Is The Disinfection Procedure of The Intensive Care Unit Able To Reduce the Bacterial Resistance and Risk of Contamination?

Background: Healthcare-associated infections (HAIs) represent one of the main causes for the morbidity and mortality in the intensive care unit (ICU). In this study, we aimed to verify the presence, type, and antimicrobial susceptibility of bacteria in ICU settings, before and after disinfection procedures, and analyze the risk of contamination related to these bacteria in each area of the ICU. Methods: The study was conducted in the ICU of a medium-sized hospital in Brazil from February 2019 to February 2020. Samples were obtained from the surfaces of beds, bathrooms, pantries, sinks, pharmacies, administrative areas, and oors, before and after the disinfection process, with 1% benzalkonium chloride and biguanide mixture (BCB). The samples were processed and inoculated in different culture media for the selective isolation of strains of clinical interest. In addition, phenotypic identication and antimicrobial susceptibility tests were performed using the VITEK ® 2 system. We grouped different environmental, temporal, and microbial factors and characteristics to calculate the risk of contamination before disinfection (R bd ) and after disinfection (R ad ), as well as the total risk (R tt ) and real risk observed (Fold risk ) in the ICU. Results: Gram-positive cocci and rods, gram-negative rods of the Enterobacteriaceae family, and non-fermenting gram-negative rods were recovered after disinfection and were found to be widely distributed. Higher bacterial diversity and frequency of resistance were observed, mainly the resistance of gram-positive bacteria to cephalosporin (p < 0.05) and lincosamides (p < 0.0001), and gram-negative bacteria to quinolones (p < 0.0001) and aminoglycosides (p < 0.05). The pantry, pharmacy, and beds showed major risks of contamination after disinfection (R ad ), ranging from intermediate to very high-risk levels. Fold risk for the pantry and beds presented a signicant increase in two and three risk levels, respectively, after the disinfection process. Conclusions: Our results


Background
Healthcare-associated infections (HAIs) are acquired during care or hospitalization procedures and represent one of the most critical public health problems worldwide, as they are associated with increased morbidity, mortality, antimicrobial resistance, length of hospital stay, and nancial costs [1,2].
The World Health Organization (WHO) has recently listed the ESKAPE group (Enterococcus faecium, Staphylococcus aureus, Klebsiella pneumoniae, Acinetobacter baumannii, Pseudomonas aeruginosa, and Enterobacter species) in the list of bacteria against which new antibiotics are urgently needed [3,4].
Antimicrobial resistance is a signi cant threat to public health [5] and has the potential to cause 10 million deaths annually up to 2050 if nothing is done to mitigate its continued advance worldwide [6,7].
The misuse as well as excessive and uncontrolled use of antibiotics has played a vital role in selecting the antibiotic-resistant bacteria [8][9][10]. In intensive care units (ICUs), patients in critical and debilitating conditions are at an increased risk of being affected by HAIs [2,11]. The spread of HAIs is complex and multifactorial; therefore, strategies that aim to reduce the contamination and spread of resistant microorganisms can impact the occurrence of HAIs [12,13]. The existence of microbial resistance pro les, especially methicillin and vancomycin resistance, expression of extended-spectrum beta-lactamases and carbapenemases, and increasing resistance to macrolide agents, have become a threat in clinical practice in the ICU [14].
In this context, inanimate surfaces and devices present in hospitals serve as shelters for the most diverse microorganisms. When there is any failure in the disinfection of the environment, they can quickly spread throughout the health team and among patients, serving as a source of contamination and contributing to the cases of HAIs [15][16][17]. Therefore, strict cleaning protocols are used as the primary procedure to reduce contamination risks [1].
Therefore, to contribute data that can serve as guides for the development of HAI control strategies, the objective of this study was to analyze the microbial pro le of the ICU in a hospital in Ilhéus (Bahia, Brazil) before and after performing the disinfection procedure on the oors and surfaces, to analyze the impact of the standard cleaning procedure established in the hospital, with particular attention to microorganisms associated with nosocomial infections and the phenotypic pro les of resistance to clinically used antimicrobials, and to analyze the risk that these bacteria represent to the patients and staff in the ICU.

