Unit:
Faculty of MedicineLibrary of the School of Health Sciences
Author:
Tansarli Giannoula
Dissertation committee:
1. Γεώργιος Λ. Δαΐκος, Καθηγητής, Ιατρική Σχολή, ΕΚΠΑ
2. Μιχαήλ Σαμάρκος, Αναπληρωτής Καθηγητής, Ιατρική Σχολή, ΕΚΠΑ
3. Ιωσήφ Παπαπαρασκευάς, Αναπληρωτής Καθηγητής, Ιατρική Σχολή, ΕΚΠΑ
4. Ελένη Ι. Γκόγκα, Καθηγήτρια, Ιατρική Σχολή, ΕΚΠΑ
5. Μήνα Ψυχογυιού, Επίκουρη Καθηγήτρια, Ιατρική Σχολή, ΕΚΠΑ
6. Ευαγγελία-Θεοφανώ Πιπεράκη, Επίκουρη Καθηγήτρια, Ιατρική Σχολή, ΕΚΠΑ
7. Γκίκας Μαγιορκίνης, Επίκουρος Καθηγητής, Ιατρική Σχολή, ΕΚΠΑ
Original Title:
Διαχρονική εξέλιξη αντοχής στην κολιμυκίνη πολυανθεκτικών στελεχών Klebsiella pneumoniae που απομονώθηκαν από το αίμα ασθενών
Translated title:
Colistin resistance in carbapenemase-producing Klebsiella pneumoniae bloodstream isolates: Evolution over 15 years and temporal association with colistin use by time series analysis
Summary:
Bacteremia caused by carbapenemase-producing Klebsiella pneumoniae (CP-Kp) is associated with increased mortality and often, colistin is the only available antibiotic against these infections. In the present study, we i) tested last-resort antibiotics against CP-Kp blood isolates from patients who were hospitalized in Laiko hospital, a 500-bed tertiary care hospital in Athens, Greece, between 2002 and 2016, ii) investigated the trends of antimicrobial resistance to colistin over time as well as its temporal association of colistin use during the study period.
In total, 313 blood isolates were tested positive for carbapenemase production with multiplex PCR during the study period. Of them, 197 (62.9%) were resistant to meropenem (MIC >8 mg/L), 79 (25.2%) exhibited meropenem MICs of 4-8 mg/L and 37 (11.8%) displayed MICs in the susceptible range (MIC ≤2 mg/L). Resistance frequencies to colistin, fosfomycin and tigecycline were 35.1% (110/313), 30.0% (94/313) and 5.4% (17/313), respectively. Polymyxin B displayed the same susceptibility pattern as colistin, with the MICs of the two antibiotics differing by one or two dilutions without difference in interpretation. Among 313 isolates containing carbapenemase-encoding gene(s), blaKPC was detected in 198 isolates (63.3%), blaVIM in 89 (28.4%), blaNDM-1 in 17 (5.4%) and blaOXA-48 in 4 (1.3%). Five isolates carried more than one carbapenemase gene (three carried blaKPC + blaVIM, one carried blaKPC + blaNDM-1 and one carried blaVIM + blaΝDM-1).
Then, we studied the trends of antimicrobial resistance to colistin in our hospital during 2002-2016 using the linear regression analysis. The analysis revealed a statistically significant increasing trend in colistin resistance over time from 0% in 2002 to 26.9% in 2016 [R2= 0.5, p< 0.01; point estimate of coefficient, 2.97, 95% confidence interval (CI) 1.20–4.75; standard error of the slope coefficient, 0.82]. Similarly, an increasing trend in the use of colistin over time was observed (R2= 0.54, p< 0.01; point estimate of coefficient, 0.66, 95% CI 0.49–0.82; standard error of the slope coefficient, 0.08), from 0.2 DDDs/100 patient-days in the first trimester of 2002 to 7.10 DDDs/100 patient-days in the second trimester of 2016.
Last, in order to examine the potential correlation between colistin use and resistance we used time series analysis. The overall incidence rate of colistin-resistant K. pneumoniae was 0.37 isolates per 10.000 patient-days and the average quarterly colistin use was 5.14 DDDs/100 patient-days. After constructing the individual time series for colistin resistance and use, we built a linear transfer function model which enabled us to quantify the dynamic relationship between colistin use and resistance. Specifically, a temporal association between colistin use and resistance was observed; an increase in colistin use by 1 DDD/100 patient-days led to a 0.05 increase in the incidence rate of colistin resistance. The time lag between the effect of colistin use on subsequent variations in colistin resistance was 3 months. Colistin use and prior levels of colistin resistance could explain 69% of colistin resistance; in the remaining 31%, other factors might have played a role.
In conclusion, the findings of the present study have important implications in implementing strategies to contain colistin resistance in our hospital.
Main subject category:
Health Sciences
Keywords:
Antimicrobial resistance, Antibiotics, Multidrug-resistant, Carbapenemases, Gram-negative, Colistin, Polymyxins, Bacteremia, Bloodstream infection, Time series analysis
Number of references:
133
