Anam Imtiaz  ( Department of Microbiology, Combined Military Hospital (CMH), Lahore, Pakistan )
Aamer Ikram  ( National Institute of Health (NIH), Islamabad, Pakistan )
Gohar Zaman  ( Department of Microbiology, Armed Forces Institute of Pathology (AFIP), Rawalpindi, Pakistan )
Luqman Satti  ( Department of Microbiology, Armed Forces Institute of Pathology (AFIP), Rawalpindi, Pakistan )
Fatima Sana  ( Department of Microbiology, Armed Forces Institute of Pathology (AFIP), Rawalpindi, Pakistan )

Objective: To evaluate a direct antibiotic susceptibility testing method for blood culture.
Methods: The cross-sectional comparative study was conducted at the Armed Forces Institute of Pathology, Rawalpindi, Pakistan, from December 2016 to October 2017. Direct antimicrobial susceptibility testing was performed from positive blood culture bottles. Bacterial identification was done by using API 10S. Different antimicrobial panels were employed for Gram-negative rods (GNRs), gram-positive cocci (like suspected Staphylococci and Enterococci). Results were compared with conventional disk diffusion testing and very major, major and minor errors were calculated. Result agreement and kappa coefficient scores were generated for categorical agreement. SPSS 24 was used for data analysis.
Results: Of the 101 bacterial isolates, 82(81.2%) were Gram negative rods and 19(18.8%) were Grampositive cocci. Among 781 bacteria-antibiotic comparisons, the number of very major errors was 3(0.4%), major errors were 7(0.9%) and minor errors were 12(1.5%), while, 759(97.2%) comparisons yielded the same results. The kappa coefficient was 0.946, showing almost perfect agreement. Direct identification of Gram negative rods was successful in 53(64.6%) cases.
Conclusion: Direct susceptibility testing of blood culture produced reliable results for majority of the antimicrobials.
Keywords: Direct susceptibility testing, DST, Disk diffusion testing, Blood stream infections, Blood culture, Antibiotic susceptibility testing, Antimicrobial resistance. (JPMA 70: 105; 2020). https://doi.org/10.5455/JPMA.10801

Introduction

Blood stream infections (BSIs) are an important cause of mortality in hospitalised patients. Early antibiotic treatment remains the cornerstone of management in cases of sepsis. The role of a clinical microbiologist is imperative in aidingthe physician in laboratory diagnosis and in being a guide to appropriate antibiotic treatment. Early institution of specific antimicrobial therapy not only reduces mortality and morbidity, but also saves the cost of treatment, decreases hospital stay and prevents development of antibiotic resistance.1,2 Routine blood culture takes 1-5 days for a positive result, and a further 24-48 hours are required for bacterial identification and antibiotic susceptibility testing. Till the time results are available, clinicians rely on empirical
antimicrobial treatment which may not be effective in cases of multi and extensively drug-resistant nosocomial infections.3 Thus, there is always a need to make the laborator y diagnoses more efficient. Antibiotic susceptibility testing performed directly on positive cultures can provide results 48 hours in advance and, thus, can save critical time in patient management. In an effort to develop methods for rapid antibiotic susceptibility testing various options, including matrix-assisted laser desorption ionisation-time of flight mass spectrometry(MALDI-TOF MS) Vitek, Sensititre, QuickFISH, BC-GP assay etc., have been evaluated with good results. All of these use sophisticated systems which are not widely available in a resource-scarce setting.3-5 There is enough ground rationale to assess the accuracy of direct susceptibility testing (DST) method for blood cultures to reduce time for preliminary reporting of antibiotic susceptibility results. Many clinical laboratories have adopted different methods, but there is no validated method. Studies from Pakistan are lacking in this regard. The current study was planned to evaluate a simple, cost-effective method for direct antibiotic sensitivity testing and bacterial identification, done directly from positive blood culture bottles.

