a b s t r a c t

Objective: community-acquired pneumonia is a common presentation to the emergency department (ED) and has high mortality rates. The aim of our study is to investigate the risk stratification and prognostic prediction value of precalcitonin (PCT) and clinical severity scores on patients with CAP in ED.

Methods: 226 consecutive adult patients with CAP admitted in ED of a Tertiary teaching hospital were enrolled. Demographic information and clinical parameters including PCT levels were analyzed. CURB65, PSI, SOFA and qSOFA scores were calculated and compared between the severe CAP (SCAP) and non-severe CAP (NSCAP) group or the death and survival group. Receiver-operating characteristic (ROC) curves for 28-day mortality were calculated for each predictor using cut-off values. Logistic regression models and area under the curve (AUC) analysis were performed to compare the performance of predictors.

Results: Fifty-one patients were classified as SCAP and forty-nine patients died within 28 days. There was signif- icant difference between either SCAP and NSCAP group or death and survival group in PCT level and CURB65, PSI, SOFA, qSOFA scores (p b 0.001). The AUCs of the PCT and CURB65, PSI, SOFA and qSOFA in predicting SCAP were 0.875, 0.805, 0.810, 0.852 and 0.724, respectively. PCT is superior in predicting SCAP and the models combining PCT and SOFA demonstrated superior performance to those of PCT or the CAP severity score alone. The AUCs of the PCT and CURB65, PSI, SOFA and qSOFA in predicting 28-day mortality were 0.822, 0.829, 0.813, 0.913 and 0.717, respectively. SOFA achieved the highest AUC and the combination of PCT and SOFA had the highest supe- riority over other combinations in predicting 28-day mortality.

Conclusion: Serum PCT is a valuable single predictor for SCAP. SOFA is superior in prediction of 28-day mortality. Combination of PCT and SOFA could improve the performance of single predictors. More further studies with larger sample size are warranted to validate our results.

(C) 2018

Introduction

Community-acquired pneumonia is one of the most common causes of mortality worldwide and accounts for a substantial use of Healthcare resources [1]. It results in 29,000 death every year in UK and is being associated with high rates of hospital admission and length of stay [2]. The mortality rate of patients with CAP varies according to its severity, treatment failure and the requirement for hospitalization and/ or intensive care unit (ICU) management [3]. Stratifying severity and

* Corresponding author at: Department of Emergency Medicine, Beijing Chao-yang Hospital, Capital Medical University, No. 8 Workers’ Stadium South Road, Chao-yang District, Beijing 100020, China.

E-mail address: gsbchaoyang@sina.cn. (S. Guo).

predicting prognosis of CAP is of vital importance as they can be helpful for hospitalization decisions.

Current guidelines recommend the application of several severity scores, such as CURB65 (confusion, urea N7 mmol/L, respiratory rate 30/min, low systolic (b90 mm Hg) or diastolic blood pressure (<=60 mm Hg) and age 65 years) and pneumonia severity index , to classify and stratify CAP patients to determine whether the patients should receive outpatient care or hospital admission [4]. The PSI score, introduced in 1997 following a study of N50,000 patients with CAP, is a 20-point score which classified patients into five risk categories ac- cording to their percent risk of death within 30 days [5]. Comparatively, the CURB65 score is significantly easier to remember and use than the PSI and is composed of only five variables with a single point awarded to each [5]. The CRB65, without requirement to measure blood urea, is

https://doi.org/10.1016/j.ajem.2018.03.050

0735-6757/(C) 2018

Table 1

Comparison of different groups in past history

Group	n	COPD	CDVD	CBVD	Diabetes	CRD	Healthy
SCAP	51	11	10	17	10	5	5
NSCAP	175	48	32	48	27	9	14
?2		0.703	0.046	0.672	0.504	0.783	0.015
p value		0.402	0.831	0.412	0.478	0.376	0.903
Death	49	10	9	15	9	6	4
Survival	177	49	33	50	28	8	15

?² 1.053 0.002 0.105 0.523 2.724 0.000

p value 0.305 0.965 0.746 0.469 0.099 1.000

SCAP severe community-acquired pneumonia; NSCAP none-severe community-acquired pneumonia; CDVD cardio-vascular disease; CBVD cerebral-vascular disease; CRD chronic renal disease.

recommended for outpatient use [5]. Previous study revealed that there were no significant differences in overall test performance between PSI, CURB65 and CRB65 for predicting mortality from CAP [5]. The Sepsis-3 Task Force updated the clinical criteria for sepsis, but the clinical impli- cation of SOFA and qSOFA remained unknown [6]. Ranzani et al. re- ported that qSOFA outperformed Systemic Inflammatory Response Syndrome and presented better clinical usefulness as prompt tools for patients with CAP in ED and PSI had the best decision-aid tool profile [6].

