Archives of Academic Emergency Medicine. 2019; 7 (1): e59 OR I G I N A L RE S E A RC H PIRO, SOFA and MEDS Scores in Predicting One-Month Mortality of Sepsis Patients; a Diagnostic Accuracy Study Ali Vafaei1,2, Kamran Heydari1,2, Seyed-Saeed Hashemi-Nazari3, Neda Izadi4, Hassan Hassan Zadeh1∗ 1. Department of Emergency Medicine, Shahid Beheshti University of Medical Sciences, Tehran, Iran. 2. Skull Base Research Center, Loghman Hakim Hospital, Shahid Beheshti University of Medical Sciences, Tehran, Iran. 3. Prevention of Cardiovascular Disease Research Center, School of Public Health and Safety, Shahid Beheshti University of Medical Sciences, Tehran, Iran. 4. Student Research Committee, Department of Epidemiology, School of Public Health and Safety, Shahid Beheshti University of Medical Sciences, Tehran, Iran. Received: June 2019; Accepted: August 2019; Published online: 20 October 2019 Abstract: Introduction: Different scoring systems based on clinical and laboratory findings are designed for prediction of short-term mortality of patients with severe sepsis and septic shock. This study aimed to compare the screening performance characteristics of PIRO, SOFA and MEDS Scores in predicting one-month mortality of sepsis pa- tients. Methods: This diagnostic accuracy study was performed on septic shock and severe sepsis patients refer- ring to emergency department of Loghmane Hakim Hospital, Tehran, Iran, from 2017 to 2018. The performance of MEDS, SOFA, and PIRO models in predicting 30-day mortality of patients was evaluated using discrimination and calibration indices. Results: 200 patients with the mean age of 71.03±15.59 years were studied (61% male). During the 30 days, 66 patients died (mortality rate=33%). The area under the ROC curve of PIRO, MEDS, and SOFA scores were 0.83 (95% CI=0.78-0.89), 0.94 (95% CI=0.91-0.97) and 0.87 (95% CI=0.81-0.92), respectively. Based on Brier, BrierScaled and Nagelkerke’s R2 of the models, the best performance in predicting one-month mortality belonged to MEDS score. C-statistic showed that MEDS score had the highest value in the differentia- tion between the survived and non-survived cases. Conclusion: This study showed that MEDS score performs better than PIRO and SOFA scores in predicting one-month mortality of patients with severe sepsis and septic shock. Keywords: Decision support systems, clinical; patient outcome assessment; mortality; sepsis; shock, septic Cite this article as: Vafaei A, Heydari K, Hashemi-Nazari S S, Izadi N, Hassan Zadeh H. PIRO, SOFA and MEDS Scores in Predicting One-Month Mortality of Sepsis Patients; a Diagnostic Accuracy Study. Arch Acad Emerg Med. 2019; 7(1): e59. 1. Introduction Sepsis is the second common cause of mortality among pa- tients in intensive care unit (ICU), and it’s one of the top ten causes of death among all hospitalized patients (1). Accord- ing to the Centers for Disease Control and Prevention (CDC) reports in United-States, at least 1.7 million people develop sepsis each year. Also, approximately 270,000 Americans die due to sepsis every year (2). Based on World Health Organi- zation (WHO) reports in 2018, burden of sepsis in low-and- middle income countries is highest and sepsis, severe sepsis, ∗Corresponding Author: Hassan Hassan Zadeh; Department of Emergency Medicine, Shahid Beheshti University of Medical Sciences, Tehran, Iran. E- mail: H.hassanzadeh88@yahoo.com, Tel: 021-22915995 and septic shock lead to 20%, 40% and 60% of deaths per year, respectively (3, 4). Nowadays, different scoring systems based on clinical and laboratory findings are applied for prediction of short term mortality in patients with critical situations (5, 6). These prediction tools can help the clinicians in selecting the best course of action for the treatment of critically ill patients to get