Skip to main content
. 2021 May 26;36(3):879–888. doi: 10.1007/s10877-021-00721-0

Table 2.

Cardiac index measurements by Starling SV compared to bolus thermodilution technique with a pulmonary artery catheter

Starling SV All
n = 579
Phase 1
n = 24
Phase 2
n = 27
Phase 3
n = 85
Phase 4
n = 104

Bias

(L min−1 m−2)

(95% CI)

0.13

(0.07 to 0.18)

0.21

(− 0.04 to 0.46)

0.21

(− 0.07 to 0.49)

− 0.36

(− 0.49 to − 0.24)

− 0.02

(− 0.16 to 0.12)

LOA lower

(L min−1 m−2)

(95% CI)

− 1.23

(− 1.45 to − 1.01)

− 0.83

(− 1.26 to − 0.41)

− 1.08

(− 1.61 to − 0.56)

− 1.45

(− 1.82 to − 1.08)

− 1.34

(− 1.79 to − 0.9)

LOA upper

(L min−1 m−2)

(95% CI)

1.51

(1.29 to 1.73)

1.38

(0.95 to 1.8)

1.57

(1.05–2.1)

0.92

(0.54 to 1.29)

1.44

(1.0 to 1.88)

Percentage error

(95% CI)

60.7%

(51.5 to 69.9)

46.0%

(31.3 to 60.7)

60.0%

(42.9 to 77.1)

53.2%

(32.1 to 74.3)

65.6%

(43.6 to 87.6)

Regression coefficient

(L min−1 m−2)

(95% CI)

0.34

(0.21 to 0.47)

0.39

(− 0.02 to 0.79)

0.31

(− 0.31 to 0.93)

0.28

(− 0.09 to 0.64)

0.09

(− 0.26 to 0.43)

Error grid
 Zone 1 29.0% 28.4% 26.0%
 Zone 2 15.3% 18.5% 15.4%
 Zone 3 34.3% 37.0% 30.8%
 Zone 4 21.4% 16.0% 27.9%

Phase 1 is before induction of anesthesia, phase 2 is during the construction of the proximal anastomoses to the ascending aorta, phase 3 is distal coronary anastomoses, phase 4 is postoperatively in the intensive care unit before extubation. Error grid analysis was not made for the phases 1 and 2 because of the small number of measurements. n number of CI measurements during the phase. LOA limits of agreement