a b s t r a c t

Objective: The objective of this study was to evaluate the association of elevated alveolar-arterial oxygen (A-a O₂) gradient with risk of mortality in hospitalized patients with Community-acquired pneumonia .

Methods: This prospective study included 206 patients diagnosed with CAP admitted to the ED. Demographics, comorbidities, arterial blood gas, Serum electrolytes, liver-renal functions, complete blood count, NLR, PLR, CRP, CAR, procalcitonin, A-a O₂ gradient, expected A-a O₂ and A-a O₂ difference were evaluated. PSI and CURB- 65 scores were classified as follow: a) PSI low risk (I-III) and moderate-high risk (IV-V) groups; b) CURB-65; low risk (0-2) and high risk (3-5) groups.

Results: The survival rates of the PSI class (I-III) were significantly higher than the ones of the PSI class (IV-V) (92.1% vs. 62.9%, respectively). The percentage of survivors of the CURB-65 score (0-2) group (81.9%) was higher than the survivors of CURB-65 score (3-5) group (27.8%). Creatinine, BUN, Uric acid, phosphorus, RDW, CRP, CAR, procalcitonin, lactate, A-a 0₂ gradient, expected A-a 0₂ and A-a 0₂ difference were significantly higher and baso- phil was lower in non-survivors. A-a O₂ gradient (AUC 0.78), A-a O₂ difference (AUC 0.74) and albumin (AUC 0.80) showed highest 30-day Mortality prediction. NLR (AUC 0.58) and PLR (AUC 0.55) showed lowest 30-day mortality estimation. Procalcitonin (AUC 0.65), PSI class (AUC 0.81) and PSI score (AUC 0.86) indicated statisti- cally significant higher 30-day mortality prediction.

Conclusion: A-a O₂ gradient, A-a O₂ difference and albumin are potent predictors of 30-day mortality in CAP pa-

tients in the ED.

(C) 2020

Introduction

community-acquired pneumonia is one of the major reasons of mortality caused by infection with a rate of approximately 20-30% in Developing countries and 3-4% in developed countries [1,2]. The inci- dence and mortality rate of this respiratory infection increases with age [3]. Approximately one-third of CAP cases are inpatient and the mortality rate is up to 50-55% for those patients who require intensive care unit admission. The incidence of pneumonia is 1.15% and this infectious disease is ranked 15th in all acute and chronic diseases in Turkey [3]. It is estimated that 75% of CAP cases present to emergency departments (EDs). Prediction of mortality and outcomes in patients with CAP is important for early intervention, hospitalization, ICU admis- sion, optimal antibiotic treatment and discharge from EDs [2]. The diag- nosis, treatment, prognosis and mortality data of CAP can vary depending on population structure of country, experience of physicians,

* Corresponding author at: Amasya University Sabuncuoglu Serefeddin Research and Training Hospital, 05100 Amasya, Turkey.

E-mail address: [email protected] (S. Avci).

resources and location of hospital, complementary tests, and social- sanitisation qualifications.

Several algorithms based upon clinical findings and laboratory re-

sults including CURB-65, pneumonia severity index \(PSI\) and well- known inflammatory markers such as neutrophil-to-lymphocyte ratio , Platelet-to-lymphocyte ratio , c-reactive protein (CRP)/al- bumin ratio (CAR) and procalcitonin have been used to predict severity, outcomes and short-long term mortality of patients with CAP [4-7]. In addition to these scores and inflammatory markers, the alveolar- arterial oxygen (Aa-O₂) gradient, which is a measure of difference be- tween the alveolar and arterial concentration of oxygen, has also been used to evaluate the outcomes of the patients with CAP [8,9]. While ven- tilation continues in patients with pneumonia, the physical barrier in the alveoli restricts oxygen diffusion to capillaries and causes low oxy- gen levels in the blood, resulting in increased A-a O₂ gradient [8,9].

The aims of the study were to compare the A-a O₂ gradient (together with expected A-a O₂ gradient, A-a O₂ gradient difference), pneumonia severity scoring systems (PSI, CURB-65), inflammatory markers (CRP, CAR, procalcitonin, NLR and PLR) in predicting CAP patients’ outcomes (30-day mortality, ICU admission) in the ED. The objective of this study was to evaluate the association of elevated A-a O₂ gradient with

https://doi.org/10.1016/j.ajem.2020.05.048

0735-6757/(C) 2020

risk of mortality in hospitalized patients with CAP. A similar previous study reported that A-a O₂ gradient can predict the outcomes of the pa- tients with CAP (length of stay, severity of the disease) [9].

