a b s t r a c t

Background: Clinical guidelines recommend blood cultures for patients suspected with sepsis and bacteremia. Sepsis-3 task force introduced the new definition of sepsis in 2016; however, the relationship between the Sepsis-3 definition of sepsis and bacteremia remains unclear. This study aimed to investigate how to detect patients who need blood cultures.

Methods: Consecutive patients who visited the emergency department in our hospital with suspected symptoms of bacterial infection and with collected blood culture were retrospectively examined between April and Septem- ber 2019. The relationship between bacteremia and Sepsis-3 definition of sepsis, and the relationship between bacteremia and Clinical scores (quick-Sequential Organ Failure Assessment [qSOFA], systematic inflammatory re- sponse syndrome [SIRS], and Shapiro’s clinical prediction rule) were investigated. In any scores used, >=2 points were considered positive.

Results: Among the 986 patients who met the inclusion criteria, 171 (17%) were complicated with bacteremia and 270 (27%) were patients with sepsis. Sepsis was more frequent (61% vs. 20%, P < 0.001) and all clinical scores were more frequently positive in patients with bacteremia than in those without (qSOFA, 23% vs. 9%; SIRS, 72% vs. 58%; Shapiro’s clinical prediction rule, 88% vs. 49%; P < 0.001). Specificity to predict bacteremia was high in sepsis and positive qSOFA (0.80 and 0.91, respectively), whereas sensitivity was high in positive SIRS and Shapiro’s clinical prediction rule (0.72 and 0.88, respectively); however, no clinical definitions and scores had both high sensitivity and specificity. The area under the receiver operating characteristic curves were 0.59 (95% confidence interval, 0.55-0.64), 0.60 (0.56-0.65), and 0.78 (0.74-0.82) in qSOFA, SIRS, and Shapiro’s clinical prediction rule, respectively.

Conclusion: Blood cultures should be obtained for patients with sepsis and positive qSOFA because of its high specificities to predict bacteremia; however, because of low sensitivities, Shapiro’s clinical prediction rule can be more efficiently used for screening bacteremia.

(C) 2021

1. Introduction

Sepsis is a life-threatening disease in emergency medicine and critical care [1]. Early recognition of sepsis and appropriate administration of anti- biotics are needed for improving their outcomes [1,2]. Sepsis sometimes complicates bacteremia or Bloodstream infections, and patients with bac- teremia have high mortality rates than those without [3]. Blood cultures are the gold standard for detecting bacteremia and for isolation of an in- fecting organism; therefore, clinical guidelines recommended for blood cultures should be performed before the Initial administration of antibi- otics for patients suspected with sepsis and bacteremia [1].

In 2016, the Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3) task force defined sepsis as a life-threatening organ dysfunction caused by infection with an increase in at least two

* Corresponding author.

E-mail address: [email protected] (T. Otani).

points in the Sequential Organ Failure Assessment score as com- pared with that of systematic inflammatory response syndrome (SIRS) [4]. In accordance with the new definition, the quick-SOFA (qSOFA) score, which was generated by analyzing large datasets, was introduced for Screening patients with suspected sepsis outside the intensive care unit, e.g., in the emergency department (ED) [5]; however, the relation- ship between Sepsis-3 definition of sepsis, including qSOFA, and bacter- emia is not well discussed and remains difficult to determine which patients require blood cultures.

Several studies have examined how to predict bacteremia by using clinical and laboratory data. Among them, the SIRS score and the clinical prediction rule reported by Shapiro et al. have been useful clinical scores for screening bacteremia with high sensitivity [6-12]; however, majority of these studies were reported before the publication of Sepsis-3 defini- tion. Therefore, this study aimed to examine the relationship between the new definition of sepsis and bacteremia, and to assess the abilities to predict bacteremia from these clinical scores in the Sepsis-3 era.

