Comparative analysis of antimicrobial susceptibility testing: Conventional versus direct from the blood culture bottle in a tertiary care center

Study design and setting

This prospective cohort study was conducted in the Department of Microbiology at UCMS and GTB Hospital, a tertiary care center in India. The study aimed to compare AST performed using conventional methods from pure culture growth with direct AST performed directly from positive blood culture bottles.

Study population

Patients admitted to the hospital with suspected BSIs were included in the study. Blood samples were collected only from patients who were clinically suspected of having BSIs and whose blood cultures showed bacterial growth. Patients with incomplete medical records or those with polymicrobial BSIs were excluded from the study.

A total of 250 positive blood cultures were included, with pathogens categorized into Enterobacterales and nonfermenters (Pseudomonas aeruginosa and Acinetobacter spp.) as per the Clinical and Laboratory Standards Institute (CLSI) guidelines. Non-fermenting Gram-negative bacilli included only P. aeruginosa and Acinetobacter spp., as these are the most frequently isolated non-fermenters in BSIs in our setting.

Sample collection and processing

Blood samples were collected aseptically and inoculated into BacT/ALERT^® blood culture bottles (bioMérieux, France). These bottles were incubated in the BacT/ALERT^® 3D-automated blood culture system (bioMérieux, France) and monitored continuously for bacterial growth. On positivity, the blood culture broth was processed for both direct and conventional AST.

AST

Direct AST

The positive blood culture broth was processed immediately for AST without prior subculture. A standardized inoculum was prepared by centrifugation and dilution to match a 0.5 McFarland standard, followed by disk diffusion testing on Mueller-Hinton agar, as per CLSI protocol for direct AST [Figure 1]. The same incubation period of 16-18 h and interpretative criteria were applied for direct AST.^[7]

Close

Figure 1: Flow chart for standardization of inoculum preparation for direct antimicrobial susceptibility testing. CLSI: Clinical & Laboratory Standards Institute

Export to PPT

For Enterobacterales, the antibiotic disks and potencies tested were ampicillin (10 µg), amoxicillin-clavulanate (20/10 µg), ceftriaxone (30 µg), ceftazidime (30 µg), cefepime (30 µg), piperacillin-tazobactam (100/10 µg), aztreonam (30 µg), ciprofloxacin (5 µg), cotrimoxazole (1.25/23.75 µg), gentamicin (10 µg), tobramycin (10 µg), and meropenem (10 µg).

For non-fermenting Gram-negative bacilli (P. aeruginosa and Acinetobacter spp.), the antibiotic panel included ceftazidime (30 µg), cefepime (30 µg), piperacillintazobactam (100/10 µg), aztreonam (30 µg), ciprofloxacin (5 µg), gentamicin (10 µg), tobramycin (10 µg), meropenem (10 µg), and cotrimoxazole (1.25/23.75 µg), as applicable. Antibiotic selection was based on the CLSI recommendations for bloodstream isolates and institutional antimicrobial stewardship protocols.

Data collection and outcome measures

The following parameters were recorded for both methods.

Statistical analysis

Descriptive statistics were used to summarize patient characteristics and laboratory findings. CA between methods was assessed using percentages, and error rates were calculated as per the CLSI guidelines. Continuous variables (e.g., TAT) were compared using the Student’s t-test or Mann-Whitney U test, while categorical variables were analyzed using the Chi-square test or Fisher’s exact test. A p <0.05 was considered statistically significant.

CA and error rates for enterobacterales

For Enterobacterales [Table 1], CA across antibiotics ranged from 94.4% (ceftriaxone) to 99.4% (aztreonam). The highest agreement was observed with aztreonam (99.4%), followed closely by cotrimoxazole (98.8%) and ciprofloxacin (98.3%). VMEs were rare, with ceftriaxone and ceftazidime showing a VME rate of 0.5% each. MEs were minimal, ranging from 0% (cotrimoxazole, aztreonam) to 1.6% (tobramycin). MinEs were most frequent for ceftriaxone (3.8%) and least for aztreonam (0.5%).

CA and error rates for non-fermenters

Among non-fermenters [Table 2], the overall CA was highest for ceftazidime, cotrimoxazole, and aztreonam, each achieving 100% agreement with no observed VMEs, MEs, or MinEs. Ciprofloxacin showed strong agreement at 98.5%, with a MinE rate of 1.4%. Meropenem demonstrated a CA of 92.8%, with slightly higher error rates, including 1.4% VMEs and MEs and 4.2% MinEs. Tobramycin showed a CA of 95.7%, with 1.4% MEs and 2.8% MinEs.

TAT analysis

In this study, the TATs for direct AST were compared with conventional AST across 250 samples [Figure 2].

Close

Figure 2: Turnaround time comparison: Direct antimicrobial susceptibility testing (AST) versus conventional AST (n = 250). TAT: Turnaround time

Export to PPT

Average TAT and variation across various infection types

Direct and conventional AST comparisons across infection types are summarized in Table 3. BSIs showed the highest concordance at 96%, with direct AST reducing the TAT from 28-32 h to 8-12 h. Respiratory infections demonstrated 94% concordance with a similar reduction from 30-36 h to 9-12 h. CNS infections showed 93% concordance, with direct AST decreasing reporting time from 32-38 h to 10-13 h. Urinary tract infections had 89% concordance, with TATs reduced from 24-30 h to 8-10 h. Gastrointestinal infections exhibited the lowest concordance at 87%, with direct AST shortening the time from 26-34 h to 9-11 h. Overall, direct AST consistently provided faster results across all infection categories while maintaining high concordance with conventional methods.