Methods

Study design and population

This prospective cross-sectional study was conducted with approval of Kafkas University Medical Faculty Ethics Committee between No- vember 2018 and March 2019. The study included 206 patients (78 fe- males, 128 males) diagnosed with CAP admitted to the ED.

Patients aged 18 years or more, admitted to ED from the community, presenting with any new opacity on chest x-ray in accordance with the diagnosis of acute pneumonia including two or more clinical signs and symptoms associated with pneumonia (cough, dyspnea, chest pain, crackles on auscultation and temperature N 38 ?C) were enrolled in the study. The exclusion criteria were as follows: patients with human immundefiency virus infection, active pulmonary tuberculosis, cystic fi- brosis, receiving routine hemodialysis, homeless people, history of nurs- ing at home or hospitalization longer than 24 h in the last 90 days, history of chemotherapy or radiotheraphy in the last 30 days, admission to the hospital or using any antibiotic in the last two weeks. Verbal and written consent were obtained from all patients.

Data collection

A questionnaire was designed for each patient, including demo- graphic profile, comorbidities, physicial examination, laboratory results and radiological findings. At admission to ED, the scores of PSI and CURB-65 were calculated and classified as follow: 1) PSI low risk group (classes I-III, total point value <=90); and PSI moderate-High risk group (class IV-V total point value 91<=); 2) CURB-65; low risk group (scores 0-2) and high risk group (scores 3-5) [10].

The laboratory findings were analyzed within 3 h after admission to ED including arterial blood gas analysis, serum electrolytes, liver and kidney function tests, complete blood count, NLR, PLR, CRP, CAR, and procalcitonin.

Arterial blood gas samples were drawn from the Radial artery in all patients while they were breathing room air to prevent any interven- tion caused by the maintenance of supplementary oxygen. atmospheric pressure (mmHg), partial oxygen pressure (PaO₂, mmHg), fraction of inspired oxygen (FiO₂, 21% for room air), partial carbon dioxide pressure (PaCO₂, mmHg) and age (for expected A-a O₂ gradient) were recorded for all patients and calculated with https://www.mdcalc.com/a-a-o2- gradient [11]. After calculation process, the A-a O₂ gradient and ex- pected A-a O₂ gradient for age were obtained. A-a O₂ difference was cal- culated as A-a O₂ gradient-expected A-a O₂ gradient. Daily atmospheric pressure (mmHg) changes of the region where the study was carried out were obtained from the Turkish State Meteorological Service. After one-month follow up, ICU admission and 30-day mortality after ED ad- mission were evaluated.

Statistical reviews

All statistical calculations were performed with SPSS 23.0 (SPSS for Windows, Chicago, IL, SA). The continuous variables were expressed as mean +- standard deviation; categoric variables were defined as per- centages (%). The categorical parameters were compared with Chi- square test and Fischer’s exact test. The normal distribution was deter- mined by histogram and Kolmogorov-Smirnov test. Mean values of con- tinuous variables were compared between the groups using Mann- Whitney U test. Student’s t-test was used in the comparison of parame- ters showing normal distribution. prediction accuracy was assessed using the area under the receiver operating characteristic (ROC) curve.

The results were evaluated as 95% confidence interval and p value

b0.05, which was considered statistically significant.

Results

Table 1 lists demographics, comorbidities and clinical characteristics of the cases. Accordingly, 128 (62.1%) of the cases were male and 78 (37.9%) were female. Their ages ranged from 19 to 95 years, with an av- erage of 68.34 +- 16.52. 34 (16.5%) of the cases were current smoker, 83 (40.3%) were non-smoker and 89 (43.2%) were ex-smoker. The fre- quency of comorbidities seen among the patients were: Hypertension (44.7%), diabetes mellitus (15.0%), coronary artery disease (8.7%),

COPD (66.5%), asthma (4.9%), neoplastic disease (5.3%), Chronic liver disease (1.9%), congestive heart failure (14.6%), cerebrovascular dis- eases (3.9%) and chronic renal diseases (6.3%). 0.5% of the cases were nursing home resident. Confusion was present in 17 cases (8.3%) and pleural effusion was present in 26 cases (12.6%). The respiratory rate of the patients ranged between 15 and 36 per minute with an average of 20.94 +- 3.45. The systolic blood pressure of cases varied between 66 and 180 mmHg, with an average of 117.13 +- 14.24. The fever of the cases ranged from 35 to 38.7 ?C, with an average of 36.55 +- 0.38. The heart rate of the cases varied between 12 and 150 per minute, with an average of 90.96 +- 14.45. ICU admission rate was 8.7%, and the overall 30-day mortality rate was 22.8%.

