Sage Journals: Discover world-class research

Abstract

Background

Critical care management of patients undergoing transcatheter aortic valve implantation (TAVI) is a major determinant of their outcome.

Aims

The aim of this study was to compare the prognostic performance of four general scoring systems [Acute Physiology and Chronic Health Evaluation (APACHE), Simplified Acute Physiology Score (SAPS), Sequential Organ Failure Assessment (SOFA), and MultiOrgan Dysfunction (MOD) scores] in TAVI patients.

Methods

Between 1 June 2008 and 30 June 2014, 75 patients (81.2 ± 6.4 years old, 36 men and 39 women) who underwent TAVI were scored during the first 24 h of their stay at the intensive care unit (ICU). The outcome measures were in-hospital and 30-day mortality and in-hospital and 30-day morbidity defined as myocardial infarction, implantation of permanent pacemaker, stroke, tamponade, major bleeding, vascular access site complications and prolonged ventilation.

Results

Four patients (5.3%) died in ICU and one more during follow-up, indicating a 30-day mortality rate of 6.6%. Regarding in-hospital mortality, the area under the ROC curve (AUC) was 0.92 for SAPS II, 0.88 for APACHE II, 0.73 for MODS and 0.74 for SOFA. Regarding 30 day-mortality, SAPS II and APACHE II performed equally higher (AUC = 0.88) than the other two scores (0.79 for MODS and 0.80 for SOFA). SAPS II had the best calibration among all four scores for in-hospital and 30-day mortality (χ²= 3.06 and χ²= 3.29, respectively). AUCs for in-hospital and 30-day morbidity were above 0.7 for SAPS II and APACHE II.

Conclusions

SAPS II and APACHE II are reliable mortality and morbidity risk stratification models for TAVI patients with high calibration and discrimination.

Keywords

Transcatheter aortic valve implantation scoring systems intensive care outcome

Introduction

Transcatheter aortic valve implantation (TAVI) is an increasingly common procedure in high-risk and inoperable patients with severe aortic stenosis.^1,2 Patients undergoing TAVI are elderly with multiple co-morbidities and considerable frailty, generating a population of high acuity with great risk of perioperative mortality and morbidity; therefore, post-TAVI Intensive Care Unit (ICU) management has become more complex and challenging.^3,4 Despite the less-invasive nature of TAVI compared with conventional open heart surgery, a number of postoperative events and complications have been identified and need to be diagnosed and managed appropriately by intensivists.⁵ Aggressive therapies and all manner of support necessary may be utilized in the acute setting, including temporary hemodialysis, intensive treatment of respiratory disorders and early use of broad-spectrum antibiotics as sepsis is often masked in the elderly.⁴

Scoring systems were introduced into intensive care medicine to provide valuable tools for judging a patient’s condition and likely outcome. TAVI patients are critically ill patients who require robust models for mortality prediction. Furthermore, the prognostic power of different performance measures has to be evaluated in order to reduce financial costs and improve quality of care.^6,7

The purpose of our study was to compare four scoring systems commonly used in critical care management [Acute Physiology and Chronic Health Evaluation (APACHE), Simplified Acute Physiology Score (SAPS), Sequential Organ Failure Assessment (SOFA), and MultiOrgan Dysfunction (MOD) scores] with regard to their immediate and short-term predictive values and clinical applicability for TAVI patients.

Methods

The study sample included all patients with symptomatic aortic valve stenosis (aortic valve area ≤ 0.6 cm²/m²), who underwent TAVI to our department between June 2008 and June 2014. All patients were deemed inoperable or at high surgical risk for conventional surgery by a multidisciplinary team consisting of cardiothoracic surgeons, cardiologists, and anesthesiologists. Screening for TAVI included medical history, physical examination, trans-esophageal echocardiography, carotid ultrasonography, right and left heart catheterization, coronary angiography, aortography and iliac-femoral arteriography, and computed tomography angiography.

