Jebmh.com Original Research Article

SOFA OR APACHE II, WHICH DESERVES MORE ATTENTION IN PATIENTS IN ICU? AN EXPERIENCE FROM A TERTIARY CARE HOSPITAL IN NORTH EAST INDIA Deepak Chaudhury1, Subhankar Paul2, Chandraprakash3, Ilias ali4

1Assistant Professor, Department of Emergency , Gauhati Medical College and Hospital, Assam. 2Postgraduate Trainee, Department of Emergency Medicine, Gauhati Medical College and Hospital, Assam. 3Former Postgraduate Trainee, Department of Emergency Medicine, Gauhati Medical College and Hospital, Assam. 4Professor and HOD, Department of Emergency Medicine, Gauhati Medical College and Hospital, Assam. ABSTRACT BACKGROUND There are several well recognised scoring systems for evaluation and prognostication of critically ill patients. While APACHE II (Acute Physiology and Chronic Health Evaluation II) scoring system uses a point score based on physiologic parameters, age and previous health status, the SOFA (Sequential Organ Failure Assessment) scoring system takes into account the organ failure in critically ill patients. In the assessment of critically ill patients with suspected multiorgan dysfunction admitted in ICU, the role of SOFA in predictive validity for in-hospital mortality is being widely discussed.

MATERIALS AND METHODS This was a prospective study undertaken in emergency ICU of a government tertiary care hospital in North East India over a period of one year to prognosticate the patients by using two different established scoring systems, e.g. SOFA and APACHE II.

RESULTS The results showed that serial measurement of SOFA score during first week is a very useful tool in predicting the outcome especially on the day 3. The APACHE II score on day of admission, though reliable, was not very effective in predicting the mortality rate in our setup.

CONCLUSION Serial measurement of SOFA score during first week is very useful tool in predicting the outcome of sepsis patients in ICU and better than admission APACHE II scoring in predicting mortality.

KEYWORDS Sepsis, MODS, SOFA, APACHE II. HOW TO CITE THIS ARTICLE: Chaudhury D, Paul S, Chandraprakash, et al. SOFA or APACHE II, which deserves more attention in sepsis patients in ICU? An experience from a tertiary care hospital in north east India. J. Evid. Based Med. Healthc. 2017; 4(29), 1700-1704. DOI: 10.18410/jebmh/2017/332

BACKGROUND familiar among them are Sequential Organ Failure Sepsis is life-threatening caused by a Assessment (SOFA) score;5 Acute Physiology and Chronic dysregulated host response to infection.1 Sepsis and septic Health Evaluation II (APACHE II);6 Simplified Acute are major healthcare problems affecting millions of Physiology Score II (SAPS II);7and Multiple Organ people around the world each year is a common cause of Dysfunction Score (MODS).8They allow a quantification of (ICU) mortality and morbidity.2 the severity of illness and a probability of in-hospital Multiorgan dysfunction syndrome is well established as mortality.4 A well-performing ICU prognostic model helps to the final stage of the continuum. Due to the high mortality make meaningful comparison of the hospital’s current associated with sepsis and its complications, it is necessary performance with the past.9 This allows the hospital to to rapidly diagnose and treat the underlying cause.3 identify the weakness and initiate interventions aimed at Scoring systems for use in the Intensive Care Unit quality improvement and allow patients to choose (ICU) have been developed from the past 30 years. They healthcare providers based on performance. The use of are widely used in the field of critical care medicine.4 Most these prognostic models help in providing meaningful information to physicians when discussing patient Financial or Other, Competing Interest: None. 4,9 Submission 04-03-2017, Peer Review 17-03-2017, prognosis with the patient’s relatives. Since the results of Acceptance 02-04-2017, Published 08-04-2017. laboratory tests like culture and serology are available only Corresponding Author: after 24 to 48 hours using scores like APACHE II and SOFA Dr. Deepak Chaudhury, House No. 3, Panjabari, Lakhimipathi, may help in predicting outcome in the crucial initial hours PO: Khanapara-781037, Guwahati, Assam. of management. Our study focussed on Acute Physiology E-mail: [email protected] and Chronic Health Evaluation II (APACHE II) and DOI: 10.18410/jebmh/2017/332 Sequential Organ Failure Assessment (SOFA) scoring n critically-ill sepsis patients starting from the day of

