PIRO concept: staging of sepsis.

INTRODUCTION: Sepsis is common presenting illness to the emergency services and one of the leading causes of hospital mortality. Researchers and clinicians have realized that the systemic inflammatory response syndrome concept for defining sepsis is less useful and lacks specificity. The predisposition, infection (or insult), response and organ dysfunction (PIRO) staging of sepsis similar to malignant diseases (TNM staging) might give better information.
MATERIALS AND METHODS: A prospective observational study was conducted in emergency medical services attached to medicine department of a tertiary care hospital in Northern India. Patients with age 18 years or more with proven sepsis were included in the first 24 hours of the diagnosis. Two hundred patients were recruited. Multivariate logistic regression analysis was done to assess the factors that predicted in-hospital mortality.
RESULTS: Two hundred patients with proven sepsis, admitted to the emergency medical services were analysed. Male preponderance was noted (M: F ratio = 1.6:1). Mean age of study cohort was 50.50 ± 16.30 years. Out of 200 patients, 116 (58%) had in-hospital mortality. In multivariate logistic regression analysis, the factors independently associated with in-hospital mortality for predisposition component of PIRO staging were age >70 years, chronic obstructive pulmonary disease, chronic liver disease, cancer and presence of foley's catheter; for infection/ insult were pneumonia, urinary tract infection and meningitis/encephalitis; for response variable were tachypnea (respiratory rate >20/minute) and bandemia (band >5%). Organ dysfunction variables associated with hospital mortality were systolic blood pressure <90mm Hg, prolonged activated partial thromboplastin time, raised serum creatinine, partial pressure of oxygen in arterial blood/ fraction of inspired oxygen (PaO 2 /FiO 2 ) ratio <300, decreased urine output in first two hours of emergency presentation and Glasgow coma scale ≤9. Each of the components of PIRO had good predictive capability for in-hospital mortality but the total score was more accurate than the individual score and increasing PIRO score was associated with higher in-hospital mortality. The area under receiver operating characteristic curve for cumulative PIRO staging system as a predictor of in-hospital mortality was 0.94.
CONCLUSION: This study finds PIRO staging as an important tool to stratify and prognosticate hospitalised patients with sepsis at a tertiary care center. The simplicity of score makes it more practical to be used in busy emergencies as it is based on four easily assessable components.

Introduction

Sepsis is one of the leading causes of death worldwide. Its incidence is increasing though mortality attributed to it has decreased over the years at least in developed world.[12] A proper operational definition was coined for sepsis in 1991 to facilitate standardized enrolment into clinical trials.[3] However, with years of experience clinicians and researchers had realized that the systemic inflammatory response syndrome concept is less useful than originally thought and lacks specificity and clinical utility.[4] International Sepsis Definitions Conference in 2001 proposed a new sophisticated but more conceptual way of looking at sepsis syndrome: The predisposition, infection (or insult), response and organ dysfunction, also called as “PIRO” staging.[5]

Sepsis is a dynamic process involving humoral and cellular immune reactions leading to systemic inflammatory and anti-inflammatory responses, and coagulation abnormalities.[6] It should be realized that sepsis is manifestation of many different infective disease states. This heterogeneity makes risk stratification for short term prognostication and response to therapeutic interventions difficult in these patients. A few scoring systems like Acute Physiology and Chronic Health Evaluation II (APACHE II), Sequential Organ Failure Assessment (SOFA), Simplified Acute Physiology Score (SAPS) have been proposed but none of them is specific for sepsis patients and most of these scoring systems have been more predictive for large populations than individual patients and are more predictive of organ dysfunction.[789] Moreover, with the recent understanding of the pathophysiology of sepsis, the limitations of these scoring systems have come into the picture. The PIRO staging system emphasizes on accurately describing the phenotype of a patient with sepsis similar to the tumour, nodes, metastasis (TNM) model in malignancies.[5]. Each individual behaves differently to the same illness/insult/injury and the response to therapy is also not similar. Equally important is that all infections are not similar and the prognosis differs based on infection site, type of organism, extent of infection and response of the patient to infection as well as therapy. Patients might have significantly different responses to treatment based on their PIRO scores. Therefore, characterization and classification of a patient's phenotype based on parameters of PIRO staging would be more useful for enrolment in to interventional studies and prognostication as well as for better understanding of the pathophysiology of sepsis.