Study setting and design
The study was carried out in an ICU of a medium-sized hospital in Ilhéus, Bahia, Brazil, from February 2019 to February 2020. The evaluated ICU houses critically ill patients of all medical specialties. This ICU accommodates 20 electric beds and four sections sharing the same geographical, administrative, and other facilities (patients' bathrooms, pantries, etc.). At the time of this investigation, visitors were allowed from 2 pm to 4 pm, and no limit to visitors was stipulated.
A 1% mixture of benzalkonium chloride and biguanide (BCB) was used as a disinfectant. The commercial product (BCB) is composed of 5.2% alkyl dimethyl benzyl ammonium chloride (benzalkonium chloride), 3.5% polyhexamethylene biguanide (PHMB), non-ionic surfactant, solvent, and water. The disinfection process was performed at least three times a day by diluting BCB at 1% at the time of use. Samples were collected in the morning at the beginning of each 24 h shift, before the cleaning and disinfection of oors and surfaces and immediately after one of the sanitizing procedures. As part of the procedure, the product was left in contact with the surfaces for at least 10 min after application and to obtain the sample when there was no more humidity over the area. During sampling, all ICU employees and facilities were fully operational. Collections were performed twice per week.

Sample collection
A total of 35 paired samples (total 70) were collected from different areas and surfaces of the ICU, including occupied beds (n = 12), administrative spaces, bathrooms, pantries, sinks, and pharmacies.
The oor samples were collected with sterile SCOTT DURAMAX sheet (Kimberly-Clark, USA), moistened with distilled water, and conditioned in sterile Zip Lock plastic bags (Talge, Brazil). In the other places (heart monitor, bed, table, sink, tile, refrigerator, microwave, bench, and glass), collections were made with a sterile swab (Absorb, China) containing Stuart medium through rm movements to ensure complete contact with the area. The collection surface covered an area of 400 cm².
All collected material was identi ed and kept at 4-8 °C until the beginning of processing, which took place within a maximum period of one hour.
Sample processing of samples collected in the hospital in Ilhéus-BA.
In the Laboratory of Microbiology at the State University of Santa Cruz, each sample was individually immersed in tubes containing the brain heart infusion (BHI) broth (HiMedia, India) and subjected to serial dilutions. After a 24 h incubation period at 37 °C in aerobiosis, samples were subcultured in the selective media: eosin methylene blue (EMB) agar (Merck, Germany), nutrient agar (KASVI, Brazil), and blood agar (MicroMed, Brazil). After an incubation period of 24 h at 37 °C, the isolated strains were subjected to standard tests for biochemical-physiological identi cation and evaluation of their morphotintorial aspect using Gram stain. Bacteria were stored at -20 °C in Luriam Bertani Miller broth (Kasvi, Brazil) with 20% glycerol.
For the samples collected using a sterile humidi ed Scott Duramax sheet, 1 mL of water was squeezed from the sheet and transferred to 9 mL for serial dilution. All procedures were performed under aseptic and level 2 biosafety conditions. After the presumptive identi cation of bacterial strains via traditional cultivation techniques and biochemical tests, con rmation of phenotypic identi cation and automated antimicrobial susceptibility determination was performed using the VITEK ® 2 system (BioMéreiux, Brazil) with GP ID (gram-positive cocci), ANC ID (Corynebacterium and anaerobic bacteria), GN ID (gram-negative rods), and AST (antimicrobial susceptibility determination) cards. The interpretation of results followed the guidelines proposed by the Clinical and Laboratory Standards Institute (CLSI) and the Brazilian Committee on Antimicrobial Susceptibility Testing (BrCAST) [18], according to the Brazilian Secretariat of Health Surveillance, resolution No. 64, on December 11, 2018 [19] .

