Professional Documents
Culture Documents
Thesis Summary
Submitted by :
ACHMAD MUSTIKA
NIM: 13/359962/PKU/14294
Thesis Summary
Submitted by :
ACHMAD MUSTIKA
NIM : 13/359962/PKU/14294
Content Supervisor
Methods supervisor
DR. Med. dr. Untung Widodo Sp. An, KIC Date : ……………………………
ABSTRACT
Background. Risk stratification plays an important role on open heart surgery all over
the world. Early morbidity and mortality has become primary outcome in many models,
which is useful and aimed to determined indications for surgery, an estimation of
resource’s need, and to determine informed consent. EuroSCORE II is a model of risk
stratification that has been used recently and validated. The aim of this research is to
validate EuroSCORE II as a predictor of major morbidity for open heart surgery patient
(valvular and congenital heart disease) at Dr. Sardjito General Hospital.
Methods. This is a cohort retrospective study. Data was taken from 92 patient post open
heart surgery in year 2006-2016 at Dr.Sardjito General Hospital. EuroSCORE II was
used to predict major morbidity. The discrimination ability was analyzed by the receiver
operating characteristics (ROC) curve. Calibration was assesed with area under the
receiver operating characteristic curve (AUC) and the Hosmer-Lemeshow test.
Results. There were 92 samples, of which 26 patients (28.26%) got major morbidity.
EuroSCORE II showed poor discrimination as predictor for morbidity (AUC=0.626)
and poor calibration outcome (Hosmer-Lemeshow: P<0.05). The predicted value of
EuroSCORE II is underestimate in comparison with actual value.
P < 0.05
In the calibration results of the Hosmer-Lemeshow test, from the curve (Figure 2)
we found that the predicted value of EuroSCORE II underestimated the actual data of
major morbidity, with p = 0.009 (Table 3) statistically p > 0.05. Therefore, it can be
concluded that the EuroSCORE II scoring system has a poor calibration capability.
Intraoperative
Logistic regression (p < 0.05) bold numbers, univariate and multivariate analysis
DISCUSSION
Risk Stratification plays an important role in cardiac surgical practice worldwide.
Early death and morbidity have been the clinical outcomes assessed by many models, as
risk stratification prediction is useful and can help determine surgical indications,
estimate resource requirements, informed consent and quality monitoring of surgeons
and institutions (4). EuroSCORE was developed to predict mortality during
hospitalization, but there have been several studies that tested the accuracy of
EuroSCORE in predicting postoperative morbidity of cardiac surgery, using different
parameters (4,9-13). Because morbidity consists of heterogeneous parameters, it is
difficult to find models that predict overall postoperative complications. However, risk
stratification and preoperative prediction of postoperative morbidity are needed,
especially if this is done with a widely used and validated risk stratification model such
as EuroSCORE (4). The validation study of EuroSCORE II as a predictor of major
morbidity aims to analyze whether the improved version of EuroSCORE can be applied
in Dr. Sardjito General Hospital.
In this study, the patients who satisfied the inclusion and exclusion criteria were
93 patients, normality test was performed with Q-Q plots to see whether the distribution
of samples was normal or not. One sample outlier was obtained so the total number of
samples were 92 patients. The number of these subjects is small compared to previous
studies, Wang et al (2013) with 146 subjects in New Zealand, Najafi et al (2014) with
570 subjects in Iran, Toumpolis et al (2005) with 5051 subjects in the United States, and
Biancari et al (2012) with 1027 subjects in Finland (4,9,11,13).
Actually, the number of cardiac surgery in Dr. Sardjito General Hospital from
2006-2016 was quite a lot, which were 442 patients. However, administrative
completeness and policy of medical record management which regulates that patients
who do not come back to the hospital within 5 years then the data will be inactive, thus
not all subjects can be analyzed.
Discrimination ability test in this study can be seen in Figure and Table 2, Figure
1 showed the result of AUC value equal to 0.626 (95% CI: 0.507-0.748), thus it can be
concluded that the discrimination power of this scoring system is poor, where similar
result was found in a study that compared discrimination test to major morbidity, which
was Toumpoulis et al (2005) to 5051 patients with cardiac surgery in the United States
(having AUC value of 0.66) and Najafi et al (2014) to 570 patients undergoing isolated
CABG surgery in Iran (AUC 0.617), while other studies showed strong discrimination
analysis results of EuroSCORE II such as studies conducted by Wang et al (2014) with
AUC 0.720, Biancari et al (2012) with AUC 0.747, Wang et al (2017) with AUC value
of 0.719 and Hirose et al (2009) with an AUC value of 0.705 (4,9,11-14).
