Comparison of the accuracy of disk diffusion zone diameters obtained by manual zone measurements to that by automated zone measurements to determine antimicrobial susceptibility

https://doi.org/10.1016/j.mimet.2008.05.020Get rights and content

Abstract

Although a variety of techniques are available for antimicrobial susceptibility testing, disk diffusion methods remain the most widely used. We compared the accuracy of disk diffusion zone diameters as obtained by manual zone measurements in a low resource country (Indonesia) to that by automated zone measurements (Oxoid aura image system) in a high resource setting (the Netherlands) to determine susceptibility categories (sensitive, intermediate susceptible or resistant). A total of 683 isolates were studied, including 294 Staphylococcus aureus, 195 Escherichia coli and 194 other Enterobacteriaceae. Antimicrobial agents included tetracycline, oxacillin, gentamicin, erythromycin, trimethoprim/sulfamethoxazole and chloramphenicol for S. aureus and ampicillin, gentamicin, cefotaxime, ciprofloxacin, trimethoprim/sulfamethoxazole, and chloramphenicol for E. coli and other Enterobacteriaceae. Of the 4098 drug–organism combinations, overall category agreement (CA), major discrepancy (MD) and minor discrepancy (mD) between the two methods were 82.4% (3379/4098), 6.0% (244/4098) and 11.6% (475/4098), respectively. One hundred and sixty three of 244 MDs were resolved using reference broth microdilution method. Overall very major error (VME), major error (ME) and minor error (mE) of manual zone measurement were 28.8%, 45.4% and 4.9%, respectively and for the aura image system 4.9%, 16.0% and 4.9%, respectively.

The results of this study indicate that the disk diffusion method with manual zone measurement in Indonesia is reliable for susceptibility testing. The use of an automated zone reader, such as the aura image system, will reduce the number of errors, and thus improve the accuracy of susceptibility test results for medically relevant bacteria.

Introduction

Antimicrobial susceptibility testing (AST) is one of the most important tasks performed by clinical microbiology laboratories. It is important for the choice of an antimicrobial drug for the treatment of a patient and for epidemiological monitoring. Although a variety of methodologies are available for detecting resistance to antimicrobials, disk diffusion techniques remain the most widely used (Felmingham and Brown, 2001). Whenever these are performed according to a standard such as the Clinical and Laboratory Standards Institute (CLSI; formerly the NCCLS) reference procedure, they are considered reliable methods. Furthermore, the disk diffusion method is cost effective, highly reproducible, and drug combinations can be changed easily (Berke and Tierno, 1996). However, measurement of zone sizes is tedious, time-consuming, and prone to transcription errors (Nijs et al., 2003).

In recent years, a variety of automated instruments for AST have been introduced. The advantages of automation include a higher degree of standardization resulting in an increased accuracy, improved data management with a concomitant reduction in transcription errors, earlier availability of results, and the possibility for the use of so-called “expert” software (Barry et al., 2003, Geiss and Klar, 2000, Korgenski and Daly, 1998, Nijs et al., 2003). Unfortunately, most automated systems may be too expensive for some laboratories in the Western world and for most laboratories in low resource countries, including Indonesia.

In medical microbiology laboratories in Indonesia the AST is still performed using the disk diffusion method and zones of inhibition are measured manually with a ruler or caliper. The aura image system (Oxoid, Basingstoke, UK) combines the advantages of disk diffusion and automation. By automatically performing zone measurements and giving interpretations against the user's chosen reference database it also obviates the aforementioned limitations of the disk diffusion method.

The present study was conducted in the course of a large population based study of antibiotic resistance among bacteria from patients and healthy individuals in Indonesia (Lestari et al., 2008). The purpose of our study was to compare the accuracy of disk diffusion zone diameters as obtained by manual zone measurements in Indonesia to that by automated zone measurements (Oxoid aura image system) in the Netherlands to determine interpretative categories. Discrepancies were analyzed using broth microdilution, a CLSI reference method.

Section snippets

Bacterial isolates

A total of 683 isolates were studied, including 294 Staphylococcus aureus, 195 Escherichia coli and 194 other Enterobacteriaceae. The collection of other Enterobacteriaceae consisted of 152 Klebsiella pneumoniae, 19 Enterobacter cloacae, 15 Klebsiella ozaenae, 4 Enterobacter amnigenus, 2 Enterobacter aerogenes, 1 Citrobacter koseri and 1 Citrobacter youngae. These isolates were collected in the framework of a population based study aimed at determining the prevalence of antimicrobial resistance

Results

Of the 4098 drug–organism combinations, overall CA, MD and mD between the two methods, manual zone measurement and aura image system, were 82.4% (3379/4098), 6.0% (244/4098) and 11.6% (475/4098), respectively. The coefficient of correlation between the test results from manual measurements and those from the aura image system was 0.63 (p < 0.001).

Table 2 shows the performance of the disk diffusion method with manual zone measurement compared to the aura image system for AST of S. aureus, E. coli,

Discussion

Automated image analysis systems for the reading of disk diffusion zone diameters have been evaluated in several countries (Geiss and Klar, 2000, Korgenski and Daly, 1998, Medeiros and Crellin, 2000, Nijs et al., 2003). In this study, we compared the accuracy of disk diffusion zone diameters as obtained by manual zone measurements in Indonesia to that by automated zone measurements (Oxoid aura image system) in the Netherlands for antimicrobial susceptibility testing of 4098 drug–organism

Acknowledgements

We thank the deans of the Medical Faculty, University of Airlangga, Surabaya, Indonesia and the Medical Faculty, University of Diponegoro, Semarang, Indonesia, the directors of the Dr. Soetomo Hospital, Surabaya, Indonesia and the Dr. Kariadi Hospital, Semarang, Indonesia, who have facilitated our work in these teaching hospitals.

