Dear Editor,

To shorten the turnaround time for bacterial identification from positive blood culture bottles, several commercial kits, such as the SepsiTyper Kit (Bruker Daltonics, Bremen, Germany) and SepsiPrep kit (ASTA Corp., Suwon, Korea), and in-house protocols for direct identification of pathogens from positive blood cultures have been used [1, 2]. However, positive blood cultures contain complex medium components as well as host cells and proteins, which can generate additional spectral peaks. Sodium dodecyl sulfate (SDS) can help remove culture medium components and host cells by disrupting membranes and denaturing proteins by breaking protein–protein interactions [3^-5]. We developed a modified protocol, incorporating an additional lysis step with a low concentration (0.1%) of SDS not to affect bacterial viability, and evaluated it for direct identification and direct antimicrobial susceptibility testing (AST) using the cell pellet.

In phase 1 of this multicenter study, we evaluated the modified protocol in comparison with the original SepsiPrep kit protocol for direct identification from positive blood culture bottles. The modified protocol has an additional lysis step with 1 mL of 0.1% SDS after the first centrifugation and discarding the supernatant [6]. This part of the study was conducted between May and June 2021 in three participating centers in Korea (Seoul St. Mary’s Hospital, Seoul; International St. Mary’s Hospital, Incheon; Samsung Medical Center, Seoul) using two types of blood culture bottles: BACTEC Plus (BD Diagnostics, Sparks, MD, USA) and BacT/Alert (bioMérieux, Marcy l’Étoile, France). The Institutional Review Board of each center approved this study (IRB No. Seoul St. Mary’s Hospital: KC18DNDI0866, International St. Mary’s Hospital, Incheon: IS18SISI0052, Samsung Medical Center: 2018-08-009). For cases identified as monomicrobial on Gram staining, direct identification using the MicroIDSys Elite system (ASTA Corp.) was performed in duplicate using both the SepsiPrep kit and the modified protocol. As a reference method for bacterial identification, each colony was identified in duplicate using matrix-assisted laser desorption ionization time-of-flight mass spectrometry [6]. Identification at the species level was considered correct when at least one of two spots matched the reference method results with an identification score of ≥140 [7].

Between December 2022 and March 2023, as phase 2, we conducted a single-center study (at Seoul St. Mary’s hospital) of direct identification combined with direct AST from two different types of positive blood cultures using the modified protocol. For direct AST, pellets were suspended in 0.45% saline, and turbidity was adjusted to McFarland 0.5. Comparison of AST between the direct and standard methods was expressed in terms of categorical agreement (CA), very major error (VME), major error (ME), or minor error.

Identification results were compared using the chi-square or McNemar’s test, with statistical significance set at a two-sided P<0.05. All statistical analyses were conducted using the online version of the GraphPad software (https://www.graphpad.com/quickcalcs/).

In phase 1, the correct identification rate was higher with the modified protocol than with the original SepsiPrep kit for both BACTEC Plus (94.4% vs. 83.3%, P=0.078) and BacT/Alert (89.1% vs. 60.0%, P=0.001) bottles. In line herewith, Bidart, et al. [4] reported that a 1.8% SDS extraction method was superior to the SepsiTyper kit for yeast identification. Jeddi, et al. [5] reported that log score values were significantly higher for an SDS protocol than for the SepsiTyper method (Table 1).

Table 1 . Percentage of correct identification through direct identification by MALDI-TOF MS depending on the type of blood culture bottle

Definitive identification	N (%) organisms correctly detected by
	BACTEC Plus bottle				BacT/Alert bottle
	Modified protocol	SepsiPrep kit	P		Modified protocol	SepsiPrep kit	P
Gram-positive
Bacillus sp.	1/1 (100)	0/1 (0)			Not available	Not available
Corynebacterium spp.	2/3 (66.7)	2/3 (66.7)			Not available	Not available
Enterococcus spp.	12/13 (92.3)	10/13 (76.9)			10/11 (90.9)	5/11 (45.5)
Staphylococcus spp.	11/11 (100)	9/11 (81.8)			5/5 (100)	5/5 (100)
Streptococcus spp.	3/3 (100)	3/3 (100)			1/3 (33.3)	1/3 (33.3)
Total gram-positive	29/31 (93.5)	24/31 (77.4)	0.074		16/19 (84.2)	11/19 (57.9)	0.074
Gram-negative
Acinetobacter spp.	1/1 (100)	0/1 (0)			0/2 (0)	1/2 (50.0)
Bacteroides spp.	Not available	Not available			1/2 (50.0)	0/2 (0)
Escherichia sp.	10/10 (100)	9/10 (90.0)			17/18 (94.4)	13/18 (72.2)
Klebsiella spp.	6/6 (100)	6/6 (100)			11/12 (91.7)	7/12 (58.3)
Pseudomonas sp.	2/2 (100)	2/2 (100)			1/1 (100)	1/1 (100)
Roseomonas sp.	0/1 (0)	1/1 (100)			Not available	Not available
Salmonella sp.	1/1 (100)	1/1 (100)			Not available	Not available
Serratia sp.	Not available	Not available			1/1 (100)	0/1 (0)
Stenotrophomonas sp.	2/2 (100)	2/2 (100)			Not available	Not available
Total gram-negative	22/23 (95.7)	21/23 (91.3)	1.000		31/36 (86.1)	22/36 (61.1)	0.016
Total	51/54 (94.4)	45/54 (83.3)	0.078		47/55 (89.1)	33/55 (60.0)	0.001

Abbreviation: MALDI-TOF MS, matrix-assisted laser desorption ionization time-of-flight mass spectrometry.