Risk analysis
To understand the risk that the ICU microbial diversity represents to the patients and staff before and after the disinfection process, we grouped different environmental, temporal, and microbial factors and characteristics, according to the water risk analysis protocol by [20], and adapted the principle of this analysis to carry out the risk analysis of the ICU environment in this study. The following aspects were considered in the risk analysis: (i) the microbial diversity found; (ii) the number of isolates; (iii) the temporal variation in microbial diversity and quantity, namely the microbial pro le before and after the ICU disinfection process; (iv) the susceptibility pro le, if sensitive and intermediate (common) or resistant (resistant) isolates based on the minimum inhibitory concentration (MIC) values; (v) the location of the bacterial isolates in the ICU environment, including the locations and surfaces from which the samples were collected; and (vi) the classi cation of these data according to risk categories: level I = very low; level II = low; level III = intermediate; level IV = high; and level V = very high. The elements above were used to calculate the risk of contamination before and after disinfection. Then, "risk analysis," or simply "R," allowed us to predict temporal risks: before (R bd ) and after (R ad ) the disinfection process and total risk (R tt ), the latter being the sum of the risks before and after disinfection.
Given the differences and clinical signi cance of the isolates, different multiplier values were established to calculate the risk for common and resistant isolates, and for bacteria isolated before and after the disinfection process, so that the resistant isolates and isolates after disinfection have a higher calculation weight, as described in the formulae below: where, I c1 and I r1 represent the number of common and resistant isolates before disinfection, respectively; R a and R d are risk multiplier for common bacteria and resistant bacteria, whose values correspond to 0.1 and 1, respectively; B c represents the risk multiplier value before disinfection, whose value is 1.
where, I c2 and I r2 represent the number of common and resistant isolates after disinfection, respectively; R a and R d are risk multiplier for common bacteria and resistant bacteria, whose values correspond to 0.1 and 1, respectively; B d represents the risk multiplier value after disinfection, whose value is 2.
Formula 3: Total risk (R tt ) R tt = R bd + R ad where, R bd represents the risk before disinfection, and R ad the risk after disinfection.
Subsequently, ve risk levels were established based on the possible temporal risk values (R) before disinfection, after disinfection, and total risk in the studied ICU environment: very low risk = R < 1.0, low risk = R ≥ 1.0 ≤ 2.5, intermediate risk = R > 2.5 ≤ 5.0, high-risk = R > 5.0 ≤ 9, and very high-risk = R > 9.0. To obtain the real risk observed after ICU disinfection, we calculated the ratio between the risk after disinfection (R ad ) and the risk before disinfection (R bd ). . Finally, to obtain an adjusted parameter, the Fold risk value obtained for each ICU environment was adjusted to the previously established risk categories (very low, low, intermediate, high, and very high) to analyze whether the increase (or decrease) in the real risk observed (Fold risk ) resulted in a change between the different risk categories established. For example, if before disinfection, the pantry oor had a risk level of R bd = 0.8 (very low), and after disinfection, a risk level of R ad = 2.0 (low), then this environment will have a Fold risk = 2.5, that is, the ICU pantry oor after disinfection will have a 2.5x greater risk than before disinfection. At the same time, it will no longer belong to the category of very low risk and move to the low-risk category. Thus, when a change in the risk category was observed, as in the example mentioned above, the risk was considered signi cant. The Fold risk was considered signi cant: if the risk level decreased one level ( -); if decreased 2 levels ( --); if decreased 3 levels ( ---); if decreased 4 levels ( ----); if increased 1 risk level ( + ); if increased 2 levels ( ++ ); if increased 3 levels ( +++ ); and if increased 4 levels ( ++++ ). When there was no change in the risk level, the risk was considered non-signi cant ( # ). Thus, in the example above, we would have a Fold risk = 2.5, with a signi cance level of ( + ) for the pantry oor after disinfection.

Statistical analysis
Statistical analyses were performed using GraphPad Prism Software version 7.04 (San Diego, USA). The Kolmogorov-Smirnov test was used to analyze the distribution of normality before choosing the appropriate statistical test. The frequency distribution is described as a percentage. The analysis between two groups with non-parametric distribution was performed using the Mann-Whitney test. Data are represented as the mean ± standard error of the mean (SEM) or mean ± min to max values, and values of p < 0.05 were considered to be statistically signi cant.