Table 2 lists 30-day mortality and ICU admission rates according to the PSI classes and CURB-65 score groups. Accordingly, there was a sta- tistically significant difference between 30-day mortality and ICU ad- mission between scoring groups (p value b0.01). The survival rates of the PSI class (I-III) were significantly higher than the ones of the PSI class (IV-V) (92.1% vs. 62.9%, respectively). The PSI class’ (I-III) ICU ad- mission rates (2.0%) were significantly lower than the PSI class’ (IV-V) ICU admission rates (15.2%). The percentage of survivors of the CURB- 65 score (0-2) group (81.9%) was higher than the survivors of CURB- 65 score (3-5) group (27.8%). The ICU admission rates (3.7%) of the

Table 1

Demographic properties, comorbidities and clinical findings of the patients

CAP patients (n = 206)

No. %

Gender	Male	128	62.1
	Female	78	37.9
Smoking status	Current smoker	34	16.5
	Non-smoker	83	40.3
	Ex-smoker	89	43.2
Hypertension		92	44.7
Diabetes mellitus		31	15.0
Coronary artery disease		18	8.7
COPD		137	66.5
Asthma		10	4.9
Nursing home resident		1	0.5
Neoplastic disease		11	5.3
Chronic liver diseases		4	1.9
Congestive heart failure		30	14.6
cerebrovascular diseases		8	3.9
Chronic renal diseases		13	6.3
Confusion		17	8.3
Pleural effusion		26	12.6
ICU admission		18	8.7
30-day mortality	Min-Max	47 Mean	22.8 SD
Respiratory rate (min)	15-36	20.94	3.45
Systolic blood pressure (mmHg)	66-180	117.13	14.24
Fever (?C)	35-38,7	36.55	0.38
Heart rate (min)	12-150	90.96	14.45
Age (years)	19-95	68.34	16.52

CAP, community acquired pneumonia; COPD, chronic obstructive pulmonary disease; min-max; minimum and maximum values.

Table 2

30-day mortality and ICU admission rates of PSI class and CURB-65 score groups.

PSI class

I-III IV-V P

No. (%) No. (%)

30-day mortality	Survivors Non-survivors	93 (92.1%) 8 (7.9%)	66 (62.9%) 39 (37.1%)	.000**
ICU admission		2 (2.0%)	16 (15.2%)	.001**
		CURB-65 score
		0-2	3-5
		No. (%)	No. (%)
30-day mortality	Survivors Non-survivors	154 (81.9%) 34 (18.1%)	5 (27.8%) 13 (72.2%)	.000**
ICU admission		7 (3.7%)	11 (61.1%)	.000**

CURB-65 score (0-2) group were significantly lower than the ICU ad- mission rates (61.1%) of the CURB-65 score (3-5) group.

Independent sample t-test analysis of A-a 0₂ gradient and A-a 0₂ dif- ference averages between PSI class and CURB-65 score groups did not reveal a statistical significance (p value N0.005).

Table 3 lists the independent sample t-test analysis results of the mean values of calcium, hematocrit, mean platelet volume, platelet, PaO₂ and albumin were given. Calcium, PaO₂ and albumin were signifi- cantly lower in non-survivors (p b 0.001).

Table 4 lists analysis of Blood parameters with Mann Whitney U test between survivors and non-survivors. Creatinine, Blood urea nitrogen , uric acid, phosphor, red cell distribution width (RDW), CRP, CAR, procalcitonin, lactate, A-a 0₂ gradient, expected A-a 0₂ gradient and A-a 0₂ difference were significantly higher (p b 0.05) and basophil count was lower in non-survivors (p = 0.013).

Table 5 and Fig. 1 demonstrate the accuracy of A-a O₂ gradient, A-a O₂ difference, albumin, CAR, CRP and CURB-65 in predicting 30-day mortality. A-a O₂ gradient (AUC 0.78, 95% Cl: 0.72-0.84), A-a O₂ differ- ence (AUC 0.74, 95% CI: 0.68-0.74) and albumin (AUC 0.80, 95% CI:

0.74-0.85) showed highest 30-day mortality prediction. CAR (AUC 0.69, 95% CI: 0.62-0.75), CRP (AUC 0.64, 95% CI: 0.57-0.71) and CURB-

65 (AUC 0.76, 95% CI: 0.70-0.82) indicated statistically significant lower 30-day mortality prediction.

