Investigation of in-vitro susceptibility of multidrug-resistant Acinetobacter baumannii strains isolated from clinical specimens to tigecycline

Th e management of infections due to A. baumannii is diffi cult because of rapidly developing resistance, however, tigecycline, a glycylcycline antimicrobial, is in use for several years. In the present study, it was aimed to determine the susceptibility rates of A. baumannii to tigecycline. A total of  A. baumanni isolates were tested using three methods such as disk diff usion, broth microdilution, and E-test. Th e MIC and MIC values and the MIC range were found as  μg/ml,  μg/ml, and .- μg/ml by microdilution; and  μg/ml,  μg/ml, and .- μg/ml by E-test, respectively. Th ere were a few major errors as well as the minor rates were all high as between .-.. Th e accuracy rates between the methods were low as . (/) between disk diff usion and E-test, . (/) between disk diff usion and microdilution, and . (/) between E-test and microdilution. In the ROC curve analysis, an inhibition zone diameter of susceptibility breakpoint of . mm had sensitivity between .-.; specifi city between .-.; and accuracy between .-.. An analysis based on EUCAST’s nonspecies breakpoints, the MIC tests showed higher accuracy with a rate of ., however, performance of disk diff usion got worse as lower than . In conclusion, we showed that the reliability of the methods even did not remain as high as the past. Our study presented that none of three methods revealed reliable results in determination of susceptibility of A. baumanni to tigecycline, so the clinical response should be followed up carefully in such cases. ©  Association of Basic Medical Sciences of FB&H. All rights reserved


INTRODUCTION
Acinetobacter baumannii is primarily associated with nosocomial infections particularly in patients with ventilatorassociated pneumoniae and bacteremia and particularly in intensive care units (ICUs).Nosocomial infections due to this species is increasing in means of frequency with high rates of mortality particularly in critically ill patients [, ].Th e management of infections due to A. baumannii is difficult because of rapidly developing resistance in this species.In addition, A. baumannii exhibits resistance to multi drug groups such as carbapenems, aminoglycosides and tetracyclines, meaning of reducing the therapeutic options.However tigecycline, a glycylcycline antimicrobial, is in use for several years with a mechanism of entering the bacterial cell through energy-dependent pathways or via passive diffusion, and then binding to the subunit S of the ribosomes, resulting the inhibition of protein syntesis of the microorganism.As a result tigecycline can escape from tetracycline effl ux mechanism of the bacteria, meaning of causing slower and lower resistance in populations of A. baumannii [, , ].Besides this, by the time tigecycline has been used in treatment of infections due to A. baumannii, resistance rates have been increasingly reported.In addition, the susceptibility breakpoints of inhibition zone diameter of tigecycline using disk diff usion test has changed over years.Th e question is whether the disk diff usion test still remains reliable [, ].In the present study, it was aimed to determine the susceptibility rates of A. baumannii to tigecycline using three methods of disk diffusion, E-test and broth microdilution.

Samples
In the present study, a total of  multidrug-resistant A. baumannii isolates were used for susceptibility tests.The isolates were obtained from cultures of respiratory tract specimens ( isolates), blood ( isolates), urine ( isolates), and wound ( isolates)  For all the tests, isolates growth overnight and fresh as less than six hours manganese cation adjusted Mueller Hinton agar media were used.Th e MIC values and zone diameters were evaluated using both the previous criteria of British Society for Antimicrobial Chemotherapy (BSAC) for A. baumannii and European Committee on Antimicrobial Susceptibility Testing (EUCAST) for non-species testing [, ].

Statistical Analysis
A statistical analysis was performed using IBM SPSS Statistics Version  (SPSS Inc., Chicago, IL, USA).Continuous variables were tested for normality using the Shapiro-Wilk test.The breakpoints of disk diffusion zone diameters for predicting the susceptibility according to E-test and microdilution were analyzed using receiver operating characteristic (ROC) curve analysis.Th e sensitivity and specifi city were presented when a signifi cant cut-off value was observed.A p value of less than . was considered statistically signifi cant.In addition, a major error was defi ned for the isolate that was found as susceptible by a method and as resistant by another method.In addition, minor error was defi ned for the isolate that was determined as intermediate by a method and as susceptible or resistant by another method.Th e accuracy rate was counted by division of true determination of the susceptibility according to each other of the methods by the total number.

RESULTS
Th e MIC  and MIC  values and the MIC range were found as  μg/ml,  μg/ml, and .- μg/ml by microdilution; and  μg/ml,  μg/ml, and .- μg/ml by E-test (Table ).Ac-cording to the methods the rates of susceptibility versus resistance were found as    An analysis based on EUCAST's non-species breakpoints, the MIC tests showed higher accuracy with a rate of ., however, performance of disk diffusion got worse as lower than  (Table ).In addition, no ROC curve analysis could be done for EUCAST criteria due to the low number of susceptible isolates.

