[Downloaded free from http://www.annals.in on Sunday, September 23, 2018, IP: 202.169.225.

53]

Janak
Mehta
The application of European system
Award for cardiac operative risk evaluation
This article is
accompanied
by an invited
II (EuroSCORE II) and Society of
Thoracic Surgeons (STS) risk-score
commentary
by Dr. Praveen
Kerala Varma

for risk stratification in Indian patients

undergoing cardiac surgery
Deepak Borde, Uday Gandhe, Neha Hargave, Kaushal Pandey1, Vishal Khullar1
Departments of Cardiac Anesthesia and 1Cardiac Surgery, P.D. Hinduja National Hospital, Lilavati Hospital,
Fortis Hospital, Mumbai, Maharashtra, India

Aims and Objectives: To validate European system for cardiac operative risk evaluation II (EuroSCORE II)
and Society of Thoracic Surgeons (STS) risk-score for predicting mortality and STS risk-score for predicting
ABSTRACT morbidity in Indian patients after cardiac surgery. Materials and Methods: EuroSCORE II and STS risk-scores
were obtained pre-operatively for 498 consecutive patients. The patients were followed for mortality and
various morbidities. The calibration of the scoring systems was assessed using Hosmer-Lemeshow test. The
discriminative capacity was estimated by area under receiver operating characteristic (ROC) curves. Results:
The mortality was 1.6%. For EuroSCORE II and STS risk-score C-statics of 5.43 and 6.11 were obtained
indicating satisfactory model fit for both the scores. Area under ROC was 0.69 and 0.65 for EuroSCORE II
and STS risk-score with P values of 0.068 and 0.15, respectively, indicating poor discriminatory power. Good
fit and discrimination was obtained for renal failure, long-stay in hospital, prolonged ventilator support and
deep sternal wound infection but the scores failed in predicting risk of reoperation and stroke. Mortality risk
was correctly estimated in low (< 2%) and moderate (2-5%) risk patients, but over-estimated in high-risk (>
5%) patients by both scoring systems. Conclusions: EuroSCORE II and STS risk-scores have satisfactory
calibration power in Indian patients but their discriminatory power is poor. Mortality risk was over-estimated
by both the scoring systems in high-risk patients. The present study highlights the need for forming a national
database and formulating risk stratification tools to provide better quality care to cardiac surgical patients in India.

Received: 23‑12‑12 Key words: European System for Cardiac Operative Risk Evaluation II, Society of Thoracic Surgeons
Accepted: 15‑05‑13 risk-score, Adult Cardiac Surgery

Access this article online INTRODUCTION could be overestimating post‑operative risk in

Website: www.annals.in
PMID:
***
The European system for cardiac operative
DOI:
risk evaluation (EuroSCORE) has been used
10.4103/0971-9784.114234 for many years since its introduction in
Quick Response Code: 1999.[1] It is widely used for predicting risk
of mortality and has served as benchmark for
assessment of cardiac surgical practices with
more than 1300 formal citations in medical
literature. However, in recent years, different
publications have highlighted that this scale
certain subgroup of patients.[2] EuroSCORE II
is available since October 2011 and has been
developed to overcome short comings of
original EuroSCORE.[3] Society of Thoracic
Surgeons (STS) risk-score in addition
to providing risk of mortality, provides
risk of post‑operative major morbidities.
A prospective observational study to compare
these two scoring systems for risk prediction
in cardiac surgery was designed.

Address for correspondence: Dr. Deepak Borde, Department of Cardiac Anesthesia, P.D. Hinduja National Hospital, Veer Savarkar Marg, Mahim, Mumbai,
Maharashtra, India. E‑mail: deepakborde2482@gmail.com

Annals of Cardiac Anaesthesia    Vol. 16:3    Jul-Sep-2013 163

[Downloaded free from http://www.annals.in on Sunday, September 23, 2018, IP: 202.169.225.53]

