Risk factors for colonisation and infection with multidrug-resistant Pseudomonas aeruginosa in intensive care unit: a systematic review and meta-analysis

Abstract

Background

There are currently two opposing hypotheses regarding the risk factors for acquisition, colonisation and infection with multidrug-resistant Pseudomonas aeruginosa (MDR-PA) in the intensive care unit (ICU). Acquisition could be either endogenous or exogenous or both. It is of great interest to estimate pooled prevalence and describe individual and environmental factors associated with the colonisation and infection with MDR Pseudomonas aeruginosa in intensive care units.

Methods

This systematic review was performed following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. The protocol of this review has been registered under CRD42021233832 in the International Prospective Register of Systematic Reviews (PROSPERO) and has been published on Systematic review on November 2022All types of studies carried out in intensive care units (ICUs) were included. MEDLINE (Pubmed), EMBASE (OVID), the Cochrane Library (Wiley), Web of Science, CINAHL (EBSCOHost), LILACS (BIREME), Google Scholar and Open Grey were searched from 1983 to 2023 and the results of electronic searches were uploaded to Rayyan software. The methodological quality of the studies was assessed using the National Heart, Lung, and Blood Institute Critical Appraisal Tools. The I² test was performed to assess the statistical heterogeneity among the included studies. The publication bias was assessed by using the funnel plot and Egger’s test. Descriptive analysis and meta-analysis were performed. Fixed effect model were used to calculate the surveyed prevalence and odds ratio (OR) with their respective 95% confidence intervals (95% CI).

Results

A total of 10,791 articles were identified, of which 13 were retained for descriptive analysis and 8 for meta-analysis. The majority of these 13 studies were conducted in Western countries. Methods were heterogeneous and few studies addressed environmental factors. The pooled prevalence of MDR-PA in the ICU was 4% (95% CI: 0%—11%). Identified risk factors were: length of stay in ICU (> 8 days), mechanical ventilation with OR: 3.19; 95% CI: 2.25—4.53, use of invasive devices with OR: 2.97; 95% CI: 2.40—3.68 (use of central venous catheter with OR: 3.16 95% CI: 1.87–5.33, the use of urinary catheters with OR: 2.65; 95% CI: 2.05–3.44, the use of parenteral nutrition with OR: 2.43; 95% CI: 1.15–5.16 and the use of arterial catheters with OR: 7.00; 95% CI: 2.77–17.68) and the use of antibiotics with OR: 3.69; 95% CI: 3.16–4.27 (carbapenem with OR: 4.12; 95% CI: 3.29–5.16, quinolones with OR: 3.31; 95% CI: 2.45–4.47, bectalactam with OR: 3.58; 95% CI: 2.4–5.24, and aminoglycosides with OR: 3.32; 95% CI: 2.33–4.73) and environmental factors.

Conclusion

All this suggest that acquisition or infection by MDR-PA in ICU could be due to endogenous and exogenous transmission. Due to the few numbers of studies analysed, further investigation with more studies is needed to draw definitive conclusions.

Systematic review registration

The protocol of this review has been registered under CRD42021233832 in PROSPERO and has been published on Systematic review on November 2022. Eyebe et al. (Syst Rev 11:270, 2022).

Background

Pseudomonas aeruginosa (PA) is a ubiquitous Gram-negative bacterium with minimal requirements for survival in the environment. It has a remarkable ability to colonise surfaces even in the most restricted environments.

P. aeruginosa can also cause infection in immunocompromised individuals [1]. It is one of the most common pathogens involved in healthcare-associated infections in intensive care units (ICU) [2]. The prevalence of hospital-acquired P. aeruginosa carriage was 15.3% in an adult French population hospitalised in ICUs [3]. P. aeruginosa resistance to various antibiotics is increasing, along with the global burden of antibiotics resistance [4]. Pseudomonas spp. was found as the most common pathogen (28%) in ICU infections. In the same study, Multidrug-resistance was observed in 80.1% of PA isolates [5]. Infection by MDR-PA is associated with increased morbidity and mortality rates [5–8]. An average length of ICU stay of 25 days and an associated mortality rate of 20% was found as associated with MDR-PA infection [5]. It is also associated with a high surgery incidence and an increased number of invasive procedures [5–7].

Infection with P. aeruginosa in the ICU is known to be either endogenous or exogenous or both, but the role of each of these contamination route remains to be clarified [9]. Endogenous transmission has been well described [1, 10]. It is the main source of contamination associated to the gut microbiota under antibiotic pressure [11]. Exogenous transmission comes from aqueous environments such as faucets, invasive medical devices, other patients and health workers through their hands [12, 13].

