Article, Infectious Diseases

The neutrophil-lymphocyte count ratio as a diagnostic marker for bacteraemia: A systematic review and meta-analysis

a b s t r a c t

Background: Bacteraemia is a common cause of increased morbidity and mortality in critically ill patients, but its early diagnosis and identification are complicated. The neutrophil-lymphocyte count ratio (NLCR) has been sug- gested as a useful indicator for the diagnosis of bacteraemia. We performed this meta-analysis to investigate the diagnostic accuracy of the NLCR for bacteraemia.

Methods: We searched the PubMed, Embase, Web of Science, and Cochrane Library databases for this meta- analysis. We calculated individual and pooled sensitivities and specificities. I2 statistics and Cochran’s Q test were used to evaluate heterogeneity, and the cause of heterogeneity was explored with sensitivity analyses.

Results: In total, 8 of 1086 eligible articles were included in the present meta-analysis. The pooled analyses re- vealed that the diagnostic accuracy of the NLCR in terms of its bacteraemia sensitivity was 0.723 [95% CI: 0.660, 0.777], and its specificity was 0.596 [95% CI: 0.556, 0.634]. The area under the summary receiver operating characteristic curve was 0.69 [95% CI 0.65-0.73].

Conclusion: The NLCR is an easy-to-collect marker for bacteraemia. However, the NLCR is inadequate, and only a combination of multiple biomarkers will improve its diagnostic accuracy for bacteraemia.

(C) 2018

  1. Introduction

Bacteraemia is a life-threatening disease with high morbidity and mortality that requires early rapid identification and prompt antibiotic treatment [1-3]. Biomarkers can play an important role in this process, because they can indicate the presence, absence, or severity of the dis- ease. Conventional inflammatory markers, including the white blood cell count, Neutrophil count, neutrophilic granulocyte percent- age, C-reactive protein level, and procalcitonin level have limited evaluation potential and are expensive options for the assess- ment of bacteraemia [4-6]. Blood culture, which is considered the “gold standard”, is the traditional approach for the detection of bacteraemia, but this method is time consuming [7]. In addition, the sensitivity of blood culture decreases significantly if fastidious or slow- growing pathogens are cultured or if antibiotic therapy has been insti- tuted prior to blood sampling [7,8]. Therefore, other biochemical

* Corresponding author at: Tianjin First Center Hospital, No. 24 Fukang Roads, Nankai District, Tianjin 300192, People’s Republic of China.

E-mail address: [email protected] (Y. Wang).

1 Jiawei Jiang and Rui Liu contributed equally to this work.

markers are needed in addition to blood culture to evaluate bacteraemia.

Neutrophils and lymphocytes are the two major inflammatory cell types in the body. Currently, the neutrophil-lymphocyte count ratio (NLCR) is used in a wide range of applications for the diagnosis, treat- ment, and prognostic evaluation of inflammation-relatED diseases, such as malignant tumors, cardiovascular diseases, renal diseases, and inflammatory bowel diseases. Recently, several studies have assessed the value of the NLCR for the diagnosis of bacteraemia. The NLCR was also recently described a marker for the diagnosis of bacterial infections in young infants [9]. Some studies have described increased neutrophil production and lymphocyte apoptosis, which are the causes of the ele- vated NLCR during inflammatory stress [10]. Neutrophils and lymphopenias have been detected in patients with systemic inflamma- tory response syndrome. Additionally, the inflammatory response re- sults in the destruction of cells at the site of inflammation, and thus, apoptosis may promote lymphopenia [11]. Therefore, the NLCR was proven to be a simple infection marker for predicting bacteraemia upon hospital admission [12]. A previous study considered the NLCR a rapid indicator of the inflammatory response in critically ill patients [13]. We conducted this meta-analysis to assess the diagnostic accuracy

https://doi.org/10.1016/j.ajem.2018.10.057 0735-6757/(C) 2018

of the NLCR for bacteraemia and to assist physicians with clinical decision-making.

  1. Methods

This meta-analysis followed the relevant guidelines (Preferred Reporting Items for Systematic Reviews and Meta-Analyses, PRISMA). An ethics statement was not required.

