The comparative and added prognostic value of biomarkers to the Revised Cardiac Risk Index for preoperative prediction of major adverse cardiac events and all‐cause mortality in patients who undergo noncardiac surgery

Lisette M Vernooij; Wilton A Klei; Karel GM Moons; Toshihiko Takada; Judith Waes; Johanna AAG Damen

doi:10.1002/14651858.CD013139.pub2

. 2021 Dec 21;2021(12):CD013139. doi: 10.1002/14651858.CD013139.pub2

The comparative and added prognostic value of biomarkers to the Revised Cardiac Risk Index for preoperative prediction of major adverse cardiac events and all‐cause mortality in patients who undergo noncardiac surgery

Lisette M Vernooij ^1,^2,^✉, Wilton A Klei ^2,³, Karel GM Moons ^1,⁴, Toshihiko Takada ¹, Judith Waes ², Johanna AAG Damen ^1,⁴

Editor: Cochrane Heart Group

PMCID: PMC8689147 PMID: 34931303

Abstract

Background

The Revised Cardiac Risk Index (RCRI) is a widely acknowledged prognostic model to estimate preoperatively the probability of developing in‐hospital major adverse cardiac events (MACE) in patients undergoing noncardiac surgery. However, the RCRI does not always make accurate predictions, so various studies have investigated whether biomarkers added to or compared with the RCRI could improve this.

Objectives

Primary: To investigate the added predictive value of biomarkers to the RCRI to preoperatively predict in‐hospital MACE and other adverse outcomes in patients undergoing noncardiac surgery.

Secondary: To investigate the prognostic value of biomarkers compared to the RCRI to preoperatively predict in‐hospital MACE and other adverse outcomes in patients undergoing noncardiac surgery.

Tertiary: To investigate the prognostic value of other prediction models compared to the RCRI to preoperatively predict in‐hospital MACE and other adverse outcomes in patients undergoing noncardiac surgery.

Search methods

We searched MEDLINE and Embase from 1 January 1999 (the year that the RCRI was published) until 25 June 2020. We also searched ISI Web of Science and SCOPUS for articles referring to the original RCRI development study in that period.

Selection criteria

We included studies among adults who underwent noncardiac surgery, reporting on (external) validation of the RCRI and:

‐ the addition of biomarker(s) to the RCRI; or

‐ the comparison of the predictive accuracy of biomarker(s) to the RCRI; or

‐ the comparison of the predictive accuracy of the RCRI to other models.

Besides MACE, all other adverse outcomes were considered for inclusion.

Data collection and analysis

We developed a data extraction form based on the CHARMS checklist. Independent pairs of authors screened references, extracted data and assessed risk of bias and concerns regarding applicability according to PROBAST. For biomarkers and prediction models that were added or compared to the RCRI in ≥ 3 different articles, we described study characteristics and findings in further detail. We did not apply GRADE as no guidance is available for prognostic model reviews.

Main results

We screened 3960 records and included 107 articles.

Over all objectives we rated risk of bias as high in ≥ 1 domain in 90% of included studies, particularly in the analysis domain. Statistical pooling or meta‐analysis of reported results was impossible due to heterogeneity in various aspects: outcomes used, scale by which the biomarker was added/compared to the RCRI, prediction horizons and studied populations.

Added predictive value of biomarkers to the RCRI

Fifty‐one studies reported on the added value of biomarkers to the RCRI. Sixty‐nine different predictors were identified derived from blood (29%), imaging (33%) or other sources (38%). Addition of NT‐proBNP, troponin or their combination improved the RCRI for predicting MACE (median delta c‐statistics: 0.08, 0.14 and 0.12 for NT‐proBNP, troponin and their combination, respectively). The median total net reclassification index (NRI) was 0.16 and 0.74 after addition of troponin and NT‐proBNP to the RCRI, respectively. Calibration was not reported. To predict myocardial infarction, the median delta c‐statistic when NT‐proBNP was added to the RCRI was 0.09, and 0.06 for prediction of all‐cause mortality and MACE combined. For BNP and copeptin, data were not sufficient to provide results on their added predictive performance, for any of the outcomes.

Comparison of the predictive value of biomarkers to the RCRI

Fifty‐one studies assessed the predictive performance of biomarkers alone compared to the RCRI. We identified 60 unique predictors derived from blood (38%), imaging (30%) or other sources, such as the American Society of Anesthesiologists (ASA) classification (32%). Predictions were similar between the ASA classification and the RCRI for all studied outcomes. In studies different from those identified in objective 1, the median delta c‐statistic was 0.15 and 0.12 in favour of BNP and NT‐proBNP alone, respectively, when compared to the RCRI, for the prediction of MACE. For C‐reactive protein, the predictive performance was similar to the RCRI. For other biomarkers and outcomes, data were insufficient to provide summary results. One study reported on calibration and none on reclassification.

Comparison of the predictive value of other prognostic models to the RCRI

Fifty‐two articles compared the predictive ability of the RCRI to other prognostic models. Of these, 42% developed a new prediction model, 22% updated the RCRI, or another prediction model, and 37% validated an existing prediction model. None of the other prediction models showed better performance in predicting MACE than the RCRI. To predict myocardial infarction and cardiac arrest, ACS‐NSQIP‐MICA had a higher median delta c‐statistic of 0.11 compared to the RCRI. To predict all‐cause mortality, the median delta c‐statistic was 0.15 higher in favour of ACS‐NSQIP‐SRS compared to the RCRI. Predictive performance was not better for CHADS₂, CHA₂DS₂‐VASc, R₂CHADS₂, Goldman index, Detsky index or VSG‐CRI compared to the RCRI for any of the outcomes. Calibration and reclassification were reported in only one and three studies, respectively.

Authors' conclusions

Studies included in this review suggest that the predictive performance of the RCRI in predicting MACE is improved when NT‐proBNP, troponin or their combination are added. Other studies indicate that BNP and NT‐proBNP, when used in isolation, may even have a higher discriminative performance than the RCRI. There was insufficient evidence of a difference between the predictive accuracy of the RCRI and other prediction models in predicting MACE. However, ACS‐NSQIP‐MICA and ACS‐NSQIP‐SRS outperformed the RCRI in predicting myocardial infarction and cardiac arrest combined, and all‐cause mortality, respectively. Nevertheless, the results cannot be interpreted as conclusive due to high risks of bias in a majority of papers, and pooling was impossible due to heterogeneity in outcomes, prediction horizons, biomarkers and studied populations.

Future research on the added prognostic value of biomarkers to existing prediction models should focus on biomarkers with good predictive accuracy in other settings (e.g. diagnosis of myocardial infarction) and identification of biomarkers from omics data. They should be compared to novel biomarkers with so far insufficient evidence compared to established ones, including NT‐proBNP or troponins. Adherence to recent guidance for prediction model studies (e.g. TRIPOD; PROBAST) and use of standardised outcome definitions in primary studies is highly recommended to facilitate systematic review and meta‐analyses in the future.

Keywords: Adult, Humans, Bias, Biomarkers, Heart Arrest, Myocardial Infarction, Peptide Fragments, Predictive Value of Tests, Prognosis, Risk Assessment

Plain language summary

Can biomarkers improve predictions of the RCRI tool to predict heart‐related complications in patients undergoing surgery other than heart surgery?

Background and review question

Although patients undergo surgery to maintain or increase life expectancy or to improve quality of life, surgery is not without risks. Some patients will develop a heart‐related complication after surgery other than heart surgery, such as a heart infarction. Several tools try to predict someone's chance of developing a heart complication after surgery using information collected in the period before surgery. The Revised Cardiac Risk Index (RCRI) is such a tool that tries to estimate the chance of developing a heart complication during hospital admission in patients undergoing surgery other than heart surgery. It uses information on whether the patient has in the past experienced a heart infarction, heart failure and/or a stroke during his/her life, their use of insulin for the treatment of diabetes mellitus, their current renal (kidney) function and whether he/she will undergo high or non‐high risk surgery. The RCRI is commonly used by physicians, but the predictions are not always very accurate. Therefore, several researchers have attempted to improve these predictions by adding extra information to this tool. This information can be derived from so‐called biomarkers, which are, for example, measurements from blood, imaging techniques or other characteristics, such as age, smoking status or physical condition of the patient.

The aim of this systematic review was to investigate whether the addition of such biomarkers to the RCRI improves predictions of heart‐related complications during hospitalisation in patients undergoing surgery other than heart surgery. In addition, we investigated whether biomarkers and other prediction tools resulted in better predictions of heart‐related complications during hospitalisation compared to the predictions of the RCRI in patients undergoing surgery other than heart surgery.

Key results

We identified 69 different predictors that were added to the RCRI tool to improve predictions of these heart‐related complications. The evidence is current to 25 June 2020. Predictions seem to improve with the addition of some biomarkers derived from blood. These are troponin (which measures muscular damage of the heart), brain natriuretic peptide (BNP) and (NT‐pro)brain natriuretic peptide (NT‐proBNP) (which both measure severity of heart failure).

In addition, there were 60 biomarkers that were studied to compare their predictions to the RCRI. Other studies included in this review suggest that BNP and NT‐proBNP alone may predict heart‐related complications even better than the RCRI. Sixty‐five prediction tools other than the RCRI tried to improve its predictions. The American College of Surgeons National Surgical Quality Improvement (ACS‐NSQIP) and ACS‐NSQIP‐MICA (myocardial infarction or cardiac arrest) surgical risk score tools could make better predictions than the RCRI, but this was only true for certain outcomes, and not for heart‐related complications. However, for all of these research questions, we are not confident in the results due to large variation in the research methods applied and signs of less accurate research approaches having been used.

Authors' conclusions

Troponin, BNP and NT‐proBNP may improve the ability of the RCRI to predict heart‐related complications. The ACS‐NSQOP‐MICA and ACS‐NSQIP surgical risk score tools seem to be better at predicting postoperative complications than the RCRI tool, but not heart‐related complications. However, due to deficiencies in how the studies were conducted, we are uncertain whether the results we found apply to all patients undergoing surgery other than heart surgery. We need more and better research on biomarkers with promising predictive performance in other settings.

Summary of findings

Summary of findings 1. Summary of findings ‐ objective 1: added value of biomarkers to the RCRI.

Population: patients undergoing noncardiac surgery Index model: Revised Cardiac Risk Index (RCRI) Comparator: RCRI extended with biomarker(s) Outcome: postoperative occurrence of (in‐hospital) major adverse cardiac events (MACE), all‐cause mortality and other adverse outcomes Timing: time point of prognostication: before surgery; prediction horizon: in‐hospital, but all time spans are included Setting: to inform physicians of the patient's risk of developing in‐hospital events after noncardiac surgery
Outcomes	Biomarker	№ of participants (studies)	Measure	Pooled result		Comments
Outcomes	Biomarker	№ of participants (studies)	Measure	Summary measure	Median (range)	Comments
MACE	Troponin	3 studies 810 patients 77 MACE	Discrimination	Delta c‐statistic	0.14 (0.06 to 0.33)	Surgical specialty was vascular and noncardiac surgery. Prediction horizon was 30‐day MACE.
		0 studies	Calibration	—	—	—
		2 studies 577 patients 70 MACE	Reclassification	NRI	0.16 (0.09 to 0.22)	Surgical specialty was vascular surgery. Prediction horizon was 30‐day MACE and long‐term MACE (> 30 days).
		1 study 122 patients 29 MACE	—	IDI	0.05	Surgical specialty was vascular surgery. Prediction horizon was long‐term MACE (> 30 days).
	NT‐proBNP	7 studies 13,687 patients 1710 MACE	Discrimination	Delta c‐statistic	0.08 (0.04 to 0.22)	Surgical specialty was vascular and noncardiac surgery. Prediction horizon was 30‐day MACE.
		1 study 10,402 patients 1269 MACE	Calibration	Calibration plot	Good calibration	Surgical specialty was noncardiac surgery. Prediction horizon was 30‐day MACE.
		2 studies 10,524 patients 1560 MACE	Reclassification	NRI	0.74 (0.26 to 1.22)	Surgical specialty was noncardiac and vascular surgery. Prediction horizon was 30‐day MACE and long‐term MACE (> 30 days).
		1 study 122 patients 29 MACE	—	IDI	0.23	Surgical specialty was vascular surgery. Prediction horizon was long‐term MACE (> 30 days).
	Troponin + NT‐proBNP	3 studies 575 patients 120 MACE	Discrimination	Delta c‐statistic	0.12 (0.1 to 0.34)	Surgical specialty was vascular and noncardiac surgery. Prediction horizon was 30‐day MACE.
		0 studies	Calibration	—	—	—
		0 studies	Reclassification	—	—	—
	BNP	0 studies	Discrimination	—	—	—
		0 studies	Calibration	—	—	—
		2 studies 874 patients unknown MACE	Reclassification	NRI	0.72 (0.47 to 0.96)	Results are based on two studies as one study did not report the total NRI. Surgical specialty was orthopaedic and vascular surgery. Prediction horizon was 30‐day MACE. For one study, the number of outcomes was not reported.
All‐cause mortality and MACE	NT‐proBNP	3 study 12,214 patients 548 events	Discrimination	Delta c‐statistic	0.06 (0.06 to 0.07)	Surgical specialty was vascular and noncardiac surgery. Prediction horizon was 30‐day events.
		1 study 411 patients 74 events	Calibration	Hosmer Lemeshow	P = 0.03	Surgical specialty was vascular surgery. Prediction horizon was 30‐day events.
		2 study 1812 patients 102 events	Reclassification	NRI	0.19 (0.13 to 0.25)	Surgical specialty was vascular and noncardiac surgery. Prediction horizon was 30‐day events.
		1 study 411 patients 74 events	—	IDI	0.06	Surgical specialty was vascular surgery. Prediction horizon was 30‐day events.
Myocardial infarction	NT‐proBNP	2 studies 2626 patients 132 MI	Discrimination	Delta c‐statistic	0.09 (0.06 to 0.11)	Surgical specialty was noncardiac surgery. Prediction horizon was within 3 days after surgery and in‐hospital events.
		0 studies	Calibration	—	—	—
		1 study 572 patients 30 MI	Reclassification	NRI	0.46	Surgical specialty was noncardiac surgery. Prediction horizon was within 3 days after surgery.