There are few studies which have shown the clinical utility of the PIRO staging in different clinical setting, mostly in the ICU setting.[101112131415161718] Sepsis is one of the leading causes of emergency visits in our hospital and we expect same situation in other hospitals in developing countries too. However, there is no study published from Indian subcontinent to address this important issue. Therefore we planned this study with the aim to develop a PIRO staging model for sepsis in the emergency setting.

Materials and Methods

This prospective observational study was conducted in emergency medical services attached to medicine department of a tertiary care hospital in Northern India over 10 month's period. Patients with age 18 years or more with proven sepsis were included in the first 24 hours of the diagnosis. Two hundred patients were recruited. Sepsis was defined based on diagnostic criteria given in the 2001 International Sepsis Definitions Conference.[5] Diagnosis of various infections was done according to the CDC criteria.[19] We included variables for development of PIRO staging defined in the 2001 SCCM/ESICM/ACCP/ATS/SIS International Sepsis Definitions Conference guideline document where mention of this staging system was made.[5] Importantly, logistics and feasibility regarding testing of many response and some organ dysfunction variables impacted the decision to include those variables in final model. The different variables which were considered for development of PIRO staging are given below:

The candidate variables for the predisposition (P) component of PIRO, included age, sex, alcoholism, history of myocardial infarction, cerebro-vascular accident (CVA), congestive heart failure (CHF), connective tissue disease (CTD), chronic obstructive pulmonary disease (COPD), dementia, diabetes mellitus (DM), presence of an indwelling foley's catheter, intravenous drug abuse, chronic liver disease (CLD), peripheral vascular disease (PVD), chronic renal insufficiency, immune deficiency (e.g., human immunodeficiency virus infection, steroid use, splenectomised etc.) and cancer.

The infection (I) category included the site of infection as respiratory tract: Pneumonia/non pneumonic lower tract respiratory infection/lung abscess/empyema; abdominal: Urinary tract infection (UTI)/biliary disease/liver abscess/others; line/catheter; central nervous system (CNS): Meningitis/encephalitis; endocarditis, skin/soft tissue infections, skeletal, and unknown; as well as type of infection i.e., the type of organism.

The candidate variables for response (R) were increased respiratory rate (respiratory rate >20 breaths/min), bandemia (>5% immature band forms on differential cell count), pulse rate >90 beats/min, and temperature ≥100.4° F.

The variables for organ (O) failure were neurologic: Alteration in mental status with Glasgow coma scale (GCS)≤9; cardiovascular: Systolic blood pressure (SBP) <90 mm Hg after a fluid challenge of 20 ml/kg over 30 min of crystalloid; hematologic: Platelet count ≤100,000/μL, prolongation of prothrombin time (PT) and activated partial thromboplastin time (APTT); renal: Serum creatinine ≥1.8 mg/dL not known to be chronic; and pulmonary: Respiratory rate >20/minute and hypoxemia defined as pulse oximetry oxygen saturation ≤90% on room air or ≤95% while breathing supplemental oxygen of ≥4 L/min or partial pressure of oxygen in arterial blood/fraction of inspired oxygen (PaO2/FiO2) <300.

The final outcome was in-hospital mortality. Various demographic, clinical examination data pertaining to patients were recorded in a predesigned instrument. Laboratory evaluation routinely done for all patients with sepsis at first contact with health system at emergency services was also recorded in proforma. All patients were followed up during entire hospital stay i.e., till discharge or death. All patients or their legally authorized representatives provided written informed consent. The study was approved by the institute ethics committee.

