Evaluation of O-POSSUM score in oesophageal surgery Romanian Journal of Anaesthesia and Intensive Care 2015 Vol 22 No 1, 7-12

ORIGINAL PAPERS

Evaluation of O-POSSUM vs ASA and APACHE II scores in patients undergoing oesophageal surgery Raluca Fodor1, Adrian Cioc2, Bianca Grigorescu2, Bogdan Lăzescu2, Sanda Maria Copotoiu1

1 2

Anaesthesiology and Intensive Care Department, University of Medicine and Pharmacy of Tîrgu Mureş, Romania Anaesthesiology and Intensive Care Clinic, Clinical County Emergency Hospital Mureş, Romania

Abstract Background and aims: Risk and prognostic scores quantify the patient’s risk of death or complication according to the severity of his illness. The aim of this study was to evaluate the predictive accuracy of O­POSSUM vs ASA and APACHE II models on patients undergoing oesophageal surgery. Material and method: In this observational retrospective study 55 patients were enrolled who had undergone surgical interventions of excision and reconstruction of the oesophagus for neoplastic oesophageal stenosis, in the Surgical Clinics (I and II) of the Clinical County Emergency Hospital Mures, between January 2011 and January 2014. By using patients file records after extracting the data we calculated the predictive mortality, according to the prognostic scores O-POSSUM, ASA and APACHE II and we analyzed its correlations with the postoperative evolution. We evaluated the discriminatory power of the three scores using the ROC (receiver-operating characteristic) curves. According to the cut-off value corresponding to each score, we compared the Kaplan Meier survival curves during the hospitalization period. Results: ROC curves analysis revealed that O-POSSUM had a better discriminatory power for mortality compared to the other two scores: AUC = 0.73 for O-POSSUM, AUC = 0.57 for APACHE II and AUC = 0.64 for ASA (p < 0.001). The cut-off value was statistically significant only in case of O-POSSUM, as it derives from the statistical analysis of the survival curves (p = 0.035). Conclusion: O-POSSUM predicts mortality more accurately compared to ASA or APACHE II in patients undergoing oesophageal surgery. Keywords: O-POSSUM, ASA, APACHE II, oesophageal surgery Rom J Anaesth Int Care 2015; 22: 7-12

Introduction Surgical interventions for the removal and reconstruction of the oesophagus represent the „corner stone” in the elective treatment of esophageal stenosis. Most patients with oesophageal cancer apply for surgery in an advanced stage of disease. Curative surgery requires considerable resources in the operating theatre Adress for correspondence:

Raluca Fodor, MD, PhD Clinical County Emergency Hospital Str. Gh. Marinescu nr 50 540136, Tîrgu Mureş, Romania E-mail: [email protected]

and in the critical care unit and is associated with important postoperative morbidity and mortality. Respiratory complications occur the most frequently and are responsible for the majority of the postoperative deaths [1]. Therefore, preoperative surgical risk assessment is a crucial part of the modern surgical practice. Risk and prognostic scores quantify the patient’s risk of death or complication according to the severity of his illness. This will allow a surgeon to anticipate the adverse events following the surgery and to facilitate the informed consent process and surgical decision making. Various surgical risk prediction models have been developed to objectively quantify the postoperative morbidity and mortality. Some of the common risk

8

Fodor et al.