Population: patients undergoing noncardiac surgery Index model: Revised Cardiac Risk Index (RCRI) Comparator: other prediction models Outcome: postoperative occurrence of (in‐hospital) major adverse cardiac events (MACE), all‐cause mortality and other adverse outcomes Timing: time point of prognostication: before surgery; prediction horizon: in‐hospital, but all time spans are included Setting: to inform physicians of the patient's risk of developing in‐hospital events after noncardiac surgery
Outcomes	Prediction model	№ of participants (studies)	Measure	Pooled result		Comments
Outcomes	Prediction model	№ of participants (studies)	Measure	Summary measure	Median (range)	Comments
MACE	ACS‐NSQIP‐MICA	3 studies 1567 patients 95 MACE	Discrimination	Delta c‐statistic	0.00 (‐0.09 to 0.04)	Surgical specialty was neurosurgery, vascular and noncardiac surgery. Prediction horizon was in‐hospital or 30‐day MACE. The prediction horizon was not reported in one study.
		1 study 870 patients 76 MACE	Calibration	Calibration plot	Poor calibration	Poor calibration for both RCRI and NSQIP MACE. Calibration improved after recalibration of NSQIP MACE. Surgical specialty was noncardiac surgery.
				Calibration intercept	0.95 for RCRI and 2.37 for NSQIP‐MICA	—
				Calibration slope	0.29 for RCRI and 0.70 for NSQIP‐MICA	—
		0 studies	Reclassification	—	—	—
	ACS‐NSQIP‐SRS	2 studies 1087 patients 26 MACE	Discrimination	Delta c‐statistic	0.06 (0.00 to 0.11)	Surgical specialty was noncardiac surgery. Prediction horizon was in‐hospital or 30‐day MACE.
		0 studies	Calibration	—	—	—
		0 studies	Reclassification	—	—	—
	Detsky	3 studies 3361 patients 191 MACE	Discrimination	Delta c‐statistic	0.05 (‐0.07 to 0.11)	Surgical specialty was orthopaedic, vascular and noncardiac surgery. Prediction horizon was in‐hospital or 30‐day MACE.
		0 studies	Calibration	—	—	—
		0 studies	Reclassification	—	—	—
	Goldman	3 studies 3361 patients 191 MACE	Discrimination	Delta c‐statistic	‐0.03 (‐0.07 to 0.08)	Surgical specialty was orthopaedic, vascular and noncardiac surgery. Prediction horizon was in‐hospital or 30‐day MACE.
		0 studies	Calibration	—	—	—
		0 studies	Reclassification	—	—	—
	VSG‐CRI	3 studies 2023 patients 208 MACE	Discrimination	Delta c‐statistic	0.03 (0.00 to 0.05)	Surgical specialty was vascular surgery. Prediction horizon was in‐hospital MACE. In one study, the prediction horizon was not reported.
		0 studies	Calibration	—	—	—
		0 studies	Reclassification	—	—	—
Myocardial infarction or cardiac arrest	ACS‐NSQIP‐MICA	6 studies 243,896 patients unknown MICA	Discrimination	Delta c‐statistic	0.11 (‐0.05 to 0.39)	Surgical specialty was general, vascular, orthopaedic and noncardiac surgery. Prediction horizon was 30‐day MICA. The prediction horizon was not reported in one study.
		2 studies 181,920 patients 1889 MICA	Calibration	Calibration plot	Poor calibration	Calibration was poor for both scores, however calibration was better for the RCRI compared to the NSQIP‐MICA. Calibration improved after recalibration of NSQIP‐MICA. Surgical specialty was noncardiac surgery. Prediction horizon was 30‐day MICA, but was not reported in one study.
		2 studies 43,047 patients 463 MICA	—	Hosmer Lemeshow	RCRI: P = 0.018 to P < 0.001 ACS‐NSQIP‐MICA P < 0.001	Surgical specialty was general and noncardiac surgery. Prediction horizon was 30‐day MICA, but was not reported in one study.
		0 studies	Reclassification	—	—	—
	ACS‐NSQIP‐SRS	2 studies 9678 patients 94 MICA	Discrimination	Delta c‐statistic	0.18 (0.13 to 0.22)	Surgical specialty was noncardiac surgery or not specified. Prediction horizon was 30‐day MICA. The prediction horizon was not reported in one study.
		1 study 9015 patients 91 MICA	Calibration	Calibration plot	RCRI: poor calibration, ACS‐NSQIP‐SRS: acceptable calibration	Surgical specialty was noncardiac surgery. Prediction horizon was not reported.
		1 study 9015 patients 91 MICA		Hosmer Lemeshow	RCRI: P < 0.001 ACS‐NSQIP‐SRS P = 0.07	Surgical specialty was noncardiac surgery. Prediction horizon was not reported.
		0 studies	Reclassification	—	—	—
All‐cause mortality	ACS‐NSQIP‐SRS	3 studies 2461 patients 155 deaths	Discrimination	Delta c‐statistic	0.15 (0.12 to 0.47)	Surgical specialty was neurosurgery or noncardiac surgery. The prediction horizon was in‐hospital or 30‐day events. In one study the prediction horizon was not reported.
		0 studies	Calibration	—	—	—
		0 studies	Reclassification	—	—	—
	CHADS₂	3 studies 35129 patients 1177 deaths	Discrimination	Delta c‐statistic	0.00 (‐0.02 to 0.01)	Surgical specialty was noncardiac surgery. The prediction horizon was 30‐day events.
		0 studies	Calibration	—	—	—
		3 studies 35129 patients 1177 deaths	Reclassification	NRI	0.07 (0.01 to 0.12)	Surgical specialty was noncardiac surgery. The prediction horizon was 30‐day events.
	CHADS₂VASc	2 studies 2969 patients 121 deaths	Discrimination	Delta c‐statistic	0.00 (‐0.02 to 0.02)	Surgical specialty was noncardiac surgery. The prediction horizon was 30‐day events.
		0 studies	Calibration	—	—	—
		2 studies 2969 patients 121 deaths	Reclassification	NRI	0.09 (0.01 to 0.17)	Surgical specialty was noncardiac surgery. The prediction horizon was 30‐day events.
	R₂CHADS₂	3 studies 35129 patients 1177 deaths	Discrimination	Delta c‐statistic	‐0.03 (‐0.03 to 0.03)	Surgical specialty was noncardiac surgery. The prediction horizon was 30‐day events.
		0 studies	Calibration	—	—	—
		3 studies 35129 patients 1177 deaths	Reclassification	NRI	0.03 (‐0.09 to 0.13)	Surgical specialty was noncardiac surgery. The prediction horizon was 30‐day events.
Stroke	CHADS₂	4 studies unknown patients unknown events	Discrimination	Delta c‐statistic	0.02 (‐0.01 to 0.11)	Surgical specialty was noncardiac surgery. The prediction horizon was 30‐day events. For one study the number of included patients and number of events were not reported.
		0 studies	Calibration	—	—	—
		2 studies 33121 patients 391 events	Reclassification	NRI	0.05 (‐0.06 to 0.17)	Surgical specialty was noncardiac surgery. The prediction horizon was 30‐day events.
	CHADS₂VASc	3 studies unknown patients unknown events	Discrimination	Delta c‐statistic	0.04 (0.00 to 0.12)	Surgical specialty was noncardiac surgery. The prediction horizon was 30‐day events. For one study the number of included patients and number of events were not reported.
		0 studies	Calibration	—	—	—
		1 studies 961 patients 11 events	Reclassification	NRI	0.07	Surgical specialty was noncardiac surgery. The prediction horizon was 30‐day events.
	R₂CHADS₂	3 studies unknown patients unknown events	Discrimination	Delta c‐statistic	0.05 (0.01 to 0.12)	Surgical specialty was noncardiac surgery. The prediction horizon was 30‐day events. For one study the number of included patients and number of events were not reported.
		0 studies	Calibration	—	—	—
		2 studies 33,121 patients 391 events	Reclassification	NRI	‐0.06 (‐0.14 to 0.01)	Surgical specialty was noncardiac surgery. The prediction horizon was 30‐day events.
All‐cause mortality or stroke	CHADS₂	3 studies 33,748 patients unknown events	Discrimination	Delta c‐statistic	0.03 (0.02 to 0.07)	Surgical specialty was noncardiac surgery. The prediction horizon was 30‐day events. For one study, the number of outcomes was not reported.
		0 studies	Calibration	—	—	—
		3 studies 33,748 patients unknown events	Reclassification	NRI	0.31 (0.14 to 0.35)	Surgical specialty was noncardiac surgery. The prediction horizon was 30‐day events. For one study, the number of outcomes was not reported.
	CHADS₂VASc	2 studies 1588 patients unknown events	Discrimination	Delta c‐statistic	0.04 (0.01 to 0.07)	Surgical specialty was noncardiac surgery. The prediction horizon was 30‐day events. For one study, the number of outcomes was not reported.
		0 studies	Calibration	—	—	—
		2 studies 1588 patients unknown events	Reclassification	NRI	0.30 (0.24 to 0.36)	Surgical specialty was noncardiac surgery. The prediction horizon was 30‐day events. For one study, the number of outcomes was not reported.
	R₂CHADS₂	3 studies 33,748 patients unknown events	Discrimination	Delta c‐statistic	0.03 (0.01 to 0.06)	Surgical specialty was noncardiac surgery. The prediction horizon was 30‐day events. For one study, the number of outcomes was not reported.
		0 studies	Calibration	—	—	—
		3 studies 33,748 patients unknown events	Reclassification	NRI	0.17 (0.11 to 0.44)	Surgical specialty was noncardiac surgery. The prediction horizon was 30‐day events. For one study, the number of outcomes was not reported.