Borde, et al.: EuroSCORE II and STS in Indian patients

MATERIALS AND METHODS to every patient. Patients were prospectively followed

up in the post‑operative period. The primary end points
This study includes 537 consecutive patients who for the study were in‑hospital mortality; and morbidity
underwent cardiac surgery between December 2011 in the form of length of stay, ventilation duration, stroke,
and October 2012. All patients were operated by the renal failure and DSW infection. The definitions for
same surgical team. Thirty nine patients were excluded morbidity were followed as directed by STS available
from analysis because STS risk-scores could not be on http://riskcalc.sts.org/STSWebRiskCalc273/support_
calculated for them; the reasons included multiple valve definitions. Based on their assigned scores patients
surgeries, tricuspid valve surgery, concomitant MAZE were categorized into low (<  2), moderate (2‑5) and
procedure, surgery on aorta, and surgery for congenital high (> 5) risk of mortality. The statistical analysis was
heart diseases. The final analysis included 498 patients done with the statistical package for windows (SPSS
who underwent isolated coronary artery bypass grafting Version 17). The calibration of both scoring systems
(CABG), mitral valve replacement (MVR)/mitral valve was assessed using the Hosmer‑Lemeshow (HL) test,
repair and aortic valve replacement (AVR) or combined which compares the observed versus expected mortality
CABG with MVR or AVR. The patient characteristics by risk decile. Calibration is considered to be poor if
and the operative data are summarized in Table 1. Based the test is significant. The discrimination measures
on history, physical examination and investigations, the the capacity of a model, in this case EuroSCORE II
operative risk was estimated by EuroSCORE II available and STS, to differentiate between the individuals of
at http://www.euroscore.org and STS risk-score available a sample that suffer an event, death or morbidity and
at http://riskcalc.sts.org/STSWebRiskCalc273/de.aspx. those who do not. The discriminative capacity of the
The EuroSCORE II assign risks of mortality while analyzed events of the two scales was estimated by
STS risk-score assign risk of mortality, length of stay, means of receiver operating characteristic (ROC) curves.
prolonged ventilation, renal failure, stroke, deep sternal Their areas under the curve (AUC) were calculated.
wound (DSW) infection and probability of reoperation
RESULTS
Table 1: Study and EuroSCORE II cohort
Variables Study cohort EuroSCORE II The study population included 498 patients eligible for
N 498 22381 calculation of EuroSCORE II and STS risk-score. The
Age  (years) 60.48 ± 07.51 64.6 ± 12.50
patients were operated at three tertiary centers by the
Weight (kg) 68.51 ± 09.16 77.90 ± 15.90
same surgical team. The age was 60.48  ±  7.51 years,
Height (cm) 162.00 ± 07.21 168.50 ± 09.60
19.89% patients were female. The prevalence of various
Female % 19.89 30.90
Diabetes mellitus % 46.58 25
risk factors in study population is shown in Table 1. The
COPD % 08.23 10.7 overall mortality was 1.6%. Predicted mortality with
CVA % 03.81 03.20 EuroSCORE II and STS was 2.01 ± 1.41 and 1.60 ± 1.23,
Extra cardiac arteriopathy % 12.85 respectively. To validate the scores, their calibration
Infective endocarditis % 0.80 2.20 power and discriminatory power were assessed. Using
Serum creatinine (mg/dL) 0.99 1.13 HL test, C‑static of 5.43 and 6.11 were obtained for
Ejection fraction %
EuroSCORE II and STS risk-score, respectively, indicating
> 50% 55.82
satisfactory model fit for both the scores. The calculated
31‑50 27.30
21‑30 12.44
area under ROC was 0.69 and 0.65 for EuroSCORE II
≤ 20 04.41 and STS risk-score, respectively. This could not achieve
Pre‑operative IABP % 01.20 01.70 statistical significance with P values of 0.068 and 0.15,
Emergency surgery % 01.60 04.30 respectively, indicating poor discriminatory power
Type of surgery [Figure 1]. The specific observed and predicted morbidity
Isolated CABG % 86.54 46.70 rates with their level of significance are summarized in
AVR % 05.22 30.20 Table 2. Both the scores accurately predicted mortality
MVR % 05.02 15.20
in low and moderate risk patients, but over‑estimated
Combined CABG+valve % 03.20
the risk in high risk category as shown in Tables 3 and 4.
REDO surgery % 2
EuroSCORE: European system for cardiac operative risk evaluation,
COPD: Chronic obstructive pulmonary disease, CABG: Coronary artery DISCUSSION
bypass grafting, AVR: Aortic valve replacement, MVR: Mitral valve
replacement, REDO: Repeat surgery, CVA: Cerebrovascular accident,
IABP: Intra-aortic-balloon-pump Ideally modern stratification system should use large

164 Annals of Cardiac Anaesthesia    Vol. 16:3    Jul-Sep-2013

[Downloaded free from http://www.annals.in on Sunday, September 23, 2018, IP: 202.169.225.53]