J. Evid. Based Med. Healthc., pISSN- 2349-2562, eISSN- 2349-2570/ Vol. 4/Issue 29/April 10, 2017 Page 1700

Jebmh.com Original Research Article admission into ICU as they are the most commonly used Committee. In our study, sepsis was diagnosed as criteria scoring system in our setup. laid by the American College of Chest Physicians/Society of The Acute Physiology and Chronic Health Evaluation Critical Care Medicine (ACCP/SCCM) Consensus Committee (APACHE) score is probably the best-known and most in 1992.18 widely used score. The original APACHE I score was first Sepsis patients above 18years of age were included in used in 1981 and scores for three patient factors that the study. Patients primarily suffering from , influence acute illness outcome (pre-existing disease, pulmonary embolism, cardiac tamponade, drug overdose, patient reserve and severity of acute illness). These , adrenal insufficiency, burns, tumour- included 34 individual variables, a chronic health evaluation associated lactic acidosis, patients with pregnancy and and the two combined to produce the severity score.10 patients on treatment with immunosuppressive agents Knaus et al developed the next generation of APACHE were excluded from the study. The detailed history, clinical scoring system- APACHE II. The APACHE II is measured examination and all the relevant laboratory investigations during the first 24 hrs. of ICU admission; the maximum were done. All the patients of sepsis included in the study score is 71. A score of 25 predicts mortality of 50% and a were prognosticated on the basis of APACHE II and SOFA score more than 35 represents a predicted mortality of scores. APACHE II was calculated on day of admission to 80%.The APACHE II severity score has shown a good predict mortality and to assess the extent of multiorgan calibration and discriminatory value across a range of dysfunction. SOFA scoring was done daily from day 1 to disease processes and remains the most commonly used the last day. international severity scoring system worldwide.11,12 The sepsis-related organ failure assessment score was Statistical Methods developed to evaluate organ dysfunction in patients with Descriptive and inferential statistical analysis has been sepsis. Later, it was renamed as Sequential Organ Failure carried out in the present study. Results on continuous Assessment (SOFA) score because it’s utility was not measurements are presented on Mean±SD (Min-Max) and restricted merely to patients with sepsis.13 The SOFA results on categorical measurements are presented in system was created in a consensus meeting of the Number (%). Significance was assessed at 5% level of European Society of in 1994 and significance. Student’s t-test (two-tailed, independent) has further revised in 1996. The SOFA is an organ-dysfunction been used to find the significance of study parameters on score measuring multiple organ failures daily. Each organ is continuous scale between two groups on metric graded from 0 to 4 providing a daily score of 0 to 24 parameters. Chi-square/Fisher exact test has been used to points. Serial assessment of organ dysfunction during the find the significance of study parameters on categorical initial few days of admission in ICU is a good prognostic scale between two or more groups. Receiver Operating indicator. Both the mean and highest SOFA scores are Characteristic (ROC) curves were plotted to define particularly useful predictors of outcome. Independent of discriminative value of scores as a prognosis of mortality. the initial score, an increase in SOFA score during the first An Area under ROC (AuROC) of 1 means a perfect 48 hours in the ICU predicts a mortality rate of at least discrimination while 0.5 is a random chance. A model is 50%.14,15 considered acceptable if the AuROC is ≥0.7 and is SOFA scoring system is better than APACHE II system considered excellent if the AuROC is ≥0.9. All data were in predicting mortality in ICU surgical patients. Serial analysed with SPSS 16.0 Microsoft word 10.0 and excel measurements of SOFA significantly improve their have been used to generate graphs and tables. predictive accuracy.16 Vital organ dysfunctions developing after the onset of RESULTS AND OBSERVATIONS sepsis influence outcome markedly. Studies have shown After applying the inclusion and exclusion criteria, 50 the APACHE II score at the onset of sepsis or the SOFA patients diagnosed with sepsis were finally studied, out of score and the number of organ dysfunctions developing which 28 were males (56%) and 22 were females (44%). thereafter are independent prognostic factors for patients Age of patients varied from 18 years to 90 years with mean with sepsis.15,17 48.36 years (SD±17.16). 18 patients (36%) died and 32 The objectives of our study were to assess mortality patients (64%) survived in our study (Figure 1). and morbidity of patients with sepsis in ICU and to For all patients, APACHE II scoring was done on day of prognosticate the patients by using SOFA and APACHE II admission (Table 1). Mean APACHE II score was high scores as well as comparison between these two scoring among non-survivors (23.28±9.65) than survivors system in predicting mortality. (18.75±7.34). However, p-value was 0.068, rendering this difference in APACHE II scoring between survivor and non- MATERIALS AND METHODS survivor groups statistically non-significant (Figure 2). This was a prospective observational study undertaken in The SOFA score was done daily on and from day 1 of the Emergency ICU in the Department of Emergency ICU admission. It was observed that day 1 mean SOFA Medicine of Gauhati Medical College and Hospital, score was significantly high (p=0.014) among non- Guwahati, a tertiary care Government Institution in North survivors (10.17±3.45) than survivors (7.94±2.64). East India over a period of one year from August 2014 to July 2015 after obtaining approval from Institutional Ethical