Predictor	Definition	Point distribution
High‐risk surgery	Intraperitoneal, intrathoracic, or suprainguinal vascular surgery	1
Ischaemic heart disease	History of myocardial infarction, positive exercise test, current complaint of ischaemic chest pain or use of nitrate therapy, or ECG with Q waves. Patients with prior CABG surgery or PTCA were included in this definition only if they had current complaints of chest pain that were presumed to be due to ischaemia.	1
History of congestive heart failure	History of congestive heart failure, pulmonary oedema, or paroxysmal nocturnal dyspnoea, physical examination showing bilateral rales or S3 gallop, or chest radiograph showing pulmonary vascular redistribution.	1
History of cerebrovascular disease	History of transient ischaemic attack or stroke.	1
Insulin therapy for diabetes mellitus	—	1
Preoperative serum creatinine > 2.0 mg/dL	—	1

Population	Patients undergoing noncardiac surgery
Index Model	Revised Cardiac Risk Index (RCRI)
Comparator	Biomarker(s) added or compared to the RCRI; other prediction models compared to the RCRI
Outcome(s)	Postoperative occurrence of (in‐hospital) major adverse cardiac events, all‐cause mortality and other adverse outcomes
Timing	Time point of prognostication: before surgery Prediction horizon: in‐hospital, but all time spans are included
Setting	To inform physicians of the patient's risk of developing in‐hospital events after noncardiac surgery

	All validations	Added value of biomarkers	Comparison of biomarkers	Comparison of prediction models
N	172	62	89	79
Geographical area (%)
Europe	51 (29.8)	22 (35.5)	24 (27.3)	28 (35.9)
North America	63 (36.8)	12 (19.4)	42 (47.7)	27 (34.6)
Asia	20 (11.7)	14 (22.6)	10 (11.4)	3 (3.8)
Africa	2 (1.2)	1 (1.6)	1 (1.1)	1 (1.3)
Australia	1 (0.6)	0 (0.0)	1 (1.1)	0 (0.0)
South America	5 (2.9)	1 (1.6)	1 (1.1)	4 (5.1)
Combination	29 (17.0)	12 (19.4)	9 (10.2)	15 (19.2)
Data collection (%)
Prospective	124 (72.5)	44 (71.0)	66 (74.2)	25 (32.1)
Retrospective	41 (24.0)	15 (24.2)	18 (20.2)	54 (68.4)
Unclear	6 (3.5)	3 (4.8)	5 (5.6)	0 (0.0)
Study design (%)
Cohort	130 (75.6)	57 (91.9)	57 (64.0)	68 (86.1)
Existing registry	35 (20.3)	2 (3.2)	26 (29.2)	9 (11.4)
Case‐control	1 (0.6)	0 (0.0)	0 (0.0)	1 (1.3)
Existing RCT	1 (0.6)	1 (1.6)	1 (1.1)	1 (1.3)
Individual patient data meta‐analysis	5 (2.9)	2 (3.2)	5 (5.6)	0 (0.0)
Surgical specialty (%)
Noncardiac	77 (44.8)	36 (58.1)	30 (33.7)	37 (46.8)
Vascular	47 (27.2)	19 (30.2)	23 (25.6)	25 (31.6)
ENT and dental	2 (1.2)	1 (1.6)	1 (1.1)	1 (1.3)
General	5 (2.9)	0 (0.0)	2 (2.2)	4 (5.1)
Neurological	25 (14.5)	0 (0.0)	24 (26.7)	1 (1.3)
Orthopaedic	8 (4.6)	3 (4.8)	5 (5.6)	5 (6.3)
Other	5 (2.9)	1 (1.6)	2 (2.2)	3 (3.8)
Not specified	3 (1.7)	2 (3.2)	2 (2.2)	3 (3.8)
Prediction horizon (%)
Intraoperative events	1 (0.6)	0 (0.0)	1 (1.1)	1 (1.3)
1 to 7 days	7 (4.1)	6 (9.7)	7 (7.9)	1 (1.3)
In‐hospital events	25 (14.5)	12 (19.4)	13 (14.6)	14 (17.7)
In‐hospital or within 30 days	10 (5.8)	8 (12.9)	2 (2.2)	2 (2.5)
30‐day events	109 (63.4)	29 (46.8)	59 (66.3)	52 (65.8)
> 30 days (long‐term)	12 (7.0)	6 (9.7)	5 (5.6)	4 (5.1)
Not reported	8 (4.6)	1 (1.6)	2 (2.2)	5 (6.3)
Outcome (%)
MACE	70 (40.7)	31 (50.0)	35 (39.3)	32 (40.5)
MICA	8 (4.7)	2 (3.2)	0 (0.0)	7 (8.9)
Myocardial infarction	5 (2.9)	3 (4.8)	3 (3.4)	0 (0.0)
Cardiovascular mortality	6 (3.5)	3 (4.8)	1 (1.1)	2 (2.5)
Troponin elevation/myocardial injury	6 (3.5)	5 (8.1)	4 (4.5)	3 (3.8)
All‐cause mortality	22 (12.8)	6 (9.7)	10 (11.2)	13 (16.5)
All‐cause mortality and MACE	15 (8.7)	8 (12.9)	7 (7.9)	6 (7.6)
Other	40 (23.3)	4 (6.5)	29 (32.6)	16 (20.3)
Number of participants (median (IQR))	922 (244 to 9267)	442 (223 to 1389)	594 (227 to 52066)	941 (251 to 2284)
Number of events (median (IQR))	49 (23 to 112)	38 (21 to 84)	39 (19 to 77)	64 (21 to 132)
Incidence (median (IQR))	0.06 (0.02 to 0.13)	0.09 (0.05 to 0.14)	0.06 (0.02 to 0.13)	0.06 (0.03 to 0.14)

	Overall	Added value of biomarkers	Comparison of biomarkers	Comparison of prediction models
n	93	41	42	45
Cardiac death	28 (30.1)	14 (34.5)	16 (38.1)	11 (24.4)
Cardiovascular death	16 (17.2)	9 (22.0)	8 (19.0)	6 (13.3)
All cause mortality	17 (18.3)	9 (22.0)	8 (19.0)	6 (13.3)
Nonfatal myocardial infarction	22 (23.7)	11 (26.8)	12 (28.6)	8 (17.8)
Fatal myocardial infarction	1 (1.1)	0 (0.0)	1 (2.4)	0 (0.0)
Myocardial infarction (not specified)	44 (47.3)	22 (53.7)	23 (54.8)	18 (40.0)
Myocardial infarction (any)	66 (70.1)	33 (80.5)	35 (83.3)	26 (57.8)
Heart failure	33 (35.5)	12 (29.3)	17 (40.5)	19 (42.2)
Cardiac arrest	27 (29.0)	6 (14.6)	8 (19.0)	18 (40.0)
Complete heart block	7 (7.5)	2 (4.9)	4 (9.5)	3 (6.7)
Pulmonary oedema	18 (19.4)	8 (19.5)	9 (21.4)	8 (17.8)
Ventricular arrhythmia	12 (12.9)	4 (9.8)	8 (19.0)	7 (15.6)
Atrial arrhythmia	4 (4.3)	1 (2.4)	3 (7.1)	2 (4.4)
Arrhythmia, not specified	17 (18.3)	7 (17.1)	8 (19.0)	8 (17.8)
Revascularisation	6 (6.5)	4 (9.8)	3 (7.1)	2 (4.4)
Acute coronary syndrome	6 (6.5)	4 (9.8)	4 (9.5)	2 (4.4)
Unstable angina	8 (8.6)	2 (4.8)	6 (14.3)	4 (8.9)
Myocardial injury	15 (16.1)	10 (24.4)	10 (23.8)	4 (8.9)
Stroke	14 (15.1)	4 (9.5)	5 (11.9)	9 (20.0)
Hypertensive crisis	2 (2.2)	0 (0.0)	2 (4.8)	0 (0.0)
ST‐T changes on ECG	1 (1.1)	0 (0.0)	0 (0.0)	1 (2.2)
Intraoperative hemodynamic adversity	1 (1.1)	0 (0.0)	1 (2.4)	1 (2.2)
Systemic embolism	1 (1.1)	1 (2.4)	1 (2.4)	0 (0.0)

	All included studies	Added value to the RCRI	Comparison of biomarkers	Comparison of prediction models
N	107	51	51	52
Performance category (%)
Discrimination	102 (95.3)	48 (94.1)	49 (96.1)	50 (96.2)
Calibration	39 (36.4)	10 (19.6)	15 (29.4)	22 (42.3)
Reclassification	23 (21.5)	18 (35.3)	2 (4.0)	5 (9.6)
C‐statistic (%)	98 (91.6)	40 (78.4)	45 (88.2)	48 (92.3)
O/E (%)	22 (20.6)	6 (11.8)	12 (23.5)	8 (15.4)
Calibration plot (%)	14 (13.1)	4 (7.8)	1 (2.0)	10 (19.2)
Hosmer Lemeshow test (%)	7 (6.5)	1 (2.0)	3 (5.9)	7 (13.5)
IDI (%)	7 (6.5)	7 (13.7)	1 (2.0)	0 (0.0)
NRI (%)	22 (20.6)	17 (33.3)	2 (3.9)	5 (9.6)
Other reported measures (%)
Sensitivity	41 (38.3)	6 (11.8)	27 (52.9)	14 (26.9)
Specificity	40 (37.4)	6 (11.8)	27 (52.9)	13 (25.0)
Negative predictive value	19 (17.8)	3 (5.9)	12 (23.5)	5 (9.6)
Positive predictive value	18 (16.8)	3 (5.9)	11 (21.6)	5 (9.6)
Accuracy	3 (2.8)	0 (0.0)	2 (3.9)	1 (1.9)

	Number of studies	Derivation
NT‐proBNP	13	Blood
Troponin	7	Blood
NT‐proBNP + troponin	5	Blood
BNP	4	Blood
Copeptin	3	Blood
Coronary artery calcium score (CACS)	2	Imaging
CRP	2	Blood
fQRS of an ECG	2	Imaging
NT‐proBNP + CRP	2	Blood
V‐POSSUM	2	Other
V‐POSSUM + NTproBNP	2	Blood
V‐POSSUM + troponin	2	Blood
6 minute walking test	1	Other
Abdominal aortic aneurysm size	1	Other
Age	1	Other
Age + abdominal aortic aneurysm size	1	Other
Age + sex + copeptin	1	Other
Age > 70 years	1	Other
Anaerobic threshold	1	Other
Anaemia	1	Other
Angina pectoris	1	Other
ASA	1	Other
ASA + SORT + NSQIP‐MICA	1	Other
Atrial fibrillation	1	Other
Copeptin + NT‐proBNP	1	Blood
Coronary CT angiography	1	Imaging
Duke Activity Status Index	1	Other
ECG abnormalities	1	Imaging
Echocardiography	1	Imaging
Echocardiography + beta blockers	1	Imaging
EE ratio of echocardiography	1	Imaging
Frailty	1	Other
Jeopardy score	1	Imaging
Left bundle branch block on ECG	1	Imaging
Left ventricular ejection fraction	1	Imaging
Male sex	1	Other
Metabolic equivalent (METS)	1	Other
METS + positive stress test with no false negatives	1	Other
METS + stress test	1	Imaging
Multi vessel disease	1	Imaging
Multi vessel disease + CACS	1	Imaging
Peak oxygen	1	Other
Polygenic risk score for coronary artery disease	1	Other
Presepsin	1	Blood
Presepsin + NT‐proBNP	1	Blood
Presepsin + troponin	1	Blood
Presepsin + troponin + NT‐proBNP	1	Blood
Reactive hyperaemia peripheral arterial tonometry	1	Other
Regulatory T cells	1	Blood
Regulatory T cells + troponin + NT‐proBNP	1	Blood
Right bundle branch block on ECG	1	Imaging
Segment involvement + Jeopardy score	1	Imaging
Segment involvement score	1	Imaging
Smoking	1	Other
ST2 + troponin	1	Blood
ST2 cardiac biomarker	1	Blood
Stenosis of CTA + CACS	1	Imaging
Stenosis on CTA	1	Imaging
Stress echocardiography	1	Imaging
Survivin	1	Blood
Thallium scan	1	Imaging
Total joint arthroplasty risk score	1	Other
Type of surgery	1	Other
Type of surgery + age	1	Other
Type of surgery + age + hypertension	1	Other
Valve sclerosis	1	Imaging
Valve stenosis	1	Imaging
V‐POSSUM + troponin + NT‐proBNP	1	Blood
Wall abnormalities on an echocardiography	1	Imaging

