To the Editor: The emergence and rapid worldwide dissemination of carbapenemase-producing bacteria (CPB), especially carbapenemase-producing Enterobacteriaceae (CPE), have prompted public health authorities to reconsider prevention strategies to control the spread of these organisms (1–5). In France, national guidelines recommend systematic screening for commensal CPE and glycopeptide-resistant enterococci (GRE) in all patients admitted to hospitals who have been hospitalized in other countries during the preceding 12 months (6,7) (repatriated patients), independently of whether transfer was direct from hospital to hospital (DT) or not (NDT). These guidelines also recommend implementation of presumptive patient isolation and contact precautions on admission (6,7). We conducted a 33-month survey at Hôpital Européen Georges Pompidou (HEGP), a university teaching hospital in Paris, of CPE and GRE in repatriated patients; we also investigated incidence of extended-spectrum β-lactamase (ESBL)–producing Enterobacteriaceae and carbapenemase-producing Acinetobacter baumannii and Pseudomonas spp. in the same patient group.

During November 2010–July 2013, a total of 541 patients who had previously been hospitalized in a total of 71 other countries were admitted to HEGP. Rectal swab specimens were taken from 510 patients; 82 (16.1%) were DT, 415 (81.4%) were NDT, and 13 (2.5%) had an unclear history of transfer. Median patient age was 61 (range 12–98) years; 70% of patients were male. Results of screening by using antibiotic-containing Luria Bertani broths for enrichment and plating on selective media were negative for 354 (69.4%) of the 510 patients surveyed; 33 (6.5%; 16 DT, 17 NDT) patients were colonized with >1 CPB and/or GRE and 123 (24.1%; 22 DT, 99 NDT, 2 unclear) with ESBL producers only. More specifically, 19.5% (16/82) of DT patients and 4.1% (17/415) of NDT patients were colonized with CPB and/or GRE (p<10⁻⁵ by χ² test); 26.8% (22/82) of DT patients and 23.9% (99/415) of NDT patients were colonized with ESBL producers only (p = 0.67). Characteristics of the 33 patients carrying CPB and/or GRE are shown in the Table. Of all isolates, 191 produced ESBLs only.

Rates of resistance for ESBL-producing Enterobacteriaceae and CPE were, respectively, 53.1% and 57.1% to gentamicin, 16.7% and 32.1% to amikacin, 77.1% and 82.1% to nalidixic acid, 63% and 75% to levofloxacin, and 70.3% and 75% to ciprofloxacin. The Pseudomonas spp. and A. baumannii isolates were also multidrug resistant; all isolates were colistin susceptible.

Among the 33 colonized patients, 13 (39.4%) were not infected; 1 of the uninfected patients died. Seven patients were infected with >1 CPB (health care–related in 2 patients, 1 of whom died), 4 patients with ESBL-producing Enterobacteriaceae (health care–related in 1 patient, who died), and 9 patients with other bacteria (health care–related in 4 patients, 1 of whom died). No patients were infected with GRE. Overall, 60.6% of colonized patients were infected and 12.1% died; 35% (7/20) of the infections were health care–related (3 urinary tract device–related infections, 2 cases of ventilator-associated pneumonia, 1 infection at the site of a portacath, and 1 case of cellulitis).

Almost 25% of the repatriated patients carried ESBL-producing Enterobacteriaceae (mostly CTX-M-15 producers; Technical Appendix); 6.7% carried CPB and/or GRE. By comparison, during the study period, only 10.8% of 2,314 systematically screened patients in the medical and general surgery intensive care units at HEGP (repatriated patients excluded) carried ESBL-producing Enterobacteriaceae; 1 carried vanA Enterococcus faecium (data not shown). For patients with no record of hospitalization abroad, no CPE isolates were found; other bacterial isolates included 1 vanA E. faecalis, 13 vanA E. faecium (all known from previous outbreaks), 4 OXA-23–producing A. baumannii, and 4 VIM- and 1 IMP-producing P. aeruginosa.

Of the repatriated patients, 19.5% of DT patients (vs. 4.1% of NDT) and 23.9% (7 DT, 4 NDT) of those who were transferred to medical and general surgery intensive care units (ICUs) were CPB and/or GRE carriers. This finding highlights the role of severe underlying disease or injury and recent antimicrobial drug treatment. Among ICU patients, 3 died, most likely from underlying conditions, findings in line with the observation that carriage of or infection with multidrug-resistant bacteria is not the only predictor of death (8). Most of the 28 CPE isolates were resistant to fluoroquinolones and aminoglycosides except amikacin; 21 carried OXA-48–type genes, 7 of which were non-ESBL producers and were detected only around an ertapenem disk on Drigalski agar (Bio-Rad, Marnes-la-Coquette, France). All CPB, irrespective of species, showed imipenem hydrolysis in a recently described test (9) that was shortened and simplified by incubating colonies directly in antibiotic solution.

Although time-consuming and certainly perfectible, implementation of strict control measures to limit CPB and GRE spread (6,7) seems justified, a conclusion supported by the occurrence, since November 2010, of just 1 cross-transmission–linked CPB outbreak in an ICU at HEGP (after urgent intervention for cardiac arrest). Of particular concern is the high proportion of OXA-48–producing isolates in persons with no documented link to repatriation in France (10). This finding could be explained in part by the historical and demographic relationships between France and North Africa, where prevalence of OXA-48 is high, reflected in results from patients repatriated from that part of the continent.