CA and error rates for Enterobacterales

Our findings indicate that CA across antibiotics for Enterobacterales ranged from 94.4% for ceftriaxone to 99.4% for aztreonam. Notably, aztreonam exhibited the highest agreement at 99.4%, followed by cotrimoxazole (98.8%) and ciprofloxacin (98.3%). VMEs were infrequent, with ceftriaxone and ceftazidime each showing a VME rate of 0.5%. MEs were minimal, ranging from 0% for cotrimoxazole and aztreonam to 1.6% for tobramycin. MinEs were most frequent for ceftriaxone (3.8%) and least for aztreonam (0.5%).

These results align with previous studies that have demonstrated high CA and low error rates for direct AST methods in Enterobacterales. For instance, a study by Jacobs et al.^[10] reported an overall CA of 92.3% for direct AST of Gram-negative bacteria from positive blood cultures, with VMEs and MEs at 0.4% and 1.9%, respectively.^[10]

CA and error rates for non-fermenters

Among non-fermenters, our study observed 100% CA for ceftazidime, cotrimoxazole, and aztreonam, with no detected VMEs, MEs, or MinEs. Ciprofloxacin also showed strong agreement at 98.5%, with a MinE rate of 1.4%. Meropenem demonstrated a CA of 92.8%, with slightly higher error rates, including 1.4% for both VMEs and MEs, and 4.2% for MinEs. Tobramycin exhibited a CA of 95.7%, with 1.4% MEs and 2.8% MinEs. These findings are consistent with other research in the field. For example, a study evaluating direct susceptibility testing protocols for Gram-negative bacilli reported high CA rates, particularly for non-fermenters, with minimal major and VMEs.^[11]

TAT analysis

The analysis of TATs revealed that conventional AST required between 18 and 32 h, with a mean of 28 h, reflecting variability due to subculturing and processing times. In contrast, direct AST significantly reduced the mean TAT to 9.7 h, ranging from 8 to 12.5 h, resulting in an average time savings of 18.3 h, equating to a 65% reduction.

This substantial decrease in TAT is clinically significant, as rapid availability of AST results can lead to more timely adjustments to appropriate antimicrobial therapy, thereby improving patient outcomes. Our findings are corroborated by other studies that have demonstrated the efficacy of direct AST methods in reducing TAT. For instance, a study by Khan et al.^[12] reported that direct AST from positive blood cultures reduced the time to appropriate antimicrobial treatment by 2 days compared to conventional methods.^[12]

When comparing our results with the study by Khan et al.,^[12]which evaluated the performance of direct AST methods, we observe similar trends in CA and error rates. Khan et al. reported CA rates ranging from 90.3% to 95.8% for various antibiotics, with VMEs and MEs at 0.4% and 1.9%, respectively.^[12] Our study’s findings of CA rates between 94.4% and 99.4% for Enterobacterales, and 92.8-100% for nonfermenters, with similarly low error rates, are in concordance with these results.

Furthermore, both studies highlight the significant reduction in TATs achieved through direct AST methods. Khan et al. reported a decrease in time from blood culture positivity to AST results from 20.0 h to 9.7 h, aligning closely with our findings of a reduction from a mean of 28 h to 9.7 h.^[12]

BSIs showed the highest clinical impact, with a 96% concordance and a reduction in AST time from 28-32 h to 8-12 h, aligning with previous studies emphasizing the importance of rapid AST in sepsis management.^[12] Similarly, respiratory infections (94% concordance) benefited from faster AST, facilitating early-targeted therapy and minimizing antibiotic resistance risks. Central nervous system infections, with a 93% concordance, particularly benefited from expedited AST, as early intervention is crucial in improving outcomes. Urinary tract (89%) and gastrointestinal infections (87%) also showed reduced TATs, supporting the clinical utility of direct AST in optimizing antibiotic use and patient care.

The present study adds important real-world evidence supporting the utility of direct disk diffusion AST from positive blood culture bottles in a routine tertiary care laboratory setting. While previous studies have demonstrated the feasibility of direct AST using automated systems or specialized rapid platforms, our findings demonstrate that a simple, low-cost disk diffusion-based direct AST approach can achieve high CA with conventional methods while significantly reducing TAT. In our study, CA exceeded 94% for Enterobacterales and 92% for non-fermenters, with very major and ME rates remaining well within CLSI-acceptable limits, findings comparable to those reported by Jacobs et al.^[10] and Khan et al.,^[12] despite methodological differences in inoculum preparation and testing platforms.

A key novel finding of this study is the consistent reduction in susceptibility reporting time by ~18 h using direct disk diffusion AST, without compromising accuracy. This magnitude of time savings is clinically meaningful in the management of BSIs, where delays in appropriate antimicrobial therapy are strongly associated with increased mortality and prolonged hospital stay. Unlike studies relying on automated AST systems, our approach demonstrates that significant gains in TAT can be achieved using existing laboratory infrastructure, making it particularly relevant for resource-limited and high-volume tertiary care settings.

Limitations and future recommendations

Although direct AST provided substantially earlier susceptibility results compared to conventional methods, data regarding subsequent clinical interventions, including antibiotic modification, de-escalation, or escalation following AST reporting, were not systematically collected in this study. Consequently, the direct impact of earlier susceptibility reporting on antimicrobial prescribing practices and patient outcomes could not be evaluated. This represents an important limitation of the present work. Future prospective studies integrating microbiological data with clinical decision-making and therapeutic changes are warranted to better define the clinical benefits of direct AST, including its effect on antimicrobial optimization, length of hospital stay, and patient survival.