Author	Outcome	Prediction horizon	N events	N total	c‐statistic RCRI	CI (95%) c‐statistic RCRI	c‐statistic ASA	CI (95%) c‐statistic RCRI
Bronheim 2018	Any noncardiac complication	30 days	3399	52,066	0.62	(0.61 to 0.63)	0.77	(0.73 to 0.82)
Bronheim 2018	Unplanned intubation	30 days	111	52,066	0.84	(0.83 to 0.84)	0.74	(0.74 to 0.75)
Bronheim 2018	Pulmonary embolism	30 days	149	52,066	0.41	(0.4 to 0.42)	0.81	(0.81 to 0.82)
Bronheim 2018	Ventilated > 48 hours	30 days	65	52,066	0.85	(0.84 to 0.85)	0.74	(0.74 to 0.75)
Bronheim 2018	Acute renal failure	30 days	36	52,066	0.88	(0.88 to 0.89)	0.79	(0.78 to 0.79)
Bronheim 2018	Cerebrovascular accident	30 days	42	52,066	0.75	(0.74 to 0.75)	0.84	(0.84 to 0.84)
Bronheim 2018	Coma > 24 hours	30 days	8	52,066	0.90	(0.87 to 0.93)	0.65	(0.65 to 0.66)
Bronheim 2018	Sepsis	30 days	259	52,066	0.83	(0.82 to 0.83)	0.91	(0.9 to 0.91)
Bronheim 2018	Septic shock	30 days	50	52,066	0.85	(0.84 to 0.85)	0.76	(0.76 to 0.76)
Bronheim 2018	Reoperation	30 days	912	52,066	0.85	(0.85 to 0.86)	0.87	(0.86 to 0.87)
Bronheim 2018	Superficial surgical site infection	30 days	452	52,066	0.72	(0.71 to 0.72)	0.84	(0.84 to 0.85)
Bronheim 2018	Deep incisional surgical site infection	30 days	297	52,066	0.88	(0.88 to 0.88)	0.95	(0.95 to 0.95)
Bronheim 2018	Organ space surgical site infection	30 days	104	52,066	0.88	(0.87 to 0.88)	0.78	(0.77 to 0.78)
Bronheim 2018	Wound dehiscence	30 days	102	52,066	0.72	(0.71 to 0.72)	0.79	(0.79 to 0.8)
Bronheim 2018	Pneumonia	30 days	177	52,066	0.74	(0.73 to 0.74)	0.82	(0.82 to 0.83)
Bronheim 2018	Progressive renal insufficiency	30 days	35	52,066	0.85	(0.84 to 0.85)	0.81	(0.81 to 0.82)
Bronheim 2018	Urinary tract infection	30 days	558	52,066	0.74	(0.73 to 0.74)	0.83	(0.82 to 0.83)
Bronheim 2018	Peripheral nerve injury	30 days	21	52,066	0.07	(0.07 to 0.08)	0.51	(0.51 to 0.52)
Bronheim 2018	Bleeding transfusions	30 days	1621	52,066	0.71	(0.71 to 0.72)	0.80	(0.8 to 0.8)
Bronheim 2018	Deep vein thrombosis/thrombophlebitis	30 days	165	52,066	0.71	(0.7 to 0.71)	0.78	(0.78 to 0.79)
Bronheim 2018	Readmission	30 days	NR	52,066	0.84	(0.83 to 0.84)	0.91	(0.9 to 0.91)
Ehlert 2016	Clavien Dindo class IV complications	In‐hospital	800	5621	0.56	NR	0.55	NR
Ehlert 2016	Clavien Dindo class IV complications	In‐hospital	541	15,354	0.59	NR	0.56	NR
Ehlert 2016	Clavien Dindo class IV complications	In‐hospital	455	8367	0.56	NR	0.57	NR
Ehlert 2016	Clavien Dindo class IV complications	In‐hospital	32	1833	0.56	NR	0.59	NR
Ehlert 2016	Clavien Dindo class IV complications	In‐hospital	835	40,803	0.69	NR	0.56	NR
Farina‐Castro 2020	Postoperative complications (CCI 0 vs CCI ≥ 1)	Not reported	179	244	0.69	(0.60 to 0.79)	0.65	(0.56 to 0.74)
James 2014	Surgical complications	30 days	40	83	0.53	(0.4 to 0.65)	0.60	(0.48 to 0.72)
Makary 2010	Surgical complications	30 days	34	594	0.72	NR	0.71	NR
Makary 2010	Discharge to a nursing facility	In‐hospital	14	594	0.75	NR	0.78	NR
Press 2006	All‐cause mortality or nonfatal stroke	30 days	64	1998	0.61	NR	0.53	NR
Press 2006	Noncardiac complications	30 days	63	1998	0.66	NR	0.62	NR
Press 2006	Minor neurological complications	30 days	138	1998	0.56	NR	0.53	NR
Press 2006	Wound complications	30 days	119	1998	0.61	NR	0.54	NR

	Number of studies
NSQIP‐MICA	10
NSQIP surgical risk score	9
CHADS2 score	4
Detsky index	4
Goldman index	4
CHADS2VASc	3
R2CHADS2 score	3
Vascular Study Group of New England Cardiac Risk Index	3
AUB‐HAS2 Cardiovascular Risk Index	2
Charlson Index	2
Glasgow Aneurysm Risk score	2
Halm score	2
Individual items of the RCRI	2
POSSUM	2
P‐POSSUM	2
RCRI without insulin use with low eGFR	2
Reiss Index	2
South African Vascular Surgical Cardiac Risk Index	2
Surgical Mortality Probability Model	2
Thoracic RCRI	2
Tu score	2
V‐POSSUM	2
Adult Comorbidity Evaluation‐27 score	1
Adult Comorbidity Evaluation‐27 score + high age	1
Age + type of admission + RCRI + arrhythmia + electrolyte disorder + hypertension	1
Age + type of admission + RCRI + arrhythmia + electrolyte disorder + hypertension + polygenic risk score for coronary artery disease	1
ANESCARDIOCAT	1
ASA + NSQIP surgical risk score	1
ASA + Surgical Outcome Risk Tool	1
Ashton	1
Biochemistry and Haematology Outcome Model	1
Coronary artery disease + atrium fibrillation + diabetes mellitus + mechanical ventilation + heart rate	1
CR‐POSSUM	1
Detsky score + type of surgery	1
Dilated cardiomyopathy + ischaemic cardiopathy + CVA	1
Eagle score	1
Geriatric Sensitive Perioperative Cardiac Risk Index	1
Insulin use + open surgery + high fibrinogen + CRP + NT‐proBNP	1
Long Term Survival Score	1
MASHOUR	1
Modified Frailty Index	1
Myocardial infarction + sex + insulin‐dependent diabetes mellitus + low BMI + high age + atrium fibrillation	1
New model 1	1
New model 2	1
New model 3	1
NSQIP score "Death"	1
NT‐proBNP + high creatinine + ischaemic heart disease + congestive heart failure	1
Patient Outcomes in Renal Transplant model	1
Preoperative risk score of the estimation of physiological ability + surgical stress score	1
RCRI with redefined high‐risk surgery	1
RCRI with redefined high‐risk surgery and clinical characteristics	1
RCRI without insulin use and creatinine > 2.0 mg/dL	1
Recalibrated NSQIP surgical risk score	1
Recalibrated RCRI	1
Regulatory T cells + age + sex + ASA + previous PCI + creatinine	1
Surgical Outcome Risk Tool	1
Surgical risk score	1
TJA individual factors	1
TJA risk score	1
Updated Cardiac Risk Score	1
Vascular Biochemistry and Haematology Outcome Model	1
Vascular Quality Initiative Cardiac Risk Index	1
Vascular Study Group of New England Cardiac Risk Index + anaemia	1
V‐POSSUM + troponin	1
V‐POSSUM + troponin + NT‐proBNP	1

*Study characteristics*
General information	Objective Biomarkers compared Journal Heart & Lung Country Turkey Study design Prospective cohort study
Participants	Number of included patients 714 Surgical specialty Orthopaedic surgery Age Mean 70.4 years Male sex 35% High‐risk surgery Not reported Insulin‐dependent diabetes mellitus Not reported History of ischaemic heart disease 12.9% History of congestive heart failure Not reported History of cerebrovascular accidents 5.9% Elevated creatinine Not reported 0 RCRI factors Not reported 1 RCRI factor Not reported 2 RCRI factors Not reported 3 or more RCRI factors Not reported
Predictors	Predictor 1: Aortic arch calcification Objective: biomarker compared Category: imaging Scale: categorical Threshold: Grade > 1, grade > 2 Assay/device: Curix HT 1.000G Plus, Agfa, Mortsel, Belgium
Outcome	Outcome category MACE Full outcome definition Acute coronary syndrome (STEMI, non‐STEMI, UAP), decompensated heart failure, new onset atrial fibrillation, stroke and cardiac death Prediction horizon 30‐day events
Analysis	Number of outcomes 33 Handling missing data No information on handling missing data Discrimination reported? Yes Calibration reported? No Reclassification reported? No
PROBAST: Applicability	Domain 1: Participant selection High Justification: Only orthopaedic patients were included and patients with malignancy and previous cardiac surgery were excluded. In addition, patients from 18 years onwards were eligible. Domain 2: Predictors Unclear Justification: No information on the definition of the individual RCRI predictor definitions. Domain 3: Outcome High Justification: Although outcome is MACE, it also includes stroke, atrial fibrillation and unstable angina pectoris. Overall judgement High Justification: Only a selected group of patients was included, there was no/unclear information on predictor definitions and outcome definition was different compared to the development study
Notes

Item	Authors' judgement	Support for judgement
Domain 1: Participant selection	Yes	Although only patients undergoing orthopaedic surgery were included, participant selection was appropriate and the RCRI model can be applied in these patients.
Domain 2: Predictors	Unclear	No information on the definition of the individual RCRI predictor definitions.
Domain 3: Outcome	Yes	Clearly defined outcome definitions and appropriate adjudication of outcomes.
Domain 4: Analysis	No	Low number of outcomes and no information on handling missing data.
Overall judgement	No	Low number of outcomes and no information on handling missing data and no information on the definition of the individual RCRI predictor definitions. However, patient selection was appropriate and outcome definitions were clearly defined and assessed.

*Study characteristics*
General information	Objective Added value biomarkers Biomarkers compared Journal Journal of American College of Cardiology Country Republic of Korea Study design Retrospective cohort study
Participants	Number of included patients 239 Surgical specialty Noncardiac surgery Age Median 69 years (IQR 62 to 75) Male sex 52.3% High‐risk surgery Not reported Insulin‐dependent diabetes mellitus 13% History of ischaemic heart disease Not reported History of congestive heart failure 3.3% History of cerebrovascular accidents 10.9% Elevated creatinine Not reported 0 RCRI factors 43.9% 1 RCRI factor 41% 2 RCRI factors 11.7% 3 or more RCRI factors 3.3%
Predictors	Predictor 1: Coronary artery calcium scores (CACS) Objective: added value, biomarker compared Category: imaging Scale: dichotomous Threshold: 113 CACS Assay/device: Brilliance 64, Philips Healthcare, Best, the Netherlands Predictor 2: Multi‐vessel disease Objective: added value Category: imaging Scale: categorical Threshold: significant stenosis (50% luminal diameter narrowing) in 1, 2 or 3 vessels Assay/device: Brilliance 64, Philips Healthcare, Best, the Netherlands Predictor 3: Coronary artery calcium scores (CACS) + multi‐vessel disease Objective: added value Category: imaging Scale: categorical Threshold: not applicable Assay/device: Brilliance 64, Philips Healthcare, Best, the Netherlands
Outcome	Outcome category MACE Full outcome definition Cardiac death, acute coronary syndrome (nonfatal myocardial infarction and unstable angina), pulmonary oedema, ventricular fibrillation, ventricular tachycardia with haemodynamic compromise, and complete heart block Prediction horizon 30‐day events
Analysis	Number of outcomes 19 Handling missing data No information on handling missing data Discrimination reported? Yes Calibration reported? Yes Reclassification reported? Yes
PROBAST: Applicability	Domain 1: Participant selection Low Domain 2: Predictors Low Domain 3: Outcome Low Overall judgement: Low Patient selection was appropriate; predictor and outcome definitions were clearly defined and comparable to the definitions used in the development study.
Notes	—

Item	Authors' judgement	Support for judgement
Domain 1: Participant selection	No	Patients with severe cardiac morbidities such as previous myocardial infarction, severe heart failure or severe valvular disease were excluded from the analysis.
Domain 2: Predictors	Yes	Clear (RCRI) predictor definitions were described.
Domain 3: Outcome	Yes	Clearly defined outcome definitions and appropriate adjudication of outcomes.
Domain 4: Analysis	No	Small number of outcomes. No information on how missing data were handled.
Overall judgement	No	Patients with severe cardiac morbidities were excluded from the analysis. In addition, there was a small number of outcomes and no information on handling of missing data. However, predictor and outcome definitions were clearly reported and assessed.