better outcomes. Sequential Organ Failure Assessment (SOFA), Mortality in Emergency Department Sepsis (MEDS), and Predisposition, Infection, Response and Organ dysfunc- tion (PIRO) are three well-known tools for assessment of ill patients with sepsis, severe sepsis and septic shock (7, 8). SOFA is an objective and simple scoring system that consid- ers the number and severity of failures in six organs including respiratory system, coagulative function, liver, cardiovascu- lar, kidney, and neurology system. This score ranges from 0 This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: http://journals.sbmu.ac.ir/aaem A. Vafaei et al. 2 to 24 and higher points predict higher mortality probability. The range of PIRO scoring is 0 to 13 (9). Usually rate of res- piratory system, bandemia, pulse rate, and temperature are being evaluated and finally for the organ dysfunction, alter- ation in the mental status, according to the Glasgow Coma Scale (GCS), systolic blood pressure, platelet count, prolon- gation of prothrombin time and etc. are being considered (10). Different studies have compared these scores for predicting mortality, but there is not a general consensus regarding the best and most accurate rule in this regard (11-13). Macdon- ald et al. reported that PIRO model performed better than SOFA score and similar to MEDS score for predicting mortal- ity in ED patients with severe sepsis and septic shock (14). In addition, in Nguyen’s study, PIRO performed equally well when compared with APACHE II and surpassed MEDS in dis- criminating survivors from non-survivors (11). Therefore, in this study, we are going to compare the performance mea- sures of SOFA, PIRO, and MEDS scoring models in predicting 30-day mortality of septic shock and severe sepsis patients. 2. Methods 2.1. Study design and setting In this diagnostic accuracy study, septic shock and se- vere sepsis patients referring to Loghmane Hakim educa- tional Hospital in Tehran province of Iran from 2017 to 2018 were examined. The performance of MEDS, SOFA, and PIRO models for predicting 30-day mortality were eval- uated using discrimination and calibration indices. The protocol of study was approved by ethics committee of Shahid Beheshti University of Medical Sciences (Ethics code: IR.SBMU.MSP.REC.1396.119). The researched adhered to the principals of Helsinki recommendations regarding the ethi- cal consideration in medical researches. 2.2. Participants All patients aged > 18 years with septic shock and severe sep- sis who were admitted to intensive care unit (ICU) during the study period were examined. Patients with heart attack, pul- monary embolism, cancer, human immunodeficiency virus (HIV ) infection, trauma and those who had recent major surgery were excluded. The instructions of scores regarding the included patients’ characteristics were considered during data collection. 2.3. Data gathering The necessary information from the clinical examination and medical records were extracted. Data collection form included age, sex, admission ward, duration of admission, transfer type (EMS or private car), history of smoking, opium abuse, history of different underlying diseases (kidney, hy- pertension, ischemic heart disease, ICU hospitalization in previous 3 months, IV antibiotic therapy in previous 30 days, previous trauma), early and final diagnosis, and vital sign findings including tachycardia, tachypnea, temperature, blood pressure, respiratory rate. Every patient was followed for at least one month (30 days). In absent cases, the research staff contacted the patient or patient’s family at certain inter- vals and attempted to collect the necessary medical informa- tion of the patient’s latest condition. A third year emergency medicine resident was responsible for data gathering, follow up, and calculation of scores for all patients, under the direct supervision of an emergency medicine specialist. 