Nonetheless, previous studies have demonstrated the limitations of existing severity scores and advised that they should be used with cau- tion and in conjunction with clinical judgment [7]. Moreover, there has been considerable interest in the development of rapid biomarkers for reliable prognosis prediction [7]. Among them, C-reactive protein (CRP) and precalcitonin (PCT) are widely used in virtue of their higher Predictive capacity [8]. Frank et al. reported that PCT is associated with the severity of illness in patients with severe pneumonia and ap- pears to be a prognostic marker of morbidity and mortality [9]. Kim et al. proved that PCT is a reliable single predictor for short-term mortal- ity [10]. Regrettably, previous studies are inconsistent in regard to whether biomarkers are superior to severity scores in predicting prog- nosis of patients with CAP [7].

The purpose of the study is to investigate the risk stratification and prognostic prediction value of PCT and clinical severity scores (CURB65, PSI, SOFA and qSOFA) for patients with CAP in the ED. The pri- mary end-point used in this study is 28-day mortality.

Methods

This study was approved by the Institutional Review Board and Medical Ethics Committee of Beijing Chao-yang Hospital, Capital Medi- cal University, which is an urban university hospital with approximately 250,000 ED admissions every year. Written informed consents were col- lected from all enrolled patients. Adult patients who fulfilled the CAP criteria [4] admitted to the Emergency Department of Beijing Chao- yang Hospital between January 2016 and October 2017 were enrolled. The following patients were excluded from this study: a) age b 18 years old, b) patients with metastatic tumor, AIDS, uremia, late stage of liver cirrhosis, active tuberculosis, refractory heart failure, pre- vious transplantation, immunosuppressive therapy and pregnancy,

c) patients who refused to participate in this study, d) patients from hospice or patients with DNR (Do-Not-Resuscitate) request.

Demographics information of all enrolled patients were collected and recorded on admission. Clinical (blood pressure, respiratory rate, pulse rate) and laboratory parameters (white blood cell count, hemo- globin level, platelet level, hematocrit, precalcitonin, renal function, he- patic function, electrolytes, arterial blood gas) were assessed and recorded. CURB65, PSI, SOFA and qSOFA scores were calculated accord- ing to international criteria and analyzed. PCT levels were analyzed from serum electrochemiluminescent immunoassay using the PCT kit and the mini-VIDAS(R) system (BIOMERIEUX SA FRANCE).

All patients were divided into severe community-acquired pneumo- nia (SCAP) group and non-severe community-acquired pneumonia (NSCAP) group according to Consensus guidelines [4]. All patients were followed up for 28 days and the primary end point was death at 28 days. Patients were divided into death group and survival group ac- cording to their prognosis. According to international consensus and guidelines [4], patients who meet the following criteria were defined as having SCAP (>=one major criteria or >=three minor criteria). Major criteria: a) mechanical ventilation. b) septic shock with the need for va- sopressors. Minor criteria: a) respiratory rate >= 30 breaths/min. b) PaO₂/ FiO₂ ratio <= 250. c) multilobar infiltrates. d) confusion/disorientation.

e) uremia (BUN level >= 20 mg/dL). f) leucopenia (WBC count

b 4000 cells/mm³). g) thrombocytopenia (platelet count

b 100,000 cells/mm³). h) hypothermia (core temperature b 36 ?C).

i) hypotension requiring aggressive fluid resuscitation.