Table 6 and Fig. 2 depict the accuracy of NLR, PLR, procalcitonin, PSI class and PSI scores in predicting 30-day mortality. NLR (AUC 0.58, 95% CI: 0.50-0.65) and PLR (AUC 0.55, 95% CI: 0.47-0.62) showed lowest 30-

day mortality estimation. Procalcitonin (AUC 0.65, 95% CI: 0.58-0.72),

PSI class (AUC 0.81, 95% CI: 0.75-0.87) and PSI score (AUC 0.86, 95%

CI: 0.80-0.91) indicated statistically significant higher 30-day mortality prediction.

Discussion

Estimating mortality in patients with CAP in the ED is important for patient

Table 3

Comparison of the mean values of calcium, hematocrit, mean platelet volume, platelet, PaO₂ and albumin survivors and non-survivors

30-day mortality

Survivors Non-survivors P

Mean +- SD Mean +- SD

Calcium	8.82 +- 0.61	8.28 +- 0.63	.000**
Hematocrit	43.6 +- 7.45	41.83 +- 7.22	.153
Mean platelet volume	9.68 +- 1.15	9.75 +- 1.36	.707
Platelet	254.96 +- 90.08	242.49 +- 110.07	.430
PaO2	56.25 +- 9.07	46.65 +- 8.64	.000**
Albumin	37.58 +- 5.51	31.21 +- 6.91	.000**

management. This study sought to evaluate the A-a O2 gradient in predicting mortality

compared to PSI, CURB-65 and various inflammatory markers. In the study, the 30-day mortality rate and ICU admission rate were 22.8% and 8.7%, respectively. Our results including mortality and ICU admission were higher compared to similar studies [10,12-14]. Advanced age, in- creased number of comorbidities and late admission to the ED may lead to these poor results.

PSI and CURB-65 are verified and frequently used clinical scoring systems for prediction of prognosis and mortality of patients with pneu- monia [12]. Increment value of scores may indicate poor prognosis, in- creased severity and risk of mortality [14]. In our study, PSI class (IV- V) and CURB-65 score (3-5) were associated with higher ICU admission rates and 30-day mortality (p-value b0.01). These expected findings were similar and consistent with several earlier studies [10,12,15]. Higher age (mean +- sd 65 yearsb), male gender (61.1% of all), increased number of comorbidities, impaired vital signs and clinical findings, which are the components of these scores (PSI and CURB-65) that are easily used in the ED, may predict poor prognosis in patients with pneumonia.

In the study, lower albumin, calcium and PaO₂ levels on admisson were related to increased 30-day mortality among CAP cases (p b 0.001). Hypoalbuminemia is an inevitable result of capillary leakage that caused by impaired endothelial permeability due to excessive re- lease of pro-inflammatory and anti-inflammatory cytokines (secreted proteins and signal molecules) including interleukin (IL)-1, IL-2, IL-6, IL-10, interferon gamma and tumor necrosis factor-alpha during inflam- matory reaction in critically ill patients such as CAP [16]. Moreover, al- bumin is the main serum protein that binds approximately 40% of calcium in the human body circulation, which is responsible for cardiac contraction, coagulation cascade and platelet activation [17]. In acute Serious diseases such as pneumonia, the decrease of circulating albumin called hypoalbuminemia, an acute phase reactant, can predict poor out- comes including morbidity, ICU admission, mortality and cause lower calcium level [16-18]. A second cause of mortality in critically ill patients with CAP is hypoxemia, in which arterial Blood oxygen saturation mea- sured by Pulse oximeter is b90% or PaO₂ is b60 mmHg [19]. Alveolar congestion triggered by inflammatory reaction, prominent intrapulmonary shunting, enlarged dead space and ventilation/perfu- sion mismatch may result in disrupted Gas exchange and lower PaO₂ [19].

