Patient characteristics
A total of 25 nonconsecutive patients were screened. Five patients were not included due to right ventricular dysfunction (1 patient), ejection fraction less than 50% (1 patient) and presence of arrythmia (3 patients). Two patients were excluded due to absence of fluid expansion performed after LRM. A total of 18 patients were analyzed (Figure 2). Nine patients (50%) were responders to volume expansion and nine were not. Patient main characteristics, hemodynamic, respiratory, pharmacological and biological variables in both Responders and Non-Responders are shown in Table 1. The PPV value was not calculated for 8 patients.
The baseline norepinephrine concentration was higher in the Responders group (R) than in the Non-Responders (NR) group (0.78 ug/kg/min vs 0.24 ug/kg/min, respectively; p=0.031). Baseline PPV was higher in the R than in the NR group (14% vs 6%, respectively; p=0.034). R and NR did not differ for baseline values of SV, CO, MAP, CVP or lactate level. Static pulmonary compliance was not statistically different between R and NR.
Table 1
Patient demographic data, hemodynamic, respiratory, pharmacological and biological variables at baseline in Responders (n=9) and Non-Responders (n=9 patients)
|
Overall population
|
Responders
|
Non-Responders
|
p value
|
|
N=18
|
N=9
|
N=9
|
|
Age (mean SD), yr
|
60 (15)
|
66 (7)
|
57 (18)
|
|
Gender (M/F)
|
10/8
|
5/4
|
5/4
|
|
BMI (mean SD), kg.m-2
|
29 (9)
|
29 (6)
|
29 (12)
|
|
Ideal body weight (mean SD), kg
|
66 (14)
|
63 (11)
|
66 (17)
|
|
SAPS 2 (mean SD)
|
61 (22)
|
69 (22)
|
57 (21)
|
|
ICU admission criteria
|
|
|
|
|
Septic shock
|
4
|
3
|
1
|
|
Cardiac failure
|
7
|
3
|
4
|
|
Respiratory failure
|
4
|
2
|
2
|
|
Hemorrhagic shock
|
1
|
1
|
0
|
|
Polytrauma
|
1
|
0
|
0
|
|
Cranio-cerebral trauma
|
1
|
0
|
1
|
|
Comorbidities
|
|
|
|
|
Arterial hypertension
|
8
|
6
|
2
|
|
Diabetes
|
4
|
2
|
2
|
|
Coronary artery disease
|
1
|
1
|
0
|
|
Hemodynamic parameters
|
|
|
|
|
Mean arterial pressure (mean SD), mmHg
|
73 (9)
|
70 (10)
|
76 (7)
|
0.110
|
Heart Rate (mean SD), HR/min
|
90 (18)
|
91 (23)
|
87 (11)
|
0.813
|
Stroke volume (mean SD), mL
|
62 (26)
|
53 (20)
|
64 (29)
|
0.177
|
Cardiac output (mean SD), L/min
|
5.3 (2.2)
|
4.5 (1.0)
|
5.5 (2.7)
|
0.150
|
PPV (mean SD), %
|
10 (8)
|
14 (10)
|
6 (2)
|
0.034
|
Central Venous Pressure (mean SD), mmHg
|
10 (4)
|
9 (3)
|
10 (4)
|
0.238
|
Respiratory parameters
|
|
|
|
|
Tidal volume (mean SD), mL
|
436 (62)
|
452 (55)
|
405 (68)
|
0.288
|
Respiratory Rate (mean SD), RR/min
|
20 (5)
|
18 (4)
|
22 (6)
|
0.218
|
Positive end expiratory pressure (mean SD), cmH2O
|
10 (3)
|
9 (2)
|
10 (3)
|
0.671
|
Plateau pressure (mean SD), cmH2O
|
22 (6)
|
19 (3)
|
23 (8)
|
0.099
|
Static pulmonary compliance (mean SD), mL/cmH2O
|
42 (15)
|
48 (11)
|
37 (15)
|
0.065
|
P/F (mean SD)
|
236 (97)
|
224 (104)
|
242 (97)
|
0.679
|
Neuromuscular blockade, n
|
9
|
4
|
5
|
|
Pharmacological parameters
|
|
|
|
|
Norepinephrine (mean SD), ug/kg/min
|
0.50 (0.64)
|
0.78 (0.76)
|
0.24 (0.35)
|
0.031
|
Dobutamine (mean SD), ug/kg/min
|
2.06 (3.37)
|
1.44 (2.96)
|
2.67 (3.81)
|
0.394
|
Biological parameters
|
|
|
|
|
Lactates (mean SD), mmol/l
|
2.2 (1.5)
|
2.7 (2.1)
|
2.0 (0.8)
|
0.215
|
Prediction of Fluid Responsiveness
SV values for baseline, maximum lung recruitment pressure, and before and after volume expansion are represented in figure 3.