Procedural and prosthesis details

Details of the device and the technical aspects of the procedure have been previously described.⁸ Briefly, the device used was the 18-french third generation, CoreValve Revalving System which is a self-expandable biological valve. It was inserted using femoral, subclavian or transapical arterial approach. Circulatory support with the insertion of an intra-aortic balloon pump was used only in one case due to severe left ventricular dysfunction. All procedures were performed with surgical back-up and under deep sedation and local anesthesia. General anesthesia and mechanical ventilation were employed only when the subclavian or the transapical arterial approaches were used or when trans-esophageal echocardiography was deemed necessary. Vascular access was performed with the aid of an artery hemostatic device, the Prostar TM XL 10-F system (Abbott Vascular, Redwood City, CA, USA).

Prophylactic antibiotic therapy was administered 1 h prior to the procedure. Pre-procedural medication consisted of aspirin 80 mg and clopidogrel 75 mg once daily, which was also recommended for six months post-implantation followed by aspirin 80 mg indefinitely.

After the procedure, all patients were transferred to the ICU for the first 48 h. The clinical and physiological data for all scores were retrospectively collected according to the criteria and definitions described by the developers for the first 24 h post-TAVI. The variables included in these four scores are shown in Table 1.

Table 1.

A comparison of the variables included in different scoring systems post-TAVI.

Parameters	APACHE II	SAPS II	MODS	SOFA
Patient data
Age	⊗	⊗
Temperature	⊗	⊗
Cardiovascular system
Systolic arterial blood pressure	⊗	⊗	⊗	⊗
Mean arterial pressure	⊗		⊗	⊗
Central venous pressure
Heart rate	⊗	⊗	⊗
Catecholamines				⊗
Respiratory system
FiO2		⊗	⊗	⊗
PaO2	⊗	⊗	⊗	⊗
Arterial pH	⊗
Respiratory rate	⊗
Patient dependence on ventilator	⊗	⊗		⊗
Renal system
Urea		⊗
Creatinine	⊗		⊗	⊗
Urine output		⊗		⊗
Acute renal failure	⊗	⊗
Chemistry
Bilirubin	⊗	⊗	⊗	⊗
Bicarbonate	⊗	⊗
Sodium	⊗	⊗
Potassium	⊗	⊗
Hematology
Hematocrit	⊗
White blood cell	⊗	⊗
Platelets			⊗	⊗
Central nervous system
Glasgow coma score	⊗	⊗	⊗	⊗
Chronic disease (severe organ system insufficiency or immunocompromised)	⊗	⊗
ICU admission
Elective surgery	⊗	⊗
Internal disease	⊗	⊗

TAVI: transcatheter aortic valve implantation.

The primary outcome was in-hospital mortality, defined as any death occurring before hospital discharge. Morbidity was defined by the following points:

intra- or post-TAVI myocardial infarction, verified by positive creatine kinase-MB in combination with either typical electrocardiographic changes or echocardiographically detected new left ventricular wall motion abnormalities;

implantation of permanent pacemaker;

stroke (verified by neurological consult or computerized tomography-scan);

tamponade;

major bleeding;

vascular access site complications and

prolonged ventilation (>48 h).

Clinical and echocardiographic follow-up was continued for 30 days post TAVI. Thirty-day mortality included any death within 30 days from TAVI.

Statistical analysis

We used t-test for two independent samples or Mann–Whitney test to investigate differences in continuous variables between groups (in-hospital or 30-day mortality/morbidity). Categorical variables were tested using chi-square (χ²) or Fisher’s exact test. Spearman correlation coefficient was used to assess the association between continuous variables. Logistic regression was used to assess the effectiveness of each of the four general scoring systems on mortality/morbidity. The Hosmer & Lemeshow goodness-of-fit test was used for calibration of the models. Α low χ² value and p > 0.05 indicate good calibration. The area under the receiver operating characteristic (ROC) curve (AUC) was used to asses the discriminative ability (the ability of a scoring system to distinguish between mortality/morbidity or not) of the models. The overall correct classification (OCC; the ratio of correctly predicted mortality/morbidity or not to the total number of patients) was also calculated. All tests were two-sided in a 5% level of statistical significance.