Literature search

We prepared a protocol before undertaking this systematic review and meta-analysis. A systemic overall search was conducted in the PubMed, Web of Science, Embase, and Cochrane Library databases to identify studies evaluating the diagnostic accuracy of the NLCR for sepsis published until March 2018 without any language restrictions. The fol- lowing search terms were used (“bacteraemia”OR “infection” OR “sep- sis” OR “critically ill patients”) AND (“neutrophil-to-lymphocyte” OR “neutrophil/lymphocyte ratio” OR “neutrophil-lymphocyte count ratio” OR “neutrophil-lymphocyte ratio” OR “NLCR”). The references of the included articles were also checked to identify additional studies.

Study selection

The original data were independently collected from the retrieved studies by two investigators (JWJ and RL) using the same standards. Any disagreement was resolved through discussion, and if no agree- ment was reached, then a third reviewer (YQW) made the final deci- sion. The studies were included if they met all of the following

criteria: the study evaluated the diagnostic accuracy of the NLCR for bacteraemia in adult patients (>=18 years old); blood culture was per- formed as a diagnostic criterion or control measure; and the sensitivity and specificity of the data were provided or could be calculated. Re- views, meetings, abstracts, letters, duplicated studies, studies without a 2 x 2 table, studies not related to the diagnostic value, and studies with pediatric patients were excluded.

Procedures

Two investigators independently extracted data from the retrieved studies using a standard form. Differences were resolved by a third au- thor. The data included the year of publication, study design, population setting, cut-off values used, and details of the NLCR assays, including true positives (TP), true negatives (TN), false positives (FP), false nega- tives (FN), sensitivity (SEN), specificity (SPE), and area under the sum- mary receiver operating characteristic (SROC) curve (AUC) for an early bacteraemia diagnosis.

The retrieved studies were assessed using the Quality Assessment of Diagnostic Accuracy Studies (QUADAS) checklist [14]. We used the guidelines to score each item of the checklist in this meta-analysis.

Statistical analysis

Review Manager, version 5.1.3 (RevMan; the Cochrane Collabora- tion, Oxford, UK), and STATA, version 14.0 (Stata Corporation, College Station, TX, USA) were used to perform the statistical analyses. Abivariate random effects regression model was used to determine the SEN, SPE, positive likelihood ratio (PLR), negative likelihood ratio

Fig. 1. Flow chart depicting the study selection procedure.

(NLR), and Diagnostic odds ratio (DOR) with 95% confidence intervals (95% CIs). Then, we constructed an SROC curve by plotting individual and summary points for SEN and SPE to assess the diagnostic accuracy of the NLCR [15,16]. All statistical tests were bilateral, with statistical significance considered for differences with p b 0.05. Heterogeneity was calculated using Cochran’s Q test and the I 2 inconsistency test. Cochran’s Q test indicated heterogeneity at p b 0.10, and an I 2 N 50% was considered to indicate significant heterogeneity. A random effects method (Der Simonian-Laird method) was used to generate the pooled mean differences and 95% CIs. In addition, sensitivity analyses were con- ducted to investigate potential sources of heterogeneity. Publication bias was evaluated using a Deek funnel plot, and a Fagan nomogram was used to calculate the post-test probability (PTP).

TN

FN

Sensitivity (95% CI)

0.77[0.67,0.85]

Specificity (95% CI)

0.63[0.52,0.73]

21

58

65

374

1080

4

28

116

68

596

1432

0.85[0.66,0.96]

0.72[0.62,0.81]

0.70[0.65,0.74]

0.51[0.42,0.60]

0.61[0.58,0.65]

0.57[0.55,0.59]

11

38

0.82[0.70,0.90]

0.63[0.50,0.75]

89

16

1031

352

0.61[0.55,0.68]

0.79[0.68,0.88]

0.66[0.64,0.68]

0.56[0.52,0.60]

30

58

0.63[0.52,0.73]

0.72[0.60,0.81]

  1. Results

FP

34

23

73

270

22

141

61

529

281

23

Literature search and study characteristics

Cut off

TP

71

49

51

The literature search found a total of 1086 articles, 8 of which [12,17- 23] satisfied the inclusion criteria. After examining their titles and ab- stracts, we excluded 1067 studies. After reviewing the full contents, we excluded another 11 studies, resulting in the inclusion of 8 eligible studies (Fig. 1). The main reasons for excluding studies were that their sensitivity and specificity data could not be retrieved, a 2 x 2 table was not reconstructed, or they were not diagnostic studies. All studies were published between 2010 and 2018, and a total of 7095 patients were included in this meta-analysis. Six studies [12,17-19,21,23] were performed in the Emergency Department, and the other two studies [20,22] had unspecified patient sources. The cut-off values for the NLCR differed between studies. The study characteristics are shown in Table 1.