Statistical analysis

Continuous variables were described as the mean +- standard devi- ation and compared using one-way analysis of variance (ANOVA) for the normally distributed data. For skewed distributions, the data are presented as the median (interquartile range) and compared using Mann Whitney-U nonparametric test. The categorical variables were described as percentages and compared using the Chi-Squared test or Fisher’s exact test. Multivariate logistic regression was performed to an- alyze the potential determinants for risk stratification and prognostic prediction of CAP. All analyses were performed using SPSS 22.0 statisti- cal software package (SPSS Inc, Chicago, IL, USA). receiver operating characteristics (ROC) analysis was carried out and area under the curve (AUC) was compared using MedCalc 15.0 Software (Acacialaan, Ostend, Belgium) to evaluate the predicting ability of PCT and CAP se- verity scores. Based on the cut-off values, sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) were also calculated. A Z-test was used for comparing the AUCs between different curves. A two-tailed value of p b 0.05 was considered statistically significant.

Results

A total of 226 patients were enrolled in this study. Of them, fifty-one patients were classified as having SCAP and forty-nine patients were dead after a 28-day follow-up (Table 1). The total mortality rate was 21.68% (49/226). The overall mean age of the patients was 65 (58,71) years old and the male to female ratio was 5.28:1 (190:36). Past history

Table 2

PCT and CAP severity scores comparison in predicting SCAP

Group	n	Age (years)	Male n (%)	PCT (ng/mL)	CURB65	PSI	SOFA	qSOFA
SCAP	51	64.7 +- 11.9	42 (82.4)	4.20 (1.60,13.78)	3 (2,3)	132.0 +- 33.7	3 (2,4)	2 (1,2)
NSCAP	175	65 (59,70)	148 (84.6)	0.27 (0.10,1.05)	1 (1,2)	94 (81,109)	6 (5,9)	3 (2,3)
Z value		-0.656	-0.380	-8.146	-7.026	-6.739	-7.745	-5.418
p value		0.512	0.704	0.000	0.000	0.000	0.000	0.000

SCAP severe community-acquired pneumonia; NSCAP none-severe community-acquired pneumonia.

Table 3

PCT and CAP severity scores comparison in predicting 28-day mortality

Group	n	Age (years)	Male (%)	PCT (ng/mL)	CURB65	PSI	SOFA	qSOFA
Death	49	68 (58,76)	41 (83.7)	3.60 (1.30,11.32)	3 (2,3)	133.9 +- 34.5	6 (6,10)	2 (1,3)
Survival	177	65 (58,70)	149 (84.2)	0.27 (0.11,1.10)	1 (1,2)	95 (81,108)	3 (2,4)	1 (1,2)
Z value		-1.616	-0.086	-6.892	-7.476	-6.708	-8.967	-5.177
p value		0.106	0.932	0.000	0.000	0.000	0.000	0.000

of enrolled patients includes COPD (26.11%), cardiovascular disease (18.58%), cerebral vascular disease (28.76%), chronic renal disease (6.19%), diabetes (16.37%) and healthy (8.41%). There was no statistical significance between the SCAP and NSCAP group or death and survival group in age, male to female ratio and past history (Tables 1 and 2).

As to PCT level and CURB65, PSI, SOFA, qSOFA scores, there were sig- nificant differences between either SCAP and NSCAP group or death and survival group (Tables 2 and 3). The PCT level and CURB65, PSI, SOFA, qSOFA scores were significantly higher in SCAP group or in death group. According to ROC curve, the AUC of PCT was higher than that of CURB65, PSI, SOFA and qSOFA in predicting SCAP (Table 4, Fig. 1). The cut-off value was 1.16 ng/mL. There was no significant difference in multiple pairwise comparisons between PCT and CURB65 (p = 0.069), PCT and PSI (p = 0.088), PCT and SOFA (p = 0.540), CURB65 and PSI (p = 0.894), CURB65 and SOFA (p = 0.249), PSI and SOFA (p =

0.303). However, significant difference was found in multiple pairwise comparisons between qSOFA and PCT (p = 0.001), qSOFA and CURB65 (p = 0.038), qSOFA and SOFA (p = 0.012). This indicated that qSOFA had the lowest superiority to other severity scores and PCT in predicting SCAP. When combination was made between PCT and se- verity scores in predicting SCAP, the combination of PCT and SOFA (PCT

+ SOFA) achieved the highest AUC (Table 4, Fig. 3). However, there was no significant difference in multiple pairwise comparisons between PCT

+ SOFA and PCT + CURB65 (p = 0.145), PCT + SOFA and PCT + PSI (p

= 0.137), PCT + CURB65 and PCT + PSI (p = 0.957), PCT + PSI and PCT

+ qSOFA (p = 0.122).