Results

A total of 75 patients underwent TAVI and admitted to the ICU for further care. All TAVI procedures were performed electively and completed successfully in all cases. Transfemoral access was used in 59 (78.6%) cases, transsubclavian in 14 (18.6%) and transaortic in 2 (2.6%) cases. Aortic valvuloplasty was performed prior to the procedure in all patients. A 29 mm CoreValve prosthesis was implanted in 36 cases (48%) and a 26 mm prosthesis in 33 cases (44%). Small 23 mm CoreValve prosthesis and large 31 mm were rarely implanted in three cases each (4%). In five cases, a second valve had to be implanted with a good final result during the same procedure, owing to the fact that the first one was implanted too high. At admission and during the first day of ICU stay, 30 (40%) patients required mechanical ventilation, two (2.6%) received therapy with vasopressors, and two (2.6%) suffered from acute renal failure.

The median length of hospital stay was 10 [25–75% interquartile range (IQR) 8, 13] days and the length of ICU stay was four (3, 6) days. Before hospital discharge, four patients died (5.3%) in ICU. Causes of death were massive pulmonary embolism (1), respiratory insufficiency (1) and multi organ failure (2). During follow-up, there was one more sudden death, indicating a 30-day mortality rate of 6.6%. Patients’ characteristics and clinical outcomes including comparison of survivors and non-survivors are shown in Table 2. The majority of clinical outcomes occurred during hospitalization except from one case of femoral artery pseudoaneurysm, which was successfully thrombosed.

Table 2.

Patients’ characteristics and clinical outcomes including comparison of survivors and non-survivors.

	Total	Survivors	Non-survivors	P value
Patients
Age, median (IQR)	82.0 (79.0, 85.0)	82.0 (79.0, 85.0)	83.0 (79.0, 84.0)	0.983
Male, n (%)	36 (48.0)	33 (47.1)	3 (60.0)	0.666
Diabetes, n (%)	31 (41.3)	29 (41.4)	2 (40.0)	>0.999
Hypertension, n (%)	70 (93.3)	66 (94.3)	4 (80.0)	0.299
Dyslipidemia, n (%)	54 (73.0)	51 (73.9)	3 (60.0)	0.607
Log Euroscore %, median (IQR)	25.8 (19.7, 36.5)	25.3 (19.5, 34.3)	30.0 (25.8, 43)	0.158
Comorbidities, n (%)
Renal insufficiency	13 (17.6)	12 (17.4)	1 (20.0)	>0.999
COPD^,	26 (34.7)	22 (31.4)	4 (80.0)	0.046
Atrial fibrillation	23 (30.7)	21 (30.0)	2 (40.0)	0.639
Peripheral vascular disease	27 (36.0)	25 (35.7)	2 (40.0)	>0.999
History of cerebrovascular disease	6 (8.0)	6 (8.6)	0 (0.0)	>0.999
Previous myocardial infarction	11 (14.7)	11 (15.7)	0 (0.0)	>0.999
Previous PTCA/CABG	31 (41.3)	29 (41.4)	2 (40.0)	>0.999
Symptoms, n (%)
NYHA II	28 (38.4)	27 (39.7)	1 (20.0)	0.732
NYHA III	41 (56.2)	37 (54.4)	4 (80.0)
NYHA IV	4 (5.5)	4 (5.9)	0 (0.0)
Angina	21 (28.4)	20 (29.0)	1 (20.0)	>0.999
Syncope	8 (10.8)	8 (11.6)	0 (0.0)	>0.999
ICU stay, d, median (IQR)	4.0 (3.0, 6.0)	4.0 (3.0, 6.0)	8.0 (7.0, 20.0)	0.063
Hospital stay, d, median (IQR)	10.0 (8.0, 13.0)	10.0 (8.0, 12.0)	19.0 (7.0, 23.0)	0.467
Days of intubation, median (IQR)	0.0 (0.0, 1.0)	0.0 (0.0, 1.0)	1.0 (1.0, 12.0)	0.001
Number of blood units transfused, median (IQR)	2.0 (1.0, 3.0)	2.0 (1.0, 3.0)	2.0 (1.0, 7.0)	0.700
Scoring systems, median (IQR)
APACHE II	11.0 (8.0,13.0)	10.0 (8.0,12.0)	15.0 (14.0,19.0)	0.004
SAPS II	32.0 (30.0,39.0)	32.0 (30.0,36.0)	43.0 (41,55.0)	0.005
MODS	3.0 (2.0,4.0)	3.0 (1.5,4.0)	7.0 (4.0, 7.0)	0.028
SOFA	4.0 (3.0, 5.0)	3.0 (3.0,5.0)	7.0 (4.0,8.0)	0.026
Clinical outcomes, n (%)
Myocardial infarction	1 (1.3)	1 (1.4)	0 (0.0)	>0.999
Stroke	2 (2.7)	2 (2.9)	0 (0.0)	>0.999
Permanent pacemaker	19 (25.3)	17 (24.3)	2 (40.0)	0.596
Major bleeding	14 (18.9)	12 (17.4)	2 (40.0)	0.237
	14 (18.7)	12 (17.1)	2 (40.0)	0.232
Vascular complications Prolonged ventilation (>48 h)	4 (5.3)	2 (2.9)	2 (40.0)	0.020