Non-survivors had higher creatinine, BUN, uric acid, phosphorus, RDW, procalcitonin, CRP, CAR, and lactate (p b 0.005). impaired renal function may contribute to infections such as pneumonia with mecha- nisms delineated by uremia-related damaged monocyte, T/B lympho- cytes, neutrophil phagocytosis and increment of cytokines, as well as increased risk of Cardiovascular complications [20]. Higher serum level of uric acid, a product of purine destruction and excreted in the urine, is associated with impaired renal function, hypertension, atherosclero- sis and Vascular abnormalities [21]. Hyperphosphatemia triggered by chronic renal failure cause calcification of vascular media layer and car- diac valves, stimulation of cardiomyocyte hypertrophy and inhibition of nitric oxide synthase [22]. In addition, RDW is a routine laboratory test used to evaluate size, form and degree of heterogeneity of erythrocyte volume that predicts prognosis of thrombotic and cardiovascular events, and it has been associated with increased mortality of patients with CAP [23]. Increased procalcitonin, CRP and CAR are inflammatory biomarkers that have been extensively studied for use in the diagnosis and prognostication of pneumonia [5,6,24]. An increase of these markers may indicate severe pneumonia and even bacteremia. In- creased lactate, which measures Severity of disease and perfusion sta- tus, particularly in septic patients, indicates higher mortality and the need for vasopressors, mechanical ventilation and ICU admission [25].

This study suggests that A-a O₂ gradient, A-a O₂ expected and A-a O₂ difference are higher in non-survivors. A-a O₂ gradient (AUC 0.78), A-a O₂ difference (AUC 0.74) and albumin (AUC 0.80) showed highest 30-

Table 4

Blood parameters of survivors and	non-survivors.
	Groups	n	Mean rank	Sum of rank	U	z	P
Creatinine	Non-survivors	47	127.77	6005.00	2596.000	-3.177	.001**
	Survivors	159	96.33	15,316.00
Sodium	Non-survivors	47	102.45	4815.00	3687.000	-0.138	.890
	Survivors	159	103.81	16,506.00
Glucose	Non-survivors	47	108.09	5080.00	3521.000	-0.600	.548
	Survivors	159	102.14	16,241.00
Aspartate aminotransferase	Non-survivors	47	116.83	5491.00	3110.000	-1.746	.081
	Survivors	159	99.56	15,830.00
Alanine aminotransferase	Non-survivors	47	109.21	5133.00	3468.000	-0.748	.454
	Survivors	159	101.81	16,188.00
Uric acid	Non-survivors	47	116.38	5470.00	3131.000	-1.687	.092
	Survivors	159	99.69	15,851.00
Magnesium	Non-survivors	47	97.53	4584.00	3456.000	-0.782	.434
	Survivors	159	105.26	16,737.00
Phospor	Non-survivors	47	126.63	5951.50	2649.500	-3.028	.002**
	Survivors	159	96.66	15,369.50
Lymphocyte	Non-survivors	47	94.18	4426.50	3298.500	-1.220	.222
	Survivors	159	106.25	16,894.50
Plateletcrit	Non-survivors	47	96.60	4540.00	3412.000	-0.904	.366
	Survivors	159	105.54	16,781.00
RDW	Non-survivors	47	130.45	6131.00	2470.000	-3.528	.000**
	Survivors	159	95.53	15,190.00
Basophil	Non-survivors	47	85.03	3996.50	2868.500	-2.484	.013*
	Survivors	159	108.96	17,324.50
Eosinophil	Non-survivors	47	99.34	4669.00	3541.000	-0.548	.584
	Survivors	159	104.73	16,652.00
Neutrophil	Non-survivors	47	111.29	5230.50	3370.500	-1.019	.308
	Survivors	159	101.20	16,090.50
Monocyte	Non-survivors	47	101.82	4785.50	3657.500	-0.155	.876
	Survivors	158	103.35	16,329.50
Lactate	Non-survivors	46	133.34	6133.50	2215.500	-4.029	.000**
	Survivors	158	93.52	14,776.50
Blood urea nitrogen	Non-survivors	47	143.61	6749.50	1851.500	-5.251	.000**
	Survivors	159	91.64	14,571.50
C-reactive protein	Non-survivors	47	126.32	5937.00	2664.000	-2.987	.003**
	Survivors	159	96.75	15,384.00
CAR	Non-survivors	47	133.59	6278.50	2322.500	-3.938	.000**
	Survivors	159	94.61	15,042.50
NLR	Non-survivors	47	113.50	5334.50	3266.500	-1.309	.190
	Survivors	159	100.54	15,986.50
PLR	Non-survivors	47	106.74	5017.00	3584.000	-0.425	.671
	Survivors	159	102.54	16,304.00
Procalcitonin	Non-survivors	35	111.73	3910.50	1759.500	-2.766	.006**
	Survivors	144	84.72	12,199.50
A-a 02 gradient	Non-survivors	47	148.64	6986.00	1615.000	-5.909	.000**
	Survivors	159	90.16	14,335.00
Expected A-a 02	Non-survivors	47	128.03	6017.50	2583.500	-3.213	.001**
	Survivors	159	96.25	15,303.50
A-a 02 difference	Non-survivors	47	141.22	6637.50	1963.500	-4.938	.000**
	Survivors	159	92.35	14,683.50