Most studies tend to favour the endogenous source, with the exogenous source more often implicated in outbreaks [2, 9, 12].

Furthermore, the source of P. aeruginosa in ICUs is not well understood, nor are the factors associated with colonisation and infection in ICUs [9]. This may be due to a lack of data on the prevalence, risk factors and environmental determinants of MDR-PA colonisation or infection. For example, in a systematic review conducted in 2018, Raman and colleagues found significant gaps and inconsistencies in the knowledge of risk factors associated with the occurrence of MDR P. aeruginosa in nosocomial infections. Raman's systematic review did not specifically address the issue of ICUs and was limited to endogenous risk factors. Raman's systematic review did not specifically address the issue of ICUs and was limited to endogenous risk factors [14].

Therefore, there is a growing interest in understanding both individual and environmental factors associated with multidrug-resistant P. aeruginosa in the ICU, where high morbidity and mortality rates have been reported [5, 6, 15].

The primary objective of this systematic review was to investigate individual (endogenous) and environmental (exogenous) factors associated with the colonisation and infection with MDR Pseudomonas aeruginosa in intensive care units. The secondary objectives of this study were to (i) estimate the prevalence of Multidrug resistant P. aeruginosa colonisation in ICU, and (ii) describe the main environmental sources of MDR-PA acquisition and infection in ICU.

Methods

This systematic review is reported according to the Preferred Reporting Items for Systematic Review and Meta-Analysis (PRISMA) [16]. The review was performed as indicated in the protocol published [17].

Data collection

Type of studies included

This review included cross-sectional studies, case–control studies (nested or not) and cohort studies (either prospective, retrospective or ambi-directional) carried out in intensive care units (ICUs). For meta-analysis only articles in which risk factors of MDR Pseudomonas aeruginosa acquisition/infection could be assessed were selected. Studies without data on MDR-PA were excluded. For multi-disease studies, when MDR-PA data was not available separately in isolation, these studies were excluded. This was also the case for studies conducted in several hospital wards where it was not possible to obtain data from patients in the ICU.

Type of participants from selected studies

Studies including all patients, admitted in an ICU, whatever the reason for admission and the origin (community, another hospital, or another unit of the same hospital) of patients, were taken into consideration. No specific definition of ICU was applied, and all studies which authors reported as having been conducted in an ICU were considered. We also included studies where it was possible to identified patients with specimens collected and tested for P. aeruginosa (positivity and antimicrobial susceptibility) during hospitalization and/or samples collected from the ICU patient’s environment (including staff hands and gloves) for P. aeruginosa testing. No specific restrictions on type of patients were applied.

Outcomes

Our study focused on people admitted to the intensive care unit in whom MDR-PA was found.

The primary outcome

(i) Factors associated with acquisition of MDR-PA. Since 2000, the most commonly used definition of MDR for Gram-positive and Gram-negative bacteria is 'resistance to three or more antimicrobial classes' [18, 19].

Secondary outcomes

(ii) Risk factors of carriage. Carriage was defined as positive screened specimens. These specimens could be collected from rectal swabs, nasal swabs or tracheal aspiration or others [20].

(iii) Risk factors of colonisation. Colonization was defined as positive clinical specimens, and infection as positive clinical specimens with clinical signs confirming infection.

The prevalence of MDR-PA was calculated as the number of people admitted to the medical intensive care unit who were found to have MDR-PA out of the total number of patients admitted to the intensive care unit during the same study.

Search strategy

Six databases were searched from inception onward, namely MEDLINE (Pubmed), EMBASE (OVID), the Cochrane Library (Wiley), Web of Science, CINAHL (EBSCOHost) and LILACS (BIREME). We also searched in grey literature using Google Scholar and Open Grey.

A search strategy using Medical Subject Headings (MeSH) and text words related to Pseudomonas aeruginosa, intensive care units, carriage, colonization or infection were used as described in Table 1 for MEDLINE (Pubmed); these were adjusted for the search in other databases cited above. For studies that were not found via internet search, the authors were directly contacted in order to obtain them. Furthermore, when a full text was not available online, the author was asked by email to provide it.

Table 1.