J. Evid. Based Med. Healthc., pISSN- 2349-2562, eISSN- 2349-2570/ Vol. 4/Issue 29/April 10, 2017 Page 1701

Jebmh.com Original Research Article

However, most significant difference was observed from day 3 onwards as shown in (Table No.2, Figure 3). We plotted ROC curves to define discriminative value of APACHE II and maximum SOFA scores (max SOFA) as a prognosis of mortality (Figure 4). Area under ROC (AuROC) of APACHE II was 0.643 while that of max SOFA score was 0.878, which demonstrated that max SOFA score had more discriminating value than APACHE score II in predicting mortality of sepsis patients in our study.

Apache II Non-Survived Survived <10 2(11.1%) 4(12.5%) Figure 3. Showing Comparison of Serial SOFA Scores 10-20 5(27.8%) 16(50.0%) in Survivors vs. Non-Survivor Sepsis Patients 20-30 8(44.4%) 10(31.3%) >30 3(16.7%) 2(6.3%) Total 18(100.0%) 32(100.0%) Mean±SD 23.28 ± 9.65 18.75 ± 7.34 Table 1. Comparison of APACHE II Score in Survivors and Non-Survivors

SOFA Score Non-Survived Survived p value Day 1 10.17 ± 3.45 7.94 ± 2.64 0.014* Day 2 11.63 ± 4.33 8.28 ± 2.62 0.002** Day 3 13.42 ± 4.06 6.84 ± 2.96 <0.001** Day 4 10.78 ± 3.77 5.94 ± 3.41 0.001** Day 5 12.25 ± 4.8 4.55 ± 3.27 <0.001** Day 6 12.29 ± 6.1 3.39 ± 2.77 <0.001** Day 7 14.2 ± 3.9 2.82 ± 2.61 <0.001** Day 8 13 ± 3.39 2.45 ± 2.5 <0.001** Day 9 13.8 ± 4.09 1.81 ± 1.72 <0.001** Day 10/last day 13.5 ± 5.69 1.33 ± 1.23 <0.001** Table 2. Comparison of Serial SOFA Scores in Survivors and Non-Survivors

Figure 4. ROC Curve of APACHE II and Max Sofa Score for Predicting Mortality in Sepsis Patients with their Respective AuROC

DISCUSSION Sepsis is the main cause of mortality from infection, especially if not recognised early and treated aggressively. Figure 1. Pie Diagram Showing Frequency Its recognition thereby mandates urgent attention. The Distribution of Survivors and Non-Survivors mortality recorded in our study was 36%. In large clinical trials, the mortality associated with severe sepsis and ranges between 13% and 50%.19There is lack of statistical data concerning incidence of sepsis in India. The incidence and mortality rates are considered higher than in the West. In a multicentre, prospective, observational study conducted in four Intensive Therapy Units (ITUs) in India from June 2006 to June 2009 where a total of 5,478 ITU admissions were studied. Systemic Inflammatory Response Syndrome (SIRS) with organ dysfunction was found in 25% of patients of which 52.77% were due to sepsis. The incidence of severe sepsis was Figure 2. Showing Comparison of APACHE II Scores 16.45% of all admissions. ITU mortality of all admissions in Survivors vs. Non-Survivor Sepsis Patients was 12.08% and that of severe sepsis was 59.26%.20