day mortality prediction. A-a O₂ gradient, A-a O₂ difference and albumin were better than CAR, CRP and CURB-65 in predicting 30-day mortality of CAP patients. These results indicate that A-a O₂ gradient, A-a O₂ dif- ference and albumin are strong predictors of 30-day mortality in pa- tients with CAP admitted to ED. Albumin is a well-known marker predicting mortality in patients [16,18], and A-a O₂ gradient and A-a O₂ difference can also be used to estimate mortality.

Table 5

Investigation of measurements effective in estimating mortality

	AUC	SE	95% CI
A-a O2 gradient	0.784	0.0408	0.721 to 0.838
A-a O2 difference	0.737	0.0468	0.672 to 0.796
Albumin	0.802	0.0384	0.741 to 0.854
CAR	0.689	0.0466	0.621 to 0.752
CRP	0.644	0.0476	0.574 to 0.709
CURB-65	0.761	0.0370	0.697 to 0.817

Studies have reported that A-a O₂ gradient, which is used to measure the etiology and oxygenation of hypoxemia assessing the functionality of the blood-air barrier within the alveoli, predicts length of stay and survival in hospitalized patients with pneumonia [9,26]. To estimate 30-day mortality, increased A-a O₂ gradient and A-a O₂ difference in pa- tients with CAP in the ED may be useful for physicians.

Lastly, NLR (AUC 0.58) and PLR (AUC 0.55) demonstrated lowest mortality estimation in this study. Procalcitonin, PSI class and PSI score were better than NLR and PLR in predicting 30-day mortality of CAP patients. In recent studies, NLR and PLR have been reported as novel cost-effective inflammatory markers that can be easily obtained from complete blood count, and can predict prognosis and mortality in various clinical conditions including CAP [4,27-29]. In the study, NLR and PLR were less predictive of 30-day mortality compared to an inflammatory marker such as procalcitonin and a clinical scoring such as PSI.

This single-centre study had some limitations. To begin with, the size of the patient population was relatively small and we did not determine

Fig. 1. The ROC curves for prediction of 30-day mortality for A-a O₂ gradient, A-a O₂ difference, albumin, CAR, CRP and CURB-65.

Table 6

Investigation of measurements effective in estimating mortality

	AUC	SE	95% CI
NLR	0.577	0.0552	0.501 to 0.650
PLR	0.545	0.0543	0.469 to 0.619
Procalcitonin	0.652	0.0457	0.577 to 0.721
PSI class	0.811	0.0401	0.746 to 0.866
PSI score	0.861	0.0380	0.801 to 0.908

accurate etiology of microorganisms causing to community-acquired pneumonia. Moreover, detailed effect of comorbidities to 30-day mor- tality was unknown. During the one-month follow-up period after the emergency department, only the 30-day mortality was recorded and the management of the treatment was not evaluated. In addition, an ar- terial blood gas sample may cause hemorrhage, local hematoma, aneu- rysm, air or Thrombus embolism, local infection, arteriovenous laceration, needle stick injuries and severe pain [30,31].

Conclusion

A-a O₂ gradient, A-a O₂ difference and albumin are potent predictors of 30-day mortality in CAP patients in the ED. In addition, A-a O2 gradi- ent and A-a O2 difference are simple, precise, and practical measure- ments for estimating 30-day mortality in these patients, better than procalcitonin, PSI score, NLR and PLR.

Financial support

None declared.

CRediT authorship contribution statement Sema Avci:Conceptualization, Data curation, Writing - original draft.

Gokhan Perincek:Data curation, Formal analysis.

Declaration of competing interest

None declared.