Borde, et al.: EuroSCORE II and STS in Indian patients

improvements and aims to reflect better current cardiac

surgical practice. The EuroSCORE II was calculated
on a consecutive sub cohort of 16,828 patients across
154 centers (four from India) in 43 countries over a
12‑week period and its validity estimated in another
sub cohort of 5553 subjects. EuroSCORE II was
capable of predicting hospital mortality after major
cardiac surgery, was well calibrated with an excellent
discriminative capacity (AUC: 0.81, 95% confidence
interval: 0.78‑0.83). Analysis of high quality national
and international database is utilized to update risk
stratification and should be available to update it
continuously. It requires professional organization
to support such an initiative, as has been set up for
Figure 1: Comparison of mortality by European system for cardiac operative STS database (maintained by STS). This not only
risk evaluation and Society of Thoracic Surgeons. Area under receiver operating
characteristic was calculated to be 0.69 (P = 0.068) for EuroSCORE II and 0.65 improves our knowledge but also provides guidance
(P = 0.15) for STS risk-score. Both were statistically not significant for quality improvement and shared decision making.
External validation study is must before incorporating
Table 2: Observed and predicted morbidity with STS such database to local clinical practice. Such studies
risk-score
for EuroSCORE II have been recently published from
Morbidity Observed Predicted AUC
parameter morbidity % morbidity % (P value)
UK,[5,6] Spain,[7] Italy[8,9] and Finland.[10] These validation
Length of stay 18.47 12.27 0.64  (< 0.001)* studies demonstrated that EuroSCORE II is a good
Prolonged ventilation 6 09.78 0.65  (0.004)* predictor of perioperative mortality and performs better
Renal Failure 4.20 03.17 0.79  (< 0.0001)* over previous versions. However, it is important to
Stroke 0.80 01.17 0.55  (0.71) remember that some of these studies are retrospective
DSW infection 0.20 0.28 0.89  (0.008)* and hence may not necessarily represent contemporary
Reoperation 0.20 06.25 0.55  (0.71)
cardiac surgical care.
*Statistically significant, STS: Society of thoracic surgeons, AUC: Areas
under the curve, DSW: Deep sternal wound
The present study being prospective, demonstrates
Table 3: Category wise distribution of EuroSCORE II that using HL test, EuroSCORE II and STS risk-score
Risk No. of Observed mortality Predicted have good calibration power (P  =  0.71 and P  =  0.63,
category patients (actual no.) mortality % respectively) indicating satisfactory model fit. However,
Low  (0‑2) 368 01.07  (4) 01.02 the area under the ROC curve 0.69 and 0.65 for
Moderate (2‑5) 099 02.02  (2) 02.93
EuroSCORE II and STS risk-score, respectively,
High  (>  5) 031 06.45  (2) 11.08
could not achieve statistical significance (P  =  0.068
EuroSCORE: European system for cardiac operative risk evaluation
and P  =  0.15 for EuroSCORE II and STS risk-score
respectively) indicating poor discriminatory power in
Table 4: Category wise distribution of STS risk-score
the present cohort.
Risk No. of Observed mortality % Predicted
category patients (actual no.) mortality %
Low  (0‑2) 401 01  (4) 00.98 There is only one study for Indian population validating
Moderate (2‑5) 081 03.70  (3) 02.90 original EuroSCORE. This study (n = 1000) revealed good
High  (> 5) 016 06.25  (1) 10.78 calibration and discrimination power of the score.[11]
STS: Society of thoracic surgeons The EuroSCORE accurately predicted mortality in low
and moderate risk Indian patients but under‑estimated
datasets representing current clinical practice and in high‑risk Indian patients. In contrast, in the present
apply a systematic approach with the contemporary study, the risk was over estimated in high‑risk patients
modeling techniques that employ advanced estimation by both the scoring systems. Apparently, the application
and validation techniques.[4] In a step towards this aim, of these scoring systems remains uncertain in high‑risk
EuroSCORE II has been published. EuroSCORE II, an patients where risk assessment is very crucial.
update of logistic EuroSCORE model, uses similar
methodology but is derived from a more current In addition to mortality, various important morbidities
data set and refined to incorporate evidence‑based indicate quality of care and have economic implications