2.4. Definitions - Severe sepsis was defined as having two or more crite- ria from the "Systemic Inflammatory Response Syndrome (SIRS)", at least one criterion from signs of circulatory shock and one criterion from the evidence of infection (15). - Septic shock patients were those diagnosed with systolic blood pressure (SBP) lower than 90 mmHg who did not respond to treatment with at least one liter of Crystalloid serum, and still had SBP<90 mmHg or lactate level ≥4 mmol. 2.5. Sequential Organ Failure Assessment (SOFA) This score is used during the stay in the ICU and is based on six different indices including: respiration (PaO2/FiO2 (mmHg) or SaO2/FIO2 (mmHg)), cardiovascular system (sta- tus of hypotension), liver function (bilirubin level (mg/dl) [µmol/L]), coagulation status (platelets count), kidney func- tion (creatinine level or urine output) and neurology status (Glasgow coma scale). SOFA score ranges from 0 to 24 points and higher scores predict higher mortality probability in in- fected patients (16). 2.6. Mortality in Emergency Department Sepsis (MEDS) MEDS score comprises of nine variables, including termi- nal illness (6 points), septic shock, tachypnea or hypoxemia, platelet count<150,000 cells/mm3, bands>5%, age>65 yrs. (3 points for each variable, respectively), lower respiratory in- fection, nursing home resident, and altered mental status (2 points for each variable, respectively). In this study we used all these variables except bandemia for calculating MEDS score because bandemia was not reported for the patient in the hospital. Hence the range of this score is from 0 to 24 points depending on whether variables were present or ab- sent (17). 2.7. Predisposition, Infection, Response, and Or- gan dysfunction (PIRO) For calculation of PIRO score we used the first table provided by H. Bryant et al, in their article (11). In this scoring system This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: http://journals.sbmu.ac.ir/aaem 3 Archives of Academic Emergency Medicine. 2019; 7 (1): e59 patients got a score between 0 to 13 according to their age, co-existence of comorbidities like chronic liver disease, con- gestive cardiomyopathy, existence of community acquired urinary tract infection (UTI) or hospital acquired UTI and also the type of culprit pathogen, presence of tachycardia and tachypnea, and the number of organ failures and also hepatic failure (18). 2.8. Statistical Analysis For quantitative variables with non-normal distribution (evaluated using the Kolmogorov-Smirnov test) median (In- terquartile Range=IQR) and for qualitative variables, count (percentage) were used to describe them. The distribution of age, weight, BMI, PIRO, MEDS and SOFA scores among alive and died subjects were compared using T-test and Mann- Whitney test. In addition, the frequency of qualitative vari- ables in the two groups was compared using Chi-square and Fisher’s exact test. We used univariate logistic regression to evaluate the asso- ciation between MEDS, SOFA and PIRO scores and 30-day mortality among the studied patients. The performance of MEDS, SOFA, and PIRO models for predicting 30-day mortal- ity were evaluated using discrimination and calibration in- dices. We calculated Brier, BrierScaled and Nagelkerke’s R2 indices for overall performance, C-statistic, discrimination slope, validity indices (sensitivity, specificity, positive predic- tive value (PPV ), negative predictive value (NPV ), positive likelihood ratio (PLR) and negative likelihood ratio (NLR)), area under the curve (AUC) and also Box plots calculated for discrimination and calibration-in-the-large, calibration slope and Hosmer-Lemeshow tests were measured for evalu- ation of calibration. Also, we calculated optimism corrected with the bootstrap method (500) for all performance indices. Data were analyzed using the R software (version 3.4.1). In this study, p<0.05 was considered statistically significant for all statistical tests; yet, we presented the exact p values for all tests. 