COPD: chronic obstructive pulmonary disease; ICU: Intensive Care Unit; IQR: interquartile range; NYHA: New York Heart Association; PTCA/CABG: percutaneous transluminal coronary angioplasty/coronary artery bypass grafting.

The median scores (25–75% IQR) of each individual prediction system among all patients and between survivors and non-survivors are presented in Table 2. The median preoperative collected logistic Euroscore is 25.8 (19.7, 36.5). Table 3 shows the results for the calibration, discrimination and OCC of the four general risk scores from the first 24-h ICU stay and for Euroscore. There were no significant differences between expected and observed mortality using the Hosmer & Lemeshow test for all risk models and for all primary outcomes (p > 0.05). ROC curves were plotted separately for mortality and morbidity for each score (Figure 1). AUC for mortality prediction were greater than 0.70 for all scores and remarkably high for APACHE II and SAPS II (>0.85). AUCs for morbidity were considerably lower than those for mortality for all scores. Regarding in-hospital mortality, SAPS II exhibited the best discriminatory power [AUC (=0.92)] and the best calibration compared with the other scoring systems. APACHE II (AUC = 0.88) performed better than MODS (AUC = 0.73) and SOFA (AUC = 0.74) and had the highest OCC among all four scores (94.67%). Euroscore had worse discrimination and calibration than SAPS II and APACHE II but equally high OCC with APACHE II. Regarding 30-day mortality, APACHE II and SAPS II performed equally higher than the other two scores in terms of discrimination (AUC = 0.88). SAPS II again had the best calibration among all four scores. SOFA and MODS performed better in terms of OCC values when compared with the APACHE II and SAPS II scores. Euroscore had the worst discrimination, but its calibration was better than SAPS II. Figure 2 shows changes in risk scores over the course of the 6-year study period.

Table 3.

Predicted value of risk scores.