Materials and Methods

The cross-sectional comparative study was conducted at the Department of Microbiology, Armed Forces Institute of Pathology (AFIP), Rawalpindi, Pakistan, from December 2016 to October 2017. After approval from the institutional ethics board, samples for blood culture received from two tertiary care hospitals and medical institutions were evaluated, and samples that were positive on direct Gram stain were included through non-probability convenience sampling. Repeat or duplicate samples from the same patients were excluded, and so were samples showing more than one microorganism and/or yeast cells on Gram stain. Sample size was calculated by taking the prevalence of positive blood cultures from a local study by using the sample size calculation software.6,7 The media used for blood culture included commercially available Bactec, BacT/ALERT and in-house prepared Brain Heart Infusion (BHI) broth (Oxoid, UK). For the standard method, in case of automated systems, Bactec and BacT/ALERT 3D, positive growth signal was indicated by the system. Blood samples collected in BHI broth bottles were incubated at 35±20C. These bottles were examined daily for any turbidity, clot formation and/or gas production for 7 days before terminal subculture. Blind sub-culturing was performed at days 1, 3 and 7.8 Standard drug susceptibility was performed as per Clinical and Laboratory Standards Institute (CLSI) guidelines by Kirby-Bauer disk diffusion method on Muller- Hilton agar (Oxoid, UK).9 For DST, the positive blood culture bottles were confirmed by Gram stain for the presence of microorganisms. The inoculum for direct testing was prepared by taking 1ml of broth from positive blood culture bottles with a needle and syringe and mixing with 1ml of BHI. This was incubated at 35±20C for 2 hours. The turbidity of the broth mixture was adjusted with 0.9% normal saline to visually match 0.5 McFarland barium sulfate (BaSO4) standards. The standardised inoculum was swabbed onto Muller-Hilton (Oxoid, UK) plate in three directions and antibiotic disks were applied. Antibiotic disk panel (Oxoid, UK) applied was selected according to the Gram stain results as follows; For Gram-negative rods (GNR)/coccobacilli representative antibiotics from -lactams, -lactam/-lactamase inhibitor combinations, aminoglycosides, fluoroquinolones, tetracyclines and folate pathway inhibitors were used. The drugs included ampicillin (10μg), ciprofloxacin (5μg), amikacin (30μg), ceftriaxone (30μg), meropenem (10μg), minocycline (30μg), amoxicillin-clavulanate (20/10μg) and tr imethoprim- sul fameth oxaz ol e (1.2 5/23.7 5μg). A different panel was used for GNR (suspected typhoidal Salmonellae from clinical history). The panel included first-line and second-line anti-typhoidal drugs; ampicillin (10μg), trimethoprim-sulfamethoxazole (1.25/23.75μg), chloramphenicol (30μg), ciprofloxacin (5μg) and pefloxacin (5μg), ceftriaxone (30μg) and azithromycin (15μg). For Gram-positive cocci (GPC), the antimicrobial groups used included penicillins, fluoroquinolones, tetracyclines, glycopeptides, aminoglycosides, lincosamides, phenicols and folate pathway inhibitors. The disks used for GPC in clusters (suspected staphylococci) were cefoxitin (30μg) for detection methicillin resistance, penicillin (10u), clindamycin (2μg), teicoplanin (30μg), trimethoprimsulfamethoxazole (1.25/23.75μg), doxycycline (30μg), ciprofloxacin (5μg) and amikacin (30μg). GPC (suspected enterococci) were tested against ampicillin (10μg), penicillin (10u), vancomycin (30μg), linezolid (30μg) and doxycycline (30μg). Salmonella was considered a separate group from enterobacteriaceae because of the different antimicrobial panel used. Clinical history was obtained from the treating physician. Zones of inhibition were recorded with transmitted light, using a ruler to the nearest 1mm for both direct sensitivity and the standard testing. CLSI9 interpretative criteria were used. The results of direct testing were compared with the standard method using the following criteria; Same (S): No change in results of both direct and standard method; Minor error (Min): Sensitive or resistant by the conventional method but intermediate by the direct method and vice versa; Major error (Maj): Sensitive by the conventional method but resistant by the direct method; Very major error (VME): Sensitive by the direct method but resistant by conventional method.10 For standard bacterial identification API 20E and 20NE (bioMérieux, France) was used for GNRs. Staphylococci were identified with coagulase and DNAse test as Staphylococcus aureus and coagulas e-negative
Staphylococci (CoNS). Enterococci were identified on biochemical reactions of API Strept (bioMérieux, France).8 For direct identification of GNRs the inoculum prepared for DST was inoculated onto API 10S (bioMérieux, France). Reactions were read after 18- 24 hours incubation at 35±20C and results were compared with the standard method. Staphylococcus aureus ATCC 25923, Escherichia coli ATCC 25922 and Pseudomonas aeruginosa ATCC 27853 were used as control strains. Data was analysed using SPSS 24. Descriptive statistics were calculated. Frequenc ies and percentages of isolation of the different species and the percentage of errors were calculated. Simple percentage agreement, as well as Kappa scores, were generated for categorical agreement. The following standards were used for the strength of agreement for the kappa coefficient: 0-0.0099, poor; 0.01-0.20, slight; 0.21-0.40, fair; 0.41-0.60, moderate; 0.61-0.80, substantial; and 0.81-1, almost perfect.