Statistical analysis

Data were expressed as percentages (%), mean ± SD, or median and 25% to 75% inter-quartile range (IQR), as appropriate. Patients were divided into hospital survivors and non-survivors. The dichotomous variables were created using clinically relevant thresholds. Continuous, normally distributed variables were compared with t-test, and for non-normally distributed, Mann-Whitney U-test was used. Categorical variables were compared by means of chi-square test. Any variable with a P value of 0.05 or less was eligible for inclusion in a logistic regression model for their corresponding individual component of the PIRO staging. Next, a stepwise logistic regression was performed to create a separate model for each component of PIRO (P, I, R, and O) to yield four final models of significant predictors of hospital mortality. Discrimination was assessed by using the area under the receiver operating characteristics curve (AUC). Finally a weighted integer score for each parameter of the PIRO score was calculated. The total PIRO score was obtained by addition of the individual P, I, R, and O integer scores. To create the weighted integer score for each parameter of the PIRO score, individual values were calculated by dividing the β-coefficient from the regression model for each independent predictor in each group by total β-coefficient and multiplying it by a multiplication factor. Multiplication factor was 5 if total beta coefficient of group was <5 and if >5 then multiplication factor was 10. Using the β-coefficient for each covariate, we created a weighted clinical decision rule by assigning a corresponding integer value for each covariate to yield the final PIRO Score. All the statistical tests performed were two tailed; P < 0.05 was considered statistically significant. Analysis was done using the statistical software “ IBM SPSS version 21.0” (IBM Corp., Chicago, USA).

Results

Two hundred adult (age ≥18 years) patients with proven sepsis, admitted to the emergency medical services were included within the first 24 hours of the diagnosis of sepsis. Male preponderance was noted (M:F ratio = 1.6:1). Mean age of study cohort was 50.50 ± 16.30 years. Out of 200 patients, 116 (58%) had in-hospital mortality. Details of demographic, clinical, microbiological and laboratory characteristics of study cohort have been further elaborated in Tables 1a–e.

Table 1a

Demographic characteristics of cohort of 200 patients with sepsis

Variables	n (%)
Age (Mean±SD) in years	50.50±16.30
Gender n (%)
Male	124 (62)
Female	76 (38)
Religion n (%)
Hindu	135 (67.50)
Sikh	54 (27)
Muslim	11 (5.50)

Table 1e

Laboratory parameters of cohort of 200 patients with sepsis at presentation

Laboratory parameters	(Mean ± SD)
Hb*(g/dl)	10.2±2.50
TLC† (per μL)	22784±18047
Band forms (%)	5.13±4.70
Platelet count (per μL)	204945±187517
PT‡ (sec)	23.10±14.40
PTI§ (%)	67.90±18.70
APTT||(sec)	35.10±10.70
RBS¶(mg/dL)	139±87.70
Urea (mg/dL)	108±78.40
Creatinine (mg/dL)	2.50±2.10
Sodium (meq/L)	136±10.80
Potassium (meq/L)	4.50±1
Serum Bilirubin (mg/dL)	2.70±1.50
pH	7.37±0.14
PaO2** (mm of Hg)	84.90±32.20
PaCO2†† (mm of Hg)	29.00±8.30
HCO3‡‡ (meq/L)	16.70±5.60
SPO2§§ (%)	93.90±4.70
PaO2/FiO2||||<300	263±86.10
Urine Output (1st2hrs) ml	38.90±16.60

*Hb = Heamoglobin; †TLC = Total leucocyte count; ‡PT = Prothrombin time; §PTI = Prothrombin time index; ||APTT = Activated partial thromboplastin time; RBS = Random blood sugar; **PaO2 = Partial pressure of oxygen in arterial blood; ††PaCO2 = Partial pressure of carbon dioxide in arterial blood; ‡‡HCO3 = Bicarbonate; §§SPO2 = Saturation of peripheral oxygen; ||||PaO2/FiO2 = Partial pressure of oxygen in arterial blood/Fraction of inspired oxygen

Various predisposing factors were noted and analysed. In univariate analysis, factors significantly associated with in-hospital mortality concerning P component were age >70 years, COPD, DM, CLD, cancer and presence of foley's catheter [Table 2]. In the multivariate logistic regression analysis predisposing factors independently associated with in-hospital mortality were age >70 years, COPD, CLD, cancer and presence of foley's catheter [Table 6].

