Introduction

Burkholderia pseudomallei is the causative agent for melioidosis, a potentially fatal infection endemic in Southeast Asia and northern Australia. The current mortality rate remains high at 35–54% [1, 2] and no vaccine available. B. pseudomallei is intrinsically resistant to multiple antibiotics. The use of appropriate antibiotic treatment at early stage including initial phase, ceftazidime (CAZ) or meropenem (MEM) or imipenem (IPM), followed by eradication phase, trimethoprim-sulfamethoxazole (SXT), showed reduced mortality rate [3].

Initial in vitro antibiotic susceptibility testing (AST) screening plays an important role to decide appropriate antibiotics administered and avoid therapeutic failure. The gold standard is broth microdilution method (BMD), and there is no interpretative criteria for DD according to the Clinical and Laboratory Standards Institute M45 (CLSI) [4]. Despite that, disk diffusion testing (DD) is preferable for routine AST due to its low cost and being less laborious. Closely related species such as B. cepacia or Pseudomonas aeruginosa were usually referred to due to absence of standardized interpretative criteria [4, 5]. In this study, we aimed to evaluate the use of recently established EUCAST DD interpretative criteria [6, 7] to determine antibiotic resistance based on inhibition zone diameter breakpoints with reference to CLSI [4] for AST reporting of Malaysian B. pseudomallei clinical isolates. In addition, we also examine if the colony morphotypes on Ashdown agar (ASH) affect antibiotic susceptibility.

Methods

A total of 143 B. pseudomallei clinical isolates that included eight additional isolates (different colony morphotypes from same sample) were collected from 135 patients between year 2013 and 2018 from hospitals in Malaysia. The isolates were collected from various clinical samples including blood and pus (Table S1). The isolates were confirmed as B. pseudomallei by biochemical tests, PCR and typical appearance on Ashdown agar. All bacterial isolates were grown and tested in biosafety level 3 (BSL3) laboratory. The isolates were tested with both DD and BMD against CAZ, MEM, AMC and doxycycline (DOX). All antibiotic powders were purchased from Sigma, MCE and Santa Cruz Biotechnology and antibiotic disks from Oxoid. The quality control strains used were Escherichia coli ATCC25922, E. coli ATCC35218 and P. aeruginosa ATCC27853. The procedure was performed in triplicate as described [1]. Inhibition zone diameters by DD were compared with MIC values by BMD following interpretative criteria by EUCAST [6] and CLSI M45 [4], respectively, except MEM that utilizes EUCAST MIC breakpoints as there is no interpretative criterion for MEM in CLSI. Categorical comparisons of the AST results were performed using simple correlation and linear regression (SPSS Statistics, Version 23.0). All isolates were cultured on ASH for 4 days at 37 °C. The colony morphotypes for each isolate were recorded and grouped (types 1 to 7) based on criteria as described by Chantratita et al. [8].

Results

The MIC-inhibition zone diameter distributions for CAZ, MEM, AMC and DOX against B. pseudomallei are displayed (Fig. 1). The mode and distributions of DD zone diameters and BMD MIC are 28, 6–32 mm and 4, 1–128 μg/ml (CAZ); 26, 17–32 mm and 1, 0.25–4 μg/ml (MEM); 25, 20–28 mm and 8, 4–16 μg/ml (AMC); and 28 and 29, 24–33 mm and 2, 0.5–8 μg/ml (DOX) (Fig. 1 and Tables 1 and 2). EUCAST recommends the use of susceptible, increased exposure (SI) instead of intermediate-resistant (I) to indicate higher dose antibiotics required for the isolates resides between susceptible (S) and resistant (R) category [7]. In this study, (S), (SI) and (I) were used to indicate non-resistant isolates. DD produced good categorical agreements to differentiate antibiotic-resistant and non-resistant isolates, exhibiting concordance of 100% with BMD for CAZ and DOX, 98.6% for MEM and 97.2% for AMC. Pearson’s correlations between DD and BMD were 0.69 and 0.45 (p < 0.01) for CAZ and MEM, respectively. Sixty-two AMC-intermediate-resistant isolates (16 μg/ml) based on CLSI MIC cutoff were termed as non-resistant and within the (SI) range (22–49 mm) except four isolates interpreted as R by DD at 20–21 mm (resistant, < 22 mm). The MIC50 was 4 μg/ml, 1 μg/ml, 8 μg/ml and 2 μg/ml and MIC90 was 4 μg/ml, 2 μg/ml, 16 μg/ml and 2 μg/ml for CAZ, MEM, AMC and DOX, respectively.