Results

Of the 101 bacterial isolates, 70(69.3%) were positive through automated systems and 31(30.7%) through BHI broth. Among the positive BHI bottles, turbidity alone was the indicator in 22(71%) isolates and turbidity with gas formation (bubbles) was noted in the rest. Distribution frequency of the isolates was noted ( Table 1).



The majority of blood stream infections were caused by members of Enterobacteriaceae 57(56.4%), with Salmonella Typhi 18(17.8%) being the most frequent isolate. Of the total isolates, 82(81.2%) were GNRs and 19(18.8%) were GPC. A total of 781 bacteria-antibiotic disk tests were carried out for the 101 isolates, and comparisons were made between direct and standard methods. The number of VMEs was 3(0.4%), Maj 7 (0.9%) and Min errors were 12(1.5%), while 759(97.2%) comparisons yielded the same results. The kappa coefficient was 0.946 (standard error[SE]=0.011) which meant almost perfect agreement. GNRs and GPCs were enumerated separately (Tables 2-3).





Overall error type distribution along with kappa coefficient was done separately (Table 4).



Of the 82 GNRs, 53(64.6%) were correctly identified by direct inoculation of API 10S. However, for Enterobacteriaceae, including Salmonella, 53(92.3%) out of 57 isolates were identified correctly.

Discussion

Early identification of the causative agent of blood stream infections not only helps the initiation of appropriate antibiotic therapy, but also earlier institution of infection control measures in case of multi and extensively drugresistant organisms. Protocols for direct sensitivity for blood culture have beenproposed in various studies with good results. We used a method that was proposed by Mirret et al., with the difference of BHI broth instead of trypticase soy broth with 1% yeast extract, for inoculum standardisation.11 In our study with the standardised suspension the percentage of VMEs reached 0.4%, Maj were 0.9% and Min were 1.5%. The percentage agreement between the two methods was 97.2% with a kappa coefficient of 0.946, which showed almost perfect agreement. A study from Singapore reported 0.03% Maj and 2.3% Min errors with inoculum standardisation.12 With a non-standardised inoculum, Noman et al. reported no VMEs, while 1.3%V MEs were reported in a Turkish study. 10, 13 With a standardised inoculum for direct antibiotic testing, Mirret et al. reported an overall 94.6% agreement between the direct and standard methods.11 In a study conducted in India that used the British Society for Antimicrobial Chemotherapy (BSAC) methods for antimicrobial susceptibility testing 2011-12, 83.7% comparisons between direct and standard method had the same result.14,15 Further studies are required which simultaneously compare the two methods of direct testing i.e., with and without the standardised inoculum, to demonstrate superiority of any one of these methods. Theoretically, a standardised inoculum should give more reliable results, as dilution of the broth during standardisation should achieve a better semi-confluent growth on the media and also dilute any inhibitors from blood. The CLSI methods development and standardisation working group tested broth directly from blood culture bottles, using 4 drops of blood culture broth to swab Muller-Hilton plates for DST.16 It reported an overall agreement of 87.4% which is lower compared to our study. It also compared the effects of bacterial colony counts on the results of DST and reported improved results with dilution of the blood culture broth inoculum.16 Edelmann et al. recommend a 0.5ml inoculum for GPCs and 0.2ml for GNRs to obtain a semi-confluent growth.17 A semi-confluent growth on the sensitivity plates appears to be an important factor for accurate results. Further studes are needed to be sure if volume has a major impact or not. We did not standardise the inoculum volume in our study. We lawned the agar plates by dipping sterile cotton swab into the standardised suspension, removing the excess by rotating against the wall of the tube above the liquid. We observed that most of the errors were with betalactam drugs and aminoglycosides, which was the case in other studies as well.10-13 An interesting finding in our study was 100% agreement in the susceptibility testing of Salmonella Typhi and Paratyphi with standard disk diffusion method. Similarly, there was 100% agreement in susceptibility results of c efoxi tin for detec tion of methic il lin -res istant Staphylococcus aureus (MRSA). This is concurrent with another study from Pakistan.12 These findings have a significant clinical implication in our setting as both salmonella and MRSA are common pathogens; early and accurate susceptibility testing by direct method may help in early initiation of therapy and infection control measures. The recent epidemic of ceftriaxone-resistant Salmonella in Hyderabad further highlights the importance of direct testing for the institution of early antibiotic treatment. API 10S was used for primary identification of GNRs, and it was successful in the identification of Enterobacteriaceae, including Salmonella, in 53(92.3%) cases. The API 10S was used for its low cost. None of the other studies have used API for direct bacterial identification. Further multi-centre studies with large number of isolates and carefully selected anti-microbial panels are needed to evaluate the usefulness before recommending it for routine use.