Primary outcomes	Scoring models	Logistic regression		Overall correct classification (OCC)	Hosmer & Lemeshow test		ROC analysis
		OR	95% CI	%	X2	P value	AUC	95% CI
In- hospital mortality	APACHE II	1.189	1.018–1.390	94.67	4.04	0.776	0.88	0.79–0.98
	SAPS II	1.107	1.022–1.199	94.29	3.06	0.801	0.92	0.84–1.00
	MODS	1.368	0.924–2.026	94.20	6.91	0.227	0.73	0.46–1.00
	SOFA	1.314	0.888–1.945	94.29	8.34	0.214	0.74	0.55–0.93
	EUROSCORE	1.079	1.000–1.164	94.67	5.62	0.689	0.77	0.55–0.99
In- hospital morbidity	APACHE II	1.326	1.114–1.577	66.67	6.50	0.482	0.74	0.62–0.85
	SAPS II	1.076	1.013–1.144	67.14	11.95	0.063	0.71	0.58–0.83
	MODS	1.253	0.994–1.579	59.42	10.28	0.068	0.62	0.48–0.76
	SOFA	1.290	1.031–1.612	65.71	5.84	0.441	0.67	0.54–0.80
	EUROSCORE	1.008	0.970–1.047	57.33	10.33	0.242	0.50	0.37–0.64
30-day mortality	APACHE II	1.178	1.019–1.362	92.00	6.51	0.482	0.88	0.79–0.96
	SAPS II	1.090	1.016–1.171	92.86	3.29	0.772	0.88	0.78–0.98
	MODS	1.657	1.115–2.462	94.20	4.59	0.469	0.79	0.56–1.00
	SOFA	1.597	1.078–2.368	94.29	7.14	0.308	0.80	0.62–0.98
	EUROSCORE	1.058	0.988–1.132	93.33	5.52	0.701	0.69	0.46–0.92
30 day-morbidity	APACHE II	1.099	0.971–1.243	88.00	6.68	0.462	0.70	0.51–0.89
	SAPS II	1.050	0.990–1.113	87.14	6.66	0.353	0.73	0.57–0.89
	MODS	1.261	0.945–1.682	86.96	8.62	0.125	0.62	0.42–0.82
	SOFA	1.260	0.949–1.672	87.14	4.04	0.671	0.65	0.47–0.83
	EUROSCORE	1.017	0.961–1.075	88.00	6.11	0.634	0.54	0.33–0.75

Figure 1.

Receiver operating characteristic curves of APACHE II, SAPS II, MODS and SOFA for in-hospital mortality (a), 30-day mortality (b), in-hospital morbidity (c) and 30-day morbidity (d).

Figure 2.

APACHE II, SAPS II, MODS and SOFA over the course of the 6-year study period.

APACHE II and SAPS II were significantly associated with ICU length of stay (r = 0.44, p < 0.001 and r = 0.36, p = 0.002, respectively), days of hospitalization (r = 0.27, p = 0.020 and r = 0.44, p < 0.001, respectively) and number of blood units transfused per patient (r = 0.44, p < 0.001 and r = 0.30, p = 0.012, respectively). MODS and SOFA were significantly associated with ICU length of stay (r = 0.27, p = 0.02 and r = 0.44, p < 0.001, respectively) and number of blood units transfused per patient (r = 0.28, p = 0.022 and r = 0.27, p = 0.022, respectively).

Discussion

The main findings of our study are: (1) SAPS II and APACHE II outperformed MODS and SOFA scores in terms of in-hospital and 30-day mortality. (2) The predicted value of all of the studied risk models for mortality were greater than that for morbidity. (3) Patients with high severity of illness, as assessed by SAPS II or APACHE II, have longer ICU and hospital stays as well as are more frequently transfused. To our knowledge, this report is the first to compare ICU scoring systems with regard to their validity in a cohort of TAVI patients.

SAPS II and APACHE II are two of the most frequently used general mortality prediction models, which have been extensively validated.^9–11 The accuracy of prognostic scoring systems depends on the assessment of both calibration and discrimination.^12,13 SAPS II and APACHE II demonstrated better calibration and highest discrimination in our study population compared to the other two scores, indicating their superior performance. Our study shows that scores using a large number of data inputs outperform simpler scores in TAVI population. Although greater score complexities may increase the barrier to calculation, as it increases the likehood some required variables may not be available, it seems that additional clinical information account for the better mortality prediction in our study population.

The good prognostic abilities of SAPS II and APACHE II in this study suggest that they could be used to prospectively identify sicker patients who may require more aggressive post procedural care and for whom increased resource utilization may be justified. SAPS II and APACHE II, which include not only risk factors but also information about the patient situation after the procedure, seem to perform better than Euroscore, a specific purpose risk model, which considers only patient preoperative conditions.