Table 2

Association of variables of predisposition component of PIRO with hospital mortality using univariate analysis

Variables, n (%)	Survivors (n = 84)	Non-survivors (n = 116)	Total (n = 200)	P value
Sex
Male	52 (61.9)	72 (62.1)	124 (62.0)	0.981
Female	32 (38.1)	44 (37.9)	76 (38.0)
Clinical
Alcoholism	43 (51.2)	54 (46.6)	97 (48.5)	0.517
Coronary artery disease	5 (6.0)	8 (6.9)	13 (6.5)	0.789
Congestive heart failure	0 (0)	5 (4.3)	5 (2.5)	0.054
Cerebro-vascular accident	10 (11.9)	10 (8.6)	20 (10.0)	0.445
Chronic obstructive pulmonary disease	2 (2.4)	12 (10.3)	14 (7.0)	0.029
Dementia	4 (4.8)	10 (8.6)	14 (7.0)	0.291
Diabetes mellitus	17 (20.2)	41 (35.3)	58 (29.0)	0.020
Drug abuse	28 (33.3)	42 (36.2)	70 (35.0)	0.674
Chronic liver disease	4 (4.8)	20 (17.2)	24 (12.0)	0.007
Chronic renal failure	9 (10.7)	23 (19.8)	32 (16.0)	0.083
Peripheral vascular disease	1 (1.2)	1 (0.9)	2 (1.0)	0.818
Cancer	1 (1.2)	12 (10.3)	13 (6.5)	0.010
Presence of foley’s catheter	43 (51.2)	99 (85.3)	142 (71)	<0.001
Connective tissue disorder	7 (8.3)	7 (6.0)	14 (7.0)	0.529
Immunocompromised
HIV*	3 (3.6)	2 (1.7)	5 (2.5)	0.650
Steroids	0 (0.0)	3 (1.5)	3 (1.5)	0.265
Nutritional status
Normal	48 (57.1)	47 (40.5)	95 (47.5)	0.170
Obese	24 (28.6)	48 (41.4)	72 (36.0)	0.062
Cachectic	12 (14.3)	21 (18.1)	33 (16.5)	0.806

*HIV = Human immunodeficiency virus

Table 6

Selection of variables significantly associated with hospital mortality using multivariate logistic regression, within each of the four components of PIRO

Variables	Odds ratio	95% confidence interval	P value
Predisposition (P)
Age >70 years	4.97	1.19-20.82	0.028
COPD*	10.34	1.90-56.10	0.007
CLD†	5.75	1.69-19.59	0.005
Cancer	20.94	2.36-186.12	0.006
Presence of foley’s catheter	6.11	2.77-13.50	<0.001
Infection (I)
Pneumonia	2.45	1.17-5.11	0.017
Meningitis/encephalitis	8.07	1.68-38.73	0.009
UTI‡	4.81	1.64-14.14	0.004
Response (R)
Band (>5%)	2.49	1.35-4.61	0.004
Respiratory rate (>20/min)	3.72	1.34-10.38	0.012
Organ dysfunction(O)
SBP§ <90 mmHg	2.13	1.08-4.22	0.028
PaO2/FiO2 ||<300	2.52	1.26-5.00	0.008
Urine output in first 2 hours <30 ml	2.93	1.34-5.93	0.006
APTT¶ >35 seconds	2.04	1.06-4.53	0.034
Creatinine >1.8 mg/dL	2.12	1.06-4.63	0.033
GCS** ≤9	6.02	2.24-16.19	0.001

*COPD= Chronic obstructive pulmonary disease; †CLD= Chronic liver disease; ‡UTI = Urinary tract infection; §SBP= Systolic blood pressure; ||PaO2/FiO2 = Partial pressure of oxygen in arterial blood/Fraction of inspired oxygen; APTT = Activated partial thromboplastin time; **GCS = Glasgow coma scale

Pneumonia, urinary tract infection and meningitis/encephalitis were the infections significantly associated with in-hospital mortality, concerning I component in univariate analysis [Table 3]. In the multivariate logistic regression analysis, same variables were associated with in-hospital mortality as found in initial univariate analysis [Table 6].