Search strategy

Search	Query
#1	Pseudomonas aerugnosa [tw] or Pseudomonas [tw] or P aeruginosa [tw] or Pseudomonadaceae [tw] or Gram-negative bacteria [tw] or gram negative [tw] or bacteria [tw] or Pseudomonas aerugnosa [mesh] or Pseudomonas [mesh] | 671 988 hits
#2	Intensive care unit [tw] or intensive care units [tw] or ICU [tw] or ICUs [tw] or intensive care [tw] or intensive care units [mesh] or respiratory care units [mesh] | 287 174 hits
#3	Carriage [tw] or colonization [tw] or colonisation [tw] or acquisition factors [tw] or infection [tw] or contamination [tw] or infected patients [tw] or asymptomatic Infections [mesh] or infections [mesh] | 4 009 872 hits
#4	Water [tw] or water tap [tw] or tap [tw] or bed [tw] or gloves [tw] or hands [tw] or environmental [tw] or environment [tw] or exogenous factors [tw] or Sink [tw] or care material [tw] or drains [tw] or surface water [mesh] or beds [mesh] or protective gloves [mesh] or environment [mesh] | 3 662 968 hits
#5	#3 or #4 | 364 087 hits
#6	#1 and #2 and #5 | 10 791 hits
#7	Date: from inception to December 31st, 2023
#8	No language restriction
#9	#6 and #7 and #8

Study selection

The search was conducted by an experienced information specialist and there were no language restrictions. For trials published in a language other than English, Deepl translate was used. The search was conducted until 31 December 2023. The results of the electronic searches were uploaded into Rayyan software [21] to identify and remove duplicates. After this stage, two reviewers independently screened the titles and abstracts of all the studies identified by the search strategy. First, they screened the studies according to the inclusion and exclusion criteria. The full text of these selected studies was then retrieved for a second analysis to decide whether the study should be included in the final analysis or not. Disagreements between the two reviewers were resolved by discussion and consensus. The PRISMA flowchart template was used to describe the number of articles retrieved and screened at each step [18, 21].

Data extraction

• Participants: number of participants, region(s) and country(ies) from which participants were recruited, study eligibility criteria, reasons for ICU admission, age, sex, length of hospital stay.

• Data on P. aeruginosa carriage, colonisation or infection: number of people who acquired MDR-PA, number of people who already had MDR-PA, antibiotic tested, number and type of environmental samples.

• Factors associated with resistant P. aeruginosa: all factors reported in the studies were recorded in "authors' own words" and with their corresponding effect estimates (odds ratios, relative risk or hazard ratios).

Data synthesis and management

Firstly, a descriptive analysis of the characteristics of the studies, the people included and the risk factors identified was performed.

Second, prevalence, odds ratio, relative risk (RR) with 95% confidence intervals were calculated from the collected data. For each study, we extracted the following information from the tables provided, the total number of individuals with multidrug-resistant Pseudomonas aeruginosa (MDR-PA), the total number of people without MDR-PA. In addition, within each of these two groups (MDR-PA positive and negative), we determined the number of exposed and unexposed individuals for the risk factor(s) of interest.

Thirdly, a meta-analysis of case–control and cohort studies with similar populations, same parameters and P. aeruginosa testing was performed. Factors influencing the estimates were evaluated.

Data were analysed using the meta package in R (R Studio). A fixed effect model was used to pool the different statistics. Specifically, the Mantel–Haenszel method was used for proportions and the inverse variance method for continuous data. We chose to use a fixed effect model because it was assumed that the included studies estimated a common true effect, reflecting a homogeneous population and consistent study designs. This approach was considered appropriate given the focus on specific, well-defined risk factors and outcomes. Heterogeneity between pooled studies was tested using the standard chi-squared test with the Q statistic (p < 0.10 statistically significant).

The degree of heterogeneity was quantified using the I2 statistic, with I2 > 50% considered to represent substantial inconsistency or significant statistical heterogeneity. Where statistical pooling is not possible, results are presented in a synthetic narrative form.

Quality of studies and evidence assessment

The methodological quality of the studies was assessed using the National Heart, Lung, and Blood Institute Critical Appraisal Tools for each appropriate study design (cross-sectional, case–control, cohort). Potential bias in the design, conduct and analysis of each study was assessed by two independent reviewers. Confidence in the evidence was discussed among the authors.

Publication bias was assessed graphically using a funnel plot. Futhermore,the statistical significance of publication bias was assessed using Egger's and Beggar's tests. A p-value of less than 0.05 was set as the threshold for determining the presence or absence of publication bias.

Results

Characteristics of selected studies

The strategy of search yielded 10,791 citations. After removing duplicates using RAYYAN [21], all titles and abstract were screened based on selection criteria and 148 items were selected. Full-text of these 148 studies were retrieved and evaluated, then 13 articles were extracted, see Fig. 1 [5, 19, 20, 22–31].

