 4.7 × 10 p = 0.0056 pre and medium 62 8.4 × 102 5.2 × 102 2.0 × 10 1.8 × 10 8.3 × 10 5.5 × 10 Lobby volume (m3) <10,000 52 6.3 × 102 4.4 × 102 p = 0.0957 1.3 × 10 1.7 × 10 p = 0.0016 7.3 × 10 5.6 × 10 p = 0.1365 ≥10,000 24 9.3 × 102 6.4 × 102 2.4 × 10 1.6 × 10 8.7 × 10 5.1 × 10 Total 76 7.2 × 102 5.3 × 102 1.7 × 10 1.7 × 10 7.7 × 10 5.4 × 10 GNB: Gram-negative bacteria Int. J. Environ. Res. Public Health 2013, 10 550 On the other hand, the reverse pattern was detected for fungi. Higher levels of fungi were detected after lobby service hours, when the HVAC system was not operating. Airborne microorganism levels were regressed against quantitative factors such as air temperature, humidity, CO levels, and the number of 2 people in the lobby during sampling (Table 4), and were compared between qualitative hospital characteristic categories (Table 5). Simple linear regression indicated that air temperature, CO level, and 2 the number of people in the lobby during sampling accounted for 23%, 25% and 37%, respectively, of the variation observed in the airborne bacteria level (p < 0.0001). The airborne GNB level was significantly associated only with air temperature (p = 0.068). Air temperature, humidity, and the number of people were found to affect the level of fungi at p < 0.25 (Table 4). Regarding GNB, all qualitative variables except cleaning time (p = 0.4623) and location of fresh air supply (p = 0.3524) were significant at p < 0.05. The season and the number of filters in the HVAC were significantly associated with all three types of airborne microorganism (p ≤ 0.0001), being markedly higher in summer than in winter. Correlation matrix analysis revealed that airborne levels of GNB and fungi were significantly correlated with one another (Table 6). Table 6. Correlation matrix analysis results between the log-transformed levels of airborne bacteria, gram-negative bacteria, and fungi levels (p-value) Bacteria GNB Fungi Bacteria 1 GNB NS 1 Fungi NS 0.43 (p = 0.0002) 1 GNB: Gram-negative bacteria; NS = not statistically significant at p = 0.05. All hospital and environmental variables identified as significant at p < 0.025 in the univariate analysis were included in the multiple regression analysis (Table 7). Season was the only factor that appeared to significantly affect airborne levels of all microorganisms. The other factors varied according to the type of microorganism. Significant statistical models were developed to predict airborne levels of bacteria, GNB and fungi accounting for 75% (p < 0.0001), 24% (p < 0.0001) and 36% (p < 0.0001) of the observed variation, respectively. 4. Discussion Most hospital-based airborne microorganism studies have been performed in clean rooms or departments, where the risk of infection is greatest [5–7]. Relatively little is known about microbial contamination in hospital lobbies. In the large hospitals included in our study, we found that the lobby air was contaminated with microorganisms, including bacteria, GNB, and fungi. The levels recorded (bacteria: 7.2 × 102 CFU/m3; fungi: 5.5 × 102 CFU/m3) were far higher than those hitherto reported in hospital clean rooms, such as operating rooms, hospital rooms, intensive care units (ICD), surgical units, hematological wards, pneumonological department and maternity wards [2,5–7]. Levels reported in hospital rooms with a high degree of cleanliness, such as operating rooms and transplant units, range from 0.1 to 10 CFU/m3. Augustowska and Dutkiewicz reported that mean monthly microflora measured in a hospital ward of the pneumonological department in Poland ranged from 257–436 CFU/m3 for airborne bacteria and from 10–96 CFU/m3, respectively [2].