Role of phenotypic testing in determining the mechanism of resistance in Gram-negative bacilli & risk factors for meropenem resistance

Sir,

Widespread use of antibiotics has led to the emergence of antimicrobial resistance (AMR) among Gram-negative bacteria (GNB). Detection of AMR is done by in vitro testing of the isolates. These susceptibility tests may sometimes fail to detect the mechanism of resistance, which will be revealed only by supplementary phenotypic tests or molecular methods1. There is a need for standardization of phenotypic testing to reveal all mechanisms of resistance that may be present using easy to perform in vitro tests. Detection of induced resistance in the presence of certain antibiotics like cefoxitin may alter treatment protocols in a significant way2. Hence knowing the mechanism by phenotypic methods, may be crucial in many clinical situations and will need to be reported in 48 h. Meropenem resistance is a recent phenomenon, and more studies are needed to determine the risk factors that may lead to this phenomenon3.

In this study, the primary objective was to find the role of available phenotypic tests for determining the mechanism of resistance, as a part of the routine susceptibility testing. The secondary objective was to determine probable risk factors for meropenem resistance. This study was done in Sree Chitra Tirunal Institute for Medical Sciences & Technology, Thiruvananthapuram, India, during July 2013 to July 2015.

All isolates from clinical specimens of in-patients with healthcare associated infections (HCAI)4, during the two years of the study were included. The study was approved by the Institutional Ethics Committee and written informed consent was taken from all participants. Standard methods that were chosen at the start of the study were followed for sensitivity testing and detection of extended-spectrum β-lactamase (ESBL), AmpC (cefoxitin disc for Enterobacteriaceae and antagonism with ceftazidime for Pseudomonas aeruginosa) and carbapenemase production [modified Hodge test (MHT) and ethylenediaminetetraacetic acid (EDTA)-imipenem]56. The phenotypic methods to detect the mechanisms of resistance included in this study were chosen for routine susceptibility testing789. Both sensitivity testing and mechanism of resistance were determined simultaneously without the need for a second day of testing. For carbapenemase detection, MHT and EDTA-imipenem test were done on the second day. When an isolate was found to be from a case of clinically proved HCAI, details of the patient were recorded. To look at the probable risk factors, patients with at least one isolate resistant to meropenem were compared with those with HCAI due to pathogens which were all sensitive to meropenem.

There were a total of 193 patients who had HCAI due to GNB, with 206 non-duplicate isolates excluding three isolates of Serratia and two isolates of Enterobacter. Of these, 91 (44.2%) isolates were sensitive to meropenem. Both Klebsiella pneumoniae and P. aeruginosa had an almost similar sensitivity rate for meropenem (40.4 and 50%, respectively). Of the 107 K. pneumoniae isolates, 65 were resistant to meropenem. Of these, only 32 were MHT positive. This included the 13 which were metallo-β-lactamase (MBL) positive and 11 had no zone in sensitivity testing. In 61 of 206 isolates (29.6%), no definite mechanism of resistance could be determined phenotypically, since there was no zone for any antibiotic and these were MHT negative (Table I). After the data collection was over, 29 isolates were tested for Klebsiella pneumoniae carbapenamase (KPC) and NDM1 genes by reverse transcription-PCR as recommended by the Centres for Disease Control, Atlanta10. Multiplex real-time PCR detection of KPC and NDM-1 was done on ABI 7500 Fast system using Taqman probes. Primers and probes used for the detection of KPC and NDM 1 were supplied by Origin Diagnostics and Research, Thiruvananthapuram, India. Both carbapenem sensitive and resistant isolates were included, to see if any seemingly sensitive ones carried this gene. Of the 29 isolates, seven were positive for the NDM-1 gene, showing the presence of this subtype of MBL. Six of these were K. pneumoniae. The phenotypic test showed three false-positives and one false-negative. The NDM-1 gene was seen in one isolate that was sensitive to imipenem by disc diffusion, for which phenotypic test was not done (as it was sensitive).