Fig. 1.

Flow Chart summarizing the selection process for 13 articles used for analysis and 8 for meta-anlysis

Of these 13 articles, there were 4 studies from the USA, [25, 27, 28, 30] 3 from France [22, 27, 29], a multi-centric study from 12 countries across the world (Europe, South America, Asia, North America), [22] one from Indonesia [32], Israel [6], Canada [26], Belgium [23] and Spain [33]. There were no studies from Africa.

Most of the studies were prospective. Two were cross-sectional [20, 28], Two other studies were retrospective cohorts [19, 24]. Six were prospective cohorts [22, 23, 25–27, 30, 31] and two were case-controls [5, 29] see Table 2.

Table 2.

Characteristics of studies selected

First author	Study title	Year	Number of centers	Study design	Country	Inclusion criteria in study	Number included	Already had MDR-PA	who acquired MDR-PA	Antibiotic tested	Type of environmental specimen
Aloush	MDR-PA-Resistant Pseudomonas aeruginosa: Risk Factors and Clinical Impact a	2006	1	Not nested case–control	Israël	none	184	13	69	ceftazidime, cefepime, aztreonam, ciprofloxacin, piperacillin, gentamicin,	none
Gudiol	Clinical Predictive Model of MDR-PA Resistance in Neutropenic Cancer Patients with Bloodstream Infection Due to Pseudomonas aeruginosa a	2020	34	Retrospective cohort	Spain (14 centers), Turkey(4), Brazil (3), Italy (3), Argentina (2), Germany (2), Chile (1), Colombia (1), Lebanon (1), Slovakia (1), Switzerland (1), and the United Kingdom (1)	1) adult neutropenic oncohematological patients, including hematopoietic stem cell transplant 2) 18 years and above	1217	309		fluoroquinolone, piperacillin-tazobactam, carbapenem, cephalosporin	none
Deplano	Molecular Characterization of an Epidemic Clone of Panantibiotic-Resistant Pseudomonas aeruginosa	2005	1	Cross-sectional	Belgium	non cystic patient, outbreak period	381	2	16	cefepime, meropenem, imipenem, gentamicin, tobramycin, and ciprofloxacin, piperacillin, piperacillin/tazobactam, ceftazidime, aztreonam, ticarcillin/clavulanic acid	tap water
Labaste	Risk factors for acquisition of carbapenem-resistance during treatment with carbapenem in the intensive care unit: a prospective study a	2019	1	Prospective cohort	France	1) having received at least 48 h of carbapenem therapy	78		14	Meropenem, imipenem	none
						2) ≥ 18 years
Harris	Risk factors for Development of Intestinal Colonization with Imipenem-resistant Pseudomonas aeruginosa in the Intensive Care Unit Setting a	2011	1	Prospective cohort	USA	None	3184		118	none	none
Hota	Outbreak of MDR-PA-Resistant Pseudomonas aeruginosa Colonization and Infection Secondary to Imperfect Intensive Care Unit Room Design	2009	1	Retrospective cohort	Canada	Admitted in ICU during outbreak	36		36		1) Sink taps and shower heads 2) Sink taps and shower head 3) medical equipment (22)
Johnson	The Role of Patient-to-Patient Transmission in the Acquisition of Imipenem-Resistant Pseudomonas aeruginosa Colonization in the Intensive Care Unit	2009	1	Prospective cohort	USA	None	7071	151	149	Imipemen, meropemen, meropemen, ceftazidime, cefepime, pipéracilin, pipéracilin tazobactam, ciprofloxacine, levofloxacine, amikacin, gentamicin	none
Levin	Pseudomonas in the sinks in an intensive care unit:relation to patients	1984	1	Prospective cohort	USA	All patients admitted in the ICU for more than 24 h	60	17	5	gentamicin, tobramycin, amikacin	sink
Nseir	Risk of acquiring MDR-PA-resistant Gram-negative bacilli from prior room occupants in the intensive care unit a	2011	1	Prospective cohort	France	Stay in ICU > 48 h	511	0	82	none	none
Obritsch	National Surveillance of Antimicrobial Resistance in Pseudomonas aeruginosa Isolates Obtained from Intensive Care Unit Patients from 1993 to 2002	2004	117	Cross-sectional	USA	none			nd	Imipenem, ceftazidime, cefepime, piperacilin, gentamycin, piperacilin, ciprofloxacin, piperacillin-tazobactam, aztreonam,amikacine, tobramycin, ticarcillin-clavulanate;	none
Paramythiotou	Acquisition of MDR-PA-Resistant Pseudomonas aeruginosa in Patients in Intensive Care Units: Role of Antibiotics with Antipseudomonal Activity a	2004	1	Not nested case–control	France	with/without MDR-PA	68	2	37	ceftazidime, imipenem, ciprofloxacin, piperacillin, ciprofloxacin,	none
Pena	Effects of Carbapenem Exposure on the Risk for Digestive Tract Carriage of Intensive Care Unit-Endemic Carbapenem-Resistant Pseudomonas aeruginosa Strains in Critically Ill Patientsa	2007	1	Prospective cohort	Spain	admitted to the unit for more than 48 h	266	12	30	Imipenem, ciprofloxacin, gentamicin, aztreonam, ceftazidime, tircacillin, piperacilin, tobramycin, meropénème	none
Saharman	Epidemiology and characterisation of carbapenem-non-susceptible Pseudomonas aeruginosa in a large intensive care unit in Jakarta, Indonesia a	2019	1	Prospective cohort	Indonesia	1) admitted for > 48 h 2) ≥ 18 years	412	17	34	Carbapenem	1) Wash basin on ICU ward 2) Swab of dish dryer in cleaning room 3) Wash basin 2 in cleaning room 4) Tap water from wash basin 1