prediction models in surgery are APACHE II (Acute Physiology and Chronic Health Evaluation) [2], POSSUM (Physiological and Operative Severity Score for the enUmeration of Mortality and Morbidity) [3] and ASA (The American Society of Anesthesiologist’) physical status classification [4]. The simplest and most used scoring system is the ASA score, based on the assessment of the patients’ physical status. The ASA score was initiated in the beginning of the 1940’s, and since then only minor changes have been added despite major progress in anesthesiology and surgery. The APACHE II score is a standard tool for the assessment of the degree of desease severity in ICU ranging between 0-71 points. Higher values are related to higher severity and mortality levels. This score is defined by 3 components: the acute physiological score involving 12 parameters, age score and chronic illnesses score. Each parameter has a coeficient ranging 0-4, according to the severity of abnormal findings. The APACHE II score cannot be directly converted to a percent risk of mortality. In order to calculate a mortality risk, the patient’s indication for ICU admission must be accounted for. The POSSUM model for predicting post-operative mortality was developed in 1991 using cohorts of general surgical patients, but has been used in clinical practice only in the last decade. It is based on a combination of physiological parameters, operative variables and age of the patients. Each of these factors is given a weighted value and the predicted risk of morbidity or death is calculated by a logarithmic formula.The POSSUM scoring system was particularly developed as an audit tool to compare postoperative outcome in general surgery, but it would appear that the POSSUM model increasingly overestimates mortality particularly in those who were at low risk [5]. Possible reasons for this are the increasing use of minimally invasive operative techniques and the better peri-operative care. In 1996 a new variant of POSSUM was developed: P-POSSUM [5]. It used the same set of variables as POSSUM but has a different logistic regression equation. Both POSSUM and P-POSSUM scoring systems use a 12 factor physiological score and a 6 factor operative severity score [6]. Attempts have been made to modify the POSSUM scoring system for specific surgical procedures. For example: V-POSSUM [7] for vascular, RAAAPOSSUM [8] for ruptured abdominal aortic aneurysm, CR-POSSUM [9] for colorectal surgery and OPOSSUM [10] for oesophageal and upper gastrointestinal surgery. In O-POSSUM, the operative severity score is modified to exclude operative blood loss, number of procedures and peritoneal soiling. Age was regressed independently from POSSUM.

Objective The aim of this study was to evaluate the predictive accuracy of O-POSSUM vs ASA and APACHE II scores in our case series of patients undergoing oesophageal surgery. In our setting, the O-POSSUM score is not utilized as a routine, but still holds a great potential for evaluating the risk severity associated with oesophageal surgery.

Material and method The study was observational and retrospective. The Institutional Review Board approval was obtained before enrolling the patients (n = 55). Our including criteria were: adult patients who had undergone surgical interventions of removal and reconstruction of the oesophagus, in the Surgical Clinics No. I and II of the Clinical County Emergency Hospital Mures, between January 2011 and January 2014 and were admitted postoperatively in ICU. The following data were collected from the patients’ files: age, diagnostic, length of postoperative stay in ICU, the ASA score - as it appeared in the patient pre-anesthetic informed consent, the physiological and the disease-related variables necessary for calculating APACHE II score on the first day of ICU admittance, the preoperative physiological parameters and operative variables necessary for calculating O-POSSUM score and outcome of the patient at their ICU discharge. The data were collected from patients’ files and were proccessed using Microsoft Excel and SPSS statistics version 22 (IBM, New York, USA). We calculated the predictive mortality, according to the prognostic scores O-POSSUM, ASA and APACHE II and we correlated it with the postoperative evolution. We evaluated the discriminatory power of the three scores using the ROC (receiver-operating characteristic) curves. According to the cut-off value corresponding to each score, we compared the Kaplan Meier survival curves during hospitalization period.

Results The present study enrolled a number of 55 patients (47 males and 8 females) diagnosed with neoplastic oesophageal stenosis. Patients’ age ranged 40-80 years, and the average was 61.32 ± 6.95 years. The median (range) lenght of hospitalization was 17 (2, 51) days. At the day of surgical intervention, values for OPOSSUM, APACHE II and ASA prognostic score systems were as they are presented in Table 1. Neoplastic stenosis was located most frequently at the lower third of the oesophagus (39.29%).

Evaluation of O-POSSUM score in oesophageal surgery Table 1. The average values of the preoperative prognostic scores O -P O S S UM

AP AC H E I I

24.77

15.57

3.54

6.809 14 43

5.095 5 32

0.602 2 4

Mean Std. Deviation Minimum Maximum

ASA

Using the Cox regression model, the site of tumor was not a predictor of mortality (p = 0.195, hazard ratio = 0.47) as presented in Table 2. Table 2. Tumor site as a predictor of mortality Cox regression model for tumor site as a predictor of mortality

Tumor Site

p value (higher Hazard than 0.05) Ratio 0.195 0.476

95.0% CI for Hazard Ratio Lower Upper 0.155 1.462

Fig. 2. Receiver-operating characteristic curve. O-POSSUM – real mortality. SS – Sensitivity, SP – Specificity, AUC – area under curve

CI – confidence interval

A number of 20 patients died after surgery (36.36%). We noticed that the highest mortality was in patients with neoplasic stenosis located in the intermediate and lower third of the esophagus (Figure 1). Surgical mortality is often seen as a surrogate of performance, to enable comparison between individual surgeons and hospitals. This method of comparison can be misleading due to differences in the case mix [11].