*Study characteristics*
General information	Objective Added biomarkers Biomarkers compared Journal Journal of Cardiovascular Computed Tomography Country Republic of Korea Study design Prospective cohort study
Participants	Number of included patients 206 Surgical specialty Noncardiac surgery Age Mean 69.2 years (SD 8.7) Male sex 52.9% High‐risk surgery 14.1% Insulin‐dependent diabetes mellitus Not reported History of ischaemic heart disease 35% History of congestive heart failure 12.6% History of cerebrovascular accidents 28.2% Elevated creatinine 16% 0 RCRI factors 51.9% 1 RCRI factor 35.4% 2 RCRI factors 11.2% 3 or more RCRI factors 1.5%
Predictors	Predictor 1: Dobutamine stress test Objective: added biomarker, biomarker compared Category: imaging Scale: dichotomous Threshold: abnormal if there is ischaemia during stress or fixed wall motion abnormalities Assay/device: Vivid E9 apparatus (GE Healthcare, Waukesha, WI) Predictor 2: Coronary artery stenosis Objective: added biomarker, biomarker compared Category: imaging Scale: dichotomous Threshold: 50% Assay/device: Brilliance 64 multidetector scanner (Philips Healthcare, Best, the Netherlands) Predictor 3: Coronary artery calcium scores Objective: added biomarker, biomarker compared Category: imaging Scale: dichotomous Threshold: 203 Assay/device: Brilliance 64 multidetector scanner (Philips Healthcare, Best, the Netherlands) Predictor 4: Coronary artery calcium scores + significant coronary artery stenosis ≥ 50% Objective: added biomarker Category: imaging Scale: dichotomous Threshold: not applicable Assay/device: Brilliance 64 multidetector scanner (Philips Healthcare, Best, the Netherlands)
Outcome	Outcome category MACE Full outcome definition Cardiovascular death, nonfatal myocardial infarction, myocardial injury after noncardiac surgery (MINS), pulmonary oedema, nonfatal stroke and systemic embolism Prediction horizon 30‐day events
Analysis	Number of outcomes 24 Handling missing data No information on handling missing data Discrimination reported? Yes Calibration reported? No Reclassification reported? No
PROBAST: Applicability	Domain 1: Participant selection High Justification: Patients with at least 1 RCRI factor were included, patients were excluded if they had active cardiac conditions including recent MI, decompensated heart failure, more than moderate valvular heart disease and significant arrhythmia. Domain 2: Predictors Unclear Justification: No information on the definition of the individual RCRI predictor definitions. Domain 3: Outcome High Justification: MACE includes MINS and pulmonary embolism and stroke Overall judgement High Justification: Only a selected group of patients was included, there was no/unclear information on predictor definitions and outcome definition was different compared to the development study
Notes	—

Item	Authors' judgement	Support for judgement
Domain 1: Participant selection	Yes	Clearly defined outcome definitions and appropriate adjudication of outcomes.
Domain 2: Predictors	Unclear	No information on the definition of the individual RCRI predictor definitions.
Domain 3: Outcome	Yes	Clearly defined outcome definitions and appropriate adjudication of outcomes.
Domain 4: Analysis	No	Small number of outcomes. No information on how missing data were handled.
Overall judgement	No	No information on the definition of the individual RCRI predictor definitions. In addition, the number of outcomes was low and there was no information on handling missing data. However, patient selection was appropriate and outcomes were clearly defined and assessed.

*Study characteristics*
General information	Objective Prediction model compared Journal Journal of the American Heart Association Country USA Study design Prospective existing registry
Participants	Number of included patients 172,905 Surgical specialty Noncardiac surgery Age Mean 74.1 years (SD 6.9) Male sex Not reported High‐risk surgery Not reported Insulin‐dependent diabetes mellitus 7.3% History of ischaemic heart disease Not reported History of congestive heart failure 1.1% History of cerebrovascular accidents Not reported Elevated creatinine Not reported 0 RCRI factors Not reported 1 RCRI factor Not reported 2 RCRI factors Not reported 3 or more RCRI factors Not reported
Predictors	Predictor 1: GSCRI Objective: Prediction model compared Category: prediction model Scale: continuous Threshold: not applicable Assay/device: not applicable Predictor 2: ACS‐NSQIP‐MICA Objective: prediction model compared Category: prediction model Scale: continuous Threshold: not applicable Assay/device: not applicable
Outcome	Outcome category Myocardial infarction and cardiac arrest (MICA) Full outcome definition Not applicable Prediction horizon 30‐day events
Analysis	Number of outcomes 1798 Handling missing data No information on handling missing data Discrimination reported? Yes Calibration reported? Yes Reclassification reported? No
PROBAST: Applicability	Domain 1: Participant selection Low Justification: Not applicable Domain 2: Predictors Low Justification: Not applicable Domain 3: Outcome High Justification: Outcome is different from MACE in the development study Overall judgement High Justification: Patient selection was appropriate and predictor definitions were clearly defined and comparable to definitions used in the development study. However, the outcome used was different from MACE in the development study.
Notes	—

Item	Authors' judgement	Support for judgement
Domain 1: Participant selection	Yes	Appropriate participant selection in which patients were selected in whom the RCRI model can be applied.
Domain 2: Predictors	Yes	Clear (RCRI) predictor definitions were described.
Domain 3: Outcome	Yes	Clearly defined outcome definitions and appropriate adjudication of outcomes.
Domain 4: Analysis	Yes	Clear methodology and appropriate number of outcomes.
Overall judgement	Yes	Appropriate patient selection and number of outcomes, clear predictor and outcome definitions and study methodology.

*Study characteristics*
General information	Objective Biomarkers compared, prediction model compared Journal Journal of Cardiothoracic and Vascular Anesthesia Country USA Study design Retrospective cohort study
Participants	Number of included patients 225 Surgical specialty Vascular surgery Age Mean 73.8 years (SD 9) Male sex 85.8% High‐risk surgery Not reported Insulin‐dependent diabetes mellitus Not reported History of ischaemic heart disease 74.2% History of congestive heart failure 10.2% History of cerebrovascular events 17.8% Elevated creatinine Not reported 0 RCRI factors 0% 1 RCRI factor 43.1% 2 RCRI factors 34.7% 3 or more RCRI factors 22.2%
Predictors	Predictor 1: Age Objective: biomarker compared Category: patient characteristic Scale: continuous Threshold: not applicable Assay/device: not applicable Predictor 2: Glasgow Aneurysm Risk score (GAS) Objective: prediction model compared Category: prediction model Scale: unclear Threshold: not applicable Assay/device: not applicable
Outcome	Outcome category MACE Full outcome definition Cardiac death, nonfatal myocardial infarction, unstable angina, now onset or worsening of chronic heart failure or coronary revascularisation Prediction horizon In‐hospital events
Analysis	Number of outcomes 14 Handling missing data No information on handling missing data Discrimination reported? Yes Calibration reported? No Reclassification reported? No
PROBAST: Applicability	Domain 1: Participant selection Low Justification: not applicable Domain 2: Predictors Low Justification: not applicable Domain 3: Outcome High Justification: Outcome is different from MACE in the development study Overall judgement High Justification: patient selection was appropriate and predictor definitions were clearly defined and comparable to definitions used in the development study. However, the outcome used was different from MACE in the development study.
Notes	—

Item	Authors' judgement	Support for judgement
Domain 1: Participant selection	Yes	Although only patients undergoing vascular surgery were included, participant selection was appropriate and the RCRI model can be applied in these patients.
Domain 2: Predictors	Yes	Clear (RCRI) predictor definitions were described.
Domain 3: Outcome	Yes	Clearly defined outcome definitions and appropriate adjudication of outcomes.
Domain 4: Analysis	No	Other (more advanced) performance measures could have been calculated and reported including confidence intervals and/or standard error; low number of outcomes.
Overall judgement	No	Appropriate patient selection and clearly defined predictors and outcomes. However, the number of outcomes was low and other performance measures should have been calculated with confidence intervals and/or standard error.

*Study characteristics*
General information	Objective Added biomarkers Journal International Journal of Cardiology Country Republic of Korea Study design Cohort study (prospective/retrospective unclear)
Participants	Number of included patients 428 Surgical specialty Vascular surgery Age Not reported Male sex Not reported High‐risk surgery Not reported Insulin‐dependent diabetes mellitus Not reported History of ischaemic heart disease Not reported History of congestive heart failure Not reported History of cerebrovascular events Not reported Elevated creatinine Not reported 0 RCRI factors 50.5% 1 RCRI factor 33.2% 2 RCRI factors 13.1% 3 or more RCRI factors 3.3%
Predictors	Predictor 1: Fragmented QRS complex (fQRS) Objective: added biomarker Category: imaging Scale: dichotomous Threshold: not applicable Assay/device: 12‐lead resting ECG
Outcome	Outcome category MACE Full outcome definition Myocardial ischaemia or scar, as detected by myocardial perfusion single photon emission computed tomography (SPECT) Prediction horizon Not reported
Analysis	Number of outcomes 87 Handling missing data No information on handling missing data Discrimination reported? Yes Calibration reported? No Reclassification reported? No
PROBAST: Applicability	Domain 1: Participant selection Unclear Justification: patients underwent noncardiac vascular surgery and no information is reported on baseline characteristics of included patients Domain 2: Predictors Unclear Justification: no information was provided how the items of the RCRI were interpreted and defined Domain 3: Outcome High Justification: outcome does not match outcome of the development study Overall judgement High Justification: no information on baseline characteristics of included patients was reported. There was no/unclear information on predictor definitions and outcome definition was different compared to the development study
Notes	—

Item	Authors' judgement	Support for judgement
Domain 1: Participant selection	Yes	Although only patients undergoing vascular surgery were included, participant selection was appropriate and the RCRI model can be applied in these patients.
Domain 2: Predictors	Unclear	No information was provided how the items of the RCRI were interpreted and defined.
Domain 3: Outcome	Unclear	No information on how the SPECT images were assessed and how the outcome was determined based on the SPECT.
Domain 4: Analysis	No	Low number of outcomes and no information on the handling of missing data.
Overall judgement	No	Patient selection was appropriate. However, predictors and outcomes definitions were unclear. In addition, the number of outcomes was low and no information on handling of missing data.

*Study characteristics*
General information	Objective Added biomarkers Journal American Journal of Cardiology Country Republic of Korea Study design Retrospective cohort study
Participants	Number of included patients 467 Surgical specialty Vascular surgery Age Mean 69.4 years (SD 9.5) Male sex 86% High‐risk surgery Not reported Insulin‐dependent diabetes mellitus Not reported History of ischaemic heart disease 15.2% History of congestive heart failure 6.2% History of cerebrovascular events 14.1% Elevated creatinine 8% 0 RCRI factors 46.9% 1 RCRI factor 35.3% 2 RCRI factors 12.4% 3 or more RCRI factors 5.4%
Predictors	Predictor 1: Fragmented QRS complex (fQRS) Objective: added biomarker Category: imaging Scale: dichotomous Threshold: not applicable Assay/device: Philips TraceMasterVue ECG management system, Philips 12‐leas algorithm, Andover, Massachusetts
Outcome	Outcome category MACE Full outcome definition Death, myocardial infarction, congestive heart failure, and percutaneous coronary intervention before noncardiac vascular surgery during index hospitalisation Prediction horizon In‐hospital events
Analysis	Number of outcomes 38 Handling missing data No information on handling missing data Discrimination reported? Yes Calibration reported? Yes Reclassification reported? Yes
PROBAST: Applicability	Domain 1: Participant selection Low Justification: patients underwent vascular surgery and underwent SPECT before being considered for inclusion Domain 2: Predictors High Justification: for some items, no information on the definition was provided. High‐risk surgery was not inserted into the RCRI and the definition of diabetes was different compared to the development paper Domain 3: Outcome High Justification: outcome does not match outcome of the development study Overall judgement High Justification: patient selection was appropriate. However, no/unclear information on predictor definitions for some items and other predictors of the original RCRI were not included or had a different definition. Furthermore, outcome definition was different compared to the RCRI development study.
Notes	—

Item	Authors' judgement	Support for judgement
Domain 1: Participant selection	Yes	Although only patients undergoing vascular surgery were included, participant selection was appropriate and the RCRI model can be applied in these patients.
Domain 2: Predictors	No	For some items, no information on the definition was provided. High‐risk surgery was not inserted into the RCRI and the definition of diabetes was different compared to the development paper.
Domain 3: Outcome	Unclear	No information how the outcomes were assessed and if the assessors were blinded for predictor values.
Domain 4: Analysis	No	Low number of outcomes and no information on the handling of missing data.
Overall judgement	No	Patient selection was appropriate. However, there was no information on how outcomes were assessed. Prediction definitions were unclear or were different compared to definitions used in the RCRI development study. In addition, the number of outcomes was low, no performance measures were reported and no information on handling of missing data.

*Study characteristics*
General information	Objective Added biomarkers, biomarkers compared Journal Anaesthesia Country South Africa Study design Prospective cohort study
Participants	Number of included patients 267 Surgical specialty Vascular surgery Age Median 61 years (IQR = 50 to 69, range = 20 to 86) Male sex 62% High‐risk surgery Not reported Insulin‐dependent diabetes mellitus Not reported History of ischaemic heart disease 34% History of congestive heart failure 4% History of cerebrovascular events 27% Elevated creatinine 3% 0 RCRI factors 35% 1 to 2 RCRI factors 54% 3 or more RCRI factors 11%
Predictors	Predictor 1: BNP Objective: added biomarker, biomarker compared Category: blood Scale: dichotomous, categorical Threshold: 69 pg/mLand tertiles Assay/device: Advia Centaur Xp (Siemens Medical, Deerfield, IL, USA)
Outcome	Outcome category Troponin elevation Full outcome definition Not applicable Prediction horizon Within 3 postoperative days
Analysis	Number of outcomes 36 Handling missing data No missing data Discrimination reported? Yes Calibration reported? No Reclassification reported? Yes
PROBAST: Applicability	Domain 1: Participant selection High Justification: patients underwent vascular surgery; no age limit was provided Domain 2: Predictors Unclear Justification: no information on the definition of the individual RCRI predictor definitions Domain 3: Outcome High Justification: outcome assessed is troponin elevation and not MACE as defined in the development study Overall judgement High Justification: only a selected group of patients was included, there was no/unclear information on predictor definitions and outcome definition was different compared to the development study
Notes	—

Item	Authors' judgement	Support for judgement
Domain 1: Participant selection	No	Post hoc decision to exclude a selective group of patients.
Domain 2: Predictors	Unclear	No information on the definition of the individual RCRI predictor definitions.
Domain 3: Outcome	Yes	Clearly defined outcome definitions and appropriate adjudication of outcomes.
Domain 4: Analysis	No	Very low number of events; calibration not assessed; limited information on discrimination.
Overall judgement	No	Inappropriate exclusion of a selective group of patients, predictor definitions were not reported and the number of events was low and no calibration measures were assessed. However, outcome definitions were clearly defined and assessed.