DISCUSSION
Determination of susceptibility of the therapeutic agent that has been in a limited number of choise against A. baumannii is crucial particularly for critically ill inpatients.
The method used on this topic has to be reliable and repeatable within the microbiology laboratories.Tigecycline has been used for treatment of this microorganism for several years as an efficient antimicrobial.However, the susceptibility rates have been reported to becoming changed [, , , ].In the present study we showed that the reliability of the methods even did not remain as high as the past.We found that none of the three methods we used in the study was as accurate as we can trust.
The MIC  values for tigecycline were reported as . μg/ ml by Souli et al. [], Th amlikitkul et al. [] and Seifert et al. [];  μg/ml by Draghi et al. [] and Scheetz et al. []; as  μg/ml by Song et al. [], Tan and Ng [], Mezzatesta et al. [], and Ratnam et al. [].Our MIC  value seems to be concordant with these reports.However, the MIC  values were found as  μg/ml by Sheetz et al. [], Halstead et al. [], Hohan et al. [], Ratnam et al. [], Mezzatesta et al. [], and Draghi et al. []; and as  μg/ml by Tan and Ng [], and Song et al. [].Our MIC  value determined by  microdilution was also  μg/ml, but E-test showed a value of  μg/ml as the highest among these reports.Th ese fi ndings show the increase in MIC values over the years.Besides this, it was reported that some of the studies had revealed an overestimation of antimicrobial activity of tigecycline if more conservative previous BSAC breakpoints (≤ μg/ml) was used compared with the one (≤ μg/ml) that is widely accepted [].
In the fi rst years of use of tigecycline in infections caused by A. baumannii, the inhibition zone diameter breakpoints were recommended as  mm and  mm.However in the study by Kulah et al. [] the recommendations for the breakpoints were  mm and  mm.Today, we seem to be so far away from these points.In the present study, we performed ROC analysis to determine a reliable zone diameter breakpoint according to either microdilution and E-test methods.However, a higher breakpoint of . mm in comparison to the past a few years' studies have still remained not so strong enough with its low sensitivities (Between .-.)and specifi cities (Between .-.).Th ese fi ndings showed that disk diff usion is not reliable any more for the determination of susceptibility to tigecycline.It was mentioned that the determination of the antimicrobial activity of tigecycline might vary with the use of diff erent methods, and disk diffusion method could give lower susceptibility rates when compared to E-test or broth microdilution [, , ].In the present study, we found so diff erent susceptibility rates amongst the methods.Th e accuracy rates within the methods were low as between approximately -.Th e disk diffusion test showed the worst performance on this topic.Besides this, the accuracy rate between microdilution and E-test methods, both of that are known to be more reliable methods, was the best with being not so high to be accepted, however, the highest major error rate was also observed between these two methods.Thamlikitkul et al.
[] also found diff erent MIC  and MIC  values by microdilution and E-test methods as being four folds as . μg/ml vs.  μg/ml, and  μg/ml vs.  μg/ml, respectively.Th ese results also showed the low reliability of all the three methods.Th e susceptibility testing criteria for tigecycline has changed in the last years.CLSI never suggested any breakpoints as well as EUCAST has now pulled back their criteria, and BSAC doesn't recommend any breakpoints with directing the reseachers to EUCAST's breakpoint for non-species testing to use [, ].In this aspect, the number of susceptible isolates were decreased as lower than , and the statistics has been biased.In this analysis, disk diff usion had accuracy rates below .However, microdilution and Etest methods showed greater concordance as ..Th ese aspects support that there has been an invalidation on susceptibility testing of tigecycline against A. baumanni.

CONCLUSION
Our study presented that neither disk diff usion test nor microdilution and E-test methods have lack of reliability in determination of susceptibility of A. baumanni to tigecycline.All the three methods revealed inaccordant results to each other.
In accord with the recommendations of BSAC, EUCAST, and CLSI, we consider that none of the susceptibility tests and interpretive criteria can give accurate results, so the clinical response should be followed up carefully in cases of giving tigecycline to a patient with an infection caused by A. baumannii.

TABLE 1 .
The Distribution of the MIC values according to the methods.

TABLE 2 .
The analysis of susceptibility within the methods according to the previous BSAC criteria.

TABLE 3 .
The analysis of susceptibility within the methods according to EUCAST criteria for non-species testing.TAS ET AL.: INVESTIGATION OF INVITRO SUSCEPTIBILITY OF MULTIDRUGRESISTANT ACINETOBACTER BAUMANNII STRAINS ISOLATED FROM CLINICAL SPECIMENS TO TIGECYCLINE S: Susceptible, I: Intermediate, R: Resistant.TEKIN