Pseudomonas aeruginosain a neonatal intensive care unit: molecular epidemiology and infection control measures

Gram-negative bacteria have become relevant causes of healthcare-associated infections in the neonatal intensive care unit (NICU) environment. [1, 2]. Outbreaks by Enterobacteriaceae have been widely described in this setting [3], and we recently reported the increased circulation of ESBL-producing Klebsiella pneumoniae and Serratia marcescens in the NICU of the Federico II University Hospital of Naples between 2002 and 2004 [4, 5].

Pseudomonas aeruginosa, a non-fermentative, gram-negative rod, is responsible for a wide variety of clinical syndromes in NICU patients, including sepsis, pneumonia, meningitis, diarrhea, conjunctivitis and skin infections [6]. Nevertheless, if compared to other gram-negative bacteria, outbreaks by P. aeruginosa in NICU settings have been much less reported up-to-date and have been associated with both environmental reservoirs and healthcare workers' carriage [7‐10].

An increased number of infections and colonisations by P. aeruginosa has been observed in the NICU of our university hospital between 2005 and 2007. The aims of this study were: i) to analyze the molecular epidemiology and antimicrobial susceptibility patterns of P. aeruginosa isolates; ii) to describe the infection control measures undertaken to limit P. aeruginosa spread in the ward.

The overall incidence rate of P. aeruginosa infections in NICUs is reported to be of approximately 10% [1, 2, 15], while infection attack rates during outbreak periods appear to be lower, ranging between 1% and 2.8% [10, 16]. Higher attack rates have been reported [9, 17], but such studies did not consider infections alone and analysed the combination of infections and colonisations. At our NICU, 11 neonates developed 11 P. aeruginosa severe infections, with an infection attack rate of 1.8%, and nearly 24% of the patients became colonised by P. aeruginosa, with an epidemic peak of 50% at 18 months from onset. Ten of the 11 infected patients were pre-term neonates (gestational age < 37 weeks) and 7 of them were of extremely low birth weight (ELBW), i.e. below 1000 grams. ELBW neonates have actually been shown to have a significantly increased risk of acquiring P. aeruginosa when compared to higher birth-weight infants [9]. Moreover, all of the four infected neonates who died weighed less than 1000 grams, therefore the crude mortality rate among ELBW patients infected by P. aeruginosa was of 57%. Although no attributable mortality rate was calculated, three of the four neonates died soon after (0–24 hours) the infection was diagnosed, thus our data indirectly confirm other Authors' findings regarding the very high mortality rates related to P. aeruginosa infections [15], especially in the lowest birth-weight categories' infants [18]. In our setting, in addition to the ELBW condition, two P. aeruginosa antibiotypes, both displaying resistance to imipenem, meropenem, gentamicin and ciprofloxacin, have probably affected the final outcome in two of the four fatal cases. We did not analyse mechanisms of resistance as no further such phenotypes were identified. Nevertheless, to our knowledge, this is one of the first accounts on two carbapenem-resistant P. aeruginosa genetically unrelated strains which caused two sepsis in a NICU.

P. aeruginosa frequently causes multi-clone outbreaks, with the concurrent isolation of genetically distinct strains among patients and healthcare workers (HCWs) and in the environment. During a 15 months-long epidemic in a NICU, Moolenaar et al. [9] identified the three main P. aeruginosa genotypes A, B, and C, isolated in 75%, 15%, and 10% of case-patients, respectively. Such genotypes were also found on three nurses' hands, while two positive environmental isolates showed two distinct genotypes, unrelated to any human isolate. Moreover, Foca et al. [17] described the circulation of multiple P. aeruginosa PFGE profiles over a 33 months-period in the NICU, showing the presence of a major clone, which was also isolated from the hands of a nurse, of two other PFGE types, and of eight unique clones. No environmental specimen proved to be positive for P. aeruginosa. Our study, covering a 24 months time span, identified a predominant PFGE type which was responsible for 36% of infections by P. aeruginosa and at least 35% of colonisations by the same pathogen. Such PFGE profile was also found in one sink, but not on any nurse's hand and circulated in the ward together with less recurrent and with sporadic strains, which caused the remaining infections and colonisations. Other five environmental samples proved to be positive for distinct P. aeruginosa PFGE types, unrelated to the ones colonising or infecting the patients. Transmission of P. aeruginosa from environmental sources to patients and HCWs has been thoroughly described [19, 20]. Our findings indicate the presence in our NICU of multiple and distinct P. aeruginosa reservoirs, both environmental and human, and, owing to the long time period between the appearance in neonates (July 2005) and the environmental isolation (2^nd quarter 2006), we are not able to understand whether the only sink sample found to be positive for P. aeruginosa PFGE profile A has been a result, rather than the origin, of the pathogen's circulation in the ward.

Healthcare workers' (HCWs) hands have been frequently implicated in the spreading of P. aeruginosa in the NICU setting [9, 10, 17]. At our institution one HCW, with short to medium-length natural fingernails, had a positive hand culture for P. aeruginosa, displaying the G genotype. The culture was drawn when the epidemic was at its peak value, with nearly 50% of neonates being colonised. The detection of HCWs who are colonised at any body site by P. aeruginosa epidemic clones may sometimes identify the pathogen's reservoir, thus enabling a successful and timely outbreak containment [10]. Owing to organizational difficulties, the HCW found to have a positive hand culture at our NICU could not be reassigned to non-clinical activities. Moreover, no further analyses have been performed to establish her contribution to the outbreak, therefore we can only hypothesize that she has been a transient carrier of one of the less recurrent epidemic clones. Actually, patient exposure within the first 14 days of NICU admission to a HCW with short natural fingernails and with one positive hand culture for a P. aeruginosa epidemic clone has not been recognized as an independent risk factor for acquiring P. aeruginosa colonization or infection [9]. In turn, exposure to two HCWs with negative hand cultures has been associated with an increased risk of colonisation by a P. aeruginosa epidemic clone on multivariate analysis [17]. This finding suggests that transient colonisations of HCWs' hands by P. aeruginosa may be underestimated during outbreaks investigations and that reinforcement of hand disinfection and of correct gloves use should always be promptly initiated when an increased number of P. aeruginosa isolations is detected. In agreement with previous data [21], our study shows that, compared with all the traditional infection control interventions undertaken to contain P. aeruginosa circulation of in the ward, the hand disinfection educational programme was the most effective one. The programme started during the fourth quarter of 2006, when the outbreak was at its peak value, and by the end of the second quarter of 2007 the outbreak was over. Owing to the programme's success, the meaning of different hand disinfection compliance rates before and after patient contact and how this may have affected P. aeruginosa circulation in the ward were not further investigated. A possible explanation for the partial ineffectiveness of the traditional infection control interventions may be the molecular heterogeneity of P. aeruginosa outbreaks. Thus, the timely identification of increased isolation of this pathogen, achieved by means of active surveillance, appears to be crucial to limit the spreading of P. aeruginosa in NICU settings.