The most common and easily detectable mechanism of resistance was ESBL among Enterobacteriaceae. The ceftazidime disc was most consistent in the double disc synergy method when approximated with the ceftazidime-clavulanic acid disc. The same combination discs were used in a study in Chandigarh for urinary isolates11. AmpC detection using cefoxitin disc was incorporated in routine sensitivity as also the antagonism test for P. aeruginosa. Confirmation of AmpC production using the three-dimensional test was not done in our study as it was difficult to incorporate in routine sensitivity testing protocols. However, it was realized that this might miss the detection of AmpC production in combined AmpC and carbapenemase production. Co-production of AmpC and carbapenemase can be tested either by tris-EDTA disc or boronic acid, but false positivity has been reported1112.

In our study, the MHT was positive in 35 of 70 (50%) isolates of meropenem-resistant Enterobacteriaceae. Of these, 46 per cent were K. pneumoniae, similar to the International Nosocomial Infection Control Consortium report for 2007 and slightly higher than the Morbidity and Mortality Weekly Report in 20181314. Of the 35 carbapenemase producers, 16 (45.7%) were positive for MBL. The imipenem-EDTA disc test may be done with the routine sensitivity testing, but needs to be interpreted only when the isolate is resistant to both meropenem and imipenem. Its interpretation is challenging, as a mere increase in zone size of more than 5 mm is not enough15. The initial small zone should increase beyond the cut-off for resistance. In some cases, though a 5 mm increase was there, the final zone size was still less than sensitivity cut-off. In such cases, MBL was ruled out as the mechanism.

There were 44 (32.64%) Enterobacteriaceae isolates which had no zone to any disc of β-lactam antibiotics (of these, 11 were MHT positive). This could be due to a combination of mechanisms. Hence in an ordinary microbiology laboratory, it will not be possible to detect the exact mechanism of resistance in about one-third of Enterobacteriaceae isolates without molecular methods.

P. aeruginosa and Acinetobacter baumanii formed 33.9 per cent (70/206 isolates) of all isolates. These are characterized by resistance mechanisms that include not only β-lactamases such as CTX-M and TEM but also Types D (OXA), C (AmpC) and Type B (Metallo)1617. Of these, Types B and C could be detected phenotypically in 21.9 per cent (7/32) of isolates of P. aeruginosa and only about 13.2 per cent (5/38) of isolates of A. baumanii because the rest showed no zone at all. AmpC detection using the ceftazidime disc antagonism test in isolates of P. Aeruginosa that routinely appear susceptible to ceftazidime can be incorporated in routine testing, otherwise many strains may be erroneously reported as ceftazidime sensitive. Ceftazidime-phenyl boronic acid disc test detects non-inducible AmpC, in both the species, but it is not commercially available as in the case of imipenem-EDTA181920. Only seven isolates of P. aeruginosa and five of A. baumanii showed resistance to both meropenem and imipenem and were proved to be MBL producers by the imipenem-EDTA disc test. One isolate of A. baumanii among these showed the classical NDM-1 gene.

When all Gram-negatives isolates were considered, meropenem resistance was higher among those isolates obtained from patients who stayed in ICU (hospital), more than seven days as compared to those who stayed less than seven days in the ICU (hospital) (64% vs. 43%, P<0.05). The relationship between the risk factors and meropenem resistance was analyzed using the Chi-square test (Table II).

There were several limitations of the study. The phenotypic tests were evolved over the two years of the study, but since these were fixed at the beginning, these could not be changed to suit the changing criteria, brought out by CLSI, as it would affect the analysis. Among risk factors, to test for prior antibiotic use, in the context of HCAI it was difficult to get controls who were not on antibiotics. In case of a specific antibiotic causing resistance, only the 3^rd generation cephalosporins were in column sufficient number for comparison, but their use did not cause a significant difference in meropenem resistance.

In conclusion, phenotypic methods to detect mechanism of resistance need to be evolved further for routine use in sensitivity testing. Non-inducible AmpC and subtypes of MBL cannot be detected routinely. About 30 per cent of isolates could not be allocated to any particular resistance mechanism phenotypically, because there was no zone. To decrease resistance rates in GNB, standardization of methods and easier tests for determining the mechanism of resistance phenotypically are needed. The major factor associated with meropenem resistance in GNB was the period of stay in a healthcare facility for more than seven days. To look for risk factors for meropenem resistance, it is necessary to focus on any one mechanism of resistance and study a large number of isolates for that particular mechanism.