PFGE Typing of Klebsiella Pneumoniae Isolates Incriminated in Repeated Outbreaks of Blood Stream Infection in the NICU of Alexandria University Hospitals

Background: Outbreaks of Klebsiella pneumoniae associated blood stream infections (BSI) had been a recurring serious problem for the inherently immuno compromised premature neonates admitted to the Neonatal Intensive Care Unit (NICU) of El Shatby Alexandria University Hospital throughout the last fifteen years. 

Objective:

Identify the source of the repeated outbreaks of K. pneumoniae as an initial step for an appropriate intervention plan.

Methods: A microbiologic survey was conducted involving blood cultures for all infants who developed CDC criteria of BSI and environmental samples from NICU & obstetric wards including inanimate surfaces, fluids and milk and the hands of healthcare providers. All K. pneumoniae isolates were subjected to antimicrobial sensitivity testing, ESBL confirmatory test and DNA “fingerprinting” using pulsed-field gel electrophoresis (PFGE).

Results:

During the study period 106 laboratory confirmed BSI cases were identified K. pneumoniae was the most common pathogen. It was isolated from 55 BSI cases (51.9%) and from 82 (9.4%) out of 869 samples from environmental surfaces, IV fluids, milk and hands of HCP. It was the most common Gram negative bacterium (GNB) isolated from these samples representing 31.8% of GNB. 

The antibiotic resistance profiles of the isolated K. pneumoniae strains revealed high frequency of resistance to almost all tested antibiotics except carbapenems to which all isolates were sensitive. 94.5% K. pneumoniae isolates from BSI (94.5%) and 70 % (47/67) of isolates from the hands, milk and NICU environment were ESBL producers. On the other hand the 15 K. pneumoniae isolates from obstetric ward were non producers

PFGE typing of the K. pneumoniae isolates revealed the complexity of the situation. PFGE patterns identified 24 different PFGE types. Eighteen BSI episodes (35%) were caused by 18 genetically unrelated strains.  In contrast six distinct PFGE types were incriminated in the aetiology of 65% (36/55) of the episodes of K. pneumoniae BSIs. Three of these 6 PFGE types were identified in the obstetric environment where they presented as non ESBL producers while strains of each of these three PFGE types isolated from either the NICU environmental samples or BSI samples were ESBL producers.

Conclusions:

The multiplicity of PFGE types of K. pneumoniae encountered during the present study revealed a continuum of influx of K. pneumoniae into and from the NICU environment to the infants, while the clusters of cases caused by the same PFGE type indicate continuous violations in infection prevention and control disciplines. On the other hand, detection of non ESBL producers and producers belonging to the same PFGE type raises the possibility of transfer of resistance genes between bacteria within the NICU microenvironment and the potential occurrence of epidemics of transferable ESBL.