Table 5 Results from Gamma GAMMs for the longitudinal effects of CVP on physiological parameters.

From: Intraoperative B-line scores increase over time and are minimally affected by central venous pressure during transurethral resection of the prostate

Model

Random intercept variance

Adjusted pseudo-R2

log-likelihood (REML)

Time effect

CVP effect

Interaction effect (time × CVP)

Dependent variable

Independent variable

PaO₂/FiO₂ ratio

CVP

 < 0.001

0.384

788.001

F = 166.789, P < 0.001

F = 2.785, P = 0.096

F = 6.322, P = 0.012

Hb

CVP

 < 0.001

0.155

987.000

F = 105.675, P < 0.001

F = 3.811, P = 0.052

F = 10.229, P < 0.001

Ppeak

CVP

0.001

0.281

743.064

F = 92.971, P < 0.001

F = 4.002, P = 0.046

F = 3.660, P = 0.008

Na+

CVP

 < 0.001

0.073

1441.777

F = 38.033, P < 0.001

F = 0.001, P = 0.980

F = 5.528, P < 0.001

K+

CVP

 < 0.001

0.283

698.027

F = 118.589, P < 0.001

F = 7.419, P = 0.007

F = 2.446, P = 0.119

  1. These models were adjusted for age, weight, height, effective tidal volume, and intraoperative use of furosemide. Random effects included only a patient-specific random intercept (random = list(patient_id =  ~ 1)). Pseudo-R2 (deviance explained) was computed as \(1-\frac{\text{residual deviance}}{\text{null deviance}}\), representing the proportion of deviance explained by the model. Models were fitted using restricted maximum likelihood (REML). GAMM, generalized additive mixed model; Hb, hemoglobin; Ppeak, peak airway pressure; Na⁺, arterial sodium concentration; K⁺, arterial potassium concentration.
Back to article page