Skip to main content

Table 1 The risk model of AKI derived through a hybrid Machine Learning approach

From: Prediction of acute kidney injury risk after cardiac surgery: using a hybrid machine learning algorithm

Characteristic

AOR, 95% CI

β coefficient

Risk score

CARE score—2

No

Ref

Ref

0

Yes

1.95 (1.61–2.36)

0.3284

3

CARE score—3

No

Ref

Ref

0

Yes

2.97 (2.25–3.92)

0.5143

5

CARE score—4

No

Ref

Ref

0

Yes

6.43 (1.58–9.70)

0.9298

8

Combined CABG/valve surgery

No

Ref

Ref

0

Yes

1.44 (1.20–1.73)

0.1854

2

Emergent operative status

No

Ref

Ref

0

Yes

1.54 (1.14–1.89)

0.2080

2

Anemia

No

Ref

Ref

0

Yes

1.77 (1.48–1.97)

0.2852

3

Atrial fibrillation

No

Ref

Ref

0

Yes

1.43 (1.17–1.63)

0.1778

2

NYHA Class 3

No

Ref

Ref

0

Yes

1.35 (1.17–1.63)

0.1501

1

Left ventricle ejection fraction—3

No

Ref

Ref

0

Yes

1.39 (1.11–1.73)

0.1644

1

Redo sternotomy

No

Ref

Ref

0

Yes

1.40 (1.09–1.79)

0.1778

2

Hypertension

No

Ref

Ref

0

Yes

1.43 (1.21–1.73)

0.2102

2

Preoperative intra-aortic balloon pump therapy

No

Ref

Ref

0

Yes

2.98 (1.66–4.89)

0.5486

5

Former smoker

No

Ref

Ref

0

Yes

1.25 (1.08–1.44)

0.1107

1

BMI

1.04 (1.03–1.06)

0.0350

 

 ≤ 18

 

 − 0.385

 − 4

19–24

 

Ref

0

25–29

 

0.1925

2

 ≥ 30

 

0.6125

6

  1. CARE score, Cardiac Anesthesia Risk Evaluation (CARE) mortality risk score; CABG, Coronary artery bypass grafting; NYHA class, New York Heart Association Classification; BMI, body mass index