References

1. Shah BA, Ahmed W, Dhobi GN, Shah NN, Khursheed SQ, Haq I. Validity of pneumonia severity index and CURB-65 severity scoring systems in community acquired pneu- monia in an Indian setting. Indian J Chest Dis Allied Sci. 2010;52:9-17.
2. Julian-Jimenez A, de los Reyes MJ Palomo, Miguez RP, Lain-Teres N, Cuena-Boy R, Lozano-Ancin A. Improved management of community-acquired pneumonia in the emergency department. Arch Bronconeumol. 2013;49(6):230-40. https://doi. org/10.1016/j.arbres.2012.12.008.
3. Bulbul Y. Community-acquired pneumonia data in Turkey. Guncel Gogus Hastalik Seri. 2014;2(1):1-10.
4. 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. https://doi.org/10.1371/jour- nal.pone.0046561.
5. Lee JH, Song S, Yoon SY, Lim CS, Song JW, Kim HS. Neutrophil to lymphocyte ratio and platelet to lymphocyte ratio as diagnostic markers for pneumonia severity. Br J Biomed Sci. 2016;73(3):140-2. https://doi.org/10.1080/09674845.2016.1209898.
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. 2015;196(10): 1287-97. https://doi.org/10.1164/rccm.201611-2262OC.
7. Nobre V, Borges I. Prognostic value of procalcitonin in hospitalized patients with lower respiratory tract infections. Rev Bras Ter Intensiva. 2016;28(2):179-89. https://doi.org/10.5935/0103-507X.20160019.
8. Diehl JL, Mercat A, Pesenti A. Understanding hypoxemia on ECCO₂R: back to the al- veolar gas equation. Intensive Care Med. 2019;45(2):255-6. https://doi.org/10. 1007/s00134-018-5409-0.
9. Moammar MQ, Azam HM, Blamoun AI, Rashid AO, Ismail M, Khan MA, et al. Alveolar-arterial oxygen gradient, pneumonia severity index and outcomes in pa- tients hospitalized with community acquired pneumonia. Clin Exp Pharmacol P. 2008;35:1032-7. https://doi.org/10.1111/j.1440-1681.2008.04971.
10. Shehata SM, Sileem AE, Shahien NE. prognostic values of pneumonia severity index, CURB-65 and expanded CURB-65 scores in community-acquired pneumonia in Zag- azig university hospitals. Egypt J Chest Dis Tuberc. 2017;66(3):549-55. https://doi. org/10.1016/j.ejcdt.2017.01.001.
11. PSI/PORT Score. Pneumonia severity index for CAP. https://www.mdcalc.com/a-a- o2-gradient. [Accessed 2 November 2018].
12. Zhang ZX, Yong Y, Tan WC, Shen L, Seong H, Fong KY. Prognostic factors for mortality due to pneumonia among adults from different age groups in Singapore and

    Fig. 2. The ROC curves for prediction of 30-day mortality for NLR, PLR, procalcitonin, PSI class and PSI score.

    mortality predictions based on PSI and CURB-65. Singapore Med J. 2018;59(4): 190-8. https://doi.org/10.11622/smedj.2017079.

    Irfan M, Hussain SF, Mapara K, Memon S, Mogri M, Bana M, et al. Community ac- quired pneumonia: risk factors associated with mortality in a tertiary care hospital- ized patients. J Pak Med Assoc. 2009;59(7):448-52.