In the calibration capability analysis of EuroSCORE II on major morbidity in
patient population in Dr. Sardjito General Hospital, p value = 0.009 (p < 0.05) was
found, thus calibration ability of EuroSCORE II to major morbidity showed poor
calibration. From the image of the Hosmer-Lemeshow test calibration curve in Figure 2,
the lower predicted (underestimated) value of the actual value of major morbidity is
shown. In other studies that tested the calibration of major morbidity, only 1 study of
Najafi et al (2014) in the patient population who underwent isolated CABG surgery
showed good calibration results p = 0.119 (p > 0.05) (9).
In this study, poor discrimination and calibration result is caused by several
limitations and weaknesses in this study, where this is a retrospective study, which can
produce minor bias which disturbed validity because of missing subject or incomplete
data, hard-to-find data (such as those on chronic lung disease), inaccurate data during
sampling (such as data on the number of procedures taken through anesthesia status,
whereas more accurate data can be extracted from operative surgery report data) and
there were some variables that were unavailable in our population such as extracardiac
arteriopathy, poor mobilization, perioperative critical status, aortic thoracic surgery,
CCS angina 4, category emergency surgery (emergency and saalvage). This study also
used a small sample size compared with other studies due to the policy of medical
record management in Dr Sardjito General Hospital. In this study, there were
differences in the characteristics of the subjects with the population of EuroSCORE II,
which is likely due to differences in demography and variation of medical care between
patients in European countries and this study population in which in this study, age,
BMI, chronic lung disease, DM with therapy insulin and chronic lung disease have a
lower population than EuroSCORE II. In this study population, all patients underwent
elective surgery and only infected endocarditis was greater in frequency than the
EuroSCORE II population, which may affect the outcome of this study. The difference
in the etiology of valvular heart disease between the EuroSCORE II population and our
study population may also have an effect on outcomes, where valve degenerative
etiology is more common in Europe whereas in our population, most causes are
rheumatic heart disease (15). This study also only took data from one data center only
so it cannot describe the condition of the population of the whole country. Differences
in the quality of human resources, facilities and infrastructure may influence the
outcomes, therefore looking for other important risk factors based on regional needs and
patient characteristics in Dr Sardjito General Hospital may be able to be used as a new
model to accurately predict the outcome.
There is only a preoperative factor in EuroSCORE II consisting of patient factor,
cardiac factor and surgical factor, so to determine other factors affecting other
outcomes, this study also observed intraoperative factors (ischemic time,
Cardiopulmonary Bypass (CPB) time, Cross Clamp (Aox) Time) which were not
present in the EuroSCORE II parameter. The result of independent T-test showed
difference of mean value of ischemic time, CPB time, Cross Clamp (Aox) Time
intraoperative factors between those who experienced and not experienced major
morbidity. Therefore, intraoperative factors may have an effect on the outcomes of
major postoperative morbidity of cardiac surgery in accordance to other studies, in
which many have reported that CPB time and Cross Clamp (Aox) time are independent
factors in predicting postoperative mortality and morbidity (16,17). However, the
purpose of this study was not to develop new model/score or to determine the impact of
other independent variables on the occurrence of postoperative major morbidity of
cardiac surgery which is known that for most risk factors, the predictive value for
morbidity is very complex and different from mortality, while the addition of
intraoperative parameter in future model may be able to increase its discrimination
ability.
Logistic regression test showed univariate analysis results of perioperative factors
that can affect the occurrence of major morbidity where age, creatinine level before
surgery, gender, creatinine clearance of moderate category, CPB time, Aox Time and
Ischemic Time were factors affecting the occurrence of major morbidity (p < 0.05).
Multivariate analysis found only age variable alone can have a significant effect on the
occurrence of major morbidity following cardiac surgery with p = 0.025 (p < 0.05). This
analysis may be used as a reference in the future to determine the development of
predictor model of the occurrence of major morbidity following cardiac surgery at Dr
Sardjito General Hospital.
CONCLUSION AND SUGGESTION
In this study, EuroSCORE II had poor discrimination ability and calibration
result. The predicted value of EuroSCORE II found in this study was underestimate in
comparison to the actual value. EuroSCORE II cannot be used to discriminating groups
experiencing major morbidity or not. EuroSCORE II cannot be used for patient
stratification based on the percentage of likelihood of major morbidity. It is necessary to
develop another model according to the characteristics of the population in Dr. Sardjito
General Hospital. Other factors need to be identified such as the addition of
intraoperative factor to predict the occurrence of major morbidity in Dr Sardjito General
Hospital. It is necessary to check the parameters contained in EuroSCORE II on the
medical records of each patient who will be undergoing open heart surgery for further
development and study.