Mean age (years)

Time of markers detection

During the observation period

Setting

ED

10

68.9 (51.6-86.2)

67 (53-78)

66

During the observation period At enrollment

during the study period

NR

ED ED ED

10

10

10

NR

9.37

NR

At enrollment

NR ED

10

10

NR

ED

13.2

NR not report, BCP blood culture positive, BCN blood culture negative, TP true positive, FP false positive, FN false negative, TN true negative, ED Emergency department,

Study quality and publication bias

66 (18-96)

62.6

(47.7-77.5)

61 (47-72)

71 (58-81)

68(50.9-85.1)

The QUADAS-2 tool was used to assess the risk of bias in the 8 [12,17-23] included studies (Fig. 2). The results revealed that 1 study

[21] had a high risk of bias according to the reference standard. The Deek funnel plot is shown in Fig. 3. Significant publication bias was ob- served (p b 0.001).

No. of patients/female/no. of patients with bacteremia

746/96/85

Diagnostic value of the NLCR for the diagnosis of bacteraemia

140/NR/27

1083/460/104

2002/1001/270

120/47/60

1807/916/230

1572/697/197

162/65/81

The following pooled parameters were calculated for all 8 stud- ies that examined the value of the NLCR for the diagnosis of bacteraemia. The overall I2 value for the bivariate model was 91.51% (95% CI 83.52-99.51), indicating significant heterogeneity.

Definition of bacteremia

BCP

The pooled SEN and SPE values were 0.723 (95% CI 0.660, 0.777)

and 0.596 (95% CI 0.556, 0.634), respectively (Fig. 4). The AUC

BCP BCP BCP

BCP

BCP BCP

BCP

was 0.69 (95% CI 0.65, 0.73). (Fig. 5). Fagan’s nomogram for the likelihood ratios indicated that utilizing the NLCR to diagnose bacteraemia increased the post-probability to 64% when the results were positive and reduced the post-probability to 32% when the re- sults were negative (Fig. 6).

Suspected community-acquired bacteremia

With 2 SIRS criteria Suspected infection Suspected infection

With SIRS criteria and suspected or documented infection Suspicion of infection

Suspicion of a community-onset sepsis

BCP and BCN

We concluded that the proportion of heterogeneity was most likely due to a threshold effect at 0.84 (tested with the STATA MIDAS module). However, no evidence of a threshold effect was found. Substantial het- erogeneity was observed in both the SEN (Cochran’s Q = 22.98, p b 0.001, I2 = 69.54%) and SPE (Cochran’s Q = 50.84, p b 0.001, I2 =

Table 1

Characteristics of the included studies.

Population

86.23%) assessments among all included studies [24]. A proportion of the heterogeneity probably was caused by population differences, small sample size bias, and different cut-off values. Therefore, we per- formed sensitivity analysis to explore the sources of potential heteroge- neity in SEN and SPE (Table 2). Due to the use of different cut-off values in the studies, we calculated the SEN and SPE of 6 studies with a cut-off value of 10.0. However the subsequently pooled performance indices were not significantly different; the SEN, SPE, PLR, NLR, and DOR were

Author

Year

De Jager

2010

Loonen Laukemann Lowsby

2014

2015

2015

Zhang HB

2016

Pan Ljungstorm

2017

2017

Camilo Vallejo

2017

Fig. 2. Summary of the methodological quality of the studies according to the QUADAS-2 (Quality Assessment of Diagnostic Accuracy Studies-2) criteria.

