Table 3 Model results for estimating the cortisol profile during HD in 28 study participants without Synacthen. Estimated effects from univariate models and a final multivariable linear additive mixed model were shown. Multiple measurements during HD per subject were accounted for including a random intercept for subjects. Serum cortisol concentration is modelled on the log-scale accounting for patient and HD variables. Time during HD is assumed to be non-linear and fitted as restricted cubic splines

Patient variables at start of HD

Age (years)

0.003 (-0.007, 0.013) [0.55]

Gender:

- Male

- Female

(reference)

-0.04 (-0.26, 0.18) [0.72]

Dry weight (kg)

0.002 (-0.003, 0.006) [0.48]

Albumin (g/l)

0.008 (-0.02, 0.03) [0.57]

0.009 (-0.01, 0.03) [0.43]

Patient variables varying over HD time

Loss of weight between start and end of HD (kg)

-0.04 (-0.12, 0.04) [0.35]

Glucose (mmol/l)

0.007 (-0.02, 0.04) [0.68]

-0.010 (-0.04, 0.02) [0.46]

Systolic blood pressure*

-0.003 (-0.007, -0.0002) [0.04]

-0.005 (-0.007, -0.002) [0.002]

Diastolic blood pressure*

0.001 (-0.005, 0.006) [0.79]

Daytime of HD:

- Morning

- Afternoon

(reference)

-0.27 (-0.44, -0.09) [0.006]

(reference)

-0.26 (-0.44, -0.09) [0.005]

Ultrafiltration amount

-0.016 (-0.04, 0.009) [0.021]

HD time (min)

- 0

- 30 vs. 0

- 60 vs. 0

- 240 vs. 0

(reference)

-0.22 (-0.29, -0.14)

-0.4 (-0.53, -0.27)

-0.35 (-0.49, -0.22)

[< 0.001]

HD time (min) at morning:

- 0

- 30 vs. 0

- 60 vs. 0

- 240 vs. 0

(reference)

-0.2 (-0.31, -0.09)

-0.38 (-0.58, -0.17)

-0.50 (-0.70, -0.29)

[< 0.001]

HD time (min) at afternoon:

- 0- 30 vs. 0- 60 vs. 0- 240 vs. 0

(reference)-0.21 (-0.31, -0.11)-0.39 (-0.58, -0.21)-0.26 (-0.45, -0.08)

[< 0.001]

Random effects

Between-subject variability (σ _b )

Within-subject variability (σ _ε )

0.18 (0.11, 0.29)

0.26 (0.21, 0.31)