https://doi.org/10.1016/j.ajem.2021.03.023

0735-6757/(C) 2021

1. Methods
   1. Study design

Consecutive patients aged >=15 years who visited the ED in our hos- pital with symptoms of suspected bacterial infection and who had blood culture collected were retrospectively investigated between April and September 2019. Blood cultures were ordered for patients suspected with bacteremia, and primary physicians decided the indica- tion for blood cultures. At least two sets of blood cultures consisting of one aerobic and one anaerobic bottle were obtained before the admin- istration of antibiotics. Patients administered antibiotics within 48 h be- fore the hospital arrival were excluded from this analysis because the sensitivity of blood culture reduces after initiation of antibiotics [13]. Pa- tients suspected with false bacteremia (contamination) were also ex- cluded. Relationships between bacteremia and the following factors were investigated: baseline characteristics, medical symptoms, vital sta- tus, blood sample results, Sepsis-3 definition of sepsis and septic shock, clinical scores (SOFA, qSOFA, SIRS, and Shapiro’s clinical prediction rule [defined as Shapiro score in this study]), sources of infection, and out- comes. Medical symptoms (chills, vomiting, and altered mental status) and vital status (temperature, systolic blood pressure, heart rate, and re- spiratory rate) upon hospital arrival were extracted from physicians’ and nurses’ electronic medical records. Blood sample results closest to the hospital arrival were also extracted from the medical records. Clini- cal scores were calculated using these data, and diagnostic

Table 1

Shapiro’s clinical prediction rule.

Major criteria minor criteria (1 point each)

performances were considered predictors of bacteremia. The study pro- tocol conformed to the concepts of the Declaration of Helsinki and its amendments and was approved by the institutional review board.

• 1. Definitions

Sepsis and septic shock were diagnosed based on the Sepsis-3 defini- tion [4]. Bacteremia was defined asatleast one positive blood culture result with clinical evidence or microbiological findings suggesting a primary focus of infection. The growth of common skin pathogens, such as coagulase-negative Staphylococcus species, Bacillus species, or micrococci in only one set of blood cultures without evidence of Infective endocarditis, was regarded as a contamination under the assessment by at least two in- vestigators [14]. The qSOFA score consists of three clinical variables: sys- tolic blood pressure of <=100 mmHg, respiratory rate of >=22 breaths/min, and altered mentation. The presence of >=2 of these criteria is considered a positive qSOFA score [5]. The SIRS score consists of four clinical variables: respiratory rate, >20 breaths/min; body temperature, >38 ?C or <36 ?C; heart rate, >90 beats/min; and white blood cell count, >12,000/uL or

<4000/uL. The presence of >=2 of these criteria is considered a positive SIRS score [15]. Shapiro score, evaluated in this study, is shown in Table 1. The original Shapiro score includes “bands proportion of >5% (bandemia)” as aminor criterion [6]; however, this criterion was excluded as part of the present analysis, because a majority of patients in this study did not have bands measured. At least one major criterion or two minor criteria (>=2 points) is considered a positive Shapiro score [6].

• 1. Statistical analysis

Data were presented as medians (interquartile range) for continu- ous variables and numbers and percentages for categorical variables.

Suspected endocarditis (3 points)

Temperature > 39.4 ?C (3 points)

Indwelling vascular catheter (2 points)

Temperature 38.3-39.3 ?C

Age > 65 years Chills

Vomiting

Hypotension (systolic blood pressure < 90 mmHg)

White blood cell count >18,000/uL Bands >5%^a

Platelets <150,000/uL Creatinine >2 mg/dL

Continuous variables were compared using a Mann-Whitney U test, and categorical variables with a chi-square or Fisher’s exact test. To ex- amine the relationship between the Sepsis-3 definition (sepsis and sep- tic shock) and bacteremia, and to assess the diagnostic abilities of clinical scores (qSOFA, SIRS, and Shapiro score) for predicting bacter- emia, we calculated sensitivity, specificity, positive likelihood rate, and negative likelihood rate with 95% confidence intervals (CI). A receiver operating characteristic (ROC) curve was created to illustrate each cut-off value of clinical scores for predicting bacteremia and calculated area under the ROC curves (AUC) to evaluate their predictive values.

