Air sampling methods to evaluate microbial contamination in operating theatres: results of a comparative study in an orthopaedics department

Air sampling methods to evaluate microbial contamination in operating theatres: results of a comparative study in an orthopaedics department


Air biocontamination and related health effects are an emerging public health problem. Airborne bacteria, fungi and viruses can cause infection in diverse living or working environments. This is particularly relevant in medical facilities where there are susceptible patients and tissues are exposed to the air during surgery. As such, there is a need for various systems to minimize the introduction, generation and retention of particles in these environments.1,2 In this context, microbiological monitoring of air quality is useful in order to determine the potential exposure of individuals at risk. The control of air biocontamination was first deemed to be necessary in order to reduce the risk of deep wound infections in prodedures such as hip arthroplasties.3,4 It is generally accepted that bacterial contamination of the air in operating theatres, predominantly caused by contaminated skin scales shed from the surgical team, is the main factor causing surgical site infection after clean operations.1,5,6 Whilst the procedures for microbiological assessment of other environmental matrices are established in European Law (e.g. methods are specified for the identification of microbes from water samples), there are no established regulations for air monitoring other than the international norm ISO 14698. This states that different sampling methods exist and different types of devices are commercially available, each with limi- tations, thus leaving the choice of system open.7,8 Microbio- logical content of the air can be monitored by two principal methods: active and passive monitoring of air flows.8 In active monitoring, a microbiological air sampler physically draws a known volume of air through or over a particle collection device, which can be liquid or solid, and the number of micro- organisms present is given in colony-forming units (cfu)/m3 of air. This system is applicable when the concentration of micro- organisms is not very high, such as in an operating theatre.2,9,10 Passive monitoring uses ‘settle plates’ (i.e standard Petri dishes containing culture media) that are exposed to the air for a given time to collect biological particles; these ‘sediment’ out and are subsequently incubated. Results are expressed in cfu/plate/time or cfu/m2/h.11 Several studies have compared the two sampling methods with discordant results; some studies have found significant correlation between the methods,11e14 while other studies have found no correlation.15,16 In particular, Friberg et al. proposed an equation that permits the transformation of the number of cfus settling on a plate over 1 h (cfu/plate/h) into air contamination units (cfu/m3).11,17 One of the most important questions that still needs to be resolved is whether the data from these monitoring systems are actually relevant to what is happening on the operating table. If one of the principal causes of air biocontamination is the surgical team, the results from the general room monitoring systems may underestimate the real risk of exposure to the patient on the surgical table, which may be surrounded by air with a higher level of micro-organisms. To evaluate this hypothesis, the present study investigated and compared the levels of air contamination measured with active and passive systems, and also evaluated the level of air contamination near the wound using nitrocellulose membranes.18,19

