INTRODUCTION

Antimicrobial resistance continues to be a major global health problem, mainly in developing countries ¹. Latin America is no stranger to this problem, where the growing spread of resistant bacteria threatens to increase treatment failures, even while using top-of-the-line antimicrobials ^{2 , 3}. The number of multidrug resistant bacteria associated with infectious diseases in humans increases constantly. The World Health Organization (WHO) highlights Klebsiella pneumoniae as one of the most important multidrug resistant bacteria for research ¹, due to its ability to accumulate and disseminate antimicrobial resistance genes, and to its high plasmid load and wide variability in guanine + cytosine (G+C) content ^{( 4}. Due to the decrease in the number of effective antimicrobials for infections caused by multidrug resistant microorganisms, polymyxins (polymyxin B and polymyxin E, also called colistin) are considered as last-resort treatments, especially in severe infections by carbapenem-resistant K. pneumoniae. Colistin is often used in combination with other antibiotics such as tigecycline, meropenem, gentamicin, or fosfomycin to improve its spectrum and effectiveness ⁵.

However, in recent years, reports of K. pneumoniae with low sensitivity to colistin are increasing worldwide ⁶. Resistance mechanisms occur due to modifications in the integrity of the cell membrane and affect bacterium anchorage, either at the lipopolysaccharide level or by ionic alteration of membrane (by changes in Mg+ and Ca++) ⁷. These modifications can be due to mutations in chromosomal genes ^{6 , 7} and to mobile resistance genes transmissible via plasmids (mcr) ⁸.

Even though in 2016 the Pan American Health Organization (PAHO) warned of the increasing prevalence of colistin resistance (based on the first reports from 2015 of mcr-1 gene presence and the importance of its surveillance) ⁸, little has been studied on the evaluation of molecular mechanisms prevalence of this type of resistance and its clinical relevance in Peru.

The aim of this study is to evaluate the levels and mechanisms of resistance to colistin and carbapenems in strains of multidrug resistant K. pneumoniae isolated during 2015-2018 at the Instituto Nacional Materno Perinatal de Lima.

THE STUDY

This is a descriptive cross-sectional study of K. pneumoniae isolates from biological samples, such as serum, cerebrospinal fluid, bronchial aspirate, and stool culture, obtained from patients admitted to the Instituto Nacional Materno Perinatal de Lima.

We included 36 K. pneumoniae isolates from different patients in the study. Isolates were selected from a collection of bacteria from the Microbiology Laboratory of Instituto Nacional Materno Perinatal de Lima. Only those with multidrug resistance (MDR) sensitivity profile were considered, i.e. resistant to at least three different families of antimicrobials, as determined by disk diffusion tests. These isolates were collected between 2015 and 2018 ⁹.

Antimicrobial sensitivity tests were performed using disk diffusion tests. This resistance type was interpreted according to Clinical and Laboratory Standard Institute (CLSI) guidance ¹⁰. Colistin resistance assessment was carried out by microdilution to determine the minimum inhibitory concentration (MIC) as recommended by the joint group CLSI- European Committee on Antimicrobial Susceptibility Testing (EUCAST) on polymyxin cut points ¹¹.

We made a 1/1000 dilution in Mueller Hinton II broth (MHII) from an inoculum at 0.5 McFarland. Initially, we prepared an antibiotic solution concentration four times the concentration needed, because the bacterial inoculum (50 µL) and 50 µL of MHII medium, when added, decrease the concentration to the desired level for the serial dilutions. Each sample included a growth control well (without antibiotic) and a negative control well (MHII without the bacterial inoculum). All strains were stored at -70 °C until testing. A MIC of at least 4 mg/mL was used as a cut-off point to determine resistant strains. Escherichia coli ATCC 25922 and Serratia spp. were used as controls.

A MDR microorganism was defined as a microorganism not susceptible to at least one antimicrobial agent from three unrelated antimicrobial groups.

Clonality between K. pneumoniae isolates was defined according to the previously reported technique of pulsed-field gel electrophoresis (PFGE) ^{( 9}. We considered that strains with an identity level of 100% were from the same clone, and those with an identity higher than 80% were clonally related ¹².