*Study characteristics*
General information	Objective Added biomarkers Journal JAMA Country Netherlands and Italy Study design Prospective cohort study
Participants	Number of included patients 1351 Surgical specialty Vascular surgery Age Not reported Male sex 78% High‐risk surgery Not reported Insulin‐dependent diabetes mellitus Not reported History of ischaemic heart disease Not reported History of congestive heart failure 5.3% History of cerebrovascular events 8.8% Elevated creatinine 4.1% 0 RCRI factors 45% 1 RCRI factor 38% 2 or more RCRI factors 17%
Predictors	Predictor 1: Dobutamine stress echocardiography (DES) + betablocker use Objective: added biomarker Category: imaging Scale: dichotomous Threshold: worsening of ≥ 1 point during the stress test using a 5‐point ordinal scale Assay/device: not reported
Outcome	Outcome category MACE Full outcome definition Cardiac death or nonfatal myocardial infarction Prediction horizon 30‐day events
Analysis	Number of outcomes 45 Handling missing data No information on handling missing data Discrimination reported? Yes Calibration reported? Yes Reclassification reported? No
PROBAST: Applicability	Domain 1: Participant selection High Justification: population very different from the development study; only high‐risk patients included Domain 2: Predictors Low Justification: Domain 3: Outcome High Justification: outcome is cardiovascular death with myocardial infarction in this study and MACE in the development study Overall judgement High Justification: only high‐risk patients were included. Predictors were clearly defined. However, the outcome used was different compared to the development study.
Notes	—

Item	Authors' judgement	Support for judgement
Domain 1: Participant selection	No	Only patients with at least one cardiac risk factor had a DSE meaning that only high‐risk patients were assessed.
Domain 2: Predictors	Yes	Clear (RCRI) predictor definitions were described.
Domain 3: Outcome	Yes	Clearly defined outcome definitions and appropriate adjudication of outcomes.
Domain 4: Analysis	No	Low number of outcomes and no information on handling missing data.
Overall judgement	No	Only high‐risk patients were included. Predictors and outcomes were clearly defined. However, the number of outcomes was low and there was no information on handling missing data.

*Study characteristics*
General information	Objective Added biomarkers Journal Korean Circulation Journal Country Republic of Korea Study design Retrospective cohort study
Participants	Number of included patients 26501 Surgical specialty Noncardiac surgery Age Not reported Male sex Not reported High‐risk surgery Not reported Insulin‐dependent diabetes mellitus Not reported History of ischaemic heart disease Not reported History of congestive heart failure Not reported History of cerebrovascular events Not reported Elevated creatinine Not reported 0 RCRI factors Not reported 1 RCRI factor Not reported 2 RCRI factors Not reported 3 or more RCRI factors Not reported
Predictors	Predictor 1: Atrial fibrillation Objective: added biomarkers Category: patient characteristic Scale: dichotomous Threshold: not applicable Assay/device: ECG
Outcome	Outcome category All‐cause mortality and MACE Full outcome definition Composite of death, ischaemic stroke and myocardial infarction Prediction horizon In‐hospital or within 30 days
Analysis	Number of outcomes 353 Handling missing data No information on handling missing data Discrimination reported? Yes Calibration reported? No Reclassification reported? No
PROBAST: Applicability	Domain 1: Participant selection Low Justification: Domain 2: Predictors Low Justification: Domain 3: Outcome High Justification: MACE definition varies from definition of MACE in development cohort Overall judgement High Justification: patient selection was appropriate and predictor definitions were clearly defined and comparable to definitions used in the development study. However, the outcome used was different from MACE in the development study.
Notes	—

Item	Authors' judgement	Support for judgement
Domain 1: Participant selection	No	Patients without cardiac evaluation were excluded (approximately 80% of sample).
Domain 2: Predictors	Yes	Clear (RCRI) predictor definitions were described.
Domain 3: Outcome	Yes	Clearly defined outcome definitions and appropriate adjudication of outcomes.
Domain 4: Analysis	Yes	However, no information on handling missing data, only c‐statistic reported. No measures of calibration or reclassification.
Overall judgement	No	Only a selected group of patients were included in the analysis and no information on the handling of missing data. In addition, only discrimination was reported and no other performance measures. However, outcomes and predictors were clearly defined and assessed.

*Study characteristics*
General information	Objective Added biomarkers Journal Heart Country Republic of Korea Study design Prospective cohort study
Participants	Number of included patients 2054 Surgical specialty Noncardiac surgery Age Median 68 years (IQR = 61 to 73 years) Male sex 60.7% High‐risk surgery 41.1% Insulin‐dependent diabetes mellitus 3.5% History of ischaemic heart disease 21.6% History of congestive heart failure 3% History of cerebrovascular events 9.3% Elevated creatinine Not reported 0 RCRI factors 27% 1 RCRI factor 41.2% 2 RCRI factors 28.2% 3 or more RCRI factors 3.6%
Predictors	Predictor 1: NT‐proBNP Objective: added biomarkers Category: blood Scale: dichotomous Threshold: 301 mg/L Assay/device: not reported Predictor 2: CRP Objective: added biomarkers Category: blood Scale: dichotomous Threshold: 3.4 mg/L Assay/device: not reported Predictor 3: NT‐proBNP + CRP Objective: added biomarkers Category: blood Scale: dichotomous Threshold: 301 and 3.4 mg/L, respectively Assay/device: not reported
Outcome	Outcome category MACE; myocardial infarction; pulmonary oedema; cardiovascular death Full outcome definition MACE was defined as myocardial infarction, development of pulmonary oedema or primary cardiovascular death. Cardiovascular death was defined as sudden death that could not be explained by any other than cardiovascular postoperative complications. Prediction horizon In‐hospital or within 30 days
Analysis	Number of outcomes 291 MACE Handling missing data No information on handling missing data Discrimination reported? Yes Calibration reported? No Reclassification reported? No
PROBAST: Applicability	Domain 1: Participant selection High Justification: patients were required to have ≥ 1 cardiovascular risk factor such as hypertension, diabetes, angina, history of revascularisation, heart failure or stroke, or abnormal preoperative electrocardiography with pathological Q wave or non‐sinus rhythm. In addition patients with creatinine > 2.0 mg/dL were excluded from the analysis. Domain 2: Predictors Low Justification: not applicable Domain 3: Outcome Low Justification: not applicable Overall judgement High Justification: only a selected group of patients was included. However, predictors and outcomes were clearly defined and comparable as used in the development study.
Notes	—

Item	Authors' judgement	Support for judgement
Domain 1: Participant selection	Yes	Appropriate participant selection in which patients were selected in whom the RCRI model can be applied.
Domain 2: Predictors	Yes	Clear (RCRI) predictor definitions were described.
Domain 3: Outcome	Yes	Clearly defined outcome definitions and appropriate adjudication of outcomes.
Domain 4: Analysis	No	Data were dichotomised for all predictors of interest; no information on the handling of missing data. No calibration or reclassification measures were reported.
Overall judgement	No	Appropriate patient selection and clearly defined predictors and outcomes. However, data were dichotomised, there was no information on the handling of missing data and no information on calibration and reclassification measures were reported.

PERMALINK

The comparative and added prognostic value of biomarkers to the Revised Cardiac Risk Index for preoperative prediction of major adverse cardiac events and all‐cause mortality in patients who undergo noncardiac surgery

Lisette M Vernooij

Wilton A Klei

Karel GM Moons

Toshihiko Takada

Judith Waes

Johanna AAG Damen

Abstract

Background

Objectives

Search methods

Selection criteria

Data collection and analysis

Main results

Authors' conclusions

Plain language summary

Summary of findings

Summary of findings 1. Summary of findings ‐ objective 1: added value of biomarkers to the RCRI.

Summary of findings 2. Summary of findings ‐ objective 2: comparison of predictive performance of biomarkers to the RCRI.

Summary of findings 3. Summary of findings ‐ objective 3: comparison of predictive performance of other prediction models to the RCRI.

Background

Description of the condition

Description of the prognostic model

1. Scoring of the Revised Cardiac Risk Index.

Health outcomes

Why it is important to do this review of these prognostic models

Objectives

Primary objective

2. PICOTS for the objectives based on the CHARMS checklist.

Other objectives

Investigation of sources of heterogeneity between studies

Methods

Criteria for considering studies for this review

Types of studies

Types of participants (target population)

Types of prognostic models

Types of outcomes

Search methods for identification of studies

Electronic searches

Searching other resources

Data collection and analysis

Selection of studies

Data extraction and management

Assessment of risk of bias of included studies

Measures of predictive performance to be extracted

Dealing with missing data

Assessment of heterogeneity

Assessment of reporting deficiencies

Data synthesis

Data synthesis and meta‐analysis approaches

Subgroup analysis and investigation of heterogeneity

Sensitivity analysis

Rating the certainty of evidence and summary of findings

Results

Description of studies

Results of the search

1.

Risk of bias and concern regarding applicability

2.

3.

PROBAST domain 1: Participants

PROBAST domain 2: Predictors

PROBAST domain 3: Outcome

PROBAST domain 4: Analysis

Included studies

Objective 1: the added predictive value of biomarkers to the RCRI

Study design and study population

3. Study characteristics of included studies.

4.

Outcomes and composition of MACE

4. Composites used to define major adverse cardiac events (MACE).

Risk of bias and concern regarding applicability

Predictive performance measures reported

5. Reporting of performance measures in included studies.

Added biomarkers

6. Biomarkers/predictors added to the RCRI.

N‐terminal pro‐B type natriuretic peptide (NT‐proBNP)

5.

Brain natriuretic peptide (BNP)

CHADS₂

CHA₂DS₂‐VASc

R₂CHADS₂

*Study characteristics*
General information	Objective Added biomarkers, prediction model compared Journal Head and Neck Country Netherlands Study design Retrospective cohort study
Participants	Number of included patients 135 Surgical specialty Head and neck surgery Age Median 59 years (range = 24 to 83 years) Male sex 59.3% High‐risk surgery Not reported Insulin‐dependent diabetes mellitus Not reported History of ischaemic heart disease Not reported History of congestive heart failure 0.9% History of cerebrovascular events Not reported Elevated creatinine Not reported 0 RCRI factors 57% 1 RCRI factor 28.9% 2 RCRI factors 11.1% 3 or more RCRI factors 3%
Predictors	Predictor 1: Age Objective: added biomarker Category: patient characteristic Scale: dichotomous Threshold: ≥ 70 years Assay/device: not applicable Predictor 2: Adult comorbidity evaluation (ACE‐27) Objective: prediction model compared Category: prediction model Scale: categorical Threshold: grade 1: mild decompensation; grade 2: moderate decompensation; or grade 3: severe decompensation Assay/device: not applicable Predictor 3: Adult comorbidity evaluation (ACE‐27) + age ≥ 70 years Objective: prediction model compared Category: prediction model Scale: categorical Threshold: not applicable Assay/device: not applicable
Outcome	Outcome category MACE Full outcome definition Cardiac death, nonfatal myocardial infarction, heart failure and cardiac arrhythmias Prediction horizon In‐hospital events
Analysis	Number of outcomes 23 Handling missing data No information on handling missing data Discrimination reported? Yes Calibration reported? Yes Reclassification reported? No
PROBAST: Applicability	Domain 1: Participant selection Low Justification: not applicable Domain 2: Predictors Low Justification: not applicable Domain 3: Outcome Low Justification: not applicable Overall judgement: Low Patient selection was appropriate, predictor and outcome definitions were clearly defined and comparable to the definitions used in the development study.
Notes	—

Item	Authors' judgement	Support for judgement
Domain 1: Participant selection	Yes	Although only patients undergoing head and neck surgery were included, participant selection was appropriate and the RCRI model can be applied in these patients.
Domain 2: Predictors	Yes	Clear (RCRI) predictor definitions were described.
Domain 3: Outcome	Yes	Clearly defined outcome definitions and appropriate adjudication of outcomes.
Domain 4: Analysis	No	Low number of outcomes and only c‐statistic reported without any confidence intervals or standard errors.
Overall judgement	No	Appropriate patient selection and clearly defined predictors and outcomes. However, the number of outcomes was low and no calibration was reported.

