Clinical impact of endemic NDM-producing Klebsiella pneumoniae in intensive care units of the national referral hospital in Jakarta, Indonesia

Abstract

Objective: A prospective observational study was performed to assess the epidemiology and clinical impact of carbapenem-non-susceptible Klebsiella pneumoniae (CNKP) in intensive care units (ICUs) of the national referral hospital in Jakarta, Indonesia.

Materials/methods: Adult patients consecutively hospitalized for > 48 h in two ICUs of the national referral hospital were included from April until October 2013 and from April until August 2014. K. pneumoniae from clinical cultures and standardized screening of rectum and throat on admission, discharge and weekly if hospitalized > 7 days were collected. Environmental niches and healthcare workers (HCWs) were also screened. Susceptibility was determined phenotypically and the presence of carbapenemase genes by PCR. Raman spectroscopy as well as multiple-locus variable number tandem repeat analysis (MLVA) were used for typing.

Results: Twenty-two out of 412 (5.3%) patients carried CNKP on admission and 37/390 (9.5%) acquired CNKP during ICU stay. The acquisition rate was 24.7/1000 patient-days at risk. One out of 31 (3.2%) environmental isolates was a CNKP. None of the HCWs carried CNKP. Acquisition of CNKP was associated with longer ICU stay (adjusted Hazard Ratio: 2.32 [CI₉₉: 1.35-3.68]). ICU survival was lower among patients with CNKP compared to patients with carbapenem-susceptible K. pneumoniae (aHR 2.57, p = 0.005). Ninety-six of the 100 (96%) CNKP isolates carried a carbapenemase gene, predominantly bla_NDM. Raman typing revealed three major clusters among 48 Raman types identified, whereas MLVA distinguished six major clusters among a total of 30 different genotypes.

Conclusions: NDM-producing CNKP are introduced into these ICUs and some strains expand clonally among patients and the environment, resulting in endemic CNKP. CNKP acquisition was associated with prolonged ICU stay and may affect ICU survival.

Trial registration: The study was registered at Netherlands Trial Register http://www.trialregister.nl. Candidate number: 23527, NTR number: NTR5541, NL number: NL5425 (https://www.trialregister.nl/trial/5424), Retrospectively registered: NTR: 22 December 2015.

Keywords: Carbapenemase; Indonesia; Intensive care unit; Klebsiella pneumoniae; Microbial drug resistance; Mortality.

Fig. 1

Rate of acquisition of carbapenem-susceptible and -non-susceptible K. pneumoniae in ICUs. Acquisition dynamics of carbapenem-susceptible and –non-susceptible K. pneumoniae during ICU stay. The blue line represents the cumulative percentage of patients by first day of culture being positive for carbapenem-susceptible K. pneumoniae during ICU stay. The red line represents the cumulative percentage of patients by first day of culture being positive for carbapenem-non-susceptible K. pneumoniae during ICU stay. P value was calculated using independent samples-Mann Whitney U test. In total, data from 100 patients are included in this figure

Fig. 2

Cumulative percentage of length of stay according to K. pneumoniae status. Cumulative length of ICU stay of patients based on their K. pneumoniae status. Length of stay (days) represent total days patients were hospitalized in the ICU. The blue line represents patients that were always K. pneumoniae negative during their ICU stay. The red line represents patients already positive for carbapenem-susceptible K. pneumoniae on the day of admission. The green line represents patients already positive for carbapenem-non-susceptible K. pneumoniae on the day of admission. The orange line represents patients that acquired carbapenem-susceptible K. pneumoniae during ICU stay and the black line represents patients that acquired carbapenem-non-susceptible K. pneumoniae during ICU stay. The length of stay of patients that became positive with carbapenem-non-susceptible K. pneumoniae during ICU stay was longer than that of the other groups (Cox regression, P < 0.001)

Fig. 3

Survival of patients according to their K. pneumoniae status. Survival of patients with carbapenem-non-susceptible K. pneumoniae (on admission or acquired during ICU stay) (red line) compared with the survival of patients that had carbapenem-susceptible K. pneumoniae (on admission or acquired during ICU stay) (blue line) in their screening and/or clinical cultures. P value was calculated using logistic regression

Fig. 4

Persistence of prevalent clones of carbapenem-non-susceptible K. pneumoniae in ICUs, as determined by Raman spectroscopy typing. Endemicity of the three largest clusters, as determined by Raman spectroscopy, of carbapenem-non-susceptible K. pneumoniae in ICUs, April–October 2013 and April–August 2014. The orange bars represent cluster CIPTOKPN24. The blue bars represent CIPTOKPN27. The green bars represent CIPTOKPN30. The x-axis indicates time periods of the study (by week, April 2013-Oktober 2013 and April 2014–August 2014). The y-axis indicates number of isolates

Fig. 5

MLVA minimum spanning trees of carbapenem-non-susceptible Klebsiella pneumoniae. Minimum spanning tree analysis of K. pneumoniae isolates based on clustering at the VNTR loci. Clusters of genotypes differing in only one marker are indicated with a grey background. Panel a: Colours correspond to specimens from which K. pneumoniae isolates were cultured. Panel b: Distribution of genotypes per patient. Each colour, except white, indicates a different patient. Only patients with 2 or more isolates are presented in this manner. Patients that had only one isolate of a carbapenem-non-susceptible K. pneumoniae are indicated by the colour white