^a Bands of >5% were not included as part of the present analysis.

All statistical analyses were performed using the R software package

Included cases, n = 986 (83%)

Exclusion criteria Antibiotics used before arrival, n = 182

Suspected false bacteremia, n = 24

blood culture collection, n = 1,192

Non-bacteremia, n = 815 (68%)

Bacteremia, n = 171 (14%)

Fig. 1. Flow chart of participant inclusion.

diagnostic performances for predict”>(version 4.0.2, R Development Core Team; https://www.r-project.org/). A P-value of <0.05 was considered to be statistically significant.

Table 2

Comparisons between patients with and without bacteremia in this study.

		All		Bacteremia		Non-bacteremia	P-value
3. Results		n = 986		n = 171		n = 815
	Age, years	71 (51-81)		76 (68-84)		70 (47-80)	<0.001
3.1. Patient characteristics	Male, sex	532 (54)		104 (61)		427 (52)	0.052
	Comorbidities
During the study period, 1192 patients visited the ED of our hospital	Diabetes mellites	158 (16)	36 (21)		122 (15)		0.052
with symptoms of suspected bacterial infection and had blood culture	end stage renal disease	10 (1)	2 (1)		8 (1)		0.69
collected and 986 (83%) of them met the inclusion criteria (Fig. 1).	Malignancy	143 (15)	33 (19)		110 (13)		0.056
Among these 986 patients, 171 (17%) were complicated with bacter-	intravenous drug use steroid use	47 (5) 53 (5)		6 (4) 12 (7)		41 (5) 41 (5)	0.55 0.35
emia. In accordance with the Sepsis-3 definition, 270 (27%) had sepsis	Suspected endocarditis	2 (0)	2 (1)		0		0.030
and 27 (3%) had septic shock. A total of 28 patients (3%) died during	Indwelling vascular	15 (2)	3 (2)		12 (1)		0.73
the hospital stay (Table 2).	catheter

3.2. Comparison between patients with and without complicated bacteremia

Chills	204 (21)	76 (44)	128 (16)	<0.001
Vomiting	112 (11)	31 (18)	81 (10)	0.003
Altered mental status	152 (15)	50 (29)	102 (13)	<0.001
Temperature, ?C	37.9	38.4	37.9	<0.001
	(37.2-38.7)	(37.4-39.3)	(37.1-38.5)
Systolic blood pressure,	135	132	135 (116-156)	0.102

Table 2 shows comparisons between patients with and without complicated bacteremia. Sepsis-3 definition of sepsis and septic shock was more frequent in patients with bacteremia (sepsis, 61% vs. 20%; septic shock, 12% vs. 1%; P < 0.001). Among itEMS used to calculate clin- ical scores (qSOFA, SIRS, and Shapiro score), the majority of these were significantly different between the two groups; however, the frequency of indwelling vascular catheter, systolic blood pressure, heart rate, and white blood cell count was not different between the two groups. All clinical scores were higher (qSOFA, 1 vs. 0; SIRS, 2 vs. 2; Shapiro score, 3 vs. 1; P < 0.001) and more frequently positive (qSOFA, 23% vs. 9%;