3. Results 3.1. Baseline characteristics of studied patients 200 patients with the mean age of 71.03 ± 15.59 (21 - 95) years were studied (61% male). During the 30-day follow up period, 66 (33.0%) patients died (mortality rate = 33.0%; all cases were admitted to ICU). Table 1 compares the baseline char- acteristics of studied patients between survived and non- survived groups. While the mean age was significantly higher in subjects who died, mean weight and BMI did not show any significant difference between the two groups. Mean PIRO, MEDS and SOFA scores were significantly higher in non-survived cases. Although variables such as the history of underlying diseases were different in the two groups, most of the variables related to admission and vital signs in dead and alive groups were not significantly different. There was a sig- nificant association between PIRO, MEDS, and SOFA scores with mortality (P<0.001). The Odds ratio of PIRO, MEDS, and SOFA scores in predicting the risk of one-month mortality were 1.9 (95% CI: 1.57 - 2.3), 2.14 (95% CI: 1.73 - 2.65), and 2.1 (95% CI: 1.71 - 2.59), respectively. 3.2. Score performance measurements Table 2 summarizes the overall performance, discrimination, and calibration of the scores in predicting the one-month mortality. 3.3. Overall performance Based on Brier, BrierScaled and Nagelkerke’s R2 of the mod- els, the best overall performance in predicting one-month mortality belonged to MEDS score. 3.4. Discrimination C-statistic showed that the MEDS score had the highest value in the differentiation between the survived and dead peo- ple. Based on Box Plots for predicted probabilities of death in MEDS, SOFA and PIRO scores, the highest discrimination slope belonged to MEDS score (0.62) (Figure 1). 3.5. Calibration Area under the ROC curve of PIRO, MEDS, and SOFA scores were 0.83 (95% CI=0.78-0.89), 0.94 (95% CI=0.91-0.97) and 0.87 (95% CI=0.81-0.92), respectively (Figure 2). The opti- mal cut-off points were 11.5, 5.5 and 6.5 for MEDS, SOFA and PIRO scores, respectively. At the cut point of 11.5, MEDS score had a sensitivity of 83.3% (95% CI: 72.1-91.4), specificity of 91.8% (95% CI: 85.8-95.8), PPV of 83.3% (95% CI: 73.0 - 91.4), NPV of 91.8 (95% CI: 85.3- 95.8), PLR of 10.15 (95% CI: 5.7-18.1) and NLR of 0.18 (95% CI: 0.1-0.31). At the cut point of 5.5, SOFA score had a sensitivity of 75.8% (95% CI: 63.6-85.5), specificity of 84.3% (95% CI: 77.0-90.0), PPV of 70.4% (95% CI: 59.7-81.7), NPV of 87.6 (95% CI: 79.8- 92.2), PLR of 4.83 (95% CI: 3.18-7.32), and NLR of 0.28 (95% CI: 0.18-0.44). At the cut point of 6.5, PIRO score had a sensitivity of 77.3% (95% CI: 65.3-86.7), specificity of 72.4% (95% CI: 64.0-79.7), PPV of 57.9% (95% CI: 48.3-72.5), the NPV of 86.6 (95% CI: 78.2-90.6), PLR of 2.8 (95% CI: 2.1-3.8), and NLR of 0.31 (95% CI: 0.19-0.49). The agreement between predicted mortality using PIRO, MEDS, and SOFA scores and the actual mortality of the study population was determined using Hosmer-Lemeshow (H-L) test. H-L was non-significant for all three scores (Table 2). This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: http://journals.sbmu.ac.ir/aaem A. Vafaei et al. 4 Figure 1: Box plots of predicted probabilities of death in MEDS, SOFA and PIRO scores. The discrimination slope is calculated as the difference between the mean predicted probability of alive and died subjects (solid dots indicate means). Figure 2: The area under the receiver operator characteristic (ROC) curves of PIRO, MEDS, and SOFA scores in predicting one-month mortality of severe sepsis and septic shock patients. PIRO = Predis- position Insult Response and Organ; MEDS = Mortality in Emergency Department Sepsis; SOFA = Sequential Organ Failure Assessment. 