Annals of Cardiac Anaesthesia    Vol. 16:3    Jul-Sep-2013 165

[Downloaded free from http://www.annals.in on Sunday, September 23, 2018, IP: 202.169.225.53]
Borde, et al.: EuroSCORE II and STS in Indian patients
in cardiac surgical patients. The STS risk-score assigns
probability of length of stay, prolonged ventilation,
renal failure, stroke, DSW infection and reoperation
to every patient. In our study, for parameters of STS
risk-scores, good fit and discriminatory power was
obtained only for renal failure, long‑stay in hospital,
prolonged ventilator support and DSW infection but
the score failed in predicting reoperation and stroke.
Apparently, this highlights the difference of risk factors
prevalent in Indian population.
STS risk values can be calculated only for selected
procedures. Large population of patients undergoing
cardiac surgery in India are excluded where double
valve replacement, concomitant tricuspid valve surgery,
or MAZE procedure is commonplace. In our study, we
had to exclude 8% patients (n = 38) from analysis who
underwent surgeries for which STS risk-score was
not available. In addition, the STS risk-score is time
consuming and less user friendly.
Limitations of the study
The small sample size and performance of all surgical
procedures by the same surgical team is a major
limitation and may limit applicability of our findings to
other surgical centers. Based on EuroSCORE II and STS
risk-scores, high‑risk patients constituted substantial
number of patients in the study population 6.2% and
3.2%, respectively; hence making generalizations of the
results, particularly for high‑risk patients, and for other
units is difficult. Our study included only 10% patients
with rheumatic heart disease, one of the most common
indications of cardiac surgery in India; in addition our
study did not include patients undergoing off‑pump
CABG. The applicability of these scoring systems hence
cannot be applied to these patient populations. A large
multicenter study is recommended to further validate
these scoring systems is thus warranted.
CONCLUSIONS
To conclude, EuroSCORE II and STS risk-scores have
satisfactory calibration power in Indian patients indicating
good model fit but their discriminatory power is poor. Both
the EuroSCORE II and STS risk-scores over estimated
mortality in high‑risk category (risk-score > 5) of patients.
There is definite scope for improvement in applying
166
EuroSCORE II and STS risk-scores to Indian population
undergoing cardiac surgery. The present study highlights
need for forming a national database and utilizing it for
formulating risk stratification tools to provide better
quality care to cardiac surgical patients in India.
ACKNOWLEDGMENTS
The authors would like to thank Dr. Shreedhar Joshi MD,
FCA, DM for his contribution in preparing this manuscript
and Dr. Dnyeshwar Gajbhare, MD for the statistical analysis.
REFERENCES
1. Roques F, Nashef SA, Michel P, Gauducheau E, de Vincentiis C, Baudet E,
et al. Risk factors and outcome in European cardiac surgery: Analysis
of the EuroSCORE multinational database of 19030 patients. Eur J
Cardiothorac Surg 1999;15:816‑22.
2. Siregar S, Groenwold RH, de Heer F, Bots ML, van der Graaf Y,
van Herwerden LA. Performance of the original EuroSCORE. Eur J
Cardiothorac Surg 2012;41:746‑54.
3. Nashef SA, Roques F, Sharples LD, Nilsson J, Smith C, Goldstone AR,
et al. EuroSCORE II. Eur J Cardiothorac Surg 2012;41:734‑44.
4. Takkenberg JJ, Kappetein AP, Steyerberg EW. The role of EuroSCORE II
in 21st century cardiac surgery practice. Eur J Cardiothorac Surg
2013;43:32‑3.
5. Grant SW, Hickey GL, Dimarakis I, Trivedi U, Bryan A, Treasure T, et al.
How does EuroSCORE II perform in UK cardiac surgery; an analysis of
23 740 patients from the Society for Cardiothoracic Surgery in Great
Britain and Ireland National Database. Heart 2012;98:1568‑72.
6. Chalmers J, Pullan M, Fabri B, McShane J, Shaw M, Mediratta N, et al.
Validation of EuroSCORE II in a modern cohort of patients undergoing
cardiac surgery. Eur J Cardiothorac Surg 2013;43:688‑94.
7. Carnero‑Alcázara M, Guisasolaa JA, Lacruza FJ, Castellanosa LC,
Carnicera JC, Medinillaa EV, et  al. Validation of EuroSCORE II on a
single‑centre 3800 patient cohort. Interact CardioVasc Thorac Surg
2013;16:293‑300.
8. Barili F, Pacini D, Capo A, Rasovic O, Grossi C, Alamanni F, et al. Does
EuroSCORE II perform better than its original versions? A multicentre
validation study. Eur Heart J 2013;34:22‑9.
9. Di Dedda U, Pelissero G, Agnelli B, De Vincentiis C, Castelvecchio S,
Ranucci M. Accuracy, calibration and clinical performance of the
new EuroSCORE II risk stratification system. Eur J Cardiothorac Surg
2013;43:27‑32.
10. Biancari F, Vasques F, Mikkola R, Martin M, Lahtinen J, Heikkinen J.
Validation of EuroSCORE II in patients undergoing coronary artery
bypass surgery. Ann Thorac Surg 2012;93:1930‑5.
11. Malik M, Chauhan S, Malik V, Gharde P, Kiran U, Pandey RM. Is
EuroSCORE applicable to Indian patients undergoing cardiac surgery?
Ann Card Anaesth 2010;13:241‑5.
Cite this article as: Borde D, Gandhe U, Hargave N, Pandey K, Khullar V.
The application of European system for cardiac operative risk evaluation II
(EuroSCORE II) and Society of Thoracic Surgeons (STS) risk-score for risk
stratification in Indian patients undergoing cardiac surgery. Ann Card Anaesth
2013;16:163-6.
Source of Support: Nil, Conflict of Interest: None declared.
Annals of Cardiac Anaesthesia    Vol. 16:3    Jul-Sep-2013