0.721 (95% CI 0.652, 0.780), 0.586 (95% CI 0.545, 0.626), 1.741 (95% CI

1.594, 1.902), 0.476 (95% CI 0.390, 0.582), and 3.654 (95% CI 2.804,

4.763), respectively. Six of the included studies comprising 5168 pa- tients were performed in the emergency care unit, and 2 studies did not specify patient sources. After omitting the latter two studies, the heterogeneity of the pooled DOR decreased from high to low, and the I2 index decreased from 91.54% to 15.84%. The pooled SEN, SPE PLR, NLR, and DOR were 0.727 (95% CI 0.666, 0.781), 0.585 (95% CI 0.544,

0.626), 1.753 (95% CI 1.593, 1.929), 0.466 (95% CI 0.385, 0.565), and

3.760 (95% CI 2.886, 4.899), respectively. The estimated AUC was 0.69 (95% CI 0.65, 0.73). A subsequent evaluation should possibly be con- ducted using subgroup analyses to rule out the possibility of small sam- ple size bias. We excluded three small-sized studies (b200 patients) [17,20,23] and calculated the SEN and SPE of the five remaining studies with relatively large numbers of patients (>=200) [12,18,19,21,22]. The SEN, SPE, PLR, NLR, and DOR were 0.702 (95% CI 0.642, 0.755), 0.597

(95% CI 0.559, 0.635), 1.742 (95% CI 1.604, 1.892), 0.499 (95% CI 0.424,

0.588), and 3.490 (95% CI 2.793, 4.464), respectively.

  1. Discussion

Bacteraemia is a common cause of increased morbidity and mortal- ity in critically ill patients, but early diagnosis and accurate identifica- tion are difficult. A series of studies explored the association between the NLCR and the diagnosis of bacteraemia. However, based on their re- sults, the current value of the NLCR for the diagnosis of bacteraemia is uncertain. The NLCR is a relative bacteraemia biomarker with no prior meta-analysis of its value. Our meta-analysis assessed its diagnostic ac- curacy for bacteraemia in 8 studies including 7095 cases, which is the first systematic study on this topic. The results indicated that the NLCR was not superior at predicting bacteraemia over a single biological marker (AUC = 0.69, SEN = 0.723, and SPE = 0.59).

An ideal biological marker should be able to distinguish among bac- terial, viral, and Fungal infections as well as between systemic sepsis and local infection. Currently, approximately 150 biomarkers have been used to evaluate sepsis in the clinic [29]. In a meta-analysis from 2015, including 58 studies published between 1999 and 2014, Hoeboer and

Fig. 3. Deek funnel plot asymmetry test to detect publication bias.

colleagues concluded that PCT is able to diagnosis bacteraemia, and the SROC curve was 0.79. He considered that low PCT levels can be used to exclude the presence of bacteraemia [30]. Qu and colleagues assessed the diagnostic value of PCT, CRP, interleukin-6 (IL-6) and serum amyloid A (SAA) levels for bacterial infections in febrile patients. The AUC values

for PCT, CRP, IL-6 and SAA were 0.804, 0.693, 0.658 and 0.687, respec- tively [31]. While some biomarkers for bacteraemia showed moderate diagnostic value, the quality, methods and sample size limitations that may affect their judgment diagnostic value [32]. Sepsis or bacteraemia is a pathophysiological process and a nonspecific syndrome that is too

Fig. 4. Forest plot of the sensitivity and specificity of the NLCR for the diagnosis of bacteraemia.

Fig. 5. Summary receiver operating characteristic graph for the included studies. AUC = area under curve, SEN = sensitivity, SPE = specificity.

complex to be explained by a single description, and thus, multiple combinations of biomarkers are needed to enhance clinical diagnosis in the ED. According to a previous study, the combination of lactate

Fig. 6. Fagan nomogram of NLCR for the diagnosis of bacteremia.