*Study characteristics*
General information	Objective Added biomarkers Journal Annals of Internal Medicine Country Canada, Hong Kong, Brazil, UK, South Africa, Australia, Malaysia, Poland, USA, Germany Study design Prospective cohort study
Participants	Number of included patients 10,402 Surgical specialty Noncardiac surgery Age Mean 66 years (SD 11.1) Male sex 50% High‐risk surgery Not reported Insulin‐dependent diabetes mellitus Not reported History of ischaemic heart disease 14.7% History of congestive heart failure 3.3% History of cerebrovascular events 6.9% Elevated creatinine Not reported 0 RCRI factors 56.7% 1 RCRI factor 30.6% 2 RCRI factors 9.3% 3 or more RCRI factors 3.4%
Predictors	Predictor 1: NT‐proBNP Objective: added biomarker Category: blood Scale: categorical Threshold: < 100, 100 to 200, 200 to 1500, > 1500 pg/ml Assay/device: Roche immunoassay analysers (Roche Diagnostics)
Outcome	Outcome category MACE; all‐cause mortality and MACE Full outcome definition MACE was defined as MINS (myocardial injury after noncardiac surgery) or vascular death; all‐cause mortality and MACE were defined as all‐cause mortality or myocardial infarction Prediction horizon 30‐day events
Analysis	Number of outcomes 1269 MACE; 446 deaths or myocardial infarction Handling missing data No information on handling missing data Discrimination reported? Yes Calibration reported? No Reclassification reported? Yes
PROBAST: Applicability	Domain 1: Participant selection Low Justification: Domain 2: Predictors Unclear Justification: no definition for the RCRI items was reported Domain 3: Outcome High Justification: composite outcome that is different from MACE in the development study. In addition, the severity of the composites is very different compared to MACE in the development study. Overall judgement High Justification: patient selection was appropriate. However, there was no information on how predictors were defined. Furthermore, the outcome used was different from MACE in the development study.
Notes	—

Item	Authors' judgement	Support for judgement
Domain 1: Participant selection	Yes	Appropriate participant selection in which patients were selected in whom the RCRI model can be applied.
Domain 2: Predictors	Unclear	No definition for the RCRI items was reported.
Domain 3: Outcome	Yes	Clearly defined outcome definitions and appropriate adjudication of outcomes.
Domain 4: Analysis	Yes	Clear methodology and appropriate number of outcomes.
Overall judgement	Unclear	Patient selection was appropriate, outcome definitions were clearly defined and assessed and proper methodology was used. However, there was no/unclear information on predictor definitions.

*Study characteristics*
General information	Objective Added biomarkers Journal Journal of Vascular Surgery Country The Netherlands Study design Retrospective cohort study
Participants	Number of included patients 500 Surgical specialty Vascular surgery Age Mean 70 years (SD 9.5 years) Male sex 86% High‐risk surgery Not reported Insulin‐dependent diabetes mellitus Not reported History of ischaemic heart disease Not reported History of congestive heart failure 5% History of cerebrovascular events 15% Elevated creatinine 6% 0 RCRI factors 0% 1 RCRI factor 41% 2 RCRI factors 33% 3 or more RCRI factors 26%
Predictors	Predictor 1: AAA size Objective: added biomarker Category: imaging Scale: continuous Threshold: not applicable Assay/device: assessment based on a CTA scan Predictor 2: AAA size + age Objective: added biomarker Category: imaging Scale: continuous Threshold: not applicable Assay/device: assessment based on a CTA scan
Outcome	Outcome category MACE Full outcome definition Cardiovascular death and nonfatal myocardial infarction Prediction horizon 30‐day events
Analysis	Number of outcomes 31 Handling missing data No information on handling missing data Discrimination reported? Yes Calibration reported? No Reclassification reported? No
PROBAST: Applicability	Domain 1: Participant selection Low Justification: patient selection was appropriate and generalisable to the population used in the RCRI development study Domain 2: Predictors Unclear Justification: no information on blinding in retrospective study and no information on the definition of CHF and IHD Domain 3: Outcome High Justification: outcome is composite of cardiovascular death and nonfatal myocardial infarction which differs from outcome in development study Overall judgement High Justification: patients selected were generalisable to the patient population used in the RCRI development study. However, there was no/unclear information on predictor definitions and outcome definition was different compared to the development study.
Notes	—

Item	Authors' judgement	Support for judgement
Domain 1: Participant selection	Yes	Although only patients undergoing vascular surgery were included, participant selection was appropriate and the RCRI model can be applied in these patients.
Domain 2: Predictors	Unclear	No information on blinding in retrospective study and no information on the definition of CHF and IHD.
Domain 3: Outcome	Yes	Clearly defined outcome definitions and appropriate adjudication of outcomes.
Domain 4: Analysis	No	Low number of outcomes, no information on handling missing values and no reporting of calibration/reclassification measures.
Overall judgement	No	Outcome was clearly defined and assessed. Patient selection was appropriate. Predictor definitions were unclear/not reported. Furthermore, the number of outcomes was low, no information on the handling of missing data and inappropriate reporting of performance measures.

*Study characteristics*
General information	Objective Added biomarkers Journal PLOS One Country Germany Study design Prospective cohort study
Participants	Number of included patients 477 Surgical specialty Vascular surgery Age Median 70 years (IQR 63 to 75 years) Male sex 79.9% High‐risk surgery Not reported Insulin‐dependent diabetes mellitus Not reported History of ischaemic heart disease 37.8% History of congestive heart failure Not reported History of cerebrovascular events Not reported Elevated creatinine Not reported 0 RCRI factors 10.7% 1 RCRI factor 40.5% 2 RCRI factors 28.3% 3 or more RCRI factors 20.5%
Predictors	Predictor 1: Copeptin Objective: added biomarker Category: blood Scale: continuous Threshold: not applicable Assay/device: BRAHMS Kryptor Assay (Thermo Fisher scientific, Waltham, MA, USA)
Outcome	Outcome category MACE Full outcome definition Myocardial infarction, cardiac death and any new rise of cardiac troponin prompted by suspicion for an acute coronary syndrome Prediction horizon 30‐day events
Analysis	Number of outcomes 41 Handling missing data No information on handling missing data Discrimination reported? Yes Calibration reported? No Reclassification reported? No
PROBAST: Applicability	Domain 1: Participant selection Low Justification: patient selection was appropriate and generalisable to the population used in the RCRI development study Domain 2: Predictors Unclear Justification: no information on RCRI predictor definitions Domain 3: Outcome High Justification: composite endpoint of MACE is very different from the outcome used in the development study Overall judgement High Justification: patients selected were generalisable to the patient population used in the RCRI development study. However, there was no/unclear information on predictor definitions and outcome definition was different compared to the development study.
Notes	—

Item	Authors' judgement	Support for judgement
Domain 1: Participant selection	Yes	Although only patients undergoing vascular surgery were included, participant selection was appropriate and the RCRI model can be applied in these patients.
Domain 2: Predictors	Unclear	No information on RCRI predictor definitions.
Domain 3: Outcome	Unclear	Individual items of MACE composite are not reported; no information on blinding.
Domain 4: Analysis	No	Low number of outcomes; no information on handling missing outcome; no reporting on calibration/reclassification measures.
Overall judgement	No	Patient selection was appropriate. However, predictor and outcome definitions were unclear/not reported including their assessment. Furthermore, the number of outcomes was low, no information on the handling of missing data and inappropriate reporting of performance measures.

*Study characteristics*
General information	Objective Added biomarkers Journal European Journal of Vascular & Endovascular Surgery Country The Netherlands Study design Retrospective cohort study
Participants	Number of included patients 2642 Surgical specialty Vascular surgery Age Mean 66 years (SD 11 years) Male sex 75% High‐risk surgery 79% Insulin‐dependent diabetes mellitus 15% History of ischaemic heart disease 30% History of congestive heart failure 5% History of cerebrovascular events 31% Elevated creatinine 6% 0 RCRI factors 0% 1 RCRI factor 51% 2 RCRI factors 30% 3 or more RCRI factors 18%
Predictors	Predictor 1: Type of surgery + age + history of hypertension (low, low‐intermediate, high‐intermediate and high risk of surgery; < 55, age 56 to 65, age 66 to 75 and > 70) Objective: added biomarkers Category: patient characteristics Scale: not applicable Threshold: not applicable Assay/device: not applicable
Outcome	Outcome category MACE Full outcome definition Cardiac death, myocardial infarction, coronary revascularisation and heart failure Prediction horizon 30‐day events
Analysis	Number of outcomes 287 Handling missing data No information on handling missing data Discrimination reported? Yes Calibration reported? No Reclassification reported? No
PROBAST: Applicability	Domain 1: Participant selection Low Justification: patient selection was appropriate and generalisable to the population used in the RCRI development study Domain 2: Predictors Low Justification: predictor definitions were clearly defined and comparable to the definitions used in the development study Domain 3: Outcome High Justification: MACE outcome is different from the MACE definition used in the development study Overall judgement High Justification: patient selection was appropriate and predictor definitions were clearly defined and comparable to definitions used in the development study. However, the outcome used was different from MACE in the development study.
Notes	—

Item	Authors' judgement	Support for judgement
Domain 1: Participant selection	Yes	Although only patients undergoing vascular surgery were included, participant selection was appropriate and the RCRI model can be applied in these patients.
Domain 2: Predictors	Yes	Clear (RCRI) predictor definitions were described.
Domain 3: Outcome	Yes	Clearly defined outcome definitions and appropriate adjudication of outcomes.
Domain 4: Analysis	No	No measures of calibration/reclassification were reported and no information on handling missing data.
Overall judgement	No	Patient selection was appropriate. Predictors and outcomes were clearly defined and assessed. However, there was no information on missing data and no calibration/reclassification measures was reported.

*Study characteristics*
General information	Objective Added biomarkers Journal Lancet Country Canada, UK, Australia and New Zealand Study design Prospective cohort study
Participants	Number of included patients 1401 Surgical specialty Vascular surgery Age Median 65 years (IQR 57 to 72 years) Male sex 61% High‐risk surgery Not reported Insulin‐dependent diabetes mellitus Not reported History of ischaemic heart disease 12% History of congestive heart failure 1% History of cerebrovascular events 4% Elevated creatinine Not reported 0 RCRI factors 45% 1 RCRI factor 45% 2 RCRI factors 8% 3 or more RCRI factors Not reported
Predictors	Predictor 1: Peak oxygen consumption Objective: added biomarkers Category: patient characteristic Scale: not applicable Threshold: not applicable Assay/device: as measured during Cardiopulmonary Exercise Testing (CPET) Predictor 2: Anaerobic threshold Objective: added biomarkers Category: patient characteristic Scale: not applicable Threshold: not applicable Assay/device: as measured during Cardiopulmonary Exercise Testing (CPET) Predictor 3: DASI Objective: added biomarkers Category: patient characteristic Scale: not applicable Threshold: not applicable Assay/device: as measured using a questionnaire on functional capacity Predictor 4: NT‐proBNP Objective: added biomarkers Category: blood Scale: continuous Threshold: not applicable Assay/device: Siemens Healthcare Diagnostics, Frimley, UK
Outcome	Outcome category All‐cause mortality or MACE; all‐cause mortality Full outcome definition All‐cause mortality or myocardial infarction Prediction horizon 30‐day events; 1‐year events
Analysis	Number of outcomes 28 deaths or MACE Handling missing data No information on handling missing data Discrimination reported? Yes Calibration reported? No Reclassification reported? Yes
PROBAST: Applicability	Domain 1: Participant selection Low Justification: patient selection was appropriate and generalisable to the population used in the RCRI development study Domain 2: Predictors Unclear Justification: RCRI predictor definitions were not reported Domain 3: Outcome High Justification: outcome used is different from the MACE definition used in the development study Overall judgement High Justification: patients selected were generalisable to the patient population used in the RCRI development study. However, there was no/unclear information on predictor definitions and outcome definition was different compared to the development study.
Notes	—

Item	Authors' judgement	Support for judgement
Domain 1: Participant selection	Yes	Although only patients undergoing vascular surgery were included, participant selection was appropriate and the RCRI model can be applied in these patients.
Domain 2: Predictors	Unclear	RCRI predictor definitions were not reported.
Domain 3: Outcome	Yes	Clearly defined outcome definitions and appropriate adjudication of outcomes.
Domain 4: Analysis	No	Complete case analysis, low number of outcomes and no reporting on calibration measures.
Overall judgement	No	Patient selection was appropriate. Outcomes were clearly defined and assessed. However, predictor definitions were unclear/not reported including their assessment. Furthermore, the number of outcomes was low, complete case analysis and inappropriate reporting of performance measures.