Table 3

Association of variables of infection component of PIRO with hospital mortality using univariate analysis

Variables, n (%)	Survivors (n = 84)	Non-survivors (n = 116)	Total (n = 200)	P value
Respiratory infection
Pneumonia	22 (26.2)	47 (40.5)	69 (34.5)	0.035
Pleural effusion	21 (25.0)	32 (27.6)	53 (26.5)	0.680
Empyema/lung abscess	4 (4.8)	1 (0.9)	5 (2.5)	0.160
Abdominal infection
UTI*	6 (7.1)	26 (22.4)	32 (16.0)	0.004
Cholangitis	8 (9.5)	12 (10.3)	20 (10.0)	0.840
Liver abscess	6 (7.1)	2 (1.7)	8 (4.0)	0.071
Other abdominal infection	27 (32.1)	24 (20.7)	51 (25.5)	0.066
Central line infection	1 (1.2)	1 (0.9)	2 (1.0)	0.818
CNS† infection
Meningitis/Encephalitis	2 (2.4)	17 (14.7)	19 (9.5)	0.003
CVS‡ infection	1 (1.2)	1 (0.9)	2 (1.0)	0.349
Soft tissue infection	17 (20.2)	18 (15.5)	35 (17.5)	0.386
Skeletal infection	10 (11.9)	10 (8.6)	20 (10.0)	0.445
Organism	16 (19.0)	19 (16.4)	35 (17.5)	0.624
Name
Acinetobactor	2 (2.4)	3 (2.6)	5 (2.5)	0.998
E. coli§	8 (9.5)	9 (7.9)	17 (8.5)	0.655
MSSA||	1 (1.2)	1 (0.9)	2 (1.0)	0.999
Pseudomonas	3 (3.6)	4 (3.5)	7 (3.5)	0.999
MRSA¶	2 (2.4)	2 (1.7)	4 (2.0)	0.999

*UTI = Urinary tract infection; †CNS = Central nervous system; ‡CVS = Cardio vascular system; §E.coli = Escherichia coli; ||MSSA = Methicillin sensitive staphylococcus aureus; MRSA = Methicillin resistant staphylococcus aureus

Response variables significantly associated with mortality in univariate and multivariate logistic regression analysis were respiratory rate >20/minute and band forms >5% [Tables 4 and 6].

Table 4

Association of variables of response component of PIRO with hospital mortality using univariate analysis

Variables, (Mean ± SD)	Survivors (n = 84)	Non-survivors (n = 116)	P value
Temperature (°C)	38.04±0.71	37.96±0.72	0.436
Pulse rate (per min)	106±17	104±21	0.584
Respiratory rate (per min)	26±5	29±7	<0.001
TLC (per μL)*	23646±22120	22159±14475	0.707
Band forms (%)	3.73±3.24	6.15±5.03	<0.001
Platelet count (per μL)	207059±129157	203413±202910	0.187

*TLC = Total leucocyte count

Organ dysfunction variables significantly associated with mortality in univariate analysis were systolic blood pressure (SBP), diastolic blood pressure (DBP), mean arterial pressure (MAP), prothrombin time (PT), activated partial thromboplastin time (APTT), serum creatinine, PaO2/FiO2 ratio, urinary output in first two hours of emergency presentation, vasopressor use and GCS ≤9 [Table 5]. In multivariate analysis factors associated with in-hospital mortality were SBP <90 mm Hg, prolonged APTT (>35 sec), raised serum creatinine (>1.8 mg/dL), PaO2/FiO2 <300, decreased urinary output in first two hours of emergency presentation (<30 ml) and GCS ≤9 [Table 6].

Table 5

Association of variables of organ dysfunction component of PIRO with hospital mortality using univariate analysis

