Direct Monitoring of Gram-negative Agents of Nosocomial Infections in Hospital Air by a PCR-based Approach

Gram-negative bacteria (GNB), including Acinetobacter baumannii, Pseudomonas aeruginosa and Legionella, have emerged as causative agents of many problematic infections in healthcare settings worldwide. Using a rapid detection method, this study was designed to investigate the presence of GNB in hospital air as a potential source for spreading and transmitting these bacteria. A total of 51 air samples were taken from different wards in four hospitals over a period of eight months. Air samples were collected using an all-glass impinger and analyzed for the presence of GNB. Detection of P. aeruginosa and Legionella spp. was performed with a nested PCR assay using specific primer sets of the 16S rRNA gene region of the bacteria. For detection of A. baumannii, PCR assay with the specific primers of the inherent blaOXA-51 gene was performed. A. baumannii was the most frequently (29/51) detected gram-negative microorganism in hospital air, followed by P. aeruginosa (15/51). The lowest detection frequency was related to Legionella (9/51), which was not found in air samples from surgical wards. Intensive care units and operating theatres were high-risk areas due to the greater presence of GNB. The results of this study revealed the presence of GNB in various hospital wards and highlight the usefulness of PCR monitoring of hospital air as a rapid and reliable tool for identifying GNB in order to prevent and control nosocomial infections, especially for the protection of vulnerable patients.


INTRODUCTION
Nosocomial infections are a significant health concern, globally.These infections lead to increased morbidity and mortality and can extend hospitalization and treatment period which is associated with additional costs (Ducel et al., 2002;Sehulster et al., 2003;De Francesco et al., 2013).
Gram-negative bacteria (GNB) including Acinetobacter baumannii, Pseudomonas aeruginosa and Legionella spp.are among the most important etiological agents of hospital acquired infections worldwide (Ducel et al., 2002;Sehulster et al., 2003;Rebmann and Rosenbaum, 2011;De Francesco et al., 2013).These ubiquitous bacteria which have been strongly implicated in nosocomial infections, account for a significant proportion of morbidity and mortality in healthcare settings.A. baumannii and P. aeruginosa can cause a variety of infections including pneumonia, urinary tract, wound and blood stream infections (Sehulster et al., 2003;Rebmann and Rosenbaum, 2011;De Francesco et al., 2013).Legionella causes Legionnaires disease, a potentially fatal form of pneumonia or an acute self-limiting, influenza-like illness (Pontiac fever) (Ducel et al., 2002;Sehulster et al., 2003).Hospital air has been indicated as a potential source for dispersion and transmission of Acinetobacter infections (Sehulster et al., 2003).Exposure to Legionella and P. aeruginosa occurs through inhalation of aerosols generated from contaminated water sources and faucets (Sehulster et al., 2003).Therefore, exposure to clinically significant gram-negative bacteria released into the air could lead to a variety of infections especially in susceptible patients.In a survey among pediatric patients, identical Acinetobacter spp.were cultured from the patients, air, and room air conditioners (Sehulster et al., 2003).A same profile of macrorestriction of genomic DNA has been reported in airborne L. pneumophila from an aerated basin and the epidemic strain detected in patients (Nhu Nguyen et al., 2006).
Considering the increased number of susceptible patients due to age, illness, immunosuppression and other risk factors (Ducel et al., 2002), and the importance of eliminating exposure of high-risk patients to potential routes of acquisition of gram-negative infections, air quality monitoring is a critical part of the hospital management protocol (Wu et al., 2011;Tang and Wan, 2013).Rapid monitoring of microbial quality of air in health-care settings promotes early awareness of the possible infection sources.In other words, implementation of appropriate risk reduction and control programs of nosocomial infections could be achievable through complete microbiological information about the hospital environments especially high-risk areas.
Detection of GNB in hospital environment samples by conventional culture methods is labour-intensive, time consuming and costly due to the need to several different types of media.Polymerase chain reaction (PCR)-based methods provide a rapid and beneficial means of detection of GNB in environmental samples with high specificity and sensitivity (Asghari et al., 2013;Baghal Asghari et al., 2013).
Based on these premises, the present study was carried out to present a rapid and sensitive detection method for gram-negative agents of nosocomial infections in hospital air and also provide more information on the distribution of these bacteria in air of various parts of hospitals.