J. Evid. Based Med. Healthc., pISSN- 2349-2562, eISSN- 2349-2570/ Vol. 4/Issue 29/April 10, 2017 Page 1702

Jebmh.com Original Research Article

The SOFA or APACHE scores are not intended to be Ferreira et al observed that AuROC was largest for used as a tool for patient management, but as means to Highest SOFA scores (0.90; SE, 0.02) than other SOFA clinically characterise a septic patient and prognosticate derived variables including mean SOFA score (0.88, SE early. Because, even patients presenting with modest 0.03) and was significantly larger than initial SOFA scores dysfunction can deteriorate further very rapidly (P<0.001). Hence, in our study, we plotted ROC curves for emphasising the seriousness of this condition and the need APACHE II against maximum SOFA score (max SOFA) for for prompt and appropriate intervention, if not already predicting outcome. being instituted. Several components of SOFA assessing AuROC of APACHE II was 0.643 while that of max SOFA systemic dysfunction require laboratory testing and thus score was 0.878, which demonstrated that max SOFA score may not be promptly available. Other elements, such as had more discriminating value than APACHE score II in the cardiovascular score, can be affected by iatrogenic predicting mortality of sepsis patients. interventions. APACHE II score also has the disadvantage The areas under ROC were found to be 0.622 and of complicated evaluations.9 However, APACHE II and 0s.705 for APACHE II and SOFA respectively for predicting SOFA scorings have widespread popularity within the mortality in patients with septic shock in a prospective critical care community and have a well-validated study by Georgescu et al,26 which is quite similar with our relationship with mortality in critically ill patients in various results. studies.4,9,14,21 However, contrary to our findings, a study by Ho KM et Many studies have shown that high APACHE II score at al27 for determining hospital mortality, APACHE II score the time of admission was associated with high mortality.4,6 showed a better calibration and discrimination (AuROC Merwe et al22 has validated the use of the APACHE II 0.858) than max SOFA (AuROC 0.829) and admission SOFA model to accurately describe the risk of ICU death of the (AuROC 0.791). Another study by Hwang et al28 among patient population in a tertiary ICU in a developing country. ICU patients showed that the area under the curve for the Arabiet al23 also had shown significantly high APACHE II SOFA score was not different from APACHE II scoring scores in non-survivors. In our study, though it was system in predicting the outcomes. However, it is statistically not-significant (p=0.068), mean APACHE II acknowledged that method for calculating SOFA scores is score on day 1 definitely high among non-survivors than easier and simpler than APACHE II. survivors (23.28 v/s 18.75). SOFA score has been validated extensively for CONCLUSION prognostification in critically ill patients. Results of daily Serial measurement of SOFA score during first week is very SOFA scoring in our study are in accordance with other useful tool in predicting the outcome of sepsis patients in similar studies around the world. Vosylius et al24 in their ICU. However, the APACHE II score on day of admission, study on 117 ICU patients with sepsis showed that the was not very effective in predicting the mortality rate in our changes in the severity of organ dysfunction were closely study, though it was found reliable in various other studies. related to the outcome of the patients admitted to ICU, The trend of SOFA score was progressively declining in where SOFA score on day 1 and day 3 was significantly survivors while non-survivors had stable higher score higher in non-survivors than those in survivors. They have during the first week and thereafter. The SOFA score on also showed that the trend of the mean SOFA score for the day 3 was better compared with SOFA score on day 1 as first seven days of ICU stay showed a progressive decrease the tool for outcome prediction. in the total number of patients and the significant REFERENCES difference of SOFA scores between survivors and non- [1] Singer M, Deutschman CS, Seymour CW, et al. The survivors in each day over the first week in the ICU, which third international consensus definitions for sepsis was also evident in our study. Vincent et al25 in their and septic shock (sepsis-3). JAMA 2016;315(8):801- multicentric study spreading over 16 countries showed that 810. the total SOFA score increased in 44% of the non- [2] Angus DC. The lingering consequences of sepsis: a survivors, but in only 20% of the survivors who stayed in hidden public health disaster? JAMA ICU for at least 1week.Fereria et al found initial SOFA score 2010;304(16):1833-1834. up to 9 predicted mortality of less than 33% while an initial [3] Irwin RS, Rippe JM. Irwin and Rippe's intensive care SOFA score of greater than 11 predicted a mortality rate of medicine. 6th edn. Lippincott Williams & Wilkins 95%. However, in a recent study by Georgescu et al26 had 2008. shown that for the APACHE II and SOFA scores the [4] Le Gall JR. The use of severity scores in the intensive differences when deceased and survivors were compared care unit. Intensive Care Medicine 2005;31(12):1618- were not statistically significant (APACHE II- 26.76±6.742 1623. vs. 23.18±8.175 respectively for SOFA- 8.029±3.099 vs. [5] Vincent JL, Moreno R, Takala J, et al. The SOFA 7.136±3.342). (Sepsis-related Organ Failure Assessment) score to Furthermore, in our study, we plotted ROC curves to describe organ dysfunction/failure. On behalf of the define discriminative value of scores as a prognosis of working group on sepsis-related problems of the mortality. Value of AuROC more than 0.7 shows good European society of intensive care medicine. discriminating value. Intensive Care Med 1996;22(7):707-710.