Platelets, x103/uL Creatinine, mg/dL	204 (157-266) 0.8 (0.7-1.1)	170 (125-235) 1.0 (0.8-1.5)	213 (162-272) 0.8 (0.6-1.0)	<0.001 <0.001
Sepsis	270 (27)	104 (61)	166 (20)	<0.001
Septic shock	27 (3)	20 (12)	7 (1)	<0.001
SOFA score	1 (0-2)	2 (1-4)	1 (0-2)	<0.001
qSOFA score	0 (0-1)	1 (0-1)	0 (0-1)	<0.001
0	599 (61)	85 (50)	514 (63)
1	277 (28)	46 (27)	231 (28)
2	87 (9)	28 (16)	59 (7)
3	23 (2)	12 (7)	11 (1)
Positive qSOFA	110 (11)	40 (23)	70 (9)	<0.001
SIRS score	2 (1-3)	2 (1-3)	2 (1-3)	<0.001
0	135 (14)	12 (7)	123 (15)
1	258 (26)	36 (21)	222 (27)
2	320 (32)	57 (33)	263 (32)
3	222 (23)	48 (28)	174 (21)
4	51 (5)	18 (11)	33 (4)
Positive SIRS	593 (60)	123 (72)	470 (58)	<0.001
Shapiro score	2 (1-3)	3 (2-4)	1 (1-2)	<0.001
0	132 (13)	2 (1)	130 (16)
1	304 (31)	18 (11)	286 (35)
2	265 (27)	44 (26)	221 (27)
3	131 (13)	35 (20)	96 (12)
4	83 (8)	32 (19)	51 (6)
>=5	71 (7)	40 (23)	31 (4)
Positive Shapiro score Source of infection Intra-abdominal	550 (56) 292 (30)	151 (88) 69 (40)	399 (49) 223 (27)	<0.001 <0.001

mmHg	(116-155)	(113-154)
Heart rate, /min	97 (84-110)	99 (85-114)	97 (84-109)	0.095
Respiratory rate, /min	18 (16-22)	19 (16-23)	18 (16-22)	0.017
White blood cell count, x102/uL	99 (70-133)	99 (69-141)	99 (70-131)	0.32

SIRS, 72% vs. 58%; Shapiro score, 88% vs. 49%; P < 0.001) in patients with bacteremia. In-hospital mortality was also higher in patients with bacteremia (7% vs. 2%; P = 0.001).

3.3. Diagnostic performances for predicting bacteremia

Table 3 shows the diagnostic performances for predicting bacter- emia according to the Sepsis-3 definition (sepsis and septic shock) and clinical scores evaluated in this study. Specificity was relatively high for sepsis, septic shock and positive qSOFA (0.80, 0.99, and 0.91, re- spectively), and their positive likelihood rates were 2.99, 13.62, and 2.72, respectively. Sensitivity was relatively high in positive SIRS and Shapiro score (0.72 and 0.88, respectively), and their negative likeli- hood rates were 0.66 and 0.23, respectively. No clinical definitions and scores had both high sensitivity and specificity.

Fig. 2 shows the ROC curves that describe the diagnostic accuracies

of qSOFA, SIRS, and Shapiro score for predicting bacteremia. AUCs were 0.59 (95% CI, 0.55-0.64), 0.60 (95% CI, 0.56-0.65), and 0.78 (95%

CI, 0.74-0.82) in the qSOFA, SIRS, and Shapiro scores, respectively.	infection Pneumonia	224 (23)	3 (2)	221 (27)
	Urinary tract infection	181 (18)	65 (38)	116 (14)
3.4. Incidence rate of bacteremia and mortality rate in each diagnostic step	In-hospital admission	619 (63)	145 (85)	474 (58)	<0.001
	In-hospital mortality	28 (3)	12 (7)	16 (2)	0.001

Others 289 (29) 34 (20) 255 (31)

Fig. 3 shows the incidence rate of bacteremia and mortality rate in each diagnostic step in the study population. Among the 986 patients, 114 met the qSOFA criteria or were complicated by septic shock. Of the 114 patients, 43 (38%) had bacteremia and a mortality rate of 17%. Of the remaining patients, 494 met the Shapiro criteria or diagnostic criteria for sepsis. Of the 494 patients, 113 (23%) patients were compli- cated with bacteremia and had a mortality rate of 1%. Among the re- maining 378 patients, 15 (4%) patients were complicated with bacteremia and had a mortality rate of 1%. However, all the causes of death in this population were not infection but underlying comorbidi- ties, and none of the patients with bacteremia died. By using this algo- rithm for decision making on blood culture collection, the sensitivity and specificity were 0.91 and 0.46, respectively, and the positive and negative likelihood rates were 1.65 and 0.20, respectively.