MATERIAL AND METHODS
A total of 51 air samples were analyzed for the presence of gram-negative bacteria from four educational hospitals in Isfahan, Iran.Samples were collected from four points in each hospital included operating theatre (OT), intensive care unit (ICU), surgery ward (SW), and internal medicine ward (IM) over a period of 8 months.Air samples were collected using an all-glass impinger, containing 10 mL of phosphate buffer solution at a flow rate of 12 L min -1 and a sampling duration of 3 hours after routine cleaning in the morning.The air sampling points were about 1-1.5 m above ground level to simulate the breathing zone.Temperature and relative humidity were recorded throughout the sampling periods and were approximately 26°C ± 2°C and 28% ± 5%, respectively.
For detection of GNB, air samples were concentrated by centrifugation at 4500 g for 15 min.To extract DNA, the resuspended pellets were frozen and heated in liquid nitrogen and boiling water, respectively for five times.The DNA was further extracted and purified using Promega DNA Extraction Kit (Promega Wizard_ Genomic DNA Purification Kit, Madison, WI) according to manufacturer's instruction.
For the detection of P. aeruginosa and Legionella spp. a nested PCR assay was applied to increase the sensitivity as well as to check the nucleic acid extraction.To check the quality of extracted DNA, the first step of PCR was carried out using universal primer set Eubac 27F and 1492R that target a fragment of the 16S rRNA gene region which is conserved across a broad spectrum of bacteria (Lane, 1991).In the second PCR step, specific primers of the 16S rRNA gene region of P. aeruginosa and Legionella spp.were used as described previously (Asghari et al., 2013;Baghal Asghari et al., 2013).For the detection of A. baumannii, PCR assay with the specific primers of the inherent blaOXA-51 gene was performed (Shamsizadeh et al., 2017).
PCR reactions were run in a final volume of 25 μL as described previously (Asghari et al., 2013;Baghal Asghari et al., 2013;Shamsizadeh et al., 2017) and positive and negative controls were included in all runs.PCR products were visualized by agarose gel electrophoresis using 1.5% gel and the size of amplicons were compared with the 100bp DNA ladder.