According to ROC curve, the AUC of SOFA was higher than that of PCT, CURB65, PSI and qSOFA in predicting 28-day mortality (Table 5, Fig. 2). The cut-off value was 5.50. There was no significant difference in multiple pairwise comparisons between PCT and CURB65 (p = 0.854), PCT and PSI (p = 0.848), CURB65 and PSI (p = 0.666), PSI and

qSOFA (p = 0.056). However, significant difference was found in multi- ple comparisons between SOFA and PCT (p = 0.014), SOFA and CURB65 (0.020), SOFA and PSI (p = 0.008), SOFA and qSOFA (p = 0.000). This

indicated that SOFA had superiority to other severity scores and PCT in predicting 28-day mortality. When combination was made between PCT and severity scores in predicting 28-day mortality, the combination of PCT and SOFA achieved the highest AUC (Table 5, Fig. 4). There was no significant difference in multiple pairwise comparisons between PCT

+ CURB65 and PCT + PSI (p = 0.502), PCT + PSI and PCT + qSOFA

(p = 0.058). However, there was significant difference in multiple pairwise comparisons in PCT + SOFA and PCT + CURB65 (p = 0.018), PCT + SOFA and PCT + PSI (p = 0.007), PCT + SOFA and PCT

+ qSOFA (p = 0.000). This indicated that the combination of PCT and SOFA had the highest superiority to other combinations in predicting 28-day mortality.

Binary multivariate logistic regression revealed that PCT, CURB65, PSI, SOFA and age were independent risk factors for predicting not only SCAP, but also 28-day mortality. However, qSOFA score was not listed in the multivariate logistic regression model and it was not risk factors for predicting not only SCAP, but also 28-day mortality (Table 6).

Discussion

Community-acquired pneumonia is one of the most common infec- tious diseases and is an important cause of mortality and morbidity worldwide, especially among elderly patients with several comorbidi- ties. Identifying patients with CAP with higher risk of mortality is crucial to anticipate prognosis and program follow-up.

Many different CAP severity scoring systems have been developed in order to evaluate the severity of CAP, but debate is still going on about their performances. pneumonia severity index and CURB65 are the two kinds of most widely used severity scores [11,12]. Previous studies revealed that they can identify individuals at low risk of death who are candidates for outpatient care [7,13], but perform less well in predicting ICU admission [14]. Moreover, compared with APACHE (Acute Physiology and Chronic Health Evaluation) IIscore, CURB65 and PSI were not found valuable in predicting mortality [15]. Chalmers et al. revealed that there were no significant differences in overall test performance between PSI, CURB65 and CRB65 for predicting mortality from CAP [5]. The Sepsis-3 Task Force updated the clinical criteria for sepsis, but the clinical implication of SOFA and qSOFA remained un- known [15]. Ranzani et al. reported that qSOFA and the CRB (Confusion, Respiratory Rate and Blood Pressure) score outperformed SIRS (sys- temic inflammatory response syndrome) criteria and presented better clinical usefulness as prompt tools for patients with CAP in ED [6]. Kabundji et al. demonstrated that the utility of CRB65 score in accu- rately determining the need for admission of patients with CAP pre- sented to ED [16]. Wang et al. developed a new method and concluded that CLCGH scoring system, including Serum creatinine (Cr) N259.5 umol/L, leukocyte (WBC) N 17.35 x 10⁹/L, C-reactive protein (CRP) N 189.4 ug/mL, GCS <= 9 and serum HCO^– <= 17.65 mmol/L, is an ef- ficient accurate and objective method to predict the early hospital mor- tality among SCAP patients and the AUC of CLCGH is similar to SOFA and better than PSI and CURB65 score [17].

