Skip to main content

Table 2 Comparison of CSF variable variances between repositioned and non-repositioned samples and medians between serial samples

From: Cerebrospinal fluid cell count variability is a major confounding factor in external ventricular drain-associated infection surveillance diagnostics: a prospective observational study

 

R-samples

C-samples

P-value

Statistical test

2.1 Outcome Variance R-samples versus C-samples

Granulocytes, \(10^6\)

    

 \(\Delta\), Range

− 3505 , 619

− 77 , 757

  

 \(\Delta\), IQR

[− 4.5 , 50.5]

[− 1 , 2.5]

\(<{\mathbf .001}\)

Fligner Killeen

 Relative \(\Delta\), IQR

[− 31 , 80]

[− 10 , 15]

\(<{\mathbf .001}\)

Fligner Killeen

 Absolute \(\Delta\), IQR

[3 , 95.5]

[0 , 8.5]

\(<{\mathbf .001}\)

Fligner Killeen

Erythrocytes, \(10^6\)

    

 \(\Delta\), Range

− 271,500 , 331,700

− 47,500 , 121,100

  

 \(\Delta\), IQR

[− 2450 , 13,150]

[− 400 , 2450]

\(<{\mathbf .001}\)

Fligner Killeen

 Relative \(\Delta\), IQR

[− 23 , 56]

[− 10 , 20]

\(<{\mathbf .001}\)

Fligner Killeen

 Absolute \(\Delta\), IQR

[900 , 20,050]

[250 , 4950]

\(<{\mathbf .001}\)

Fligner Killeen

Albumin, mg/L

    

 \(\Delta\), Range

− 222 , 603

− 68 , 378

  

 \(\Delta\), IQR

[− 3 , 72]

[− 1 , 28]

\(<{\mathbf .01}\)

Fligner Killeen

 Relative \(\Delta\), IQR

[− 1.9 , 20]

[− 0.5 , 11]

.02

Fligner Killeen

 Absolute \(\Delta\), IQR

[8 , 90.5]

[3 , 30.5]

\(<{\mathbf .001}\)

Fligner Killeen

Lactate, mmol/L

    

 \(\Delta\), Range

− 1.1 , 0.8

− 0.3 , 0.4

  

 \(\Delta\), IQR

[− 0.1 , 0.1]

[0 , 0.05]

\(<{\mathbf .01}\)

Fligner Killeen

 Relative \(\Delta\), IQR

[− 3.2 , 4]

[0 , 1.1]

\(<{\mathbf .01}\)

Fligner Killeen

 Absolute \(\Delta\), IQR

[0.1 , 0.2]

[0 , 0.1]

.1

Fligner Killeen

Glucose, mmol/L

    

 \(\Delta\), Range

− 0.6 , 0.8

− 1.6 , 0.2

  

 \(\Delta\), IQR

[− 0.2 , 0.1]

[− 0.1 , 0]

\(<{\mathbf .01}\)

Fligner Killeen

 Relative \(\Delta\), IQR

[− 2.7 , 2.4]

[− 2.1 , 0]

\(<{\mathbf .01}\)

Fligner Killeen

 Absolute \(\Delta\), IQR

[0.1 , 0.2]

[0 , 0.1]

.04

Fligner Killeen

LE Ratio

    

 \(\Delta\), Range

− 72.1 , 40.1

− 2.1 , 3.9

  

 \(\Delta\), IQR

[− 0.55 , 0.64]

[− 0.25 , 0.38]

.02

Fligner Killeen

 Relative \(\Delta\), IQR

[− 15 , 35]

[− 19 , 17]

.06

Fligner Killeen

 Absolute \(\Delta\), IQR

[0.15 , 2]

[0.066 , 1.3]

\(<{\mathbf .01}\)

Fligner Killeen

 

R-Samples

C-Samples

 

Sample 1

Sample 2

P-value

Sample 1

Sample 2

P-value

Statistical test

2.2 Outcome Median Sample 1 versus Sample 2

Granulocytes, \(10^6\)

40

54

.33

13

14

.81

Clustered Signed rank

Erythrocytes, \(10^6\)

22,100

22,300

.25

7700

6700

.37

Clustered Signed rank

Albumin, mg/L

318

333

\(<{\mathbf .001}\)

235

251

.02

Clustered Signed rank

Lactate, mmol/L

3.2

3.3

.56

3.1

3.1

.53

Clustered Signed rank

Glucose, mmol/L

4.5

4.5

.19

4.5

4.5

.35

Clustered Signed rank

LE ratio

2.16

2.24

.25

2.51

2.31

.76

Clustered Signed rank

  1. Data are presented as IQR (square brackets) and in the case of the \(\Delta\), the range is also included (no brackets) in 2.1. Data are presented as medians in 2.2. Sample medians were evaluated with the clustered Wilcoxon signed rank, clustered by Patient ID. Variances were evaluated with the Fligner–Killeen test. The results of \(p<.01\) are bolded. As patients are represented in the R-samples and C-samples unequally, mixed effects regression models are required for further interpretation of this data. \(\Delta\): pair-wise difference, including absolute and relative derivatives. R-samples: Repositioned samples. C-Samples: Non-repositioned control samples