We confirmed the presence of carbapenemase-coding genes by the multiple polymerase chain reaction test (PCR Multiplex) for bla _NDM, bla _VIM, bla _IMP, bla _KPC, and bla _OXA-48 genes ¹³ in samples with low sensitivity to carbapenems. We carried out a PCR test to identify mcr-1 to mcr-5 genes, which are related to plasmid colistin resistance ¹⁴.

The Institutional Scientific Committee of Universidad Científica del Sur and the Scientific Committee of Instituto Nacional Materno Perinatal reviewed and approved this study.

FINDINGS

Molecular analysis of resistance to carbapenems and colistin

Two genes (bla _KPC and bla _NDM) were detected in the strains with low sensitivity to carbapenems. The strains were isolated from stool culture (code K80), and secretions (K81), respectively, both from 2018 (Figure 1).

Regarding the colistin resistant strains, PCR did not show the presence of any mcr gene.

Figure 1 Dendrogram of the clonal relation and antimicrobial sensitivity profile of the Klebsiella pneumoniae strains analyzed in this study. 

DISCUSSION

This study demonstrates the presence of multidrug resistant strains of K. pneumoniae that are resistant to colistin, isolated from patients of Instituto Materno Perinatal during 2018. We found that this is not due to plasmid genes (mcr), at least not from mcr1-5.

Overall, antimicrobial resistance rates among common pathogens in Lima hospitals are high ¹⁵. Hospital outbreaks of ESBL-producing Klebsiella, as well as enterobacteria, have been steadily increasing in Latin America. High incidence of ESBL (50%) has been recently described in bacteremia isolates from hospitalized patients in Peru. And the data for community-acquired infections is similar, incidence is close to 40% ¹⁵, which suggests, according to the reported findings, that presence of ESBL in the population is high ¹⁵. It should be noted that among the different pathogens reported, Klebsiella spp. leads with reported rates of up to 75% of resistance to third generation cephalosporins in the region ¹⁶.

Resistance to carbapenems is increasing in Peru, consistent with the growing trend reported worldwide ^{17 - 18}. We found the bla _KPC gene in several strains of this study, which was reported for the first time in Peru in 2013 in a Klebsiella strain ¹⁷. We also identified an isolate with the bla _NDM gene, which is increasingly reported in Peru ^{( 18}.

No colistin-resistant strains have been reported in the period 2015-2017, but in 2018, 5 isolates were found and were classified in 4 unrelated clones/pulse types. These results are aligned with the increased use of this antimicrobial in the hospital, and alert that resistant strains are already circulating in the institution. Previous studies in Peru report a level of 5% of resistance to this antimicrobial ⁷, and outbreaks of bla _KPC-producing and colistin-resistant Klebsiella strains have already been reported in the region ¹⁹.

Generally, resistance to colistin correlates with decreased affinity of polymyxin antibiotics for the lipid A of lipopolysaccharides. In this study, no plasmid-associated genes for colistin resistance were found, at least those tested for mcr-1 through 5. This suggests that colistin resistance is due to chromosomal mutations associated with changes in the two components (pmrAB and phoPQ) and their regulator mgrB ²⁰. In a Peruvian study from 2018, the mcr-1 gene was detected in E. coli strains from urine cultures, but with low frequency ²⁰; however, we suggested to monitor plasmid mechanisms for their ability to spread between different bacterial species. Currently, not all public hospitals monitor colistin resistance, so it seems likely that levels of resistance in Peru are underestimated.

Although this study contains a reduced sample size from a single institution, its importance lies in describing the presence of colistin-resistant strains, not by transferable mechanisms, but by widespread antimicrobial use as suggested by the clonality study. Recently, the Peruvian Ministry of Agriculture and Irrigation has prohibited the use, import, marketing, manufacture, or processing of veterinary products containing colistin (polymyxin E) to avoid the selective pressure of this antimicrobial, because it is a top priority for medical use.

In conclusion, we found a high percentage of colistin-resistant strains, not clonally related. It is important to note that this antimicrobial is part of the last-resort treatment alternatives. Therefore, it is urgent to establish screening systems and epidemiological surveillance of colistin resistance in all hospitals in the country.