Study	Reason for exclusion
Abbott 2017	No prediction
Abbott 2019	No prediction
Abdelmalak 2018	No external validation of RCRI
Abdullah 2017	No prediction
Abdullaha 2018	No external validation of RCRI
Abelha 2009	No prediction
Abelha 2010	No prediction
Abelha 2012	No prediction
Ackland 2007	No prediction
Ackland 2010	Other prediction model
Ackland 2011	No prediction
Ackland 2018	No prediction
Agarwal 2013	No prediction
Albaladejo 2011	No prediction
Alcock 2012	No prediction
Alcock 2013	No prediction
Alvarez 2016	No prediction
Ambler 2014	No prediction
Andersson 2015	External validation only without added value or comparison
Anghelescu 2018	No external validation of RCRI
Arain 2016	No prediction
Armstrong 2017	No prediction
Azevedo 2017	No prediction
Bae 2014	No external validation of RCRI
Baer‐Bositis 2018	No prediction
Bajaj 2013	No prediction
Bakker 2012	No prediction
Bakker 2013	No prediction
Barisione 2016	No prediction
Barrett 2007	No prediction
Batsis 2009	No prediction
Belmont 2014	No external validation of RCRI
Bertges 2010	No external validation of RCRI in the same cohort
Biccard 2007	No prediction
Biccard 2010	No prediction
Biccard 2012a	No external validation of RCRI
Biccard 2013	Non‐original research (review, comment, guideline etc.)
Biccard 2014	Non‐original research (review, comment, guideline etc.)
Biccard 2015	Non‐original research (review, comment, guideline etc.)
Biteker 2011	No prediction
Biteker 2011a	No prediction
Biteker 2012	No prediction
Biteker 2014	No prediction
Biteker 2014a	No prediction
Bolliger 2009	No external validation of RCRI
Bolliger 2012	No prediction
Borges 2013a	No external validation of the RCRI
Butt 2009	No prediction
Calvillo‐King 2010	No external validation of RCRI
Canter 2008	No external validation of RCRI
Cassagneau 2012	No external validation of RCRI
Chan 2018	Other prediction model
Chang 2019	No external validation of RCRI
Chen 2002	No external validation of RCRI
Christiansen 2017	No external validation of RCRI
Cicarelli 2001	External validation only without added value or comparison
Cloney 2017	No prediction
Cook 2017	No external validation of RCRI
Crowther 2018	No external validation of RCRI
Cullen 2020	No external validation of RCRI
Cuthbertson 2007a	No external validation of RCRI
Cuthbertson 2007b	No external validation of RCRI
Davies 2015	No prediction
Davies 2015a	No prediction
Davis 2018	No prediction
de Campos 2012	No external validation of RCRI
Dernellis 2006	Other prediction model
Devereaux 2011	External validation only without added value or comparison
de Virgilio 2009	No prediction
Dover 2013	External validation only without added value or comparison
Drake 2016	No external validation of RCRI
Drudi 2016	No external validation of RCRI
Duceppe 2018	No external validation of RCRI
Duceppe 2019	No external validation of RCRI
Edelmuth 2018	No external validation of RCRI
Ekeloef 2017	No prediction
Ekeloef 2020	No external validation of RCRI
Ekeloef 2020a	Postoperative biomarker was evaluated
Erol 2019	No prediction
Eyraud 2000	No external validation of RCRI
Faggiano 2012	No prediction
Fayad 2011	No external validation of RCRI
Feringa 2006	No prediction
Feringa 2006a	No prediction
Feringa 2007a	No external validation of RCRI
Feringa 2009	No prediction
Ferrante 2018	No external validation of the RCRI
Filipovic 2003	No prediction
Filipovic 2005	No external validation of RCRI
Flu 2009	No external validation of RCRI
Flu 2010	Other prediction model
Flu 2010a	No external validation of RCRI
Galal 2010	No prediction
Garcia 2009	External validation only without added value or comparison
Garcia 2013	No external validation of RCRI
Ghadri 2012	No external validation of RCRI
Ghazali 2017	No external validation of RCRI
Gibson 2007	No external validation of RCRI
Gillmann 2019	No external validation of RCRI
Go 2017	No external validation of RCRI
Goei 2009	No external validation of RCRI
Goh 2000	No external validation of RCRI
Gómez 2012	No prediction
Goodman 2015	No prediction
Gu 2018	No prediction
Gundes 2017	No prediction
Halm 2005	No prediction
Halm 2009	No prediction
Halm 2009a	No prediction
Hammill 2008	No prediction
Hansen 2016	No prediction
Hanss 2008	RCRI was part of the inclusion criteria
Harland 2020	No external validation of RCRI
Hawn 2013	No prediction
Hennis 2012	No external validation of RCRI
Hietala 2014	No prediction
Hirano 2014	No prediction
Hirpara 2019	No external validation of RCRI
Hoeks 2007	No prediction
Hoeks 2008	No prediction
Hoeks 2009	No prediction
Hoeks 2009a	No prediction
Hoeks 2010	No external validation of RCRI
Hofer 2018	Other prediction model
Hoftman 2013	External validation only without added value or comparison
Hokari 2015	Other prediction model
Holcomb 2016	No prediction
Holcomb 2016a	No prediction
Hollis 2016	No prediction
Huang 2017	No prediction
Jakobson 2014	No prediction
Kamber 2018	No external validation of RCRI
Kanakaraj 2017	Other prediction model
Karakas 2013	Non‐original research (review, comment, guideline etc.)
Kazimierczak 2015	Other prediction model
Kerry 2011	Non‐original research (review, comment, guideline etc.)
Kertai 2004	No prediction
Khambalia 2015	No prediction
Kikura 2008	No prediction
Kim 2013	No prediction
Kim 2016	No external validation of RCRI
Kim 2016a	No external validation of RCRI
Kim 2018	No external validation of RCRI
Kim 2019	No external validation of RCRI
Kistan 2018	No external validation of RCRI
Koh 2012	No prediction
Kougias 2013	No prediction
Kougias 2017	No prediction
Kronzer 2016	Non‐original research (review, comment, guideline etc.)
Kronzer 2016a	No prediction
Kumar 2017	No prediction
Küpper 2015	Other prediction model
Ladha 2018	No external validation of RCRI
Lau 2013	No external validation of RCRI
Lee 1999	Development study, external validation only without added value or comparison
Leibowitz 2009	No prediction
Levitan 2016	No prediction
Li 2016	No external validation of RCRI
Licker 2011	Other prediction model
Licker 2013	No external validation of RCRI
Liem 2018	No external validation of RCRI
Lin 2005	No prediction
Lin 2016	Other prediction model
Lin 2017	No prediction
Lindenauer 2004	No prediction
Lindenauer 2005	No prediction
Liu 2013	No external validation of RCRI
Lo 2014	No prediction
Long 2016	No prediction
Lucreziotti 2007	No prediction
Lupei 2014	No prediction
Maas 2007	No external validation of RCRI
MacIntyre 2018	No prediction
Mahmoud 2016	No prediction
Mann 2020	No external validation of RCRI
Marinho 2018	No prediction
Marsman 2020	No external validation of RCRI
Marston 2013	No prediction
Martins 2011	No external validation of RCRI
Mases 2014	No prediction
Matsumoto 2016	No prediction
May 2019	No prediction
McIlroy 2015	No prediction
Meershoek 2020	No external validation of RCRI
Mendonca 2014	No prediction
Mitropoulos 2006	Other prediction model
Moitra 2011	No external validation of RCRI
Moodley 2015	No prediction
Moodley 2015a	No prediction
Mooney 2016	No prediction
Moran 2008	External validation only without added value or comparison
Moses 2018	Other prediction model
Mureddu 2017	Non‐original research (review, comment, guideline etc.)
Nagayoshi 2012	No prediction
Nepogodiev 2015	Non‐original research (review, comment, guideline etc.)
Noordzij 2010	No external validation of RCRI
Noordzij 2015	No external validation of RCRI
Nordling 2016	No external validation of RCRI
Nutt 2012	No external validation of RCRI
O'Neill 2016	No prediction
Oberweis 2015	No prediction
Ochroch 2006	No prediction
Oliveros 2005	No external validation of RCRI
Oscarsson 2009	No external validation of RCRI
Oscarsson 2009a	No external validation of RCRI
Oshin 2013	No prediction
Padayachee 2018	No prediction
Paladugu 2020	Non‐original research (review, comment, guideline etc.)
Parente 2013	No prediction
Parikh 2020	No external validation of RCRI
Park 2018	No external validation of RCRI
Patel 2018	No prediction
Patorno 2015	No prediction
Patorno 2016	No prediction
Payne 2011	No external validation of RCRI
Payne 2013	External validation only without added value or comparison
Pereira 2016	No external validation of RCRI
Pili‐Floury 2012	No external validation of RCRI
Pinho 2016	No prediction
Puelacher 2018	No prediction
Rajagopalan 2008	No external validation of RCRI
Rao 2012	External validation only without added value or comparison
Redman 2014	No prediction
Reeh 2016	No prediction
Reeve 2018	No prediction
Reis 2018	No external validation of RCRI
Richards 2015	No prediction
Richardson 2018	No external validation of RCRI
Rinfret 2004	No prediction
Rodriguez 2018	No prediction
Rodseth 2014	No external validation of RCRI
Rosenberg 2016	No prediction
Roshanov 2017	Non‐original research (review, comment, guideline etc.)
Roxburgh 2011	No prediction
Sakuma 2010	No prediction
Salinas 2012	No prediction
Sankar 2014	Other prediction model
Sankar 2019	No external validation of RCRI
Schier 2012	No external validation of RCRI
Schier 2013	No prediction
Shalaeva 2016	No external validation of RCRI
Silva 2020	No prediction
Simeoni 2016	No prediction
Skaro 2016	No prediction
Smilowitz 2016	No prediction
Smilowitz 2018	No prediction
Smolock 2012	No prediction
Snowden 2010	No external validation of RCRI
Snowden 2013	No external validation of RCRI
Sousa 2016	No prediction
Stevens 2017	No prediction
Sunny 2018	External validation only without added value or comparison
Tao 2008	Other prediction model
Tashiro 2014	External validation only without added value or comparison
Tavakoli 2009	Other prediction model
Teixeira 2014	No prediction
Toda 2018	No prediction
Tong 2015	Wrong population
Toyonaga 2017	No external validation of RCRI
Valentijn 2013	No prediction
Valentijn 2013a	No prediction
van Kuijk 2009	No prediction
Vanniyasingam 2016	No external validation of RCRI
van Waes 2017	Postoperative biomarker measurement
Vanwagner 2012	No prediction
VanWagner 2014	No external validation of RCRI
Veiga 2012	No prediction
Venkatraghavan 2015	No prediction
Vetrugno 2018	Non‐original research (review, comment, guideline etc.)
Waliszek 2011	No prediction
Ward 2006	No prediction
Warnakulasuriya 2017	No prediction
Weissman 2011	No external validation of RCRI
Widmer 2018	No prediction
Wijeysundera 2010	No prediction
Wijeysundera 2011	No prediction
Wijeysundera 2012	No prediction
Wijeysundera 2020	No external validation of RCRI
Wilson 2010	No external validation of RCRI
Xara 2015	No external validation of RCRI
Yun 2008	No external validation of RCRI
Yurtlu 2016	No prediction

Study name	To predict 30‐day in hospital mortality and morbidity using preoperative hand grip strength and comparing it with existing Revised Cardiac Risk Index and Modified Frailty Index
Starting date	February 2019
Contact information	Kompal Jain, Department of Anaesthesia and Intensive Care, Government Medical College and Hospital, Sector ‐ 32, Chandigarh, India
Notes	To predict 30‐day in hospital mortality and morbidity using preoperative hand grip strength and comparing it with existing Revised Cardiac Risk Index and Modified Frailty Index

Study name	Preoperative biochemical predictors of outcome in patients with hip fracture
Starting date	January 2011
Contact information	Peter Bentzer, MD, PhD, Skane University Hospital
Notes	The objective of the study is to identify biochemical predictors of morbidity and mortality in patients suffering from hip fracture. Biochemical predictors include pro‐brain natriuretic peptide, lactate, pro‐calcitonin, adrenomedullin, copeptin, cystatin c. The predictive value of the potential markers will be compared to that of ASA, RCRI and POSSUM.

Study name	TEAMS (Troponin Elevation After Major Surgery) Study (TEAMS)
Starting date	August 2014
Contact information	University Health Network, Toronto
Notes	This study will compare postoperative health‐related quality of life of patients who did or did not experience perioperative myocardial injury (defined by troponin‐I > 0.07 ng/ml) after noncardiac surgery. Clinically based risk stratification tools used in noncardiac surgery (e.g. Revised Cardiac Risk Index) are of moderate utility and assign patients only to broad risk categories. This study will examine the usefulness of pre‐operative biomarkers (BNP, HbA1c and others) in supporting cardiac risk stratification and will address the question: Is there a set of preoperative criteria that can accurately inform the decision to monitor troponin postoperatively?