Fig. 1
figure 1

Antibiotic MIC-inhibition zone diameter distributions of B. pseudomallei for A ceftazidime, B meropenem, C amoxicillin-clavulanate and D doxycycline. Corresponding MIC values are shown through the colouring of bars. Red bar indicates resistant (R) which correspond to CLSI M45 MIC breakpoints for B. pseudomallei except meropenem by EUCAST MIC breakpoint

Full size image
Table 1 Inhibition zone diameter distributions for B. pseudomallei clinical isolates against antibiotics, n = 143
Full size table
Table 2 Minimum inhibitory concentration (MIC) distributions for B. pseudomallei clinical isolates, n = 143
Full size table

In this study, 19 of 143 B. pseudomallei isolates do not grow on ASH. The most frequently observed colony morphotypes were types VI, III and IV each at > 20% with smooth surface in the centre of colony (Table 3). Further investigation on the antibiotic susceptibility profile showed that reduced antibiotic susceptibility for AMC was observed in all colony morphotypes except type V with highest frequency in type VI. CAZ-R, CAZ-I and MEM-R were observed in each of types VII, I and IV, respectively. Among 135 patients’ samples, an additional eight different B. pseudomallei colony morphotypes were observed on ASH from eight patients’ samples. The isolates were cultured from IMRS54 (A and B), IMRS71 (A and B), IMRS101 (M and W), IMRS103 (A and B), IMRS131 (P and D), IMRS166 (P and D), IMRS187 (A and B) and IMRS195 (A and B) (Table S1). Among the eight samples, combination of colony morphotypes III-VI were most commonly observed (37.5%) followed by III-IV (25%), II-VI (25%) and IV-VI (12.5%). All showed similar antibiotic susceptibility pattern between both colony morphotypes. The details of each colony morphotype and antibiotic susceptibility are described in Table S1.

Table 3 Colony morphotypes of B. pseudomallei clinical isolates on Ashdown agar, n = 124
Full size table

Discussion

This study compared EUCAST DD interpretative criteria to reference CLSI BMD cutoffs to determine antibiotic resistance for Malaysian clinical B. pseudomallei isolates. CLSI is the primary reference for bacterial AST reporting in local hospital setting [9]. There is no CLSI DD breakpoints for B. pseudomallei; hence, this data will be beneficial to standardize EUCAST DD method as the more feasible way of testing for melioidosis in local settings. The antibiotic susceptibility profile of different colony morphotypes was also evaluated. In Malaysia, CAZ is the first-line antibiotic for melioidosis treatment and MEM as backup option for severe infections followed by eradication therapy by SXT whilst AMC as alternative for both treatment phases [10]. We observed that the inhibition zone diameter proposed by EUCAST is in concordance with CLSI MIC breakpoints and suitable to be used to interpret AST by DD in our local setting. In comparison to earlier study by Ahmad et al., the mode MIC of CAZ and AMC has increased > twofold from 2 to 4 and 8 μg/ml, respectively. The AMC MIC90 has escalated to 16 μg/ml (AMC-I) in this study from 4 μg/ml (AMC-S) [5]. The reducing susceptibility to AMC (AMC-I, 44%) over the years is alarming and requires attention. The improper use of AMC may induce development of antibiotic resistance [11]. A recent survey from several centres (Table S2) [12, 13] showed similar inhibition zone diameter and MIC distributions with results generated from this study for all antibiotics tested except AMC and DOX that showed slightly higher MIC mode in this study. EUCAST employed CAZ disk content of 10 μg instead of 30 μg by CLSI for DD [7]. Despite that, we observed that CAZ breakpoint (resistant, < 18 mm) by EUCAST is able to differentiate wild-type from resistant isolates based on CLSI MIC cutoffs. We also observed that if EUCAST MIC breakpoints were applied for AMC, the AMC-R isolates (> 8 μg/ml) increase to 44% and only 59.4% agrees between DD and BMD. The discrepancy could be due to fixed amount of clavulanate at 2 μg/ml regardless of amoxicillin concentration according to EUCAST BMD whereas AMC were performed in 2:1 ratio of amoxicillin/clavulanate by CLSI BMD [4, 7].

The 19 non-growing B. pseudomallei isolates on ASH were from Sarawak in line with previous report on the presence of gentamicin-sensitive isolates in Sarawak. A non-synonymous amino acid change in amrB of the AmrAB-OprA efflux pump is reported to be the cause of the alteration in gentamicin susceptibility [14]. Majority of B. pseudomallei in this study (70%) present smooth centre-surface differed with Thailand observing predominantly rough centre-surface colony [8, 12]. Small and slow-growing B. pseudomallei colonies have been associated with decreased susceptibility to CAZ, MEM, DOX and trimethoprim-sulfamethoxazole [1, 15]. However, we do not observe the presence of these colonies in this study. The presence of > 1 colony morphotypes in clinical B. pseudomallei as observed in this study is not uncommon. Despite low number of samples, we have observed two different colony morphotypes from blood, pus and sputum which is similar to the report of Chantratita et al. [8]. The genetic characteristics of these isolates will warrant a further study. A limitation to this study is the limited information of the patients’ history and varying distributions of colony morphotypes to be associated with antibiotic resistance.

Conclusions

The outcome of this study supports the use of zone diameter breakpoints proposed by EUCAST as complementary to CLSI MIC cutoffs and applicable for B. pseudomallei AST interpretation in local setting where DD is used routinely. Overall, the prevalence of antibiotic resistance against melioidosis first-line antibiotics is low in Malaysian clinical B. pseudomallei isolates. Smooth centre-surface is the dominant colony morphotypes in Malaysia.