^a Selected articles for meta-analysis

The studies were published between 1984 and 2023. Of the 13 articles extracted, risk factors for MDR Pseudomonas aeruginosa acquisition/infection were assessed in 8 studies selected for meta-analysis, see Fig. 1 [19, 22, 23, 27, 29–31, 34].

Environmental data found in only 4 studies [20, 24, 26, 31]. There were no standard definitions of MDR-PA. Many definitions of MDR-PA were used from 1983 to 2021. For example, in their studies, Aloush et al. defined MDR-PA as resistance to all the agents studied (ceftazidime, cefepime, aztreonam, ciprofloxacin, piperacillin and gentamicin) [6], while Kumar et al. defined it as bacteria "resistant to three or more antimicrobial classes" [18]. In addition, Jung et al. had defined it as a bacterium that is resistant to at least three of the following four drugs: ceftazidime, imipenem, ciprofloxacin and tobramycin [35]. Of the 13 extracted articles, risk factors for acquisition/colonisation of MDR Pseudomonas aeruginosa were assessed in 8 studies selected for meta-analysis (Table 2).

Descriptive analysis of main risk factors in MDR-PA acquisition and infection in the ICU

These risk factors could be divided into four groups. The first group were factors related to clinical status; the second group were factors related to medications received; the third group were factors related to invasive procedures or the use of some medical devices; and finally, environmental factors.

Factors related to clinical status

Previous ICU stay is a factor associated with acquisition and carriage of MDR-PA. In the selected articles, the clinical status of patients was assessed using many types of scores. These were either the APACHE II (Acute Physiology and Chronic Health Evaluation II) score, the SAPS II (Simplified Acute Physiology Score II), the TISS-28 (Therapeutic Intervention Scoring System), the QSOFA (Quick Sequential Organ Failure Assessment) or the Omega score. It appears that patients with a score reflecting a severe condition, such as bedridden patients [6] or those with a QSOFA > 2 [32], are at higher risk of MDR-PA infection. In addition, some specific conditions are associated with a high risk of MDR-PA infection. These included severe haematological disease [22], a history of infection or colonisation with Pseudomonas aeruginosa [24], and underlying cerebral disease [32].

Factors related to drugs received

The most cited and apparently most important factor is "antibiotic pressure" [5, 19, 20, 22–26]. The authors define it in three ways. First, it is related to the number of antibiotics received by a patient, as defined by Aloush et al. [6], or the number of days on antibiotics [24], or the percentage of days spent on antibiotics, or the number of days on antibiotics [29]. Second, the type of antibiotic is of primary importance for MDR-PA acquisition. The most commonly used antibiotics are carbapenems (imipenem), quinolones and fluoroquinolones, beta-lactams (cefepime, piperacillin-tazobactam, penicillin), aminoglycosides and metronidazole [5, 19, 20, 22–26]. Thirdly, some therapies such as sedation [29], vasopressor treatment and immunosuppressive therapy [6] have been associated with the risk of MDR-PA infection.

Factors related to invasive procedures or the use of some medical devices

Mechanical ventilation has been identified as a risk factor of MDR-PA infection [6, 24, 29, 32, 33]. Some authors suggest that the duration of mechanical ventilation is an important determinant of MDR-PA infection [29, 32]. For example Saherman et al. found a duration of more than 5 days for the acquisition of MDR-PA [32]. Haemodialysis [31], central venous catheters [23, 26], urinary catheters [23, 26], Foley catheters [6], arterial catheters [29], parenteral nutrition [33], and tracheostomy [29] are risk factors for MDR-PA infection.

