Emergence and rapid spread of carbapenem resistance during a large and sustained hospital outbreak of multiresistant Acinetobacter baumannii

Abstract

Beginning in 1992, a sustained outbreak of multiresistant Acinetobacter baumannii infections was noted in our 1,000-bed hospital in Barcelona, Spain, resulting in considerable overuse of imipenem, to which the organisms were uniformly susceptible. In January 1997, carbapenem-resistant (CR) A. baumannii strains emerged and rapidly disseminated in the intensive care units (ICUs), prompting us to conduct a prospective investigation. It was an 18-month longitudinal intervention study aimed at the identification of the clinical and microbiological epidemiology of the outbreak and its response to a multicomponent infection control strategy. From January 1997 to June 1998, clinical samples from 153 (8%) of 1,836 consecutive ICU patients were found to contain CR A. baumannii. Isolates were verified to be A. baumannii by restriction analysis of the 16S-23S ribosomal genes and the intergenic spacer region. Molecular typing by repetitive extragenic palindromic sequence-based PCR and pulsed-field gel electrophoresis showed that the emergence of carbapenem resistance was not by the selection of resistant mutants but was by the introduction of two new epidemic clones that were different from those responsible for the endemic. Multivariate regression analysis selected those patients with previous carriage of CR A. baumannii (relative risk [RR], 35.3; 95% confidence interval [CI], 7.2 to 173.1), those patients who had previously received therapy with carbapenems (RR, 4.6; 95% CI, 1.3 to 15.6), or those who were admitted into a ward with a high density of patients infected with CR A. baumannii (RR, 1.7; 95% CI, 1.2 to 2.5) to be at a significantly greater risk for the development of clinical colonization or infection with CR A. baumannii strains. In accordance, a combined infection control strategy was designed and implemented, including the sequential closure of all ICUs for decontamination, strict compliance with cross-transmission prevention protocols, and a program that restricted the use of carbapenem. Subsequently, a sharp reduction in the incidence rates of infection or colonization with A. baumannii, whether resistant or susceptible to carbapenems, was shown, although an alarming dominance of the carbapenem-resistant clones was shown at the end of the study.

FIG. 1

Temporal trends in incidence of new patients colonized or infected with CR A. baumannii (CR-Ab) and CS A. baumannii (CS-Ab) and carbapenem consumption from January 1997 to June 1998.

FIG. 2

Patterns obtained by PFGE for A. baumannii after digestion with SmaI. Lanes 1 to 4, CS isolates belonging to clone A (lane 1), clone B (lanes 2 and 3), and clone C (lane 4); lanes 5 to 12, CR isolates belonging to clone D (lanes 5 to 10) and clone E (lanes 11 and 12); lanes mm, molecular size marker.

FIG. 3

Repetitive PCR patterns for A. baumannii. Lanes 1 and 2, CS isolates of clone A; lanes 3 and 4, other CS sporadic clones previously isolated during the endemic; lanes 5 to 10, CR isolates of clone D (lanes 7 and 8) and clone E (lanes 5, 6, 9, and 10); lane mm, 100-bp molecular size marker.

FIG. 4

Temporal trends in environmental contamination with A. baumannii clones before and after interventions.

FIG. 5

Temporal trends in clonal spread of carbapenem resistance among A. baumannii isolates by using molecular characterization of epidemic and endemic clones. Differences before and after the interventions were analyzed by comparing CR A. baumannii (clones D and E) and CS A. baumannii (clone A) groups for periods 1, 2, and 3 by using linear trend analysis with proportions. Six levels of exposition were selected per each period; these corresponded to months 1 to 6 for each period compared. Differences reached significant differences when periods 1 and 2 were compared (chi-square for linear trend, 6.14; P = 0.013) and periods 1 and 3 were compared (chi-square, 5.38; P = 0.02) but not when periods 2 and 3 were compared (chi-square, 1.13; P = 0.28).