Conclusion

Direct susceptibility testing of blood culture had reliable results for majority of the antimicrobials. Preliminary reporting of DST results can help guide the clinician towards the more effec tive and less expensive antimicrobial treatment, until the results of standard method are available.

Disclaimer: None.
Conflicts of interest: None.
Source of Funding: None.

References

1. Perez KK, Olsen RJ, Musick WL, Cernoch PL, Davis JR, Land GA, et al. Integrating rapid pathogen identification and antimicrobial stewardship significantly decreases hospital costs. Arch Pathol Lab Med 2013;137:1247-54.
2. deCueto M, Ceballos E, Martinez-Martinez L, Perea EJ, Pascual A. Use of positive blood cultures for direct identification and susceptibility testing with the Vitek 2 system. J Clin Microbiol 2004;42:3734-8.
3. Machen A, Drake T, Wang YF. Same day identification and full panel antimicrobial susceptibility testing of bacteria from positive blood culture bottles made possible by a combined lysis-filtration method with MALDI-TOF Vitek mass spectrometry and the vitek2 system. PLoS One 2014;9:e87870.
4. Martinez RM, Bauerle ER, Fang FC, Butler-Wu SM. Evaluation of three rapid diagnostic methods for direct identification of microorganisms in positive blood cultures. J Clin Microbiol 2014;52:2521-9.
5. Clerc O, Prod'hom G, Vogne C, Bizzini A, Calandra T, Greub G. Impact of matrix-assisted laser desorption ionization time-of-flight mass spectrometry on the clinical management of patients with gramnegative bacteremia: a prospective observational study. Clin Infect Dis 2013;56:1101-7.
6. Qureshi M, Aziz F. Prevalence of microbial isolates in blood cultures and their antimicrobial susceptibility profiles. Biomedica 2011;27:136- 9.
7. Raosoft Inc. Sample size calculator. [Online] 2004 [Cited 2016 December 10]. Available from URL: www.raos oft.com/ samplesize.html.
8. Leber AL. Clinical microbiology procedures handbook. 4th ed. Washington DC: ASM Press, 2016.
9. CLSI. M100-S26 performance standards for antimicrobial susceptibility testing. 26th ed. Wayne, USA: CLSI, 2016.
10. ?ahin I, Sencan I, Kaya D, Öksüz ?, Gülcan A. Evaluation of direct and standard antimicrobial susceptibility testing methods on some bacteria isolated from automated blood cultures. Turk J of infect 2003;17:301-6.
11. Mirret S, Reller B. Comparison of direct and standard antimicrobial disk susceptibility testing for bacteria isolated from blood. J Clin Microbiol 1979;10:482-7.
12. Tan TY, Ng LSY, Kwang LL. Evaluation of disc susceptibility tests performed directly from positive blood cultures. J Clin Pathol 2007;61:343-6.
13. Noman F, Jehan A, Ahmed A. Reliability of direct sensitivity determination of blood cultures. J Coll Physicians Surg Pak 2008;18:660-1.
14. Goel G, Das D, Mukherjee S, Bose S, Das K, Mahato R, et al. A method for early detection of antibiotic resistance in positive blood cultures: Experience from an oncology centre in eastern India. Indian J Med Microbiol 2015;33:53-8.
15. Andrews JM. BSAC Working Party on Susceptibility Testing. BSAC standardized disc susceptibility testing method (version 5). J Antimicrob Chemother 2006;58:511-29.
16. Chandrasekaran S, Abbott A, Campeau S, Zimmer BL, Weinstein M, Thrupp L, et al. Direct-from-blood-culture disk diffusion to determine antimicrobial susceptibility of gram-negative bacteria: preliminary report from the clinical and laboratory standards institute methods development and standardization working group. J Clin Microbiol 2018;56:e01678-17.
17. Edelmann A, Pietzcker T, Wellinghausen N. Comparison of direct disk diffusion and standard microtitre broth dilution susceptibility testing of blood culture isolates. J Med Microbiol 2007;56:202-7.