Main environmental sources with MDR-PA acquisition and colonization in ICU

Main environmental sources with MDR-PA infection and colonization in ICU

Some studies point to a contaminated environment as a potential factor in the acquisition of MDR-PA. This factor would be related to the pressure of colonisation of objects or surfaces, especially previously occupied bed or room [25]. Thus, the type of environment may be a factor of an acquisition in close relation to the length of stay in the ICU. Objects such as sinks, faucets, water taps and nursing equipment are those most at risk of producing carriage, colonisation and infection [25, 28, 29, 32].

Prevalence infection of MDR aeruginosa in ICU

A total of 13 studies reported the prevalence of MDR-PA in ICU patients. There was a high degree of heterogeneity between the results of these studies (I2 = 98%, p < 0.001). However, the pooled prevalence was 4% (95% CI: 0%—11%). Figure 2 shows the prevalence with 95% CI of all the studies reporting the prevalence of MDR-PA.

Fig. 2.

Pooled prevalence of MDR-PA colonization across included studies

Association of risk factors with MDR-PA acquisition and infection in ICU: meta-analysis

Only 8 studies assessed risk factors by case–control or cohort studies. In addition, the definitions of cases or exposures and controls in these studies vary widely. For the meta-analysis, cases or exposures were defined as any person with carriage or colonisation with MDR-PA or carbapenem-resistant Pseudomonas aeruginosa strain admitted to an intensive care unit. The control was defined as any person without MDR-PA carriage or colonisation.

The mean difference in ICU length of stay was 8.64 (95% CI: 6.78—10.51) days, i.e. one week longer for patients with MDR-PA see Fig. 3.

Fig. 3.

Weighted mean difference (WMD) of length of stay of hospital (days) between the two groups

Although patients with MDRPA infection appeared to be older in most of the studies selected for the meta-analysis, there was no significant difference in age compared with patients without MDRPA (MD: -0.83; 95% CI: -2.19; 0.55), as shown in Fig. 4.

Fig. 4.

Weighted mean difference of age (years) between the two groups

The effect of mechanical ventilation on MDR-PA infection was reported in four studies, with high heterogeneity between the results of these studies (I2 = 86%; p < 0.01). Overall, mechanical ventilation significantly increased the risk of MDR-PA carriage, colonisation or infection (OR: 3.19; 95% CI: 2.25—4.53). See Fig. 5 for more details on the pooled effect of mechanical ventilation.

Fig. 5.

Odds ratio (OR) pooled effect of mechanical ventilation on MDR-PA positive test (either carriage, colonization, or infection)

The effect of different foreign bodies (central venous catheters, urinary catheters, parenteral feeding tubes and arterial catheters) on MDR-PA carriage, colonisation or infection was reported in selected studies. Overall, foreign bodies significantly increased the risk of MDR-PA (OR: 2.97; 95% CI: 2.40—3.68). Specifically, the use of a central venous catheter was OR: 3.16 95% CI: 1.87–5.33, urinary catheter use OR: 2.65; 95% CI: 2.05–3.44, parental nutrition use OR: 2.43; 95% CI: 1.15–5.16, and arterial catheter use OR: 7.00; 95% CI: 2.77–17.68. Less heterogeneity between results was observed in the urinary catheter subgroup (I2 = 35%; p = 0.20). Further details are shown in Fig. 6.

Fig. 6.

Odd ratio pooled (OR) effect of catheters and subtypes of foreign body on positive test of MDRPA

There was a significant overall association between the use of antibiotics OR: 3.69; 95% CI: 3.16—4.27. This was for carbapenems with OR: 4.12; 95% CI: 3.29-5.16, quinolones with OR: 3.31; 95% CI: 2.45—4.47, bectalactam with OR: 3.58; 95% CI: 2.45-5.24, and aminoglycosides with OR: 3.32; 95% CI: 2.33-4.73) see Fig. 7

Fig. 7.

Odds Ratio pooled effect of antibiotics exposure on MDR-PA colonisation or carriage or infection

Assessment of quality of publications

The NHLBI (National Heart, Lung, and Blood Institute) tools were used to assess the quality of the selected articles. The tool used to assess quality included 14 criteria. Amoung the 13 studies, 5 had at least half of their criteria deemed unsatisfactory. Among the 8 studies selected for the meta-analysis, six studies met more than 10 quality criteria. All studies selected for meta-analysis were rated as "good". see Table 3. The degree of heterogeneity, quantified by the I² statistic, was up to 50% in the majority of cases. Figure 8 shows the funnel plot showing the asymmetry of the 13 studies.