REFERENCES
1. Warner CD, Weintraub WS, Craver JM, Jones EL, Gott JP, Guyton RA. Effect of
Cardiac Surgery Patient Characteristics on Patient Outcomes From 1981 Through
1995. Circulation. 1997 Sep 2;96(5):1575–9.
2. Shahian DM, Edwards FH, Ferraris VA, Haan CK, Rich JB, Normand S-LT, et al.
Quality measurement in adult cardiac surgery: part 1--Conceptual framework and
measure selection. Ann Thorac Surg. 2007 Apr;83(4 Suppl):S3-12.
3. Nashef SA, Roques F, Michel P, Gauducheau E, Lemeshow S, Salamon R.
European system for cardiac operative risk evaluation (EuroSCORE). Eur J Cardio-
Thorac Surg Off J Eur Assoc Cardio-Thorac Surg. 1999 Jul;16(1):9–13.
4. Toumpoulis I, Anagnostopoulos C, Swistel D, Derosejr J. Does EuroSCORE
predict length of stay and specific postoperative complications after cardiac
surgery? Eur J Cardiothorac Surg. 2005 Jan;27(1):128–33.
5. Dupuis J-Y, Wang F, Nathan H, Lam M, Grimes S, Bourke M. The Cardiac
Anesthesia Risk Evaluation ScoreA Clinically Useful Predictor of Mortality and
Morbidity after Cardiac Surgery. J Am Soc Anesthesiol. 2001 Feb 1;94(2):194–
204.
6. García JLC, Roque ue FJV, Rodríguez ez RM, Yera ra GJB, Bravo vo EC, Plana
na YM, et al. Temporary validation of EuroSCORE model for assessing the results
of mitral valve replacement surgery. CorSalud Rev Enfermedades Cardiovasc.
2014 Jun 30;5(2):170–5.
7. Dahlan M. Langkah-Langkah Membuat Proposal Penelitian Bidang Kedokteran
dan Kesehatan. 2nd ed. Jakarta: Sagung Seto; 2014. (Evidence Based Medicine).
8. Nashef SAM, Roques F, Sharples LD, Nilsson J, Smith C, Goldstone AR, et al.
EuroSCORE II. Eur J Cardiothorac Surg. 2012 Apr 1;41(4):734–45.
9. Najafi M, Sheikhvatan M, Sheikhfathollahi M. Discriminative Power of
EuroSCORE in Predicting Morbidity and Prolonged Hospital Stay in an Iranian
Sample Population. J Tehran Univ Heart Cent. 2014 Jan 12;9(1):15–9.
10. Wang C, Zhang G, Lu F, Li B, Zou B, Han L, et al. A local risk prediction model
for prolonged ventilation after adult heart valve surgery in a Chinese single center.
Heart Lung J Crit Care. 2013;42(1):13–8.
11. Wang L, Han Q-Q, Qiao F, Wang C, Zhang X-W, Han L, et al. Performance of
EuroSCORE II in patients who have undergone heart valve surgery: a multicentre
study in a Chinese population. Eur J Cardiothorac Surg. 2014 Feb 1;45(2):359–64.
12. Wang TKM, Harmos S, Gamble GD, Ramanathan T, Ruygrok PN. Performance of
contemporary surgical risk scores for mitral valve surgery: Risk scores and mitral
surgery. J Card Surg. 2017 Mar;32(3):172–6.
13. 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 Jun;93(6):1930–5.
14. Hirose H, Inaba H, Noguchi C, Tambara K, Yamamoto T, Yamasaki M, et al.
EuroSCORE predicts postoperative mortality, certain morbidities, and recovery
time. Interact Cardiovasc Thorac Surg. 2009 Oct 1;9(4):613–7.
15. Iung B. A prospective survey of patients with valvular heart disease in Europe: The
Euro Heart Survey on Valvular Heart Disease. Eur Heart J. 2003 Jul;24(13):1231–
43.
16. Al-Sarraf N, Thalib L, Hughes A, Houlihan M, Tolan M, Young V, et al. Cross-
clamp time is an independent predictor of mortality and morbidity in low-and high-
risk cardiac patients. Int J Surg. 2011;9(1):104–109.
17. Salis S, Mazzanti VV, Merli G, Salvi L, Tedesco CC, Veglia F, et al.
Cardiopulmonary Bypass Duration Is an Independent Predictor of Morbidity and
Mortality After Cardiac Surgery. J Cardiothorac Vasc Anesth. 2008
Dec;22(6):814–22.