(Lac), PCT, The sequential organ failure assessment score (SOFA), and the NLCR is a useful tool for the early diagnosis of sepsis [34]. Studies showed that the NLCR and the currently available organ failure scoring systems, such as the Acute Physiology and Chronic Health Evaluation II (APACHE II) and SOFA, had weak but significant relationships, which could be useful for the assessment of severity in critical ill patients [35].

positive blood cultures have been used as the gold standard to diag- nosis bacteraemia. However, blood cultures are prone to errors. In par- ticular, the relationship between the amount of blood cultured and the timing of blood collection and the start of antibiotic treatment are im- portant factors [38]. When blood samples are collected after antibiotic treatment has begun, their sensitivity decreases significantly [7]. The NLCR is a potentially meaningful parameter for prediction of bacteraemia in patients suspected of having a community-acquired in- fection. A study showed that the NLCR measured after admission to the ICU was associated with short- and long-term mortality in critically ill patients [39]. Discussing the link between this marker and critical ill- ness should be the focus of our future research.

Furthermore, this meta-analysis has limitations. First, this meta- analysis included only 8 studies and the number of included studies was small. Second, sensitivity analyses were not performed to reduce and interpret heterogeneity. Third, variations in the sampling times across studies may have affected the results, leading to bias in this anal- ysis. Fourth, publication bias is the most common limitation of any sys- tematic review. Unpublished data were not discovered or used. Fifth, we included some small sample studies, which may affect consistency of studies and cause bias. Finally, research addressing the diagnostic accu- racy of the examined biomarker in patients of other ethnicities and from other regions may be required [28].

  1. Conclusions

The NLCR is an easy-to-collect and inexpensive diagnostic marker for bacteraemia. However, the NLCR cannot be recommended as the only definitive test for the diagnosis of sepsis, due to its low sensitiv- ity and specificity. It must be evaluated in the context of a detailed medical history, physical examination, and viable microbial assess- ment. In addition, reassessment should be continued during the dis- ease course. Only the use of a combination of multiple biomarkers will improve the diagnostic accuracy of bacteraemia in a clinical setting.

Abbreviations

Deek’s funnel test (p value)

b0.001 0.152 b0.001 b0.001

NLCR Neutrophil-lymphocyte count ratio

SROC the summary receiver-operating characteristic WBC white blood cell count

Test for heterogeneity (I2)

91.51

15.84

90.65

88.53

CRP C-reactive protein

PRISMA Preferred Reporting Items for Systematic Reviews and Meta- Analyses

TP true positives

FP false positives

FN false negatives

TN true negatives

DOR (95% CI)

3.839[3.007,4.901]

3.760[2.886,4.899]

3.654[2.804,4.763]

3.490 [2.791, 4.364]

PTP post-test probability

SEN sensitivity

SPE specificity

PLR positive likelihood ratio NLR negative likelihood ratio DOR diagnostic odds ratio

0.69[0.65,0.73]

0.69[0.65,0.73]

0.68[0.64,0.72]

0.69 [0.65-0.73]

Ethics approval and consent to participate

AUC (95% CI)

Not applicable.

Consent for publication

NLR (95% CI)

0.466[0.387,0.560]

0.466[0.385,0.565]

0.476[0.390,0.582]

0.499 [0.424, 0.588]

Not applicable.

Availability of data and materials

ED Emergency department, PLR Positive likelihood ratio, NLR Negative likelihood ratio, DOR Diagnostic odds ratio, AUC = area under curve,

a Number of patients >= 200.

The datasets analyzed during the current study are available in the PubMed, Web of Science, EMBASE and the Cochrane Library databases.

PLR (95% CI)

1788[1.642,1.946]

1.753[1.593,1.929]

1.741[1.594,1.902]

1.742 [1.604, 1.892]

Competing interests

The authors have no competing interests to declare.

Specificity (95% CI)

0.596[0.556,0.634]

0.585[0.544,0.626]

0.586[0.545,0.626]

0.597 [0.559, 0.635]

Funding

This work was not supported by any funding agency.

Authors’ contributions

Sensitivity (95% CI)

0.723[0.660,0.777]

0.727[0.666.0.781]

0.721[0.652,0.780]

0.702 [0.642, 0.755]

JWJ and RL contributed equally to this work. JWJ participated in the study design, selected trials, extracted data, performed the statistical analyses, and drafted the manuscript. RL participated in the study de- sign, selected trials, performed the statistical analyses, and drafted the manuscript. XY helped draft the manuscript and assessed the risk of bias of the trials. RY helped draft the manuscript and assisted with inter- pretation of the data. HX contributed to data collection. ZM participated in manuscript preparation. YQW collected the data, performed the sta- tistical analyses, and supervised the study. All authors read and ap- proved the final manuscript.