Variables, (Mean ± SD)	Survivors (n = 84)	Non-survivors (n = 116)	P value
SBP*(mmHg)	105±33	92±41	0.017
DBP†(mmHg)	64±25	54±31	0.017
MAP‡(mmHg)	77±27	67±33	0.035
PT§ (sec)	20.28±6.75	25.30±17.91	0.015
PTI|| (%)	70.13±16.94	66.37±19.98	0.167
APTT¶ (sec)	32.35±10.11	37.53±13.98	0.004
RBS** (mg/dL)	143.75±91.02	136.36±85.58	0.558
Urea (mg/dL)	97.96±82.77	116.47±74.56	0.100
Creatinine (mg/dL)	2.15±1.98	2.84±2.30	0.028
Serum bilirubin (mg/dL)	1.88±4.23	3.32±6.28	0.071
pH	7.37±0.21	7.33±0.14	0.241
PaO2†† (mm of Hg)	84.37±31.16	85.37±33.07	0.830
PaCO2‡‡ (mm of Hg)	28.57±7.55	29.37±8.91	0.509
HCO3§§ (meq/L)	17.63±5.38	16.07±5.75	0.053
SPO2||||(%)	94±4	93±5	0.284
PaO2/Fio2¶¶ <300	288.64±79.07	246.07±86.91	<0.001
Urine Output (1st2hours) ml	43±15	35±16	<0.001
Vasopressor use, n (%)	28 (33.30)	66 (56.90)	0.001
GCS***≤9, n (%)	7 (8.30)	39 (33.60)	0.007

*SBP = Systolic blood pressure; †DBP = Diastolic blood pressure; ‡MAP = Mean arterial pressure; §PT = Prothrombin time; ||PTI = Prothrombin time index; APTT = Activated partial thromboplastin Time; **RBS = Random blood sugar; ††PaO2 = Partial pressure of oxygen in arterial blood; ‡‡PaCO2 = Partial pressure of carbon dioxide in arterial blood; §§HCO3 = Bicarbonate; ||||SPO2 = Saturation of peripheral oxygen; PaO2/FiO2 = Partial pressure of oxygen in arterial blood/Fraction of inspired oxygen; ***GCS = Glasgow coma scale

Using the β-coefficient for each covariate, we created a weighted clinical decision rule by assigning a corresponding integer value for each covariate to yield the final PIRO Score [Table 7]. Each of the components of PIRO had good predictive capability for in-hospital mortality but the total score was more accurate than the individual score. We grouped the patients into logical categories by PIRO score; an increasing PIRO score was associated with higher in-hospital mortality [Figure 1]. The area under the curve for cumulative PIRO score as a predictor of in-hospital mortality was 0.94 [Table 8].

Table 7

Creation of the weighted integer score for each parameter of the PIRO score

Sig. variable	Sub-group	β-coefficient	Score value	Integer score
Predisposition (P)-group
Age (yrs)	>70	1.604	1.52	2
COPD*		2.335	2.27	2
CLD†		1.750	1.66	2
Cancer		3.042	2.89	3
Presence of foley’s catheter		1.820	1.72	2
Total		10.551		11
Infection (I)-group
Pneumonia		0.894	0.97	1
CNS	Meningitis/Encephalitis	2.089	2.29	2
Abdomen	UTI‡	1.571	1.71	2
Total		4.554		5
Response (R)-group
Band (%)	>5	0.915	2.05	2
Respiratory rate (/min)	>20	1.315	2.95	3
Total		2.230		5
Organ dysfunction (O)-group
SBP§ (mmHg)	<90	0.761	1.28	1
PaO2/FiO||2	<300	0.924	1.55	2
Urine output first in first 2 hours (mL)	<30 ml	1.042	1.75	2
APTT¶ (second)	>35	0.785	1.32	1
Creatinine (mg/dl)	>1.8	0.801	1.34	1
GCS**	≤9	1.643	2.76	3
Total		5.956		10

*COPD= Chronic obstructive pulmonary disease; †CLD = Chronic liver disease; ‡UTI = Urinary tract infection; §SBP = Systolic blood pressure; ||PaO2/FiO2 = Partial pressure of oxygen in arterial blood/Fraction of inspired oxygen; APTT = Activated partial thromboplastin time; **GCS = Glasgow coma scale

Figure 1

Performance of the predisposition, infection, response, organ dysfunction (PIRO) score in predicting in-hospital mortality

Table 8

Discrimination of the predisposition, infection, response and organ dysfunction staging system

Component	Area under receiver operating characteristic curve	95% confidence interval
Predisposition	0.79	0.744-0.856
Infection	0.74	0.674-0.808
Response	0.74	0.668-0.806
Organ dysfunction	0.81	0.745-0.867
Final predisposition, infection, response and organ dysfunction	0.94	0.900-0.971

Discussion

This prospective observational study undertaken in patients with sepsis admitted to emergency medical services of a large tertiary care center identified a group of variables associated with each component of the PIRO staging system independently associated with hospital mortality.