Results
Disinfection process increases the microbial diversity in the ICU Gram-positive cocci and rods, gram-negative rods of the Enterobacteriaceae family, and non-fermenting gram-negative rods were recovered after disinfection and were widely distributed. From the 70 samples collected, 119 isolates of bacteria were recovered, with at least one bacterium per sample. Beds (n = 44) and pantries (n = 17) were the areas with the highest numbers of bacteria (Fig. 2). For pantry, most of them were isolated before disinfection (n = 12 before; n = 5 after), and for bed, most of them were isolated after disinfection (n = 15 before; n = 29 after). It is noteworthy that from all the gram-positive rods, only Corynebacterium jeikei could be identi ed at the species level using the adopted methodology.
Considering all bacteria recovered before and after the sanitization process, gram-positive bacteria were the ones that prevailed (n = 65). Between them, especially before the sanitization procedure, grampositive rods were the most frequently (54.62%) observed. Before disinfection, they were isolated from all bed and surfaces ( oor, accessory T-serving table, and cardiac monitor). After the disinfection procedure, they were isolated only from the accessory T-serving table and cardiac monitor (Fig. 2). With the exception of the sink, K. pneumoniae was isolated from all areas.
After the disinfection process, all surfaces tested showed higher microbial diversity. This also included a higher frequency of phenotypic resistance among the microbes (Fig. 2).
Bacteria isolated after the disinfection process present a resistant pro le to antimicrobial drugs To understand the susceptibility pro le, we analyzed the MIC values of different classes of antimicrobial drugs used against the bacteria isolated from nosocomial environments and analyzed the individual drugs. Regarding the antimicrobial classes in gram-positive bacteria, a resistance pro le to cephalosporin (p < 0.05) and lincosamides (p < 0.0001) prevailed after disinfection (Fig. 3A), and a susceptibility pro le against lipopeptide (p < 0.05) and oxazolidinones (p < 0.01; Fig. 3A) was observed. Gram-negative bacteria showed antimicrobial resistance and signi cant variation in MIC values after disinfection with quinolones (p < 0.0001) and aminoglycosides (p < 0.05), when compared to isolates after disinfection (Fig. 3B), and a susceptible pro le for polymyxins (p < 0.01) after disinfection (Fig. 3B).
Disinfection process increases the risk of contact with pathogenic bacteria in the ICU Risk analyses showed a higher risk of contact/exposure (and possible contamination) with bacterial isolates from the ICU settings after the disinfection process, which was associated with an increased number of resistant strains in the environment. When the risk was analyzed for each location and surface, the administrative area, bathroom, and ICU sink presented very low or low contamination risks before (R bd ) and after (R ad ) disinfection, while the pantry, pharmacy, and beds showed major risks of contamination after (R ad ) disinfection process, being intermediate to very high-risk levels (Fig. 5A). In general, 85.7% of the areas and surfaces had a smaller R bd (0.2-0.5) with a very low risk level (Fig. 5B); however, these percentages decreased to 35.7% for very low risk level, and 21.4% and 14.3% of ICU areas present an intermediate or very high-risk level to R ad (2.8-10.8; Fig. 5C). Evaluating the risk of contamination independent of the disinfection situation, it was shown that approximately 42.8% of the total risk (R tt ) to which patients and professionals are exposed in the ICU environment, represent intermediate, high, and very high-risk levels with R tt values varying from 3.1-11.1 ( Fig. 5D; Table 1).
We considered making the risk analysis more robust and for that we assessed the real risk observed (named Fold risk ), the ratio between the risk after disinfection and that before disinfection and considered the Fold risk signi cant only if a change in the risk level was noted (see Methods, session Risk Analysis).
The values of the risk analysis and their signi cance are listed in Table 1. The Fold risk was signi cant for different ICU locations (Fig. 6). Thus, it was possible to observe that for different places evaluated, there was an increase in the real risk of contamination by bacterial isolates after the usual disinfection procedure in the ICU, especially for the pantry and beds with an increase in the two risk levels (Fold risk = 3.7 ++ ) and three risk levels (Fold risk varying from 5.6 +++ to 36.0 ++++ , with a median of 17.9 +++ ) ( Table 1; Fig. 6). These data suggest that the disinfection process with BCB was not e cient in decreasing the contamination risk in the ICU studied. The only place in the ICU area where the disinfection process was very e cient and did not show a change level was the sink surface ( # ).