Data are presented as the number (column %) of patients or median (interquartile range). SOFA, Sequential Organ Failure Assessment; qSOFA, quick Sequential Organ Failure As- sessment; SIRS, systematic inflammatory response syndrome.

4. Discussion

The major findings of this study were as follow: 1) Patients with bac- teremia were more likely to meet the diagnostic criteria of sepsis and septic shock, and more frequently would meet positive clinical scores (qSOFA, SIRS, and Shapiro score) than in those without. 2) For the diag- nostic accuracy of bacteremia, specificity was high in the definition of sepsis and septic shock and positive qSOFA, whereas sensitivity was high in the positive SIRS and Shapiro score.

Table 3

Diagnostic performances for the prediction of bacteremia.

	Sensitivity	Specificity	Positive likelihood rate	Negative likelihood rate
Sepsis	0.61 (0.53-0.68)	0.80 (0.77-0.82)	2.99 (2.49-3.58)	0.49 (0.41-0.60)
Septic shock	0.12 (0.07-0.18)	0.99 (0.98-1)	13.62 (5.85-31.69)	0.89 (0.84-0.94)
qSOFA >=2	0.23 (0.17-0.31)	0.91 (0.89-0.93)	2.72 (1.92-3.87)	0.84 (0.77-0.91)
SIRS >=2	0.72 (0.65-0.79)	0.42 (0.39-0.46)	1.25 (1.12-1.39)	0.66 (0.52-0.85)
Shapiro score >= 2	0.88 (0.83-0.93)	0.51 (0.48-0.55)	1.80 (1.65-1.97)	0.23 (0.15-0.35)

Data are presented with 95% confidence intervals.

qSOFA, quick Sequential Organ Failure Assessment; SIRS, systematic inflammatory response syndrome.

0.8

1.0

0.6

0 0.2 0.4 0.6 0.8 1.0

qSOFA SIRS

Shapiro score

Sensitivity

0

0.2

0.4

1 - specificity

Fig. 2. Receiver operating characteristic curves for qSOFA, SIRS, and Shapiro score for the prediction of bacteremia. The area under the receiver-operative characteristic curves were 0.59 (95% confidence interval, 0.55-0.64) in qSOFA, 0.60 (0.56-0.65) in SIRS, and

0.78 (0.74-0.82) in Shapiro score as a predictor for bacteremia. qSOFA: quick-Sequential Organ Failure Assessment; SIRS: systematic inflammatory response syndrome.

Early identification of sepsis and bacteremia, i.e., associated with mortality, is important for the medical decision-making of bacterial in- fection, and obtaining blood cultures is recommended before the initial administration of antibiotics [1]; however, identifying which patients require blood cultures is difficult. Quick-SOFA, a screening tool for patients likely to have sepsis that was introduced in 2016, has been re- ported to have high specificity in the early detection of sepsis and in- hospital mortality [5,16,17]; however, qSOFA was criticized for its low sensitivity in identifying sepsis and in-hospital mortality [18-21]. Fur- thermore, relationships between qSOFA and Sepsis-3 definition of sep- sis, and bacteremia are not well discussed [11,17]. In this study, patients with bacteremia were more likely to meet the diagnostic criteria of sepsis (61% vs. 20%, P < 0.001) and septic shock (12% vs. 1%, P < 0.001), and positive qSOFA was more frequent than in those without (23% vs. 9%, P < 0.001). Due to their high specificity in predicting bacter- emia (sepsis, 0.80; septic shock, 0.99; positive qSOFA, 0.91), testing blood cultures and initiating early administration of antibiotics should be carried out for patients with sepsis, septic shock, and positive qSOFA. In particular, the calculation of the qSOFA score does not require blood sample test results and may be helpful in achieving an early deci- sion making [5]; however, owing to the low sensitivity, these definitions and qSOFA should not be used for screening bacteremia. Thus, bacter- emia should be screened with other tools with high sensitivity.