All studied prediction models in TAVI patients demonstrated a similar pattern for morbidity – worse discrimination and poorer calibration compared to mortality. Only SOFA showed better calibration for morbidity compared to mortality, as expected, since it was specifically designed for the prediction of morbidity in patients with sepsis.¹⁴ Mortality is the most frequently reported outcome parameter in evaluating risk scores due to objectivity in data acquisition, whereas the objective parameters for morbidity are harder to define.¹⁵ Candidates for TAVI have several comorbidities and are at greater risk for postoperative morbidity; therefore, prompt recognition of complications is critical to improve patient outcome. Several postoperative events, such as atrioventricular blocks, myocardial infarction, stroke or vascular complications, have significant impact not only on patients’ outcome after TAVI but on health care cost and quality of life as well.¹⁶ The heterogeneity of morbidity parameters, like the need for mechanical device or reoperation for bleeding, along with the initial lack of standardized definitions makes it difficult to find common risk factors for a statistically sound prediction of overall morbidity.¹⁷ However, risk stratification using SAPS II and APACHE II and for at least certain morbidity events appear to be acceptable according to our findings.

The length of ICU stay is commonly used as measure of cost and for assessing adequate unit efficiency, despite variations which may be attributed to patient and selected institutional characteristics.¹⁸ It is shown that all studied risk scores could be reliable in predicting ICU length of stay in TAVI population, likely due to the standardized therapeutic management during their ICU stay. SAPS II and APACHE II could also identify patients at risk for a prolonged hospital stay post-TAVI. It seems that these two scores are better indicators of the severity of complications arising during prolonged ICU stay compared to MODS and SOFA. Furthermore, these risk scores can be used so that patients at risk are not discharged too early potentially leading to ICU readmission and/or prolonged hospital stay, both of which are associated with higher mortality rates.

Blood transfusions are frequently required after TAVI even in the absence of overt bleeding or vascular complications and have been significantly associated with impaired short- and mid-term outcome.^19,20 Based on our results, total transfusions indicate TAVI patients of higher risk and worse outcome as reflected by their correlation with the studied severity-of-illness scores evaluated over the first 24 h of ICU admission. Since multiple blood transfusions are associated with considerable risk, including immune suppression and increased risk of nosocomial infections, therapeutic interventions and prevention methods to reduce transfusions will potentially benefit post TAVI patient care.

It is important to recognize inherent limitations, which exist in the application of prognostic scoring systems to individual patients. Our study is also limited due to its retrospective nature, small sample size, long data collection period and single institution design. It seems likely that TAVI complications, management and approach to ICU care – which may be important determinants of mortality – have changed during the observation time period, although no significant change was observed over the course of time. Similarly, it is probable that patient selection and the perioperative handling could vary between different ICUs. Further prospective studies with a larger sample size and comparison among medical centers are needed to estimate the prognostic ICU performance of risk scoring in TAVI patients.

Conclusions

As the actual TAVI procedure becomes increasingly streamlined in high-risk elderly patients, it stands to reason that patients’ comorbidities and acute illness following TAVI will be the most important predictors of morbidity and mortality. Therefore, accurate clinical risk predictors will likely have an important role in assessing outcomes and quality. Although accurate risk predictors will probably rise out of the TAVI registries, it is also important to assess the performance of current models as they pertain to TAVI. SAPS II and APACHE II seem to be the most appropriate general ICU scoring systems for immediate and short-term mortality prediction in TAVI population allowing for earlier more aggressive therapeutic interventions.

Footnotes

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

References

Leon

Smith

Mack

. Transcatheter aortic-valve implantation for aortic stenosis in patients who cannot undergo surgery. N Engl J Med 2010; 363: 1597–1607.

Kapadia

Tuzcu

Makkar

. Long-term outcomes of inoperable patients with aortic stenosis randomly assigned to transcatheter aortic valve replacement or standard therapy. Circulation 2014; 130: 1483–1492.