Discussion
Environmental contamination associated with nosocomial infection risk has been increasingly observed [21,22]. It is well reported that most nosocomial pathogens may remain on inanimate surfaces for months. In this context, the emergence of resistance to disinfectants is a fact, as well as cross-resistance to antimicrobials [23][24][25][26].
Here, we report the persistence of contamination with multidrug-resistant bacteria on surfaces after disinfection. Although many aspects can be involved in the persistence of bacteria in this environment, the use of 1% BCB must be considered. Most reports available in the literature show the effect of these substances alone and in other environments, such as food industries, cosmetics, and water treatment, among others [27][28][29][30]. The results suggest that disinfection of surfaces with high bacterial contamination may lose its effectiveness. In addition, a study performed in an ICU with the same disinfectant suggested bacterial adaptation to the environment [1].
It is essential to note the presence of multidrug-resistant pro les of some bacteria recovered from the ICU surfaces, such as ESKAPE-type bacteria, such as Enterobacter cloacae and K. pneumoniae [31]. Their presence represents a signi cant challenge for nosocomial infections in hospitals. While the ESKAPE phenotype, for example, can survive for long periods in the environment and characterizes a continuous source of contamination, thereby making effective therapeutic strategies di cult [24,[32][33][34][35], the others show resistance related to the most common antimicrobials used for critical infection treatment, such as nitrofurantoin (NIT) and trimethoprim/sulfamethoxazole (SUT), ampicillin (AMP), and ampicillin/sulbactam (APS) [14,[36][37][38].
New investigations must be conducted to identify the cause of the resistant bacteria recovering from this ICU investigated here; however, some facts already reported in the literature may also be involved. The prescription and excessive use of antimicrobials associated with the adaptive resistance mechanism of bacteria, several broad-spectrum agents are continually used at the same time, and the lack of antimicrobial interchange or cyclic usage to treat nosocomial infections are some of them [39,40].
We were able to show that the microbial diversity within the ICU changes and may present different pro les over time before and after disinfection. Microbiological analyses allow us to assess the microbial diversity in the ICU environment to which patients and hospital staff are potentially exposed, before and after the disinfection process; however, it is important to understand the possible impact of this diversity, which is the risk that microbial diversity found in the ICU is. Therefore, bacterial characteristics, whether the isolates are composed of pathogenic bacteria, their susceptibility pro le, and identi cation of resistant isolates, among others, are crucial [41].
We proposed a risk analysis (R) that grouped different environmental, temporal, and microbial factors and characteristics. Based on these analyses, the data showed that the risk before the disinfection process (R bd ) was generally lower than the risk after disinfection (R ad ), and that accumulating risk or total risk (R tt ) varied between very low and very high in the studied ICU. The real risk observed (Fold risk ) was incredibly high to the pantry oor (Fold risk = 3.7) and to all bed surfaces (Fold risk varying from 5.6 to 36.0 with media of 17.9) of ICU after disinfection process, indicating that these areas represent a risk to the health of patients and/or staff professionals, and that disinfection processes need to be reviewed and improved to minimize the risk of surviving pathogenic bacteria persisting and may contact humans. An important limitation is that our risk analysis (R), however, did not evaluate the presence of other microorganisms that may be present in the ICU, such as fungi.
In the ICU, the risk of contamination showed that the disinfection process used did not eliminate the presence of common and multidrug-resistant bacteria in the environment. On the contrary, our evaluation showed an increase in bacterial contamination, with a high presence of multidrug-resistant bacteria after disinfection. In particular, beds, pharmacies, and pantry areas have a risk of contamination. It is important to note that although the presence of multidrug-resistant bacteria could be detected after the disinfection process, this does not mean that they appear as a direct result of the disinfection process used, because we cannot rule out the microbial exchange that takes place between health and cleaning professionals and the ICU environment [41,42,46]. Recontamination after the cleaning protocol, frequent manipulation of the devices because the team was active during sample collection, or the resistance of microorganisms to environmental changes and stress [1,[43][44][45] should be considered as they may increase the risk of bacterial contamination in the ICU.
In this study, there is evidence that the disinfection process used may lead to bacterial resistance selection, since a more signi cant number of resistant bacteria were recovered after the disinfection process. In addition, exposure of bacteria to sublethal concentrations of biocidal agents can activate adaptive stress response mechanisms, such as e ux pumps, favoring survival in hospitable environmental conditions and the emergence of bacteria that are resistant to the main antimicrobial agents used in the therapy of infectious diseases [46,47].

Conclusions
In conclusion, our data demonstrate the persistence of bacterial pathogens of clinical interest with multidrug-resistant pro les in different areas of the studied ICU, even after disinfection with BCB. In addition, the pantry oor and beds were found to be the major risk areas for ICUs. Therefore, further studies are necessary to decipher the roles of these biocides used in the hospital in the development of the wide pro les of bacteria recovered from the respective surfaces as well as the possibility of these bacteria forming bio lms. These data will help to control the nosocomial infection levels in the ICUs of hospitals.

Declarations
Ethics approval and consent to participate Not applicable.

Consent for publication
Not applicable.

Availability of data and materials
All data generated or analyzed during this study are included in this published article.

Competing interests
The authors declare that they have no competing interests. R bd -risk before disinfection; R ad -risk after disinfection; R tt -total risk; Fold risk -real risk observed; Ssigni cance; + increase in 1 risk level; ++ increase in 2 risk levels; +++ increase in 3 risk levels; ++++ increase in 4 risk levels; # no change in risk level; level I = very low; level II = low; level III = intermediate; level IV = high; and level V = very high. Figure 1 Intensive care unit (ICU) area and collection sites scheme. The ICU settings and surfaces identi ed included the administrative area, bathroom, pantry, pharmacy, beds, and sink. Frequency distribution of the bacterial strains isolated from the ICU setting. Data are expressed as the percentage of the strain isolated before (left) and after (right) the disinfection process. The ICU settings and surfaces were identi ed by colors: yellow -administrative area; orange -bathroom; green -pantry; blue -pharmacy; purple -beds; and red -sink.  Heatmap of the antimicrobial susceptibility of bacteria, before and after the disinfection process in the ICU setting. The plotted MIC values represent the mean of each strain, before and after the disinfection of the ICU environment. The strains are grouped according to the location of isolation in the ICU: administrative area (yellow), bathroom (orange), pantry (green), pharmacy (blue), beds (purple), and sink