The Shapiro score was reported to detect bacteremia with high sensitivity [6-11]. In our study, a positive Shapiro score was more fre- quent in the patients with bacteremia (88% vs. 49%, P < 0.001), and the AUC that described the diagnostic accuracy of bacteremia was

0.78 (95% CI, 0.74-0.82). This AUC was approximately similar to those of previous studies (AUC, 0.75-0.83) [6-8]. The difference be- tween our results and those of previous studies was sensitivity and specificity of the positive Shapiro score for detecting bacteremia. The sensitivity (in our study, 0.88) is slightly lower (previous re- ports, 0.94-0.98) and the specificity (in our study, 0.51) is slightly higher (previous reports, 0.27-0.48) than those of previous reports [6-10]. This difference may be explained by the fact that our study excluded bandemia (band proportion, >5%) from the Shapiro score and undercalculated it compared with previous studies. If we use Shapiro score for decision-making by obtaining blood cultures in this study population, 416/986 (42%) blood cultures could be re- duced (130 with Shapiro score 0, and 286 with Shapiro score 1); however, 20/171 (12%) bacteremia would be misdiagnosed (2 with Shapiro score 0 and 18 with Shapiro score 1). Seigel et al. reported that band measurement was useful for detecting bacteremia even in the setting of normal White blood cell counts [22]; therefore, bandemia should be evaluated if possible when using Shapiro score as a decision-making tool to obtain blood cultures. In addition, de- spite its high sensitivity for detecting bacteremia, its specificity is low; hence, its clinical application may be limited and higher accu- racy models should be investigated. A combination of Shapiro score and procalcitonin level has been reported to be more effective in the identification of bacteremia than Shapiro score alone [10,11]. This method could be a candidate for a higher-accuracy model; how- ever, most cases in our study were not tested for procalcitonin. Therefore, we could not assess the usefulness of measuring procalcitonin levels, and we do not think that it is cost-effective to measure the procalcitonin level in all cases for decision making on whether to evaluate blood cultures or not. SIRS has also been re- ported to be associated with bacteremia [9,12]. In our study, SIRS was more frequently positive in patients with bacteremia (72% vs. 58%, P < 0.001); however, the AUC, which describes the diagnostic accuracy of bacteremia, was only 0.60 (95% CI, 0.56-0.65). Previous studies also reported its poor diagnostic accuracy of bacteremia [9,12]. According to these results, Shapiro score may be better for screening bacteremia than SIRS.

Notably, calculating Shapiro score and the diagnosis of sepsis require blood sample testresults anda longertimethan qSOFA. Early appropriate administration of antibiotics is needed for improving outcomes of sepsis, especially in Serious conditions [1-3]. The use of the Shapiroscoreand the diagnosis of sepsis for clinical decision-making should be limited to pa- tients who do not meet the qSOFA criteria and who do not need fluid re- suscitation to maintain blood pressure (suspected septic shock), i.e., relativelylessserious. Basedonourstudyfindings, thefollowingmed- ical decision-making for obtaining blood cultures is recommended (Fig. 3). When patients with suspected bacterial infection meet the qSOFA criteria and are suspected with septic shock (i.e., low blood pres- sure and requiring fluid resuscitation), physicians should obtain blood cultures and initiate appropriate antibiotic administration as soon as

Study population, n = 986 Bacteremia, n = 171 (17%) In-hospital mortality, n = 28 (3%)
			qSOFA >= 2 or septic shock, n = 114 Bacteremia, n = 43 (38%) In-hospital mortality, n = 19 (17%)
			Shapiro score >= 2 or sepsis, n = 494 Bacteremia, n = 113 (23%) In-hospital mortality, n = 6 (1%)
Excluded from algorithm, n = 378 Bacteremia, n = 15 (4%) In-hospital mortality, n = 3 (1%)

Fig. 3. The incidence rate of bacteremia and mortality rate in each diagnostic step. qSOFA: quick-Sequential Organ Failure Assessment.

possible (before confirming the blood sample test results). If patients do not meet the qSOFA criteria and do not require fluid resuscitation, physi- cians should calculate the Shapiro score and evaluate whether the pa- tients meet the diagnosis of sepsis, and consider the necessity for obtaining blood cultures. Fig. 3 shows the incidence rate of bacteremia and mortality rate in each diagnostic step in this study population. Among the 378 patients excluded from the algorithm, 15 patients were complicated with bacteremia; however, the severity in this population was low. This algorithm was not superior to the Shapiro score for detect- ing bacteremia (sensitivity, 0.91 and specificity, 0.46); however, it was useful for early recognition of patients with severe conditions.