Methods This study was undertaken between January 2010 and January 2011 in the operating theatre of the Department of Orthopaedics and Traumatology of the University Hospital ‘Policlinico Consorziale’, Bari, Italy. Microbial sampling was undertaken during 60 total hip arthroplasties, all performed in the same operating theatre (dimensions 8.7 m 4.3 m, height 3.14 m, total volume 117.467 m3). The operating theatre is equipped with a turbulent ventilation system of four low-level inlets with air forced through 0.3-mm H14 99.995 high- efficiency particulate air filters, and three high-level outlets; this provides an air change rate of 19.3 air changes/h and a pressure differential of 7 Pa. Microbial air contamination was evaluated through active sampling (to measure the concentration of micro-organisms in the air) and passive sampling (to measure the rate at which viable particles settle on surfaces). Total viable count and the presence of Staphylococcus aureus were recorded by active and passive sampling using tryptone soya agar (TSA) and maltose salt agar (MSA), respectively, with plates incubated at a mean of 36 [standard deviation (SD) 1] C for 48 h. The Health Protection Agency’s standard method was used for the identification of S. aureus.20 The presence of fungi was eval- uated by both active and passive sampling using plates con- taining Sabouraud chloramphenicol dextrose agar (SDA). The plates were incubated for seven days at a mean of 27 (SD 1)C, and filamentous fungi were identified in accordance with standard methods on the basis of their macroscopic and microscopic morphological features.21 Active sampling was performed using a surface air system (SAS) (International PBI, Milan, Italy), with a flow rate of 180 L/ min and a suction volume of 500 L. The sampler was placed in the operating theatre approximately 1 m above the floor and approximately 1 m from the operating table. The number of cfus was adjusted using the conversion table provided by the manufacturer, and the value was expressed in cfu/m3. Passive sampling was performed to determine the index of microbial air contamination (IMA).22 This corresponds to the number of cfus counted on a Petri dish with a diameter of 9 cm placed according to the 1/1/1 scheme (i.e. for 1 h, 1 m above the floor, 1 m away from walls or any major obstacles), and the results were expressed in cfu/m2/h. Sampling near the wound was performed using three nitro- cellulose membranes with a diameter of 47 mm (Millipore, Billerica, MA, USA), placed horizontally near the incision. Adhesive incise drape film (Steri-DrapeTM; 3M Health Care, St Paul, MN, USA) was applied to the skin before incision. This provides a sterile operative surface all the way to the wound edge, and a barrier to bacteria on the skin for the duration of the operation. The nitrocellulose membranes were positioned on the lower adhesive border of the film; this was folded forwards so the membranes were held in place below the wound and not beside it. In this position, they did not interfere with the surgical procedure. Figure 1 illustrates the experi- mental set-up. The membranes were exposed for 1 h from the start of the operation. Each membrane was subsequently transferred to a specific nutrient pad (i.e. one to TSA, one to MSA and one to SDA), and incubated as described previously for IMA and SAS procedures.18,19 All laboratory tests were performed at the Operative Unit ‘Hygiene’ Laboratory (quality certified accord- ing to ISO 9001:2008) at the University Hospital ‘Policlinico Consorziale’, Bari, Italy. In order to evaluate a possible corre- lation between the number of people in the operating theatre and the level of microbial contamination, the number of people present was recorded for each surgical intervention.
The results from the three sampling methods were loaded into a database created with File Maker software (Santa Clara, CA, USA), and data elaboration was performed using Epi-Info Version 6.00 (Centers for Disease Control and Prevention, Atlanta, GA, USA). Student’s t-test for independent samples was used to compare continuous variables. Categorical vari- ables were expressed as percentages, and c2 test was used to compare percentages. Pearson’s correlation was used to measure the relationships between the results obtained using different sampling methods. Linear regression was used to analyse the relationship between the number of people present in the operating theatre and the bacterial load for each method. A P-value
Results The mean bacterial load for the IMA samplings was 2232.9 (SD 859.7) cfu/m2/h (range 786e4246), which was significantly lower than the results obtained with the nitrocellulose membranes [mean 2768.2 (SD 1325.4) cfu/m2/h, range 1153e6344; t 1⁄4 2.62, P 1⁄4 0.0049]. A mean value of 123.2 (SD 58.7) cfu/m3 (range 40e288) was measured with the SAS sampler. A significant correlation was found between the IMA values and the SAS values (r2 1⁄4 0.73, P < 0.0001, Figure 2), and both of these were significantly correlated with the results from the nitrocellulose membranes (IMA: r2 1⁄4 0.73, P < 0.001, Figure 3; SAS: r2 1⁄4 0.67, P < 0.05, Figure 4). S. aureus was detected in 12 of 60 operations with the membranes [20%; 95% confidence interval (CI) 6.7e33.3], five operations with the SAS method (8.3%; 95% CI 0.87e17.54) and three operations with the IMA method (5%; 95% CI 2.26e12.26). The number of samples that tested positive for S. aureus was significantly higher for the membranes compared with the other two methods (c2 1⁄4 7.53, P 1⁄4 0.02). Using the membrane results as the reference, the IMA method to detect the presence of S. aureus had sensitivity of 25% (95% CI 1e100%), specificity of 100%, and positive and negative predictive values of 100% and 84%, respectively (95% CI 75e94). The SAS method had sensitivity of 42% (95% CI 14e70), specificity of 100%, and positive and negative predic- tive values of 100% and 87%, respectively (95% CI 78e96).

The mean number of people in the operating theatre during the 60 operations was 6.4 (SD 1.44, range 4e10). Linear regression revealed a relationship between the number of people present and the total viable count with all three methods: IMA (r2 1⁄4 0.23, P < 0.0001), SAS (r2 1⁄4 0.28; P < 0.001) and nitrocellulose membranes (r2 1⁄4 0.12, P < 0.001). For the IMA method, the mean number of people in the operating theatre did not differ significantly between operations when S. aureus was detected [7.7 (SD 1.5)] compared with operations when S. aureus was not detected [6.3 (SD 1.4); t 1⁄4 1.62, P>0.05]. However, the difference was significant for the SAS method, with a mean of 7.8 (SD 1.3) people in the operating thatre when S. aureus was detected compared with 6.2 (SD 1⁄4 1.4) people when S. aureus was not detected (t 1⁄4 2.03, P 1⁄4 0.04). Similarly, the mean number of people in the oper- ating theatre when S. aureus was detected on the nitrocellu- lose membranes was 7.3 (SD 1.3), compared with 6.1 (SD 1.4) people when S. aureus was not detected (t 1⁄4 2.48, P 1⁄4 0.01). Fungi were detected during two separate surgical opera- tions: Aspergillus spp. was detected with the IMA method, and 
Alternaria spp. was detected with the nitrocellulose membranes.