4. Discussion The results of the present study showed that MEDS scor- ing system has better discrimination and performance com- pared to the other two scoring systems in predicting 30-day mortality of septic shock and severe sepsis patients (area un- der the curve 0.94). On the best cut-off point, (score=11.5), specificity and sensitivity of this score were 91.8% and 83.3%, respectively. Results of the present article are in accordance with the AUC of 0.78-0.81 found in previous studies among patients with sepsis (8, 12, 14, 19). In addition, our results are in contrast with a recent study of emergency department sepsis patients (20), which found that MEDS score had an AUC of 0.61 and another study (11) that found an AUC of 0.63 for MEDS in a registry database of patients in the emergency department. Differences in the results of various studies could be due to differences in the methods of study. A systematic review of scoring systems in the emergency department showed that there are considerable variation between studies in the mor- tality rates and inconsistency in the definition of sepsis, se- vere sepsis, and septic shock. These variations can make valid comparisons problematic (21). The concept of MEDS score is similar to PIRO and SOFA scores, except that it is specifically designed for emergency patients. In MEDS score calculation, organ dysfunctions re- ceives greater score. In addition to these organ dysfunction parameters, constant data such as age, rapidly terminal co- morbid illness, presence of a lower respiratory infection and nursing home residence, is considered in MEDS score (8). Therefore, MEDS score has higher clinical importance (22). However, this scoring system has some limitations, for ex- ample, some data required for MEDS score calculation such as the presence of lower respiratory infection and the num- ber of platelets is not available at the time of triage manage- ment. An additional limitation is that in MEDS score calcu- lation, subjective assessment of short-term mortality by the in-charge clinician has a large weighting (14). The present study showed that a cut off of more than 11.5 points for MEDS score effectively stratified septic shock and severe sepsis pa- tients into two groups, which were significantly different in mortality rate. This cut off point was close to the cutoff point that Chen et al. had calculated in their study (23). This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: http://journals.sbmu.ac.ir/aaem 5 Archives of Academic Emergency Medicine. 2019; 7 (1): e59 Table 1: Comparing the baseline characteristics of studied patients between survived and non-survived groups Variable Survived (n = 134) Died (n = 66) P-value* Demographics Gender (male) 87 (71.31) 35 (28.69) 0.10 Age (year) 66.02 (15.75) 81.18 (8.98) <0.001 Weight (kg) 68.7 (11.48) 65 (10.32) 0.02 BMI (kg/m2 ) 23.85 (3.28) 23.05 (3.4) 0.10 ICU admission No 98 (73.13) 36 (26.87) 0.009 Yes 36 (54.55) 30 (45.45) Duration of Admission (day) <5 31 (54.39) 26 (45.61) 5-10 50 (70.42) 21 (29.58) 0.052 >10 53 (73.61) 19 (26.39) History Living in a nursing home 9 (47.37) 10 (52.63) 0.07 Smoking 41 (69.49) 18 (30.51) 0.62 Cardiovascular failure 8 (44.44) 10 (55.56) 0.03 Previous trauma 12 (57.14) 9 (42.86) 0.31 Kidney diseases 20 (68.97) 9 (31.03) <0.001 End stage disease 68 (57.63) 50 (42.37) 0.001 Hypertension 89 (63.57) 51 (36.43) 0.11 Ischemic Heart Disease 26 (50) 26 (50) 0.002 ICU admission (3 month ago) 43 (64.18) 24 (35.82) 0.54 Serum Lactate level <2 64 (92.75) 5 (7.25) <0.001 2-2.9 45 (86.54) 7 (13.46) 3-3.9 6 (60.00) 4 (40.00) ≥4 19 (27.54) 50 (72.46) Drug history Steroids 11 (61.11) 7 (38.89) 0.57 Beta blocker 34 (51.52) 32 (48.48) 0.001 Opium 25 (69.44) 11 (30.56) 0.73 IV Antibiotic (> 30 days ago) 47 (54.65) 39 (45.35) 0.001 Vital signs / SIRS Tachycardia 107 (63.69) 61 (36.31) 0.02 Tachypnea 93 (64.58) 51 (35.42) 0.24 Temperature (>38 or <35.5) 110 (65.87) 57 (43.13) 0.44 SBP<90 mmHg or MAP<70 17 (27.42) 45 (72.58) <0.001 WBC (>15000 or <4000) 84 (62.69) 50 (37.31) 0.06 Respiratory rate > 20 95 (62.91) 56 (37.09) 0.03 Acidosis 89 (64.03) 50 (35.97) 0.17 Platelet < 150000 38 (56.72) 29 (43.28) 0.02 Septic Shock 21 (29.58) 50 (70.42) < 0.001 Data are presented based of mean ± standard deviation or frequency (%).BMI=Body Mass Index; ICU: intensive care unit; IV: intravenous; SIRS: systemic inflammatory response syndrome; SBP: systolic blood pressure; MAP: mean arterial pressure; WBC: white blood cell count. In present study, PIRO and SOFA scores also showed high val- ues of discrimination in predicting 30-day mortality in septic shock and severe sepsis patients (area under the curve 0.835 and 0.872 respectively). However, the discrimination value of these two scorings system was less than MEDS. Other stud- ies have evaluated the prognostic value of PIRO and SOFA scoring systems in patients with sepsis (13, 23, 24). In Chen’s study, PIRO model had an AUC of 0.82 for 28-day mortality (23). In the de Groot’s study, among low-risk sepsis patients, PIRO scoring system had an AUC of 0.83; but in higher risk patients, it had an AUC of 0.68 (24). One of the reasons for the lower prognostic value of PIRO and SOFA scoring systems is that, PIRO model does not require knowledge of the infecting organism and has been adapted specifically for use in the emergency department (25). In this study, we used univariate models to evaluate prognostic val- ues of these three indices. It is recommended to evaluate their performance in multivariate models and also externally validate these models in larger studies. This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: http://journals.sbmu.ac.ir/aaem A. Vafaei et al. 6 Table 2: Comparing the baseline characteristics of studied patients between survived and non-survived groups (continue) Variables Survived (n = 134) Died (n = 66) P-value* Scores PIRO 5.08 ± 2.18 8.01 ± 1.9 <0.001* MEDS 8 (7 - 10) 16 (13 - 17) <0.001** SOFA 3 (3 - 5) 7 (6 - 9) <0.001** Source of infection Urosepsis 35 (66.04) 18 (33.96) 0.86 Wound sepsis 15 (83.33) 3 (16.67) 0.12 Pneumosepsis 35 (49.3) 36 (50.7) <0.001 Others 49 (84.48) 9 (15.52) 0.001 Peripheral blood smear Gram negative 42 (61.76) 26 (38.24) 0.25 Gram positive 15 (57.69) 11 (42.31) 0.27 Transfer type Own car 57 (69.51) 25 (30.49) 0.52 EMS 77 (65.25) 41 (34.75) Data are presented as mean ± standard deviation, median (Q1 - Q3) and frequency (%); PIRO: Predisposition Insult Response and Organ; MEDS: Mortality in Emergency Department Sepsis; SOFA: Sequential Organ Failure Assessment. *Based on t-test; ** Based on Mann-Whitney test. Table 3: Performance characteristics of PIRO, MEDS, and SOFA scores in predicting 30-day mortality of sepsis patients Characteristics PIRO PIRO* MEDS MEDS* SOFA SOFA* Overall Brier 0.153 0.156 0.086 0.089 0.128 0.132 BrierScaled (%) 30 30 61 61 41 41 R2 (Nagelkerke)(%) 41 40.9 71.5 70.9 50 49.3 Discrimination C-Statistic 0.835 0.836 0.941 94 0.872 0.872 Slope 0.31 - 0.62 - 0.42 - Calibration In- the-large 0 0.07 0 0.01 0 -0.01 Slope 1 1 1 0.98 1 0.99 H-L tests, X2 (P) 2.92(0.93) - 4.82(0.77) - 7.03(0.53) - H-L=Hosmer-Lemeshow; *Optimism Corrected with bootstrap method. 