3

Table 4

statistical data of ROC curve between PCT and CAP severity scores in predicting SCAP

	AUC (95%CI)	p value	Cut-off value	Sensi	Speci	PPV (%)	NPV (%)
PCT	0.875 (0.830-0.920)	0.000	1.16	0.902	0.783	54.78	96.48
CURB65	0.805 (0.738-0.873)	0.000	1.50	0.843	0.646	40.97	93.38
PSI	0.810 (0.747-0.873)	0.000	105.50	0.745	0.726	44.21	90.71
SOFA	0.852 (0.786-0.917)	0.000	5.50	0.706	0.891	65.37	91.23
qSOFA	0.724 (0.638-0.809)	0.000	1.5	0.725	0.731	44.00	90.12
PCT + CURB65	0.886 (0.838-0.935)	0.000	0.18	0.882	0.783	54.27	95.79
PCT + PSI	0.885 (0.839-0.931)	0.000	0.14	0.922	0.726	49.56	96.96
PCT + SOFA	0.921 (0.883-0.959)	0.000	0.30	0.784	0.909	71.52	93.52
PCT + qSOFA	0.840 (0.777-0.903)	0.000	0.22	0.765	0.817	54.97	92.25

SCAP severe community-acquired pneumonia; Sensi sensitivity; Speci specificity; PPV positive predictive value; NPV negative predictive value.

Fig. 1. ROC curve of PCT and CAP severity scores in predicting SCAP. Fig. 2. ROC curve of PCT and CAP severity scores in predicting 28-day mortality.

In view of the limitations of existing CAP severity scores, consider- able interest has been centering on the development of reliable bio- markers that may confer reliable prognostic prediction information. Various kinds of biomarkers have been examined and validated for ap- plication in CAP. Traditional biomarkers such as white blood cell count and erythrocyte sedimentary rate (ESR) are less reliable because of their lower sensitivity and specificity, compared with the most widely-used C-reactive protein (CRP) and PCT [18]. CRP responses on the third day is reported to be a valuable predictor of 30 days mortality in hospitalized CAP patients and failure to decline in CRP was associated with a poor prognosis irrespective of the actual level of CRP or CURB-65 [19]. In addition, Colak et al. showed that PCT and CRP level were signif- icantly higher in patients with CAP in ED with indications for hospitali- zation than in patients with COPD and CRP is a more valuable marker [20]. Kim et al. proved that serum PCT is a reliable single predictor for short-term mortality and inclusion of PCT and/or CRP could significantly improve the performance of PSI and IDAS/ATS guidelines [10]. However, Christ-Crain held the opinion that the Prognostic accuracy of PCT is low as it yields overlapping values for different CAP severity [21].

Nowadays, new attention has shifted to adreno-medullin and its more stable mid-regional fragment pro-adrenomedullin (MR-pro- ADM). Preliminary studies have demonstrated its superiority in mortal- ity and prognosis prediction for CAP [22-24]. Assessment of MR-pro- ADM levels in CAP patients in addition to CURB65 scores could increase the prognostic accuracy of CURB65 alone [24]. Moreover, MR-pro-ADM in combination with PSI is helpful in individual risk stratification for short-term poor outcome of CAP patients, allowing a better reclassifica- tion of patients compared with PSI alone [25]. Nonetheless, more future

studies are warranted to determine the prognostic accuracy of MR-pro- ADM in combination with other severity scores or biomarkers [26].

In recent years, fatty-acid-binding proteins (FABPs) have also attracted attention as potential biomarkers of tissue injury in acute and chronic inflammatory circumstances. Chen et al. reported that H- FABP was found to be an independent predictor of 28-day mortality and the predictive accuracy of H-FABP was improved when used in combination with either CURB65 or PSI [27]. Tsao et al. also concluded that the urinary levels of adipocyte-FABP can serve as a new biomarker in assessing the severity of CAP and predicting the outcome of pneumo- nia in critically-ill patients who required admission in ICU [28].

The neutrophil-lymphocyte ratio (NLR) at admittance has already been described as a better Prognostic biomarker in CAP than traditional infection markers in severity and outcome prediction [29]. Neutrophil count percentage (NCP) and NLR are reported as promising candidate predictors of mortality for hospitalized CAP patients, and both of them are cheaper, easier to perform and as reliable as the new biomarkers [30].