Table 3.

NHLBI study quality assessment tools 2021

Criteria						1	2		3	4		5		6	7		8	9	10		11	12		13
1. Was the research question or objective in this paper clearly stated?						y	y		y	y		y		y	y		y	y	y		y	y		y
2. Was the study population clearly specified and defined?						y	y		y	y		y		y	y		y	y	n		y	y		y
3. Was the participation rate of eligible persons at least 50%?						y	y		y	y		y		y	y		y	y	y		y	y		y
4. Were all the subjects selected or recruited from the same or similar populations (including the same time period)? Were inclusion and exclusion criteria for being in the study prespecified and applied uniformly to all participants?						y	y		n	y		y		n	n		y	y	n		y	y		y
5. Was a sample size justification, power description, or variance and effect estimates provided?						na	na		na	na		na		na	na		na	n	na		na	n		n
6. For the analyses in this paper, were the exposure(s) of interest measured prior to the outcome(s) being measured?						y	y		y	n		y		n	n		n	y	n		y	y		y
7. Was the timeframe sufficient so that one could reasonably expect to see an association between exposure and outcome if it existed?						n	y		n	n		y		n	n		n	y	n		y	y		y
8. For exposures that can vary in amount or level, did the study examine different levels of the exposure as related to the outcome (e.g., categories of exposure, or exposure measured as continuous variable)?						y	y		na	na		na		n	na		n	n	n		y	n		n
9. Were the exposure measures (independent variables) clearly defined, valid, reliable, and implemented consistently across all study participants?						y	y		n	y		y		n	n		n	y	n		y	y		y
10. Was the exposure(s) assessed more than once over time?						n	y		n	n		y		n	n		n	y	n		n	n		y
11. Were the outcome measures (dependent variables) clearly defined, valid, reliable, and implemented consistently across all study participants?						y	y		n	y		y		n	n		n	y	n		y	y		y
12. Were the outcome assessors blinded to the exposure status of participants?						n	n		n	n		y		n	y		n	n	n		n	n		n
13. Was loss to follow-up after baseline 20% or less?						y	y		y	y		y		y	y		y	y	y		y	y		y
14. Were key potential confounding variables measured and adjusted statistically for their impact on the relationship between exposure(s) and outcome(s)?						y	y		n	y		n		n	n		n	y	n		y	y		y
1*	2*	3	4*	5*	6	7		8			9*		10			11*				12*			13*
Aloush, 2006	Gudiol, 2020	Deplano, 2005	Labaste, 2019	Harris, 2011	Hota, 2009	Johnson, 2009		Levin, 1984			Nseir, 2011		Obritsch, 2004			Paramythiotou, 2004				Pena, 2007			Saharman, 2019

Y Yes, n no, na non applicable

^*selected for meta-analysis

Fig. 8.

Funnel Plot of analysis of publication biais of 13 studies selcted

Discussion

The pooled prevalence was 4% (95% CI: 0%-11%). This result seems low compared to many authors who found more than 15% [19, 24, 36].

In our study, we found that the use of some antibiotics during the ICU stay, the use of invasive devices and procedures were risk factors for infection or acquisition of MDR-PAs. Some authors such as Raman et al., Hirsch and Tam found the same results [14, 15]. In addition, the length of stay in the ICU (more than 8 days) and inanimate objects and surfaces susceptible to high bacterial colonisation pressure (sinks, taps…) were also identified as risk factors. All this strongly suggests the role of endogenous and exogenous transmission in the infection or acquisition of MDR-PA in ICUs.

One explanation could be that ICU patients, who are often critically ill, require intensive contact with caregivers, treatment and invasive devices. All this tends to compromise protective barriers and increase the risk of exogenous contamination [6]. Furthermore, repeated sterilisation in the ICU favours Pseudomonas aeruginosa strains, which are more likely to thrive in this environment, to the detriment of other pathogens [1]. In addition, the selection pressure exerted by broad-spectrum antibiotics, which are widely used in intensive care units, and the long-term or repeated use of antibiotics could lead to the development of MDR-PAs. It could be a possible consequence of the eradication of competing flora, which is replaced by MDR-PA strains [6, 37].