5. Limitation Our study has some limitations. Firstly, this was a mono- center study. Secondly the endpoint was defined as death in 30 days, but the death might have occurred for reasons other than sepsis; and finally, we just compared the perfor- mance of these three scores in univariate models. It is recom- mended to compare these three scores in multivariate pre- diction models controlling for other patients’ variables to ob- tain better prediction models for predicting short term mor- tality. 6. Conclusion This work shows that MEDS score had an acceptable accu- racy in predicting 30-day mortality of patients with severe sepsis and septic shock. 7. Appendix 7.1. Acknowledgements The authors would like to thank the clinical research devel- opment unit (CRDU) of Loghmane Hakim hospital, Shahid Beheshti University of Medical Sciences, Tehran, Iran for their support, cooperation and assistance throughout the pe- riod of study. 7.2. Author contribution A. V: Contribution to study concept and design, acquisition, analysis and interpretation of data, drafting of manuscript K. H: Contribution to study concept and design, acquisition, analysis and interpretation of data, drafting of manuscript SS. HN: Contribution to analysis and interpretation of data, drafting of manuscript N. I: Contribution to drafting of manuscript This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: http://journals.sbmu.ac.ir/aaem 7 Archives of Academic Emergency Medicine. 2019; 7 (1): e59 HH. Z: Contribution to study concept and design, ac- quisition, analysis and interpretation of data, drafting of manuscript Authors ORCIDs Ali Vafaei: 0000-0002-7129-6457 Kamran Heydari: 0000-0001-8538-4645 Hassan Hassan Zadeh: 0000-00017251-2548 7.3. Funding/Support There is no funding. 7.4. Conflict of interest The authors declare that there is no conflict or interest. References 1. Martin GS, Mannino DM, Eaton S, Moss M. The epidemi- ology of sepsis in the United States from 1979 through 2000. N Engl J Med. 2003;348(16):1546-54. 2. Rivara FP. Introduction: the scientific basis for injury control. Epidemiologic reviews. 2003;25:20-3. 3. Mehrdad R, Seifmanesh S, Chavoshi F, Aminian O, Izadi N. Epidemiology of occupational accidents in iran based on social security organization database. Iranian Red Crescent medical journal. 2014;16(1):e10359. 4. Brun-Buisson C. The epidemiology of the systemic in- flammatory response. Intensive Care Med. 2000;26 Suppl 1:S64-74. 5. 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. 6. Lim WS, van der Eerden MM, Laing R, Boersma WG, Kar- alus N, Town GI, et al. Defining community acquired pneumonia severity on presentation to hospital: an international derivation and validation study. Thorax. 2003;58(5):377-82. 7. Vincent JL, de Mendonca A, Cantraine F, Moreno R, Takala J, Suter PM, et al. Use of the SOFA score to assess the incidence of organ dysfunction/failure in intensive care units: results of a multicenter, prospective study. Working group on "sepsis-related problems" of the Eu- ropean Society of Intensive Care Medicine. Critical care medicine. 1998;26(11):1793-800. 8. Shapiro NI, Wolfe RE, Moore RB, Smith E, Burdick E, Bates DW. Mortality in Emergency Department Sep- sis (MEDS) score: a prospectively derived and vali- dated clinical prediction rule. Critical care medicine. 2003;31(3):670-5. 9. Marshall JC. The PIRO (predisposition, insult, response, organ dysfunction) model: toward a staging system for acute illness. Virulence. 2014;5(1):27-35. 10. Rathour S, Kumar S, Hadda V, Bhalla A, Sharma N, Varma S. PIRO concept: staging of sepsis. Journal of postgradu- ate medicine. 2015;61(4):235-42. 11. Nguyen HB, Van Ginkel C, Batech M, Banta J, Corbett SW. Comparison of Predisposition, Insult/Infection, Re- sponse, and Organ dysfunction, Acute Physiology And Chronic Health Evaluation II, and Mortality in Emer- gency Department Sepsis in patients meeting criteria for early goal-directed therapy and the severe sepsis resusci- tation bundle. Journal of critical care. 2012;27(4):362-9. 