There is an enormous heterogeneity in the patients with sepsis presenting to emergency medical services of a large tertiary care center. Importantly, severity of illness is due to combination of the type and intensity of the initial insult, impacting on a patient with comorbidities and individual genetic backgrounds. The combination of these factors may result in variable degree and type of organ dysfunction. This calls for development of uniform risk assessment score for patients with sepsis which could incorporate factors like demographic profile, co-morbidities, type and intensity of infection; and lastly pattern and degree of organ involvement. It has been suggested that staging of sepsis similar to cancer staging (TNM staging) might give better information to assess risk and predict outcome in these patients, help in enrolment of patients into clinical studies and might also help in assessing the likely patient response to specific therapeutic interventions. The PIRO concept of classification scheme for sepsis includes predisposing condition, nature and the extent of insult, the nature and magnitude of host response, pattern and the degree of organ dysfunction. Recently this concept has been validated in studies done in western countries and it could successfully predict the risk of mortality in patients with sepsis. This proposed staging system is unique in that it considers multiple different known independent predictors of outcome. This prospective derivation study is possibly one of the first few studies done in Indian subcontinent. It confirmed findings from previous studies concerning predisposing factors, procedures, type of infections and; type and severity of organ dysfunction associated with increased risk of mortality in patients with sepsis. Rello et al. created a severity assessment score based on the PIRO concept in patients with severe community-acquired pneumonia in a historical cohort of 529 patients from the CAPUCI study. They compared the performance of the PIRO score with the APACHE II score and 2007 American Thoracic Society/Infectious Disease Society of America criteria as a prognostic index. It performed better to identify patients with higher risk of 28-day mortality.[10] Lisboa et al. derived a PIRO score for 441 ICU patients with ventilator associated pneumonia. It performed better than APACHE II score.[11] These studies though, were restricted to a specific cohort of patients admitted in ICUs. Our study developed a clinical staging system in patients presenting in emergency medical services with features of sepsis with various types of infective disease conditions and wide range of disease severity, widening its application to the broad range of infected patients. Howell et al. did secondary analysis of three prospectively collected, observational cohorts of patients with clinically suspected infection admitted to the hospital from the emergency department to derive a sepsis staging system based on the PIRO concept that risk stratifies patients. Validation of the staging system was undertaken in independent internal and external cohorts. There was a stepwise increase in mortality with increasing PIRO score. They however included patients with suspected infection and did not procure subsequent information from their hospital course. Considering the low mortality rate, there is high possibility that patients without sepsis might have been included in this study.[12] Our study included patients with proven sepsis only, thus minimizing selection bias. Studies undertaken by Rubulotta and Moreno et al. on the other hand were based on secondary analysis of cohorts with different primary objectives.[1314] In another prospective, multicenter observational study from Portugal by Granja et al., included biomarkers as well in the response category and took a dynamic view of the patient's daily clinical course to formulate the score and finally concluded that this novel approach to PIRO concept and overall score can be a better predictor of mortality for patients with community-acquired sepsis admitted to ICUs.[15] In study by Howell et al. in patients admitted to emergency services AUC of PIRO in predicting in-hospital mortality was 0.90 in the derivation cohort, 0.86 in the internal validation cohort, and 0.83 in the external validation cohort.[12] Nguyen et al. compared the performance of PIRO score with APACHE II and MEDS scores in patients admitted into the emergency department with sepsis with hospital mortality as primary outcome. The discrimination power of PIRO (AUC = 0.71) was better than MEDS but similar to APACHE II score.[16] Chen et al. carried a prospective observational study in emergency department involving 680 patients with sepsis to assess the performance of PIRO in predicting multiple organ dysfunction, intensive care unit admission, and 28-day mortality. The AUC of PIRO in predicting 28 day mortality was 0.90.[17] In our study, AUC for cumulative PIRO staging system as a predictor of hospital mortality was 0.94, higher as compared to previous studies with similar patient population i.e. from emergency area with sepsis. In another study by Cardoso et al., patients with infections (n = 1035) admitted in various wards of a large tertiary care hospital over a period of one year were analysed. The combined PIRO model as a predictor for mortality had an AUC of 0.85 in the derivation cohort and 0.84 in the validation cohort.[18]

The other merits of our study are the simplicity of score making it more practical to be used in busy emergencies as it is based on four easily assessable components. This staging system can be used for risk stratification in patients with sepsis. It represents a highly effective and easy to perform tool applicable for categorization and prognostication in emergency medical services patients.