Our study revealed that PCT, CURB65, PSI, SOFA and qSOFA score levels increased with the aggravation of the disease severity (all p b 0.001) and there were significant difference between either SCAP and NSCAP group or death and survival group in PCT level and CURB65, PSI, SOFA, qSOFA scores. This indicated that higher PCT level and CURB65, PSI, SOFA, qSOFA scores could discriminate patients with poor prognosis. The ROC analysis demonstrated that PCT is superior in predicting SCAP, but there was no significant difference in pairwise comparisons. Comparatively, SOFA not only had the highest AUC than CURB65, PSI, qSOFA and PCT in predicting 28-day mortality, but also

Table 5

Statistical data of ROC curve between PCT and CAP severity scores in predicting 28-day mortality

	AUC (95%CI)	p value	Cut-off value	Sensi	Speci	PPV (%)	NPV (%)
PCT	0.822 (0.758-0.886)	0.000	1.16	0.763	0.857	60.91	92.53
CURB65	0.829 (0.766-0.893)	0.000	1.50	0.650	0.878	60.87	89.57
PSI	0.813 (0.747-0.880)	0.000	116.0	0.831	0.673	42.60	93.17
SOFA	0.913 (0.864-0.962)	0.000	5.50	0.915	0.816	59.22	97.05
qSOFA	0.717 (0.629-0.805)	0.000	1.50	0.723	0.714	42.47	89.83
PCT + CURB65	0.877 (0.828-0.926)	0.000	0.18	0.780	0.898	69.07	93.32
PCT + PSI	0.861 (0.805-0.916)	0.000	0.16	0.757	0.837	57.56	92.19
PCT + SOFA	0.932 (0.899-0.965)	0.000	0.24	0.870	0.857	63.98	95.76
PCT + qSOFA	0.796 (0.720-0.871)	0.000	0.22	0.802	0.735	46.91	92.71

AUC area under the curve; Sensi sensitivity; Speci specificity; PPV positive predictive value; NPV negative predictive value; CI confidence interval.

Fig. 3. ROC curve of multiple PCT and CAP severity scores combinations in predicting SCAP.

had advantages in pairwise comparisons. According to previous studies, it is suggested that combinations of CAP severity scoring systems and biomarkers will be of greater significance than individual biomarkers. Our study proved that the combination of PCT and SOFA achieved the highest AUC in predicting not only SCAP but also 28-day mortality and the sensitivity and the negative predictive value was also improved (Tables 4 and 5). Moreover, there were also significant difference in multiple pairwise comparisons in PCT + SOFA and PCT + CURB65 (p

= 0.018), PCT + SOFA and PCT + PSI (p = 0.007), PCT + SOFA and

PCT + qSOFA (Table 5). These indicated that the combination of PCT and SOFA had the highest superiority to other combinations in predicting not only SCAP, but also 28-day mortality. Different from pre- vious study [31], our research demonstrated that qSOFA had the lowest superiority than other severity scores and PCT in predicting SCAP and it was not an independent risk factor by multivariate logistic regression. In term of current research situations, further multi-center studies with large sample size are warranted to corroborate the additive value of

Fig. 4. ROC curve of multiple PCT and CAP severity scores combinations in predicting 28- day mortality.

Table 6

Independent predictive variables analysis by multivariate logistic regression

Variables ? ?2 p					Adjusted	95%CI
				value	OR
SCAP	PCT	0.122	14.101	0.000	1.129	1.060-1.204
	CURB65	0.678	3.954	0.047	1.970	1.010-3.844
	PSI	0.033	10.245	0.001	1.033	1.013-1.054
	SOFA	0.674	20.349	0.000	1.963	1.464-2.631
	Age	-0.072	7.111	0.008	0.931	0.883-0.981
	Constant	-5.511	11.663	0.001	0.004
28-day	PCT	0.071	6.081	0.014	1.074	1.015-1.137
mortality	CURB65	0.919	6.331	0.012	2.506	1.225-5.127
	PSI	0.034	9.931	0.002	1.034	1.013-1.056
	SOFA	0.880	24.748	0.000	2.410	1.704-3.408
	Age	-0.059	4.039	0.044	0.943	0.890-0.999
	Constant	-7.826	16.601	0.000	0.000

SCAP severe community-acquired pneumonia; OR odds ratio; CI confidence interval.

these markers to clinical prediction scores to provide a safer and more effective assessment tool for clinicians [3].

The limitation of our study lies in the relatively small sample size and the retrospective non-randomized single-center design, as these can re- sult in selection bias and do not allow for analysis of all clinical data from all patients. Further randomized, multicenter studies with larger sample size are warranted to validate our results.