Number of studies

8

6

6

5

Patients admitted in ED Cut-off value of 10.0 Relatively large samplea

Acknowledgements

Categories

Not applicable.

Table 2

Results of sensitivity analysis.

References

All studies Sensitivity analysis

  1. Kumar A, Roberts D, Wood KE, Light B, Parrillo JE, Sharma S, et al. Duration of hypo- tension before initiation of effective antimicrobial therapy is the critical determinant of survival in human septic shock. Crit Care Med 2006;34(6):1589-96.
  2. Vincent JL. Clinical sepsis and septic shock–definition, diagnosis and management principles. Langenbecks Arch Surg 2008;393(6):817-24.
  3. Emonet S, Schrenzel J. How could rapid bacterial identification improve the manage- ment of septic patients? Expert Rev Anti Infect Ther 2011;9(9):707-9.
  4. Shapiro MF, Greenfield S. The complete blood count and leukocyte differential count. An approach to their rational application. Ann Intern Med 1987;106(1): 65-74.
  5. Seebach JD, Morant R, Ruegg R, Seifert B, Fehr J. The diagnostic value of the neutro- phil left shift in predicting inflammatory and infectious disease. Am J Clin Pathol 1997;107(5):582-91.
  6. Wyllie DH, Bowler IC, Peto TE. Bacteraemia prediction in emergency medical admis- sions: role of C reactive protein. J Clin Pathol 2005;58(4):352-6.
  7. Grace CJ, Lieberman J, Pierce K, Littenberg B. Usefulness of blood culture for hospitalized patients who are receiving antibiotic therapy. Clin Infect Dis 2001;32(11):1651-5.
  8. McKenzie R, Reimer LG. Effect of antimicrobials on blood cultures in endocarditis. Diagn Microbiol Infect Dis 1987;8(3):165-72.
  9. Hamiel U, Bahat H, Kozer E, Hamiel Y, Ziv-Baran T, Goldman M. Diagnostic markers of Acute infections in infants aged 1 week to 3 months: a retrospective cohort study. BMJ Open 2018;8(1):e018092.
  10. Le Tulzo Y, Pangault C, Gacouin A, Guilloux V, Tribut O, Amiot L, et al. Early circulat- ing lymphocyte apoptosis in human septic shock is associated with poor outcome. Shock 2002;18(6):487-94 (Augusta, Ga).
  11. An X, Ding PR, Li YH, Wang FH, Shi YX, Wang ZQ, et al. Elevated neutrophil to lym- phocyte ratio predicts survival in advanced pancreatic cancer. Biomarkers 2010;15 (6):516-22.
  12. de Jager CP, van Wijk PT, Mathoera RB, de Jongh-Leuvenink J, van der Poll T, Wever PC. Lymphocytopenia and neutrophil-lymphocyte count ratio predict bacteremia better than conventional infection markers in an emergency care unit. Crit Care 2010;14(5):R192 (London, England).
  13. Yoldas H, Karagoz I, Ogun MN, Velioglu Y, Yildiz I, Bilgi M, et al. Novel mortality markers for critically ill patients. J Intensive Care Med 2018. https://doi.org/10. 1177/0885066617753389 885066617753389.
  14. Whiting P, Rutjes AW, Reitsma JB, Bossuyt PM, Kleijnen J. The development of QUADAS: a tool for the quality assessment of studies of diagnostic accuracy included in systematic reviews. BMC Med Res Methodol 2003;3:25.
  15. Reitsma JB, Glas AS, Rutjes AW, Scholten RJ, Bossuyt PM, Zwinderman AH. Bivariate analysis of sensitivity and specificity produces informative summary measures in di- agnostic reviews. J Clin Epidemiol 2005;58(10):982-90.
  16. Arends LR, Hamza TH, van Houwelingen JC, Heijenbrok-Kal MH, Hunink MG, Stijnen