Our study also has certain limitations. The sample size was small. This staging system needs validation in larger set of patients. As this was a prospective observational study, bias associated with these studies e.g. ascertainment bias, informational bias cannot be excluded. Our model did not include information on variables reflecting genetic polymorphism known to place patients at increased risk of severe infection and adverse outcomes. We included limited variables in R component. Various biomarkers of response such as CRP, procalcitonin, inflammatory cytokines and coagulation protein should be included in future studies to more accurately characterize this component. Many of these biomarkers are not routinely available to clinicians and researchers especially those from developing world at present. Sepsis is a dynamic process, therefore sequential changes in biomarkers and patterns of variation in organ dysfunction during hospital stay might be more important than single values at the time of emergency presentation.

Conclusion

In conclusion, this prospective observational study finds PIRO staging as an important tool to stratify and prognosticate hospitalised patients with sepsis at a tertiary care center. The simplicity of score makes it more practical to be used in busy emergencies as it is based on four easily assessable components.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.

Table 1b

Comorbidities and predisposing factors in cohort of 200 patients with sepsis at presentation

Co-morbidities and predisposing factors	n (%)
Presence of foley’s catheter	142 (71)
Alcoholism	97 (48.50)
Drug abuse	70 (35)
Diabetes mellitus	58 (29)
Chronic renal failure	32 (16)
Chronic liver disease	24 (12)
Cerebro-vascular accident	20 (10)
Chronic obstructive pulmonary disease	14 (7)
Dementia	14 (7)
Connective tissue disorder	14 (7)
Cancer	13 (6.50)
Congestive heart failure	5 (2.50)
Peripheral vascular disease	2 (1)
Immunocompromised
HIV*	5 (2.50)
Steroid intake	3 (1.50)
Nutritional status
Normal	95 (47.50)
Obese	72 (36)
Cachectic	33(16.50)

*HIV = Human immunodeficiency virus

Table 1c

Vital parameters of cohort of 200 patients with sepsis at presentation

Vital parameter	(Mean ± SD)
Temperature,°C	38.0±0.72
SBP*(mmHg)	105±33
DBP†(mmHg)	64±25
MAP‡(mmHg)	77±27
Pulse rate/min	107±13
Respiratory rate/min	26±4
GCS§ ≤9, n(%)	46 (23)

*SBP = Systolic blood pressure; †DBP = Diastolic blood pressure; ‡MAP = Mean arterial pressure; §GCS = Glasgow coma scale

Table 1d

Infection site and causative organisms of cohort of 200 patients with sepsis at presentation

Variables	n (%)
Respiratory infection
Pneumonia	69 (34.50)
Pleural effusion	53 (26.50)
Empyema/lung abscess	5 (2.50)
Abdominal infection
UTI*	32 (16)
Cholangitis	20 (10)
Liver abscess	8 (4)
Other abdominal infection	51 (25.50)
CNS† infection
Meningitis/encephalitis	19 (9.50)
CVS‡ infection
Infective endocarditis	2 (1)
Soft tissue infection	35 (17.50)
Skeletal infection	20 (10)
Organisms name
E. coli§	17 (8.50)
Pseudomonas	7 (3.50)
Acinetobactor	5 (2.50)
MRSA||	4 (2)
MSSA¶	2 (1)

*UTI = Urinary tract infection; †CNS = central nervous system; ‡CVS = Cardio vascular system; §E.coli = Escherichia coli; ||MRSA = Methicillin resistant staphylococcus aureus; ¶MSSA = Methicillin sensitive staphylococcus aureus