Georgiadou

Kontodima

Sbarouni

. Long-term quality of life improvement after transcatheter aortic valve implantation. Am Heart J 2011; 162: 232–237.

Raiten

Gutsche

Horak

. Critical care management of patients following transcatheter aortic valve replacement. F1000Res 2013; 2: 62.

Walther

Hamm

Schuler

. Perioperative results and complications in 15,964 transcatheter aortic valve replacements: prospective data from the GARY registry. J Am Coll Cardiol 2015; 65: 2173–2180.

Reynolds

Magnuson

Wang

. Cost-effectiveness of transcatheter aortic valve replacement compared with standard care among inoperable patients with severe aortic stenosis. Circulation 2012; 125: 1102–1109.

Iannaccone

Marwick

. Cost effectiveness of transcatheter aortic valve replacement compared with medical management or surgery for patients with aortic stenosis. Appl Health Econ Health Policy 2015; 13: 29–45.

Grube

Schuler

Buellesfeld

. Percutaneous aortic valve replacement for severe aortic stenosis in high-risk patients using the second- and current third-generation self-expanding CoreValve prosthesis: device success and 30-day clinical outcome. J Am Coll Cardiol 2007; 50: 69–76.

Le Gall

J-R

Lemeshow

Saulnier

. A new simplified acute physiology score (SAPS II) based on a European/North American multicenter study. JAMA 1993; 270: 2957–2963.

10.

Knaus

Draper

Wagner

. APACHE II: a severity of disease classification system. Crit Care Med 1985; 13: 818–829.

11.

Martínez-Alario

Tuesta

Plasencia

. Mortality prediction in cardiac surgery patients: comparative performance of Parsonnet and general severity systems. Circulation 1999; 99: 2378–2382.

12.

Predicting outcome in ICU patients. 2nd European Consensus Conference in Intensive Care Medicine. Intensive Care Med 1994; 20: 390–397.

13.

Cook

. Use and misuse of the receiver operating characteristic curve in risk prediction. Circulation 2007; 115: 928–935.

14.

Vincent

Moreno

Takala

. The SOFA (Sepsisrelated Organ Failure Assessment) score to describe organ dysfunction/failure. On behalf of the Working Group on Sepsis-Related Problems of the European Society of Intensive Care Medicine. Intensive Care Med 1996; 22: 707–710.

15.

Geissler

Hölzl

Marohl

. Risk stratification in heart surgery: comparison of six score systems. Eur J Cardiothorac Surg 2000; 17: 400–406.

16.

Minutello

Wong

Swaminathan

. Costs and in-hospital outcomes of transcatheter aortic valve implantation versus surgical aortic valve replacement in commercial cases using a Propensity Score Matched Model. Am J Cardiol 2015; 115: 1443–1447.

17.

Généreux

Cohen

Mack

. Incidence, predictors, and prognostic impact of late bleeding complications after transcatheter aortic valve replacement. J Am Coll Cardiol 2014; 64: 2605–2615.

18.

Knaus

Wagner

Zimmerman

. Variations in mortality and length of stay in intensive care units. Ann Intern Med 1993; 118: 753–761.

19.

Seiffert

Conradi

Terstesse

. Blood transfusion is associated with impaired outcome after transcatheter aortic valve implantation. Catheter Cardiovasc Interv 2015; 85: 460–467.

20.

Nuis

Sinning

Rodés-Cabau

. Prevalence, factors associated with, and prognostic effects of preoperative anemia on short- and long-term mortality in patients undergoing transcatheter aortic valve implantation. Circ Cardiovasc Interv 2013; 6: 625–634.

Prognostic performance of critical care scores in patients undergoing transcatheter aortic valve implantation

Abstract

Background

Aims

Methods

Results

Conclusions

Keywords

Introduction

Methods

Procedural and prosthesis details

Statistical analysis

Results

Discussion

Conclusions

Footnotes

Declaration of conflicting interests

Funding

References