We thank Oxoid (Basingstoke, UK) for kindly providing the aura image system for this research fellowship. The excellent technical assistance of Irma Kershof is

References (23)

  • FelminghamD. et al.

    Instrumentation in antimicrobial susceptibility testing

    J. Antimicrob. Chemother.

    (2001)
  • Cited by (15)

    • Intra- and inter-laboratory agreement of the disc diffusion assay for assessing antimicrobial susceptibility of porcine Escherichia coli

      2019, Preventive Veterinary Medicine
      Citation Excerpt :

      Knowledge of the extent of laboratory to laboratory variation is valuable as this information can be used to standardise the assay further to minimise measurement error. For instance, the findings presented here add weight to the call to find new ways to consistently measure zone-diameters, such as the adoption of automated zone readers which have been previously reported to eliminate much of the measurement uncertainty associated with manual reading (Lestari et al., 2008; Hombach et al., 2013; Idelevich et al., 2016). Reducing the level of variation in zone-diameter measurements will increase confidence in the assay, particularly for isolates exhibiting decreased susceptibility to important antimicrobials.

    • Shortening the incubation time for antimicrobial susceptibility testing by disk diffusion for Enterobacteriaceae: How short can it be and are the results accurate?

      2017, International Journal of Antimicrobial Agents
      Citation Excerpt :

      However, earlier reading is possible for every laboratory. In addition, because we show the errors in categorisation over time, each laboratory can make its own decision as to when to read and which errors are acceptable and whether reading is by an automated or conventional method [16,17]. When deciding on the acceptance of certain error rates, it should be realised, however, that we used, on purpose, a set of challenge strains with all different classes of β-lactamases, including carbapenemases.

    • Antibiogramj: A tool for analysing images from disk diffusion tests

      2017, Computer Methods and Programs in Biomedicine
      Citation Excerpt :

      There are two kinds of automated or semi-automated inhibition zone readers: integrated instruments that combine hardware and software; and software tools that work with plate-images captured with a camera device. For the first kind of inhibition zone readers, a variety of instruments have been introduced over the years to automate the measurement of inhibition zones in order to determine antimicrobial susceptibility [11,21–23]. Nowadays, these systems usually consists of three components: a holding area for inoculated plates, a robotic handling mechanism for the plates, and an image capture station.

    • Development of an automatic identification algorithm for antibiogram analysis

      2015, Computers in Biology and Medicine
      Citation Excerpt :

      However, strategies such as the use of a disc dispenser can help to standardize this procedure and simplify the task, mainly to standardize the discs in the same position when there is the need for numerous antibiograms. Nonetheless, a critical aspect of this procedure is the antibiogram reading step, after all while the manual zone measurement is reliable, the use of automated approach can reduce the number of errors and improve the accuracy of susceptibility test [8]. The measurement of the inhibition zone diameter is usually performed manually by specialists using a millimeter-scale ruler (Fig. 1A).

    • Predicting and Identifying Antimicrobial Resistance in the Marine Environment Using AI &amp; Machine Learning Algorithms

      2023, 2023 IEEE International Workshop on Metrology for the Sea; Learning to Measure Sea Health Parameters, MetroSea 2023 - Proceedings
    View all citing articles on Scopus

    Preliminary results of the study were presented at the European Congress of Clinical Microbiology and Infectious Diseases in Prague, 2004 (abstract R2171).

    1

    Members of the AMRIN study group. Dr. Soetomo Hospital, School of Medicine, Airlangga University Surabaya, Indonesia: Prof. Widjoseno Gardjito, M.D.; Erni P. Kolopaking, MPPM; Prof. Karjadi Wirjoatmodjo, M.D.; Prof. Djoko Roeshadi, M.D., Ph.D.; Prof. Eddy Suwandojo, M.D.; Prof. Eddy Rahardjo, M.D., Ph.D.; Prof. Ismoedijanto M.D., Ph.D.; Prof. Paul Tahalele, M.D., Ph.D.; Prof. Hendromartono, M.D., Ph.D.; Hari Parathon, M.D.; Usman Hadi, M.D.; Nun Zairina, Hosp. Pharm.; Mariyatul Qibtiyah, Hosp. Pharm.; Endang Isbandiati, M.D., Ph.D.; Kartuti Deborah, M.D.; K. Kuntaman, M.D. Ph.D.; Ni Made Mertaniasih, M.D., Ph.D.; Marijam Purwanta, M.Sc.; Lindawati Alimsardjono, M.D.; Maria Inge Lusida, M.D., Ph.D. Dr. Kariadi Hospital, School of Medicine, Diponegoro University Semarang, Indonesia : Prof. Ariawan Soejoenoes, M.D.; Budi Riyanto, M.D.; Hendro Wahjono, M.D., Ph.D.; Musrichan Adhisaputro, M.D.; Winarto, M.D.; Subakir, M.D.; Bambang Isbandrio, M.D.; Bambang Triwara, Hosp. Pharm.; Johnny Syoeib, M.D.; Endang Sri Lestari, M.D.; Bambang Wibowo, M.D.; Muchlis AU Sofro, M.D.; Helmia Farida, M.D., M.Sc.; M.M.D.E.A.H. Hapsari, M.D.; Tri Laksana Nugraha, M.D., M.Sc. Leiden University Medical Center, Leiden, The Netherlands: Prof. Peterhans van den Broek, M.D., Ph.D.; D. Offra Duerink, M.D., M.A. Erasmus MC University Medical Center, Rotterdam, The Netherlands: Prof. Henri A. Verbrugh, M.D., Ph.D.; Inge C. Gyssens, M.D., Ph.D. Radboud University Medical Center, Nijmegen, The Netherlands: Monique Keuter, M.D., Ph.D.

    View full text