Six data points were available for each parameter for STEP-UP and STEP-DOWN. Figure 4 shows individual variations in hemodynamic parameters during LRM. STEP-UP LRM induced a decrease of SAP, PP, DAP, MAP, and SV and an increase of CVP. Fluid Responders demonstrated a greater decrease of SAP, PP, DAP, MAP, and SV as well as a greater increase of CVP compared to Non-Responders.
Evaluation of all hemodynamic variables suggested the use of a linear regression model, especially during STEP-UP LRM. Slope calculations are shown in Table 2. Slopes are reported as αSAP, αDAP and αMAP for systolic, diastolic and mean arterial pressure, αPP for pulse pressure, αCVP for central venous pressure and αSV for stroke volume.
Slope calculations showed greater absolute values for the Responder compared to the Non-Responder group for STEP-UP and STEP-DOWN LRM (Table 2). The ability of αSAP, αPP, αDAP, αMAP, αSV and αCVP to predict fluid responsiveness and the results of AUC analysis are shown in Table 3. The best predictive variables for fluid responsiveness during LRM were αPP and αCVP during STEP-UP, with Youden indices of 0.888 and 0.777 respectively.
αPP during STEP-UP was strongly predictive of fluid responsiveness with an AUC of 0.926 (95% CI, 0.78 to 1.00), and a sensitivity and a specificity of 100% and 89% respectively. Cut-off value was -42.8°. Inconclusive values ranged from -42.8° to -52.1° using the grey zone approach (35% of the patients).
αCVP during STEP-UP was also strongly predictive of fluid responsiveness with an AUC of 0.901 (95% CI, 0.76 to 1.00) and a sensitivity and a specificity of 78% and 100% respectively. The cut-off value was 20.1°. Inconclusive values ranged from 13.8° to 20.1° using the grey zone approach (44% of the patients).
By combining sensitivity of αPP and specificity of αCVP if both measures are available, taking for each angle the specificity value of αCVP and sensitivity value of αPP, fluid responsiveness prediction can be obtained with 100% sensitivity and 100% specificity during STEP-UP LRM (AUC=0.96; 95% CI, 0.90 to 1.00). One patient (5.5%) showed inconclusive values using the grey zone approach (Figure 4). Index combination has already been proposed to optimize the sensitivity and specificity of a parameter (23).
Absolute variations for SAP, PP, MAP, DAP, SV and CVP between baseline (PEEP=5mmHg, Inspiratory pressure=20mmHg) and maximum pressure level (PEEP=30mmHg, Inspiratory pressure=45mmHg) are reported as ∆SAP, ∆PP, ∆MAP, ∆DAP, ∆SV and ∆CVP. Their ability to predict fluid responsiveness is shown in table 4.
PPV at baseline was available for 11 patients. A PPV of more than 12% before LRM predicted responders with an AUC of 0.711 (95% CI, 0.42 to 1.00) and a sensitivity and a specificity of 63% and 100% respectively. Inconclusive values ranged from 4–10% using the grey zone approach (45% of the patients) (Figure 5).
Table 2
Slopes (degrees) for systolic arterial pressure (αSAP), pulse pressure (αPP), diastolic arterial pressure (αDAP), mean arterial pressure (αMAP), stroke volume (αSV) and central venous pressure (αCVP) for STEP-UP and STEP-DOWN during lung recruitment maneuver
|
STEP-UP
|
STEP-DOWN
|
|
Responders
|
Non-Responders
|
p value
|
Responders
|
Non-Responders
|
p value
|
|
N=9
|
N=9
|
|
N=9
|
N=9
|
|
α SAP (SD)
|
-59.4° (5.79)
|
-37.5° (13.1)
|
0.022
|
57.1° (8.25)
|
40.8° (12.6)
|
0.034
|
α PP (SD)