12. Safari S, Shojaee M, Rahmati F, Barartloo A, Hahshemi B, Forouzanfar MM, et al. Accuracy of SOFA score in predic- tion of 30-day outcome of critically ill patients. Turkish journal of emergency medicine. 2016;16(4):146-50. 13. de Groot B, Lameijer J, de Deckere ER, Vis A. The prognostic performance of the predisposition, infec- tion, response and organ failure (PIRO) classification in high-risk and low-risk emergency department sep- sis populations: comparison with clinical judgement and sepsis category. Emergency medicine journal : EMJ. 2014;31(4):292-300. 14. Macdonald SP, Arendts G, Fatovich DM, Brown SG. Com- parison of PIRO, SOFA, and MEDS scores for predicting mortality in emergency department patients with severe sepsis and septic shock. Academic emergency medicine : official journal of the Society for Academic Emergency Medicine. 2014;21(11):1257-63. 15. Bone RC, Sibbald WJ, Sprung CL. The ACCP-SCCM con- sensus conference on sepsis and organ failure. Chest. 1992;101(6):1481-3. 16. Jones AE, Trzeciak S, Kline JA. The Sequential Organ Failure Assessment score for predicting outcome in pa- tients with severe sepsis and evidence of hypoperfusion at the time of emergency department presentation. Crit- ical care medicine. 2009;37(5):1649-54. 17. Sankoff JD, Goyal M, Gaieski DF, Deitch K, Davis CB, Sabel AL, et al. Validation of the Mortality in Emergency Department Sepsis (MEDS) score in patients with the systemic inflammatory response syndrome (SIRS). Crit- ical care medicine. 2008;36(2):421-6. 18. Rubulotta F, Marshall JC, Ramsay G, Nelson D, Levy M, Williams M. Predisposition, insult/infection, response, and organ dysfunction: A new model for staging severe sepsis. Critical care medicine. 2009;37(4):1329-35. 19. Shojaee M, Safari S, Sabzghabaei A, Alavi-Moghaddam M, Arhami Dolatabadi A, Kariman H, et al. Pro-BNP ver- sus MEDS Score in Determining the Prognosis of Sep- sis Patients; a Diagnostic Accuracy Study. Emergency (Tehran, Iran). 2018;6(1):e4-e. 20. Jones AE, Saak K, Kline JA. Performance of the Mortal- ity in Emergency Department Sepsis score for predicting This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: http://journals.sbmu.ac.ir/aaem A. Vafaei et al. 8 hospital mortality among patients with severe sepsis and septic shock. Am J Emerg Med. 2008;26(6):689-92. 21. Calle P, Cerro L, Valencia J, Jaimes F. Usefulness of severity scores in patients with suspected infection in the emer- gency department: a systematic review. The Journal of emergency medicine. 2012;42(4):379-91. 22. Geier F, Popp S, Greve Y, Achterberg A, Glockner E, Ziegler R, et al. Severity illness scoring systems for early identifi- cation and prediction of in-hospital mortality in patients with suspected sepsis presenting to the emergency de- partment. Wiener klinische Wochenschrift. 2013;125(17- 18):508-15. 23. Chen YX, Li CS. Risk stratification and prognostic perfor- mance of the predisposition, infection, response, and or- gan dysfunction (PIRO) scoring system in septic patients in the emergency department: a cohort study. Critical care (London, England). 2014;18(2):R74. 24. de Groot B, de Deckere ER, Flameling R, Sandel MH, Vis A. Performance of illness severity scores to guide dis- position of emergency department patients with severe sepsis or septic shock. European journal of emergency medicine : official journal of the European Society for Emergency Medicine. 2012;19(5):316-22. 25. Howell MD, Talmor D, Schuetz P, Hunziker S, Jones AE, Shapiro NI. Proof of principle: the predisposition, infec- tion, response, organ failure sepsis staging system. Criti- cal care medicine. 2011;39(2):322-7. This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: http://journals.sbmu.ac.ir/aaem Introduction Methods Results Discussion Limitation Conclusion Appendix References