In conclusion, we investigated the risk stratification and prognostic prediction value of PCT and several clinical severity scores on patients with community-acquired pneumonia in ED. PCT is a valuable single predictor for SCAP. SOFA is superior in prediction of 28-day mortality. Combination of PCT and SOFA achieved the highest superiority to other combinations in predicting not only SCAP, but also 28-day mortal- ity. Combination of PCT and SOFA could improve the performance of single predictors.

Competing interest

The authors declare that they have no competing interest.

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

References

1. Heron M, Hoyert DL, Murphy SL, Xu J, Kochanek KD, Tejada-Vera B. Deaths: final data for 2006. Natl Vital Stat Rep 2009;57(14):1-134.
2. Chalmers J, Campling J, Ellsbury G, Hawkey PM, Madhava H, Slack M. Community- acquired pneumonia in the United Kingdom: a call to action. Pneumonia (Nathan) 2017;9:15.
3. Khan F, Owens MB, Restrepo M, Povoa P, Martin-Loeches I. Tools for outcome pre- diction in patients with community acquired pneumonia. Expert Rev Clin Pharmacol 2017;10(2):201-11.
4. Mandell LA, Wunderink RG, Anzueto A, Bartlett JG, Campbell GD, Dean NC, et al. In- fectious Diseases Society of America/American Thoracic Society consensus guide- lines on the management of community-acquired pneumonia in adults. Clin Infect Dis 2007;44(Suppl. 2):S27-72.
5. Chalmers JD, Singanayagam A, Akram AR, Mandal P, Short PM, Choudhury G, et al. Severity assessment tools for predicting mortality in hospitalized patients with community-acquired pneumonia. Systematic review and meta-analysis. Thorax 2010 Oct;65(10):878-83.
6. Ranzani OT, Prina E, Menendez R, Ceccato A, Cilloniz C, Mendez R, et al. New Sepsis definition (Sepsis-3) and community-acquired pneumonia mortality. A validation and clinical decision-making study. Am J Respir Crit Care Med 2017;196(10): 1287-97.
7. Viasus D, Simonetti A, Garcia-Vidal C, Carratala J. Prediction of prognosis by markers in community-acquired pneumonia. Expert Rev Anti Infect Ther 2013;11(9): 917-29.
8. Christ-Crain M, Opal SM. Clinical review: the role of biomarkers in the diagnosis and management of community-acquired pneumonia. Crit Care 2010;14(1):203.
9. Bloos F, Marshall JC, Dellinger RP, Vincent JL, Gutierrez G, Rivers E, et al. Multina- tional, observational study of procalcitonin in ICU patients with pneumonia requir- ing mechanical ventilation: a Multicenter observational study. Crit Care 2011;15(2): R88.
10. Kim MW, Lim JY, Oh SH. Mortality prediction using serum biomarkers and various clinical risk scales in community-acquired pneumonia. Scand J Clin Lab Invest 2017 Nov;77(7):486-92.
11. Fine MJ, Auble TE, Yealy DM, Hanusa BH, Weissfeld LA, Singer DE, et al. A prediction rule to identify low-risk patients with community-acquired pneumonia. N Engl J Med 1997;336(4):243-50.
12. Bauer TT, Ewig S, Marre R, Suttorp N, Welte T, CAPNETZ Study Group. CRB-65 pre- dicts death from community-acquired pneumonia. J Intern Med 2006;260(1): 93-101.
13. Lim WS, van der Eerden MM, Laing R, Boersma WG, Karalus N, Town GI, et al. Defin- ing community acquired pneumonia severity on presentation to hospital: an inter- national derivation and validation study. Thorax 2003;58(5):377-82. [15] Chalmers JD, Mandal P, Singanayagam A, Akram AR, Choudhury G, Short PM, et al. Severity assessment tools to guide ICU admission in community- acquired pneumonia: systematic review and meta-analysis. Intensive Care Med 2011;37(9):1409-20.

    Aydogdu M, Ozyilmaz E, Aksoy H, Gursel G, Ekim N. Mortality prediction in community-acquired pneumonia requiring mechanical ventilation; values of pneu- monia and intensive care unit severity scores. Tuberk Toraks 2010;58(1):25-34.