Fig. 3. Receiver-operating characteristic curve. APACHE II – real mortality. SS – Sensitivity, SP – Specificity, AUC – area under curve

Fig. 1. Deaths according to tumor location

We evaluated the discriminatory power for mortality of the three scores using the ROC (receiveroperating characteristic) curves (Figures 2, 3, 4). ROC curves analysis revealed that O-POSSUM had a better discriminatory power for mortality as compared to the other two scores: AUC = 0.73 for OPOSSUM, AUC = 0.57 for APACHE II and AUC = 0.64 for ASA (p < 0.001).

Fig. 4. Receiver-operating characteristic curve. ASA – real mortality. SS – Sensitivity, SP – Specificity, AUC – area under curve

9

10

Fodor et al.

According to the cut-off value corresponding to each score, we compared the Kaplan Meier survival curves during the hospitalization period. The cut-off value was establised in order to gain an optimal ratio between specificity and sensitivity, our target being the highest specificity (Figures 5, 6, 7).

Fig. 7. Survival curve during hospitalization period for ASA. SS – Sensitivity, SP – Specificity

Fig. 5. Survival curve during hospitalization period for O-POSSUM. SS – Sensitivity, SP – Specificity

Fig. 6. Survival curve during hospitalization period for APACHE II (cut off value = 14.5). SS – Sensitivity, SP – Specificity

The cut-off value was statistically significant only in the case of O-POSSUM, as it derives from the statistical analysis of survival curves (p = 0.035). If we change the cut-off values of the APACHE II score, in order to increase specificity, the prognostic power of this score will be superior to O-POSSUM’s (p = 0.014 vs p = 0.035) (Figure 8).

Fig. 8. Survival curve during hospitalization period for APACHE II (cut off value= 19.5). SS – Sensitivity, SP – Specificity

Evaluation of O-POSSUM score in oesophageal surgery

Discussions O-POSSUM predicts mortality more accurately than ASA or APACHE II in patients undergoing oesophageal surgery. But, in the case of patients with multiple associated comorbidities, the APACHE II score has a superior statistical significance compared to the POSSUM score. The ASA score is easy to use but it is not very precise and it does not consider surgical insult in predicting postoperative outcome. Its main advantage consists of its simplicity, but it is also its major limitation.The ASA score does not make a difference either between the illnesses of different systems nor the different abnormalities in the same system and does not rank asymptomatic patients (eg. patients with severe coronary disease). Values of the ASA score cannot guide preoperative management of the patients, because it does not make a difference in the preoperative approach for an asthmatic patient to a patient with severe renal impairment or metastatic tumors [12]. The ASA score has no capacity to delimit or to add the risk imposed by multiple diseases. The evaluation of ASA score reflects only the patient’s physical condition, with no respect to the magnitude of the surgical insult [13]. The APACHE scoring system is quite complex and time consuming. Moreover, all the parameters are not always easily obtainable, particularly outside the intensive care setting. Therefore, missing values in the data are a potential source of error in the APACHE scoring system [14].The APACHE II score is mainly useful for monitoring patients’ evolution in ICU but it is also accurate in predicting perioperative complications in the surgical patients [15, 16]. The POSSUM scoring system is an intermmediate score between ASA and APACHE regarding its objectivity and its simplicity. O-POSSUM was designed to provide a dedicated model for prediction of mortality after oesophago-gastric resections. According to our study, the O-POSSUM score system has the best discriminatory capacity in the assessment of postoperative mortality (AUC = 0.73 for O-POSSUM, AUC = 0.57 for APACHE II and AUC = 0.64 for ASA), but it could not identify patients with a higher risk for in-hospital mortality. For patients presenting multiple comorbidities, the APACHE II score is more accurate and it has a superior statistic significance than the O-POSSUM score. The high mortality is explained by the pressure to operate due to rejected patients or deferred from other hospitals, sometimes patients exceeding the indications for surgery. There were four studies of O-POSSUM on postoperative mortality in patients undergoing resection for gastro-oesophageal cancer [17-20]. Out of these four