Regarding MDR-PA, there are two opposing hypotheses. The first hypothesis is endogenous transmission of MDR-PA infection [38] and the second is exogenous transmission [39]. This review was intended to be as broad as possible in order to obtain as much information as possible to better contribute to this debate. The debate is particularly unbalanced because many authors have focused on endogenous acquisition factors. The two previous reviews cited [15, 16] did not address the exogenous aspect of MDR-PA acquisition/infection. In this review, the evidence clearly suggests that these two routes of transmission (endogenous and exogenous) are likely to operate together. Unfortunately, insufficient data and methodological weaknesses in selected studies did not allow us to perform a meta-analysis to estimate the association between MDR-PA acquisition/infection and objects and surfaces.

Limitations

Despite many articles retrieved via open keywords, only 13 articles were selected. This means that information on antimicrobial resistance in ICUs is scarce, especially for environmental data. It is therefore necessary to increase the number of studies that address environmental issues.

Geographically, the data gap is still large, especially in certain continents such as Latin America, the Middle East, Asia and Africa. It's necessary to carry out larger cohort or case–control studies and surveillance around the world [40].

The studies eligible for inclusion were heterogeneous. Although many authors chose case–control and cohort studies, the definition of cases or controls, exposure and risk factors varied between authors. In our meta-analysis, the I2 different statistic calculated shows that there was a significant heterogeneity between studies. The lack of standard definitions in the primary studies could be considered a limitation of the systematic review and meta-analysis. We tried to address this issue by using similar comparison groups in the meta-analysis. Our definitions of cases and controls were intended to be as inclusive as possible, considering patients with MDR-PA, regardless of the author's definition, and patients without MDR-PA, regardless of whether they had comorbidities or not.

The funnel plot shows an asymmetry. This may be explained by the small number of studies included, but more importantly by the large differences in the sample sizes of these studies. [37], ranging from 36 to 7071. In addition, funnel plots are most effective in detecting publication bias when at least 20 studies are analysed [40].

Conclusion

The increasing prevalence of multidrug-resistant Pseudomonas aeruginosa among hospitalised patients in intensive care units remains a challenge for practitioners. Both individual (endogenous) and environmental (exogenous) factors are associated with colonisation and infection with MDR Pseudomonas aeruginosa in ICUs. The pooled prevalence of MDR-PA infection in ICUs was 4% (95% CI: 0%—11%). Mechanical ventilation, parenteral nutrition, catheters, previous use of antibiotics, especially carbapenems, and length of stay in the ICU (> 8 days), inanimate objects and surfaces exposed to high bacterial contamination pressure were significant risk factors for acquisition or infection with MDR-PA in the ICU. Our review did not allow us to correctly estimate the contribution of all exogenous factors. This is an insufficient data worldwide with a weak methodological approach. It is therefore necessary to carry out larger studies (cohort, case–control and surveillance) worldwide with standard definitions of case/exposure and that estimate the association of exogenous factors in the acquisition/infection of MDR-PA.

Abbreviations

CDBPH

Center for Development of Best Practices in Health

CRFilMT

Center for Research on Filariasis and other Tropical Diseases

GRADE

Grading of Recommendations Assessment, Development and Evaluation

ICU (s)

Intensive Care Unit (s)

MDRPA

Multidrug- Resistant strain of Pseudomonas aeruginosa

PA

Pseudomonas aeruginosa

PRISMA

Preferred Reporting Items for Systematic Review and Meta-Analysis Statement

PROSPERO

International Prospective Register of Systematic Reviews

SE

Serge Eyebe

GSW

Guy Sadjeu Wafeu

DCE

Danielle Christine Evina

EL

Emilia Lyongha

POZ

Pierre Ongolo Zogo

AM

Arthur Mbida

ABD

Alain Beranrd Dongmo

Authors’ contributions

SE,AM, ABD, POZ and JRZ initiated the protocol, designed the study and drafted the final protocol. CDE, SE, GSW made the critical review and adjusted the search strategy and the articles selection process. SE and GSW proceed to selection of articles. SE, GSW AM, ABD and EL proceed to the quality assessment of the publication. GSW and SE performed meta-analysis. POZ have contributed to design and reviewed all the process.

Funding

This study is supported by the “Bourse du mérite” of Ecole Doctorale Galillée of the University of Sorbonne Paris Nord (Paris, France). The funding body has no role in the design of the study and collection, analysis, and interpretation of data and in writing the manuscript.

Data availability

The datasets used and/or analysed during the current study will be available and could be shared with the readers.

Declarations

Ethics approval and consent to participate

Not applicable as all the data will be secondary data.

Consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.