ABSTRACT
The spread of carbapenemase-producing Enterobacteriaceae (CPE) is
an emerging problem in children undergoing cancer-directed chemotherapy.
A previous Italian survey reported a threefold increase of CPE
colonization rate, and a fourfold increase of CPE bloodstream infections
in a 2 year-observation time interval. To assess the efficacy of the
measures put in place to control this emergency, a second survey was
performed on the years 2016-2017. The results showed that the number of
patients colonized by CPE remained stable, while the number of
bloodstream infections decreased, as well as the resulting mortality. We
conclude that children undergoing chemotherapy are at risk for CPE
colonization/infection, but the awareness of their colonization status
may allow reducing CPE morbidity and mortality.
INTRODUCTION
The rapid spread of carbapenemase-producing Enterobacteriaceae(CPE), multidrug resistant (MDR), became a serious threat for patients
receiving antineoplastic chemotherapy. In 2015, we reported the results
of a nationwide survey among centers participating in the Italian
network of pediatric hematology-oncology (Associazione Italiana
Ematologia Oncologia Pediatrica; AIEOP). During a 2-year observation
period (2012-2013), a threefold increase in the colonization rate, and a
fourfold increase of bloodstream infectious episodes caused by CPE were
observed, with a 90-day mortality of 14%. As a result, all centers
implemented measures of hospital infection control, including the
screening for intestinal MDR strains by rectal swab in every inpatient,
on admission.(1) In order to monitor the epidemiological trend of
MDR/CPE infectious episodes, we performed a second survey and the
results are here reported.
PATIENTS AND METHODS
The study was a retrospective data collection on a 24-month period (from
January 2016 to December 2017); 13 of the 40 Italian pediatric cancer
centers of the AIEOP network participated in the study. Of them, eight
were also included in the previous study.
The following data were collected: total number of children cared per
year, either newly diagnosed or on treatment; type of tumor
(leukemia/lymphoma or solid tumor); total number of days of hospital
admission; routine monitoring of MDR/CPE strains on rectal swab; number
of patients colonized by MDR/CPE; number of patients with bacteremia;
cause of bacteremia, and number of patients who died following
bacteremia. Strains were considered as “non-susceptible” to
carbapenems when they tested resistant or non-susceptible according to
the interpretative criteria recommended by European Committee on
Antimicrobial Susceptibility Testing (EUCAST) (2), or by the Clinical
and Laboratory Standards Institute (CLSI)(3). Incidence of colonization
and bacteremia were calculated as rates (episodes/1,000 days of hospital
admission) during each year of observation. Mortality of bacteremic
patients was evaluated at day 90 after the first positive blood culture.
All the analyses were performed with SAS vers. 9.4 (SAS Institute Inc.,
Cary, NC, USA).
RESULTS
During the study period the total number of episodes of admission was
8,257: 3,361 (41%) for patients with solid tumor, and 4,896 (59%) for
patients with leukemia or lymphoma. The median number of patients cared
by each center was 396 (range, 109 to 1,824). A screening program for
detection of MDR/CPE carrier was in place in 12 of the 13 (92%)
participating centers during the study period. Table 1 reports data on
colonization and bacteremia. Overall, in 53 of the 8,257 episodes
(0.64%) the patient was found to be carrier of MDR/CPE. This finding
accounted for an overall rate of colonization of 0.5/1,000 days of
hospitalization (95% CI, 0.37-0.65). The comparison of the results of
the current survey with those of the previous study (table 1) shows that
the rate of colonization was comparable between the two study periods:
0.50% (2016-2017) vs. 0.48% (2012-2013). MDR/CPE bacteremia was
reported in 13 children from six centers. The rate of MDR/CPE bacteremia
was 0.12 for 1,000 days of hospitalization (95% CI, 0.07-0.21) and
compared favorably with that observed in the previous survey (0.42 for
1,000 days of hospitalization; 95% CI, 0.31-0.57; p<0.0001).
The number of centers participating in the study was similar: 15 in
2012-2013 vs. 13 in 2016-2017. When the comparison between the two eras
(2012-2013 vs. 2016-2017) was restricted to the eight centers
participating in both surveys, no difference was observed.
The strains responsible for bacteremia were Klebsiella pneumoniaein 9/13 children, Escherichia coli in 2/13, Citrobacter
spp. and Stenotrophomonas maltophilia in one case each. The antibiotic
susceptibility was available for 10 MDR/CPE strains. All strains were
resistant to third-fourth generation cephalosporin,
piperacillin/tazobactam, and carbapenems; they were susceptible to
amikacin in seven of nine tested, colistin in five of five, fosfomycin
in six of seven and tigecycline in five of eight. Death from any cause
occurred in two of the 13 patients with bacteremia (15%). Both
patients, a 6-year old female with severe aplastic anemia, undergoing
allogeneic hemopoietic stem cell transplant from a matched unrelated
donor, and a 10-year old female with acute myeloid leukemia, had a
septic shock by MDR Klebsiella pneumoniae and died. Death
occurred after 11 and 15 days from the first positive blood culture,
respectively.
DISCUSSION
The fast and wide spreading of CPE/MDR infections among immune
compromised patients, observed in the last years, represented an excess
of “preventable deaths” on which much attention was paid. The results
of a survey performed in the period 2012-2013 had showed a threatening
number of MDR/CPE episodes of bacteremia in children treated for cancer
(4,5). This prompted the centers to enforce measures of hospital
infection control. They included early identification of MDR/CPE
colonized patients by rectal swab screening; adoption of hospital
contact precautions by caregivers of colonized patients; single room or
cohort isolation for colonized patients; empirical antibiotic therapy
guided by the antibiotic susceptibility of colonizing germ; rapid
identification of the causative germ of febrile episodes during
neutropenia.
To assess the efficacy of such measures, we performed a second,
follow-up survey on the years 2016-2017. The differences between the
study populations in the two surveys were limited. About one-half of the
participating centers were the same in the two studies. All Italian
geographical macro-areas were represented in the 2016-2017 survey, with
six centers from Northern Italy, three from the center of Italy, and
three centers from Southern Italy and islands.
By the time of the follow-up survey, routine screening for colonization
of patients was in place in all but one participating centers (92%),
while it had been progressively introduced during the years of the
initial survey (one quarter of the centers in 2012, 60% in 2013) (1).
Overall, the colonization rate observed by routine screening was
0.5/1,000 days of hospitalization, which was not different from what
observed in 2012-2013 survey, although on lower numbers of patients.
Interestingly, the MDR/CPE bacteremia rate was significantly lower in
2016-2017 than in 2012-2013, with 0.12 vs. 0.48/1,000 days of
hospitalization (p<0.0001). This difference was confirmed when
the analysis was focused on the eight centers which participated in both
surveys. Mortality by MDR/CPE remained in the range of 15% in both
studies.
In conclusion, circulation of MDR/CPE strains remained stable in
children admitted for cancer chemotherapy in the Italian centers between
2012-2013 and 2016-2017, suggesting that the “epidemic” of MDR/CPE,
observed at the beginning of the present decade, remained then under
control. Furthermore, this was associated with a lower number of
bacteremia episodes, thus suggesting that higher awareness of this
problem may have resulted in a closer application of preventive
measures. The spectrum of antibiotic susceptibility of isolates from
blood cultures showed that amikacin maintains a good susceptibility as
well as colistin, fosfomycin and tigecyclin (6).