studies, three were oesophageal studies and one gastrooesophageal study. In all of them, O-POSSUM appeared to have poor predictive accuracy and showed significant lack of goodness of fit. However, a deficiency in the current POSSUM models is that they are not based on a good understanding of the patho-physiological process that results in morbidity and mortality following cancer surgery. Recently it has become clear that the systemic inflammatory response, as evidenced by an inflammation based prognostic score (Glasgow Prognostic Score), is strongly associated with long term survival of both inoperable [21] and operable [22] cancer patients. Given that the GPS is much simpler (two factors) compared with physiological POSSUM (12 factors) further work is required to compare its value in predicting morbidity and mortality following cancer surgery. These risk assessment tools have been used to allow for the comparative audit of surgical mortality although they fail to address the prediction of morbidity and mortality in individual patients. As a retrospective study this has several limitations. Of these the most important are the relative small number of subjects included (55 patients) and the existence of bias in collecting the necessary data from patients files that may adversely impact the reliability of the results.

Conclusions Our results showed that O-POSSUM predicts mortality more accurately than the ASA or the APACHE II scores in patients undergoing oesophageal surgery. In patients with multiple comorbidities, the APACHE II score has increased statistical significance. Conflict of interest Nothing to declare

References 1. Atkins BZ, D’Amico TA. Respiratory complications after esophagectomy. Thorac Surg Clin 2006; 16: 35-48 2. Knaus WA, Draper EA, Wagner DP, Zimmerman JE. APACHE II: a severity of disease classification system. Crit Care Med 1985; 13: 818-829 3. Copeland GP, Jones D, Walters M. POSSUM: a scoring system for surgical audit. Br J Surg 1991; 78: 355-360 4. Owens WD, Felts JA, Spitznagel EL Jr. ASA physical status classifications: a study of consistency of ratings. Anesthesiology 1978; 49: 239-243 5. Whiteley MS, Prytherch DR, Higgins B, Weaver PC, Prout WG. An evaluation of the POSSUM surgical scoring system. Br J Surg 1996; 83: 812-815