Several limitations were identified in this study. First, this was a single-center, retrospective study with a relatively small sample size. Second, not all patients with suspected bacterial infection received blood cultures, and the decision whether blood cultures should be eval- uated or not was made by individual physicians. This selection bias probably eliminated several patients with less serious infections and weakened the prediction modeling. Third, the vital status was assessed only upon hospital arrival, and its changes during ED stay were not eval- uated. Fourth, the SOFA score was calculated using blood sample test re- sults closest to hospital arrival, and data after hospital admission were not evaluated. Fifth, sepsis was diagnosed using the initial SOFA score only. Sixth, the timing of blood culture collection is known to affect its positivity; however, we did not evaluate this point [23]. Finally, our study excluded bandemia from the Shapiro score calculation. We should recognize the impact of not evaluating bandemia while interpreting our study, as bandemia has been shown to predict bacteremia in some stud- ies results [6-12,22]. Further prospective validation that includes the as- sessment of bandemia should be performed to confirm its usefulness for predicting bacteremia based on the scores and definitions.

In conclusion, evaluating blood cultures and initiating early adminis- tration of antibiotics should be carried out for patients with sepsis, sep- tic shock, and positive qSOFA; however, it is better to use the Shapiro score for screening bacteremia.

Competing interests

None declared.

Funding

This research did not receive any specific grant from funding agen- cies in the public, commercial, or not-for-profit sectors.

Patient consent for publication

Not required.

Ethics approval

This study was approved by the Hiroshima City Hiroshima Citizens Hospital Ethics Board, which considers the ethical aspects of research studies involving human participants at Hiroshima City Hiroshima Citi- zens Hospital (reference 2020-94).

Acknowledgments

We thank Enago (www.enago.jp) for English language editing.