The results of the phase 2 study are summarized in Table 2. High overall CA for direct AST using cell pellet has also been reported by Watanabe, et al. [8] and Jo, et al. [9], with CA of 97.0% and 97.2% for gram-positive isolates, respectively, and 99.1% and 98.6% for gram-negative isolates, respectively. Lopez-Pintor, et al. [10] compared direct AST with an automated system MicroScan (Beckman, West Sacramento, CA, US) using the cell pellet of Enterobacterales or P. aeruginosa prepared with 5% SDS; the VME and ME rates were 0.3% and 1.3%, respectively. In our study, VME was mainly found for the Staphylococcus epidermidis/trimethoprim-sulfamethoxazole and Enterococcus faecium/glycopeptide combination. This finding is in line with previous findings that Staphylococcus epidermidis is a major cause of disagreement [9] and that most VME in gram-positive isolates are observed with teicoplanin [6] (Table 2).

Table 2 . Results of identification and antimicrobial susceptibility testing with the modified protocol using BACTEC Plus and BacT/Alert bottles

Microorganisms	N correct identifications/N total isolates (%)			Antimicrobial susceptibility test
	BACTEC Plus bottles	BacT/Alert bottles		N categorical agreement			Total N antibiotics showing discrepancy
				BACTEC Plus bottles	BacT/Alert bottles		Very major error	Major error	Minor error
Gram-positive	58/63 (92.1%)	14/15 (93.3%)		679/704 (96.4%)	215/221 (97.3%)		9	4	18
Staphylococcus aureus	9/9	5/6		175/181	120/121		Azithromycin (1)	Azithromycin (1) Quinupristin/dalfopristin (1) Trimethoprim/sulfamethoxazole (1)	Erythromycin (1) Gentamicin (1) Tobramycin (1)
Staphylococcus epidermidis	11/12	-		193/204	-		Trimethoprim/sulfamethoxazole (3) Clindamycin (1)	Telithromycin (1)	Gentamicin (2) Teicoplanin (2) Ciprofloxacin (1) Clindamycin (1)
Staphylococcus haemolyticus	3/3	1/1		47/51	17/17		Clindamycin (1)		Erythromycin (2) Nitrofurantoin (1)
Staphylococcus capitis	1/1	1/1		17/17	16/17				Gentamicin (1)
Staphylococcus hominis	1/1	-		17/17	-
Streptococcus gordonii	1/1	-		13/13	-
Streptococcus mitis/oralis	4/4	-		13/13	-
Streptococcus parasanguinis	0/1	-		12/12	-
Enterococcus faecium	11/12	5/5		127/130	53/55		Vancomycin (2) Teicoplanin (1)		Erythromycin (2) Teicoplanin (2)
Enterococcus faecalis	5/5	1/1		54/55	11/11				Linezolid (1)
Enterococcus gallinarum	1/1	-		11/11	-
Kocuria rhizophila	-	1/1		-	-
Atopobium rimae	0/1	-		-	-
Bacillus spp.	5/5	-		-	-
Bifidobacterium longum	1/1	-		-	-
Clostridium tertium	1/1	-		-	-
Corynebacterium striatum	2/2	-		-	-
Granulicatella adiacens	1/1	-		-	-
Lactobacillus gasseri	1/1	-		-	-
Phycicoccus dokdonensis	1/1	-		-	-
Gram-negative	49/50 (98.0%)	24/25 (96.0%)		740/750 (98.7%)	415/424 (97.9%)		2	2	15
Escherichia coli	25/25	18/19		422/425	322/327				Ampicillin/sulbactam (2) Cefepime (5) Ciprofloxacin (1)
Klebsiella aerogenes	1/1	-		17/17	-
Klebsiella pneumoniae	6/6	3/3		101/102	50/50			Gentamicin (1) Trimethoprim/sulfamethoxazole (1)	Cefepime (1) Cefoxitine (1)
Klebsiella oxytoca	1/1	-		17/17	-
Enterobacter cloacae	4/4	-		68/68	-
Proteus mirabilis	-	2/2		-	30/30
Serratia marcescens	1/2	-		33/34	-				Tigecycline (1)
Morganella morganii	1/1	-		17/17	-
Aeromonas hydrophila	1/1	-		12/12	-
Achromobacter xylosoxidans	1/1	-		11/14	-		Tigecycline (1)		Amikacin (1) Cefepime (1)
Acinetobacter baumannii	-	1/1		-	13/14		Amikacin (1)
Pseudomonas aeruginosa	4/4	-		42/44	-				Aztreonam (1) Piperacillin/tazobactam (1)
Bacteroides spp.	4/4	-		-
Total	107/113 (94.7%)	38/40 (95.0%)		1419/1454 (97.6%)	630/645 (97.7%)		11	6	33

Our study had a limitation in that the number of isolates was small, and we could not calculate the error rates for each isolate/antibiotic combination. However, this study demonstrated that additional lysis with SDS improves the performance of the SepsiPrep kit and that direct AST using cell pellet can guide clinicians in implementing early treatment adjustment.

We would like to thank the clinical microbiology laboratory technologists for technical assistance with this study.

Yoo IY and Park YJ designed the study. Park YJ and Yoo IY analyzed the data and wrote the manuscript. Ha SI and Shim HJ collected the samples and conducted the experiments. Kim TY, Lee HY, Kim J, Huh HJ, Lee NY, and Park YJ reviewed the manuscript. Park YJ supervised the study and reviewed the manuscript. All authors read and approved the final manuscript.

None declared.

This research was supported by a grant from the Korea Health Technology R&B Project through the Korea Health Industry Development Institute (KHIDI) funded by the Ministry of Health & Welfare, Korea (grant No. HI18C2318 and HI22C0595).