14. Kabundji DM, Musekiwa A, Mukansi M, Feldman C. Determining need for hospitalisation: evaluation of the utility of the CRB-65 score in patients with community-acquired pneumonia presenting to an emergency department. S Afr Med J 2014;104(11):769-72.
15. Wang X, Jiao J, Wei R, Feng Y, Ma X, Li Y, et al. A new method to predict hospital mor- tality in severe community acquired pneumonia. Eur J Intern Med 2017;40:56-63.
16. Shaddock EJ. How and when to use common biomarkers in community-acquired pneumonia. Pneumonia (Nathan) 2016;8:17.
17. Andersen SB, Baunbaek Egelund G, Jensen AV, Petersen PT, Rohde G, Ravn P. Failure of CRP decline within three days of hospitalization is associated with poor prognosis of Community-acquired Pneumonia. Infect Dis (Lond) 2017;49(4):251-60.
18. Colak A, Yilmaz C, Toprak B, Aktogu S. Procalcitonin and CRP as biomarkers in dis- crimination of community-acquired pneumonia and exacerbation of COPD. J Med Biochem 2017;36(2):122-6.
19. Christ-Crain M, Opal SM. Clinical review: the role of biomarkers in the diagnosis and management of community-acquired pneumonia. Crit Care 2010;14(1):203.
20. Christ-Crain M, Morgenthaler NG, Struck J, Harbarth S, Bergmann A, Muller B. Mid- regional pro-adrenomedullin as a prognostic marker in sepsis: an observational study. Crit Care 2005;9(6):R816-4.
21. Travaglino F, De Berardinis B, Magrini L, Bongiovanni C, Candelli M, Silveri NG, et al. Utility of procalcitonin and Mid regional pro-Adrenomedullin (MR-proADM) in risk stratification of critically ill febrile patients in Emergency Department (ED). A comparison with APACHE II score. BMC Infect Dis 2012;12:184.
22. Legramante JM, Mastropasqua M, Susi B, Porzio O, Mazza M, Miranda Agrippino G, et al. prognostic performance of MR-pro-adrenomedullin in patients with commu- nity acquired pneumonia in the Emergency Department compared to clinical sever- ity scores PSI and CURB. PLoS One 2017;12(11):e0187702.
23. Gordo-Remartinez S, Calderon-Moreno M, Fernandez-Herranz J, Castuera-Gil A, Gallego-Alonso-Colmenares M, Puertas-Lopez C, et al. Usefulness of midregional proadrenomedullin to predict poor outcome in patients with community acquired pneumonia. PLoS One 2015;10(6):e0125212.
24. Liu D, Xie L, Zhao H, Liu X, Cao J. Prognostic value of mid-regional pro- adrenomedullin (MR-pro-ADM) in patients with community-acquired pneumonia: a systematic review and meta-analysis. BMC Infect Dis 2016;16:232.
25. Chen YX, Li CS. The prognostic and risk-stratified value of heart-type fatty-acid- binding protein in communityacquired pneumonia in emergency department. Biomed Res Int 2014;2014:753070.
26. Tsao TC, Tsai HC, Chang SC. Clinical usefulness of urinary fatty acid binding proteins in assessing the severity and predicting Treatment response of pneumonia in criti- cally ill patients: a cross-sectional study. Medicine (Baltimore) 2016;95(19):e3682.
27. de Jager CP, Wever PC, Gemen EF, Kusters R, van Gageldonk-Lafeber AB, van der Poll T, et al. The neutrophil-lymphocyte count ratio in patients with community- acquired pneumonia. PLoS One 2012;7(10):e46561.
28. Curbelo J, Luquero Bueno S, Galvan-Roman JM, Ortega-Gomez M, Rajas O, Fernandez-Jimenez G, et al. Inflammation biomarkers in blood as mortality predic- tors in community-acquired pneumoniaadmitted patients: importance of compari- son with neutrophil count percentage or neutrophil-lymphocyte ratio. PLoS One 2017;12(3):e0173947.
29. Kolditz M, Scherag A, Rohde G, Ewig S, Welte T, Pletz M, et al. Comparison of the qSOFA and CRB-65 for risk prediction in patients with community-acquired pneu- monia. Intensive Care Med 2016;42(12):2108-10.