Discussion Infections can be debilitating complications of primary hip arthroplasty, and have been cited as the most common cause of implantation failure.23 Many contributing factors have been implicated such as patient factors, surgical technique and postoperative factors,24 but the environment, progress of the operation, clothes worn, etc. are also important. In this context, the microbiological quality of the air in operating theatres is a significant parameter to control surgical infection, and sampling may be undertaken to assess both the quality of air and the ability of the air systems to dilute or exclude contamination.25 The present study compared active and passive collection methods, and the results showed correlation between the two. Both methods can be used to measure air contamination, but if air sampling is performed during surgery to monitor the risk of microbial wound contamination, passive methods should be used to predict the likely contamination rate at the surgical site rather than volumetric sampling, as this allows direct measurement of the number of micro-organisms settling on surfaces.8,26,27 Active measurement also provides results on the concentration of particles in the air.8 In addition to traditional sampling methods, this study also used nitrocellulose membranes in order to evaluate microbial contamination near the wound. The use of nitrocellulose membranes is a passive technique, but comparison of the microbial sedimentation on the settle plates (IMA), which were near the operating table but outside the active operating area, and the microbial sedimentation on the membranes, which were actually placed on the operating table, revealed higher micro- bial loads on the membranes, confirming that the surgical team is a source of contamination. Contamination levels of the air in operating theatres are normally measured near the table but not near the wound, thus underestimating bacterial numbers. The membrane method was also useful for the detection of S. aureus, which, together with streptococci, causes 65e85% of all prosthetic joint infections.28e30 The membranes detected S. aureus during 12 of 60 operations, compared with five operations for the SAS method and three operations for the IMA method. The membranes revealed that S. aureus was more likely when more people were present, which was not the case with the IMA method. This study only looked specifically for S. aureus, but other staphylococcal species are considered to be opportunistic pathogens, and are increasingly recognized as causes of clini- cally significant nosocomial infections.31 The efficiency of nitrocellulose membrane sampling may be influenced by the electrostatic potential of these membranes, as they have been found to have a general negative electric charge at neutral pH.32 However, the authors believe that the differences in air contamination measured by the IMA and SAS methods compared with the membrane method were more a result of their positioning than the efficiency of the method used. When using passive sampling methods, Friberg et al. re- ported higher counts on settle plates near the wound than in the periphery of the operating theatre.17 A possible explanation for this is that heavier bacteria-carrying particles (i.e. those more likely to settle) tend to settle close to their main origin (i.e. the
surgical team).17 Moreover, from a practical point of view, to assess microbial contamination close to the wound, membranes are much easier to position than settle plates. This study shows that it is important to note that the limits of sampling systems are not exclusively influenced by their sensitivity and specificity (active or passive), but also by external factors such as the point of collection. Standardized sampling protocols should be developed for use within a hospital, for different clinical settings and for comparisons between different institutions and countries. The two passive systems detected two different strains of fungi, but the SAS system did not detect any fungi. This is in contrast to a study by Verhoeff et al., which found that active sampling was better for the detection of fungi.33 Furthermore, in a recent study by Asefa et al. comparing the SAS super 180 air sampler and settle plates, the SAS sampler detected more fungal species and a higher mean cfu/plate compared with the settle plates. Qualitatively, both methods detected similar predominant fungal genera and species.34 In conclusion, this study found that an air sampling system near the surgical wound may be the best way to identify the risk of exposure to pathogens, and to develop infection measures for operations with a high risk of contamination (e.g. arthroplasty). Cooperation of the surgical team is necessary, and during this study, various team members were unhappy with the presence of the sampling equipment, particularly the membranes, which had to be positioned in the sterile operating field before the first incision. Therefore, personnel need to be trained in sampling techniques, but it is also necessary to make staff aware of the importance of environmental surveillance programmes in order to plan corrective action in the case of any anomalies, thus allowing the avoidance of infective complications.

Article history: Received 6 May 2011 Accepted 11 October 2011 by J.A. Child
Available online 3 December 2011
Article in The Journal of hospital infection · December 2011 DOI: 10.1016/j.jhin.2011.10.011 · Source: PubMed
2011 The Healthcare Infection Society. Published by Elsevier Ltd. All rights reserved.