RESULTS AND DISCUSSION
Gram-negative bacteria in hospital air can cause a variety of infections in vulnerable patients via direct inhalation, inhalation of bioaerosols or indirectly through deposition of airborne microorganisms on inanimate surfaces.
The results of this study revealed the presence of GNB in various hospital environments (Tables 1-3).Although detection of some GNB such as Acinetobacter, Pseudomonas, Klebsiella and E. coli were reported in hospital air (Wu et al., 2011;Maujean et al., 2012;Huang et al., 2013;Tang and Wan, 2013;Luksamijarulkul et al., 2014), gram-positive bacteria are the most commonly found bacteria in the air of hospitals as well as indoor environments (Maujean et al., 2012;Tang and Wan, 2013;Luksamijarulkul et al., 2014;Mihoseini et al., 2016).GNB have less resistance and cannot withstand the inactivating effects of drying (Sehulster et al., 2003) and therefore may be found in viable but nonculturable (VBNC) state that are alive and may still cause infection.High detection rate of GNB in the present study may be related in part to the VBNC bacteria which couldn't be detected by culture method but could be detected by PCR assay.Analysis of aerosols produced by  industrial cooling towers showed no cultivable Legionella bacteria using the standard culture method, even when the concentrations of ~1 × 10 3 Legionella cells m -3 were detected by use of fluorescent in situ hybridization (FISH) (Mathieu et al., 2006).We used PCR assay as a qualitative method that only detects the presence of GNB in hospital air.However, application of quantitative techniques such as real-time PCR should be taken into account in order to further understand the prevalence of bacteria in hospital environments.
The results showed that the highest detection rate of GNB was related to A. baumannii (29/51) followed by P. aeruginosa (15/51) (Tables 1-2). A. baumannii was detected in various hospital wards with a frequency of 25-100%. A. baumannii has become increasingly common cause of nosocomial infections with about 50% mortality which has the highest rate in patients admitted to ICU (Rebmann and Rosenbaum, 2011).In the study, A. baumannii was detected in 54% of ICU air samples (Table 1).Detection of A. baumannii in ICU air samples has also been reported in other studies (Wang et al., 2003;Huang et al., 2013;Munoz-Price et al., 2013;Shamsizadeh et al., 2017).In air samples of a trauma ICU, 23% (12/50) of patient air zones contained A. baumannii (Munoz-Price et al., 2013).Shamsizadeh et al. (2017) have reported the detection of A. baumannii in air samples of ICU but they couldn't detect it in OT (Shamsizadeh et al., 2017).However, in the present study, A. baumannii was detected by PCR in the OT and SW with a higher rate than ICU which indicates that other parts of hospitals than ICU may be high-risk areas for A. baumannii infections.This difference may be related to the VBNC state of microorganism or higher detection sensitivity of PCR assay than culture method.Furthermore, we used the liquid impingement sampling as an efficient method for collecting GNB bioaerosols.Previous studies have reported a great capacity of the liquid impingement technique for collecting airborne Legionella, particularly in combination with molecular or FISH detection methods (Deloge-Abarkan et al., 2007;Montagna et al., 2017).Study of Deloge-Abarkan et al. (2007) showed that the culturable fraction of airborne L. pneumophila recovered with the liquid impingement was 4 and 700 times higher compared to the impaction and filtration techniques, respectively.
P. aeruginosa was also found in various hospital environments with the highest and lowest frequency in OT and IM wards, respectively.No detection of P. aeruginosa was reported for hospital D (Table 2).In agreement with our study, Tang and Wan (2013) detected Acinetobacter spp. in post-operative recovery room and operating theatres and locations surrounding operating theatres with a higher rate than P. aeruginosa in a medical center in Taiwan.However, P. aeruginosa was the most abundant and frequently detected species (37.5-42.5%) in two intensive care units of a medical center in central Taiwan.Whereas, detection rates of 7.5-13% were reported for A. baumannii (Huang et al., 2013).One study in a hospital in Bangkok also showed that the predominant isolates on the outside of the 230 used surgical masks of hospital personnel were Staphylococcus aureus (41%) and Pseudomonas spp.(38%) (Luksamijarulkul et al., 2014).Pseudomonas spp.(4.8%) was isolated from air samples of a university-hospital autopsy room in France (Maujean et al., 2012).Study of Dettori et al. (2014) on environmental investigation in an intensive care unit of a tertiary hospital in northern Sardinia, Italy, showed the detection of Pseudomonas spp. in air samples, but they couldn't detect any A. baumannii in air samples (Dettori et al., 2014).Since the density of airborne bacteria in hospital environments is affected by several factors such as the number and type of occupants, people activities and ventilation conditions (Wu et al., 2011;Huang et al., 2013), the difference in detection frequency of GNB might be caused in part by the different types of infected patients that act as dispersion sources of microorganisms in hospital air.Study of Shimose et al. (2015) showed that the ambient air of Acinetobacterpositive patients was Acinetobacter positive for an average of 21% of the days in samples of up to 10 consecutive days.Their results showed that 4 pairs of 6 air/clinical isolate pairs were closely related according to rep-PCR (Shimose et al., 2015).On the other hand, as air sampling methods are not standardized (Sehulster et al., 2003), comparison of bioaerosol results between different studies may be ambiguous (Pasquarella et al., 2012).
P. aeruginosa and Legionella can also be frequently found in water systems because of the biofilm formation.Analysis of hospital water samples showed a high rate detection of Legionella and P. aeruginosa (Asghari et al., 2013;Baghal Asghari et al., 2013).Tap water contamination to these bacteria could lead to hospital air contamination through bioaerosols released to the air by water systems (Sehulster et al., 2003;Deloge-Abarkan, 2007).Montagna et al. (2016) detected Legionella pneumophila (Lpn) in air and water samples of three health-care facilities in Italy.Molecular investigation of Lpn isolates showed the same allelic profile for strains in the water and air samples.In an investigation in two health-care centres in northern France, Legionella spp.were detected in 72.2% of air samples of hot water shower aerosols (Deloge-Abarkan, 2007).Legionella detected with the lowest frequency in hospital air (9/51) and analysis of SW air samples showed no Legionella detection (Table 3).Pasquarella et al. (2012) reported more frequently detection of Legionella spp. in water samples when P. aeruginosa was absent, probably due to the growth limitation of Legionella in the presence of Pseudomonas.Although, contaminated water systems have been considered as the sources of nosocomial legionellosis, Mathieu et al. (2006) believe that detection of airborne Legionella offers better information than water analysis.However, to minimize the exposure of patients to contaminated bioaerosols, growth of microorganisms in hospital water systems could be controlled using disinfection methods such as UV irradiation (Nourmoradi et al., 2012).The results of study also showed higher positive samples of GNB in ICU and OT (Fig. 1).Due to the importance of these units in transmission of nosocomial infections, improvement of control measures is needed in order to protect high-risk patients from GNB which may be transmitted through contaminated hospital air.

CONCLUSION
Gram-negative bacteria were frequently detected in the air of various hospital wards, especially the ICU and OT.The introduced rapid and sensitive approach of directly detecting bacteria leads to more information about human exposure to these nosocomial infectious agents.Rapid monitoring of hospital environments provides insights into the proper management of nosocomial infections, especially in high-risk areas.Since humidity is an important factor that affects the survival of GNB, more studies in humid regions are needed.

Fig. 1 .
Fig. 1.Percentage of positive air samples which contained at least one type of the detected gram-negative bacteria in each hospital ward.ICU: Intensive Care Unit; OT: Operating Theatre; SW: Surgery Ward; IM: Internal Medicine Ward.

Table 1 .
Detection frequency (positive samples/total samples) of Acinetobacter baumannii in air of different hospital wards.

Table 2 .
Detection frequency (positive samples/total samples) of Pseudomonas aeruginosa in air of different hospital wards.

Table 3 .
Detection frequency (positive samples/total samples) of Legionella in air of different hospital wards.