J. Evid. Based Med. Healthc., pISSN- 2349-2562, eISSN- 2349-2570/ Vol. 4/Issue 29/April 10, 2017 Page 1703

Jebmh.com Original Research Article

[6] Knaus WA, Draper EA, Wagner DP, et al. APACHE II: [19] Balk RA. Severe sepsis and septic shock: definitions, a severity of disease classification system. Crit Care epidemiology, and clinical manifestations. Crit Care Med 1985;13(10):818-829. Clin 2000;16(2):179-192. [7] Le Gall JR, Lemeshow S, Saulnier F. A new Simplified [20] Todi S, Chatterjee S, Sahu S, et al. Epidemiology of Acute Physiology Score (SAPS II) based on a severe sepsis in India: an update Crit Care European/north American multicenter study. JAMA 2010;14(Suppl 1):P382. 1993;270(24):2957-2963. [21] Halim DA, Murni TW, Redjeki IS. Comparison of [8] Marshall JC, Cook DJ, Christou NV, et al. Multiple Apache II, SOFA, and modified SOFA scores in organ dysfunction score: a reliable descriptor of a predicting mortality of surgical patients in intensive complex clinical outcome. Crit Care Med care unit at Dr Hasan Sadikin general hospital. Crit 1995;23(10):1638-1652. Care & Shock 2009;12:157-169. [9] Vincent JL, Moreno R. Clinical review: scoring systems [22] van der Merwe E, Kidd M, Meltzer S, et al. Validating in the critically ill. Crit Care 2010;14(2):207. the use of the APACHE II score in a tertiary south [10] Knaus WA, Zimmerman JE, Wagner DP, et al. African ICU. Southern African Journal of Critical Care APACHE-Acute Physiology and Chronic Health 2005;21(1):46-54. Evaluation: a physiologically based classification [23] Arabi Y, Al Shirawi N, Memish Z, et al. Assessment of system. Crit Care Med 1981;9(8):591-597. six mortality prediction models in patients admitted [11] Wong DT, Crofts SL, Gomez M, et al. Evaluation of with severe sepsis and septic shock to the intensive predictive ability of APACHE II system and hospital care unit: a prospective cohort study. Critical Care outcome in Canadian intensive care unit patients. Crit 2003;7(5):R116-122. Care Med 1995;23(7):1177-1183. [24] Vosylius S, Sipylaite J, Ivaskevicius J. Sequential [12] Gupta R, Arora VK. Performance evaluation of organ failure assessment score as the determinant of APACHE II score for an Indian patient with outcome for patients with severe sepsis. Croat Med J respiratory problems. Indian J Med Res 2004;45(6):715-720. 2004;119(6):273-282. [25] Vincent JL, de Mendonca A, Cantraine F, et al. Use of [13] Aggarwal AN, Agarwal R, Gupta D, et al. Non- the SOFA score to assess the incidence of organ pulmonary organ dysfunction and its impact on dysfunction/failure in intensive care units: results of a outcome in patients with acute . multicenter, prospective study. Critical Care Medicine Chest 2007;132(3):829-835. 1998;26(11):1793-1800. [14] Ferreira FL, Bota DP, Bross A, et al. Serial evaluation [26] Georgescu AM, Szederjesi J, Copotoiu SM, et al. of the SOFA scores to predict outcome in critically ill Predicting scores correlations in patients with septic patients. JAMA 2001;286(14):1754-1758. shock-a cohort study. Rom J AnaesthInt Care [15] Pittet D, Thiévent B, Wenzel RP, et al. Bedside 2014;21(2):95-98. prediction of mortality from bacteremic sepsis. A [27] Ho KM, Lee KY, Williams T, et al. Comparison of dynamic analysis of ICU patients. Am J RespirCrit acute physiology and chronic health evaluation Care Med 1996;153(2):684-693. (APACHE) II score with organ failure scores to predict [16] Azoulay E, Adrie C, De Lassence A, et al. hospital mortality. Anaesthesia 2007;62(5):466-473. Determinants of postintensive care unit mortality: a [28] Hwang SY, Lee JH, Lee YH, et al. Comparison of the prospective multicenter study. Crit Care Med sequential organ failure assessment, acute physiology 2003;31(2):428-432. and chronic health evaluation ii scoring system, and [17] Degoricija V, Sharma M, Legac A, et al. Survival trauma and injury severity score method for analysis of 314 episodes of sepsis in medical intensive predicting the outcomes of intensive care unit trauma care unit in university hospital: impact of intensive patients. The American Journal of Emergency care unit performance and antimicrobial therapy. Medicine 2012;30(5):749-753. Croat Med J 2006;47(3):385-397. [18] Bone RC, Balk RA, Cerra FB, et al. Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. The ACCP/SCCM consensus conference committee. American college of chest physicians/society of critical care medicine. Chest 1992;101(6):1644-1655.

J. Evid. Based Med. Healthc., pISSN- 2349-2562, eISSN- 2349-2570/ Vol. 4/Issue 29/April 10, 2017 Page 1704