T. Bivariate random effects meta-analysis of ROC curves. Med Decis Making 2008;28 (5):621-38.

  1. Loonen AJ, de Jager CP, Tosserams J, Kusters R, Hilbink M, Wever PC, et al. Bio- markers and molecular analysis to improve bloodstream infection diagnostics in an emergency care unit. PLoS One 2014;9(1):e87315.
  2. 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 2015;94(49):e2264.
  3. Lowsby R, Gomes C, Jarman I, Lisboa P, Nee PA, Vardhan M, et al. Neutrophil to lym- phocyte count ratio as an early indicator of blood stream infection in the emergency department. Emerg Med J 2015;32(7):531-4.
  4. Zhang H-B, Chen J, Lan Q-F, Ma X-J, Zhang S-Y. diagnostic values of red cell distribu- tion width, platelet distribution width and neutrophil-lymphocyte count ratio for sepsis. Exp Ther Med 2016;12(4):2215-9.
  5. Ljungstrom L, Pernestig AK, Jacobsson G, Andersson R, Usener B, Tilevik D. Diagnostic accuracy of procalcitonin, neutrophil-lymphocyte count ratio, C-reactive protein, and lactate in patients with suspected bacterial sepsis. PLoS One 2017;12(7): e0181704.
  6. Pan Y-P, Fang Y-P, Xu Y-H, Wang Z-X, Shen J-L. The diagnostic value of procalcitonin versus other biomarkers in prediction of bloodstream infection. Clin Lab 2017;63(2): 277-85.
  7. Vallejo C, Londono HF, Vargas Zabala D, Solano AF, Tavera K, Maya L. Neutrophil-to- lymphocyte ratio in adult patients with bacteremia at the emergency service. Repertorio de Medicina y Cirugia 2017;26(3):138-43.
  8. Wu CC, Lan HM, Han ST, Chaou CH, Yeh CF, Liu SH, et al. Comparison of diagnostic accuracy in sepsis between presepsin, procalcitonin, and C-reactive protein: a sys- tematic review and meta-analysis. Ann Intensive Care 2017;7(1):91.
  9. Backes Y, van der Sluijs KF, Mackie DP, Tacke F, Koch A, Tenhunen JJ, et al. Usefulness of suPAR as a biological marker in patients with systemic inflammation or infection: a systematic review. Intensive Care Med 2012;38(9):1418-28.
  10. Wacker C, Prkno A, Brunkhorst FM, Schlattmann P. Procalcitonin as a diagnostic marker for sepsis: a systematic review and meta-analysis. Lancet Infect Dis 2013; 13(5):426-35.
  11. Hoeboer SH, van der Geest PJ, Nieboer D, Groeneveld AB. The diagnostic accuracy of procalcitonin for bacteraemia: a systematic review and meta-analysis. Clin Microbiol Infect 2015;21(5):474-81.
  12. Qu J, L X, Liu Y, Wang X. Evaluation of procalcitonin, C-reactive protein, interleukin-6 & serum amyloid A as Diagnostic biomarkers of bacterial infection in febrile patients. Indian J Med Res 2015;141(3):315-21.
  13. Liu Y, Hou JH, Li Q, Chen KJ, Wang SN, Wang JM. Biomarkers for diagnosis of sepsis in patients with systemic inflammatory response syndrome: a systematic review and meta-analysis. Springerplus 2016;5(1):2091.
  14. Boran OF, Yazar FM, Boran M, Urfalioglu A, Bakacak Z, Yildiz M, et al. The preseptic period and inflammatory markers in the prediction of the course of sepsis. Med Sci Monit 2018;24:3531-9.
  15. Velissaris D, Pantzaris ND, Bountouris P, Gogos C. Correlation between neutrophil- to-lymphocyte ratio and severity scores in septic patients upon hospital admission. A series of 50 patients. Rom J Intern Med 2018;56(3):153-7.
  16. Weinstein MP. Current blood culture methods and systems: clinical concepts, tech- nology, and interpretation of results. Clin Infect Dis 1996;23(1):40-6.
  17. Salciccioli JD, Marshall DC, Pimentel MAF, Santos MD, Pollard T, Celi LA, et al. The as- sociation between the neutrophil-tolymphocyte ratio and mortality in critical ill- ness: an observational cohort study. Crit Care 2015;19.