|
-51.9° (7.24)
|
-31.6° (10.7)
|
0.014
|
48.8° (8.26)
|
31.8° (10.6)
|
0.010
|
α DAP (SD)
|
-23.6° (8.77)
|
-11.4° (10.0)
|
0.126
|
24.3° (9.44)
|
18.5° (10.1)
|
0.528
|
α MAP (SD)
|
-40.3° (8.50)
|
-24.2° (11.9)
|
0.089
|
39.6° (9.51)
|
30.1° (11.5)
|
0.228
|
α SV (SD)
|
-43.5° (9.65)
|
-27.4° (8.64)
|
0.038
|
25.9° (20.1)
|
29.0° (9.60)
|
0.837
|
α CVP (SD)
|
19.8° (2.66)
|
13.1° (3.03)
|
0.006
|
-18.2° (2.05)
|
-13.4° (2.60)
|
0.003
|
Table 3
Diagnostic performance of slopes for systolic arterial pressure (αSAP), pulse pressure (αPP), diastolic arterial pressure (αDAP), mean arterial pressure (αMAP), stroke volume (αSV), central venous pressure (αCVP) and relative variations from baseline of systolic arterial pressure (∆SAP), pulse pressure (∆PP), diastolic arterial pressure (∆DAP), mean arterial pressure (∆MAP), stroke volume (∆SV) and central venous pressure (∆CVP) between baseline (PEEP=5, Inspiratory Pressure=20cmH2O) and maximum pressure level (PEEP=30, Inspiratory Pressure=45cmH2O) to predict fluid responsiveness during STEP-UP and STEP-DOWN lung recruitment maneuver
STEP-UP
|
αSAP
|
αPP
|
αDAP
|
αMAP
|
αSV
|
αCVP
|
Cut-off value (degrees)
|
-47.8°
|
-42.8°
|
-10.1°
|
-20.1°
|
-34.2°
|
20.1°
|
ROC AUC
|
0.864
|
0.926
|
0.765
|
0.777
|
0.854
|
0.901
|
Sensitivity
|
1
|
1
|
1
|
1
|
0.875
|
0.777
|
Specificity
|
0.666
|
0.888
|
0.666
|
0.666
|
0.777
|
1
|
Positive predictive value
|
0.75
|
0.9
|
0.75
|
0.75
|
0.777
|
1
|
Negative predictive value
|
1
|
1
|
1
|
1
|
0.875
|
0.818
|
Youden index
|
0.666
|
0.888
|
0.666
|
0.666
|
0.653
|
0.777
|
Grey zone (degrees)
|
[47.8°-59.6°]
|
[42.8°-52.1°]
|
[10.1°-34.2°]
|
[20.1°-42.8°]
|
[24.3°-46.1°]
|
[13.8°-20.1°]
|
|
∆SAP
|
∆PP
|
∆DAP
|
∆MAP
|
∆SV
|
∆CVP
|
Cut-off value (mmHg)
|
23
|
21
|
4
|
9
|
16
|
8
|
ROC AUC
|
0.901
|
0.920
|
0.777
|
0.790
|
0.753
|
0.883
|
Sensitivity
|
1
|
1
|
1
|
1
|
0.777
|
0.777
|
Specificity
|
0.666
|
0.777
|
0.666
|
0.666
|
0.777
|
0.888
|
Positive predictive value
|
0.75
|
0.818
|
0.75
|
0.75
|
0.777
|
0.875
|
Negative predictive value
|
1
|
1
|
1
|
1
|
0.777
|
0.8
|
Youden index
|
0.666
|
0.777
|
0.666
|
0.666
|
0.555
|
0.666
|
Grey zone (mmHg)
|
[23-39]
|
[21-28]
|
[4–14]
|
[9–23]
|
[0-25]
|
[5–9]
|
STEP-DOWN
|
αSAP
|
αPP
|
αDAP
|
αMAP
|
αSV
|
αCVP
|
Cut-off value (degrees)
|
55.4°
|
47.8°
|
20.4°
|
34.7°
|
43.2°
|
-14.1°
|
ROC AUC
|
0.777
|
0.815
|
0.666
|
0.685
|
0.666
|
0.877
|
Sensitivity
|
0.777
|
0.666
|
0.777
|
0.777
|
0.375
|
0.888
|
Specificity
|
0.777
|
0.888
|
0.666
|
0.666
|
1
|
0.777
|
Positive predictive value
|
0.777
|
0.857
|
0.7
|
0.7
|
1
|
0.8
|
Negative predictive value
|
0.777
|
0.727
|
0.75
|
0.75
|
0.643
|
0.875
|
Youden index
|
0.555
|
0.555
|
0.444
|
0.444
|
0.375
|
0.666
|
Grey zone (degrees)
|
[38.7°-67.8°]
|
[32.4°-52.3°]
|
[3.9°-36.0°]
|
[9.76°-50.6°]
|
[13.2°-43.2°]
|
[13.8°-20.1°]
|
|
∆SAP
|
∆PP
|
∆DAP
|
∆MAP
|
∆SV
|
∆CVP
|
Cut-off value (mmHg)
|
32
|
21
|
8
|
17
|
15
|
5
|
ROC AUC
|
0.809
|
0.852
|
0.741
|
0.735
|
0.642
|
0.888
|
Sensitivity
|
0.777
|
0.777
|
0.777
|
0.777
|
0.666
|
1
|
Specificity
|
0.888
|
0.888
|
0.777
|
0.777
|
0.666
|
0.666
|
Positive predictive value
|
0.875
|
0.875
|
0.777
|
0.777
|
0.666
|
0.75
|
Negative predictive value
|
0.8
|
0.8
|
0.777
|
0.777
|
0.666
|
1
|
Youden index
|
0.666
|
0.666
|
0.555
|
0.555
|
0.333
|
0.666
|
Grey zone (mmHg)
|
[11-46]
|
[13-30]
|
[0-16]
|
[4-29]
|
[0-25]
|
[5–8]
|