References

1. Rhodes A, Evans LE, Alhazzani W, Levy MM, Antonelli M, Ferrer R, et al. Surviving sepsis campaign: international guidelines for management of sepsis and septic shock: 2016. Intensive Care Med. 2017;43:304-77.
2. Lee CC, Lee CH, Hong MY, Tang HJ, Ko WC. Timing of appropriate empirical antimi- crobial administration and outcome of adults with Community-onset bacteremia. Crit Care. 2017;21:119.
3. Phua J, Ngerng W, See K, Tay C, Kiong T, Lim H, et al. Characteristics and outcomes of culture-negative versus culture-positive severe sepsis. Crit Care. 2013;17:R202.
4. Singer M, Deutschman CS, Seymour CW, Shankar-Hari M, Annane D, Bauer M, et al. The third international consensus definitions for sepsis and septic shock (Sepsis-3). JAMA. 2016;315:801-10.
5. Seymour CW, Liu VX, Iwashyna TJ, Brunkhorst FM, Rea TD, Scherag A, et al. Assess- ment of clinical criteria for sepsis: for the third international consensus definitions for sepsis and septic shock (Sepsis-3). JAMA. 2016;315:762-74.
6. Shapiro NI, Wolfe RE, Wright SB, Moore R, Bates DW. Who needs a blood culture? A prospectively derived and validated prediction rule. J Emerg Med. 2008;35:255-64.
7. Jessen MK, Mackenhauer J, Hvass AM, Ellermann-Eriksen S, Skibsted S, Kirkegaard H, et al. Prediction of bacteremia in the emergency department: an external validation of a clinical decision rule. Eur J Emerg Med. 2016;23:44-9.
8. Hodgson LE, Dragolea N, Venn R, Dimitrov BD, Forni LG. An external validation study of a clinical prediction rule for Medical patients with suspected bacteraemia. Emerg Med J. 2016;33:124-9.
9. Coburn B, Morris AM, Tomlinson G, Detsky AS. Does this adult patient with suspected bacteremia require blood cultures? JAMA. 2012;308:502-11.
10. Laukemann S, Kasper N, Kulkarni P, Steiner D, Rast AC, Kutz A, et al. Can we reduce negative blood cultures with clinical scores and blood markers? Results from an ob- servational cohort study. Medicine (Baltimore). 2015;94:e2264.
11. Wyss G, Berger S, Haubitz S, Fankhauser H, Buergi U, Mueller B, et al. The Shapiro- Procalcitonin algorithm (SPA) as a Decision tool for blood culture sampling: valida- tion in a prospective cohort study. Infection. 2020;48:523-33.
12. Wildi K, Tschudin-Sutter S, Dell-Kuster S, Frei R, Bucher HC, Nuesch R. Factors asso- ciated with positive blood cultures in outpatients with suspected bacteremia. Eur J Clin Microbiol Infect Dis. 2011;30:1615-9.
13. Cheng MP, Stenstrom R, Paquette K, Stabler SN, Akhter M, Davidson AC, et al. Blood culture results before and after antimicrobial administration in patients with severe manifestations of sepsis: a diagnostic study. Ann Intern Med. 2019;171:547-54.
14. Lee CC, Lin WJ, Shih HI, Wu CJ, Chen PL, Lee HC, et al. Clinical significance of potential contaminants in blood cultures among patients in a medical center. J Microbiol Immunol Infect. 2007;40:438-44.
15. Bone RC, Balk RA, Cerra FB, Dellinger RP, Fein AM, Knaus WA, et al. Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. The ACCP/SCCM consensus conference committee. American College of Chest Physi- cians/Society of Critical Care Medicine. Chest. 1992;101:1644-55.
16. Freund Y, Lemachatti N, Krastinova E, Van Laer M, Claessens YE, Avondo A, et al. Prognostic accuracy of Sepsis-3 criteria for in-hospital mortality among patients

with suspected infection presenting to the emergency department. JAMA. 2017; 317:301-8.

1. Ramos JGR, da Hora Passos R, Teixeira MB, Gobatto ALN, Coutinho RVDS, Caldas JR, et al. Prognostic ability of quick-SOFA across different age groups of patients with suspected infection outside the intensive care unit: a cohort study. J Crit Care. 2018;47:178-84.
2. Churpek MM, Snyder A, Han X, Sokol S, Pettit N, Howell MD, et al. Quick sepsis- related organ failure assessment, systemic inflammatory response syndrome, and early warning scores for detecting clinical deterioration in infected patients outside the intensive care unit. Am J Respir Crit Care Med. 2017;195:906-11.
3. Hwang SY, Jo IJ, Lee SU, Lee TR, Yoon H, Cha WC, et al. Low accuracy of positive qSOFA criteria for predicting 28-day mortality in critically ill septic patients during the early period after emergency department presentation. Ann Emerg Med. 2018; 71:1-9.e2.
4. Song JU, Sin CK, Park HK, Shim SR, Lee J. Performance of the quick Sequential (sepsis- related) Organ Failure Assessment score as a Prognostic tool in infected patients out- side the intensive care unit: a systematic review and meta-analysis. Crit Care. 2018; 22:28.
5. Anand V, Zhang Z, Kadri SS, Klompas M, Rhee C. Epidemiology of quick sequential organ failure assessment criteria in undifferentiated patients and association with suspected infection and sepsis. Chest. 2019;156:289-97.
6. Seigel TA, Cocchi MN, Salciccioli J, Shapiro NI, Howell M, Tang A, et al. Inadequacy of temperature and white blood cell count in predicting bacteremia in patients with suspected infection. J Emerg Med. 2012;42:254-9.
7. Taniguchi T, Tsuha S, Shiiki S, Narita M. High positivity of blood cultures obtained within two hours after shaking chills. Int J Infect Dis. 2018;76:23-8.