11

12

Fodor et al. 6. Brooks MJ, Sutton R, Sarin S. Comparison of Surgical Risk Score, POSSUM and p-POSSUM in higher-risk surgical patients. Br J Surg 2005; 92: 1288-1292 7. Prytherch DR, Ridler BM, Beard JD, Earnshaw JJ; Audit and Research Committee, The Vascular Surgical Society of Great Britain and Ireland. A model for national outcome audit in vascular surgery. Eur J Vasc Endovasc Surg 2001; 21: 477-483 8. Boyle JR, Prytherch DR, Payne SP, Pemberton RM, Sutton GL. P-POSSUM models for abdominal aortic aneurysm surgery. Br J Surg 2001; 88(4): 621 9. Tekkis PP, Prytherch DR, Kocher HM, Senapati A, Poloniecki JD, Stamatakis JD, et al. Development of a dedicated riskadjustment scoring system for colorectal surgery (colorectal POSSUM). Br J Surg 2004; 91: 1174-1182 10. Tekkis PP, McCulloch P, Poloniecki JD, Prytherch DR, Kessaris N, Steger AC. Risk-adjusted prediction of operative mortality in oesophagogastric surgery with O-POSSUM. Br J Surg 2004; 91: 288-295 11. Markus PM, Martell J, Leister I, Horstmann O, Brinker J, Becker H. Predicting postoperative morbidity by clinical assessment. Br J Surg 2005; 92: 101-106 12. Holt N, Silverman DG. Modeling perioperative risk: can numbers speak louder than words? Anesthesiol Clin 2006; 24: 427-459 13. Slim K, Panis Y, Alves A, Kwiatkowski F, Mathieu P, Mantion G, et al. Predicting postoperative mortality in patients undergoing colorectal surgery. World J Surg 2006; 30: 100-106 14. Chandra A, Mangam S, Marzouk D. A review of risk scoring systems utilised in patients undergoing gastrointestinal surgery. J Gastrointest Surg 2009; 13: 1529-1538 15. Arvidsson S, Ouchterlony J, Nilsson S, Sjöstedt L, Svärdsudd K. The Gothenburg study of perioperative risk. Acta Anaesthesiol Scand 1994; 38: 679-690 16. Shuhaiber JH. Quality measurement of outcome in general surgery revisited: commentary and proposal. Arch Surg 2002; 137: 5254 17. Lai F, Kwan TL, Yuen WC, Wai A, Siu YC, Shung E. Evaluation of various POSSUM models for predicting mortality in patients undergoing elective oesophagectomy for carcinoma. Br J Surg 2007; 94: 1172-1178 18. Nagabhushan JS, Srinath S, Weir F, Angerson WJ, Sugden BA, Morran CG. Comparison of P-POSSUM and O-POSSUM in predicting mortality after oesophagogastric resections. Postgrad Med J 2007; 83: 355-358 19. Lagarde SM, Maris AK, de Castro SM, Busch OR, Obertop H, van Lanschot JJ. Evaluation of O-POSSUM in predicting inhospital mortality after resection for oesophageal cancer. Br J Surg 2007; 94: 1521-1526 20. Internullo E, Moons J, Nafteux P, Coosemans W, Decker G, De Leyn P, et al. Outcome after esophagectomy for cancer of the esophagus and GEJ in patients aged over 75 years. Eur J Cardiothorac Surg 2008; 33: 1096-1104

21. McMillan DC. An inflammation-based prognostic score and its role in the nutrition-based management of patients with cancer. Proc Nutr Soc 2008; 67: 257-262 22. McMillan DC. Systemic inflammation, nutritional status and survival in patients with cancer. Curr Opin Clin Nutr Metab Care 2009; 12: 223-226

Evaluarea scorurilor O-POSSUM vs ASA şi APACHE II la pacienţii supuşi chirurgiei esofagiene Rezumat Obiectiv: Scorurile de risc şi prognostic cuantifică riscul de deces sau complicaţie a pacientului pe baza severităţii bolilor. Obiectivul acestui studiu a constat în evaluarea acurateţei predictive a scorurilor OPOSSUM vs ASA şi APACHE II la pacienţii supuşi chirurgiei esofagiene. Material şi metodă: În acest studiu observaţional, retrospectiv au fost înrolaţi 55 de pacienţi care au nece­ sitat intervenţii chirurgicale de rezecţie şi reconstrucţie esofagiană pentru stenoze de etiologie neoplazică în Clinicile Chirurgie I şi II ale Spitalului Clinic Judeţean de Urgenţă Tîrgu Mureş în perioada 2011 – ianuarie 2014. Am calculat mortalitatea predictivă potrivit scorurilor de prognostic O­POSSUM, ASA şi APACHE II şi am corelat­o cu evoluţia postoperatorie a pacienţilor. Evaluarea capacităţii discriminatorii a celor trei scoruri s-a realizat utilizând curbele ROC (receiveroperating characteristic). În funcţie de valoarea de cut­ off aferentă fiecărui scor, am comparat curbele de supravieţuire Kaplan Meier pe perioada internării. Rezultate: Analiza curbelor ROC a arătat o mai bună capacitate discriminatorie în ceea ce priveşte mortalitatea pentru O-POSSUM (AUC = 0,73 pentru O-POSSUM, AUC = 0,57 pentru APACHE II şi AUC = 0,64 pentru ASA) (p < 0,001). Valoarea de cut-off a fost semnificativă statistic numai pentru O­POSSUM, informaţie derivată din analiza curbelor de supravieţuire (p = 0.035). Concluzii: Scorul O-POSSUM a prezentat cea mai bună capacitate de discriminare în predicţia mortalităţii la pacienţii supuşi chirurgiei esofagiene. Cuvinte cheie: O-POSSUM, ASA, APACHE II, chirurgie esofagiană