13. Madhu S, Augustine S, Kumar R, Kauser M, Kumar V, Jayaraju BS. Comparative study of CURB-65, pneumonia severity index and IDSA/ATS scoring systems in community acquired pneumonia in an Indian tertiary care setting. Int J Adv Med. 2017;4(3): 693-700. https://doi.org/10.18203/2349-3933.ijam20172088.
14. Alavi-Moghaddam M, Bakhshi H, Rezaei B, Khashayar P. Pneumonia severity index compared to CURB-65 in predicting the outcome of community acquired pneumo- nia among patients referred to an Iranian emergency department: a prospective sur- vey. Braz J Infect Dis. 2013;17(2):179-83. https://doi.org/10.1016/j.bjid.2012.
15. Damayanti N, Abidin A, Keliat EN. The correlation between Albumin levels with 30 days mortality in community acquired pneumonia patients. Earth Environ Sci. 2018;125:e012141. https://doi.org/10.1088/1755-1315/125/1/012141.
16. Alsancak Y, Sivri S, Bastug S, Bozkurt E. Role of calcium-albumin ratio in severity of coronary artery disease assessed by angiographic SYNTAX score. Arch Clin Exp Med. 2018;3(3):174-8. https://doi.org/10.25000/acem.450950.
17. Vincent JL, Dubois MJ, Navickis RJ, Wilkes MM. Hypoalbuminemia in acute illness: is there a rationale for intervention? A meta-analysis of cohort studies and controlled trials. Ann Surg. 2003;237(3):319-34. https://doi.org/10.1097/01.SLA.0000055547. 93484.87.
18. Kuti BP, Adegoke SA, Ebruke BE, Howie S, Oyelami OA, Ota M. Determinants of oxy- gen therapy in childhood pneumonia in a resource-constrained region. ISRN Pediatr. 2013;2013:435976. https://doi.org/10.1155/2013/435976.
19. Chou CY, Wang SM, Liang CC, Chang CT, Liu JH, Wang IK, et al. Risk of pneumonia among patients with chronic kidney disease in outpatient and inpatient settings. A nationwide population-based study. Medicine (Baltimore). 2014;93(27):e174. https://doi.org/10.1097/MD.0000000000000174.
20. Zhu HC, Cao RL. The relationship between serum levels of uric acid and prognosis of infection in critically ill patients. World J Emerg Med. 2012;3(3):186-90. https://doi. org/10.5847/wjem.j.issn.1920-8642.2012.03.005.
21. Gutierrez OM. The connection between dietary phosphorus, cardiovascular disease, and mortality: where we stand and what we need to know. Adv Nutr. 2013;4(6): 723-9. https://doi.org/10.3945/an.113.004812.
22. Bello S, Fandos S, Lasierra AB, Minchole E, Panadero C, Simon AL, et al. Red blood cell distribution width and long-term mortality after community-acquired pneumonia. A comparison with proadrenomedullin. Respir Med. 2015;109:1193-206. https:// doi.org/10.1016/j.rmed.2015.07.003.
23. Schmidt de Oliveira-Netto AC, Morello LG, Dalla-Costa LM, Petterle RR, Fontana RM, Danieli Conte D, et al. Procalcitonin, c-reactive protein, albumin, and blood cultures as early markers of sepsis diagnosis or predictors of outcome: a Prospective analysis. Clin Pathol. 2019;12 [2632010X19847673].
24. Chen YX, Li CS. Lactate on emergency department arrival as a predictor of mortality and site-of-care in pneumonia patients: a cohort study. Thorax. 2015;70:404-10. https://doi.org/10.1136/thoraxjnl-2014-206461.
25. Isoda K, Takeuchi T, Kotani T, Hata K, Shoda T, Ishida T, et al. Pre-treatment ferritin level and alveolar-arterial oxygen gradient can predict mortality rate due to acute/ subacute interstitial pneumonia in dermatomyositis treated by Cyclosporine A/ glucocorticosteroid Combination therapy: a case control study. PLoS ONE. 2014;9 (2):e89610. https://doi.org/10.1371/journal.pone.0089610.
26. Kartal O, Kartal AT. Value of neutrophil to lymphocyte and platelet to lymphocyte ra- tios in pneumonia. Bratisl Med J. 2017;118(9):513-6. https://doi.org/10.4149/BLL_ 2017_099.
27. Curbelo J, Rajas O, Arnalich B, Galvan-Roman JM, Luquero-Bueno S, Ortega-Gomez M, et al. Neutrophil count percentage and neutrophil-lymphocyte ratio as prognos- tic markers in patients hospitalized for community-acquired pneumonia. Arch Bronconeumol. 2019;55(9):472-7. https://doi.org/10.1016/j.arbres.2019.02.005.
28. Pantzaris ND, Platanaki C, Pierrako C, Karamouzos V, Velissaris D. Neutrophil-to- lymphocyte ratio relation to sepsis severity scores and inflammatory biomarkers in patients with community-acquired pneumonia: a case series. J Transl Int Med. 2018;6(1):43-6. https://doi.org/10.2478/jtim-2018-0009.
29. Dev SP, Hillmer MD, Ferri M. Arterial puncture for blood gas analysis. N Engl J Med. 2011;364:e7. https://doi.org/10.1056/NEJMvcm0803851.
30. Kozaci N, Gungor F, Ay MO, Beydilli I, Bozdemir N, Akyol C. Can venous blood gas values be used instead of arterial blood gas values in respiratory alkalosis? Turk J Biochem. 2014;39(1):113-8. https://doi.org/10.5505/tjb.2014.87049.