Skip to main content

Table 3 Multivariable regression models regarding headache outcome

From: MRI features of idiopathic intracranial hypertension are not prognostic of visual and headache outcome

 

Headache improvementa

Freedom of headachea

 

Univariable

Multivariable

Univariable

Multivariable

 

OR (95% CI)b

p-value

OR (95% CI)b

p-value

OR (95% CI)b

p-value

OR (95% CI)b

p-value

Age (per 5 years increase)

0.95 (0.91–0.99)

0.047

0.96 (0.91–1.01)

0.135

0.99 (0.95–1.04)

0.708

not retained

n.a.

BMI (per point)

0.98 (0.92–1.05)

0.619

not retained

n.a.

1.00 (0.94–1.06)

0.991

not retained

n.a.

CSF opening pressure (per 5cmH2O)

1.05 (0.84–1.33)

0.653

not retained

n.a.

1.15 (0.94–1.41)

0.175

not retained

n.a.

Chronic headache at baseline

1.34 (0.95–2.21)

0.105

1.39 (0.91–2.35)

0.197

0.43 (0.13–0.92)

0.021

0.48 (0.17–0.91)

0.013

Baseline headache severity

1.15 (0.68–1.95)

0.588

not retained

n.a.

0.69 (0.36–1.32)

0.258

not retained

n.a.

Empty sella

0.37 (0.11–1.22)

0.103

0.53 (0.15–1.91)

0.328

0.38 (0.13–1.09)

0.070

0.34 (0.11–1.09)

0.070

Optic nerve sheath distension

1.14 (0.34–3.78)

0.836

1.11 (0.32–3.88)

0.868

1.10 (0.39–3.11)

0.861

1.11 (0.39–3.14)

0.849

Optic nerve tortuosity

1.70 (0.52–5.59)

0.382

1.85 (0.54–6.29)

0.326

0.74 (0.28–1.98)

0.546

0.75 (0.28–2.01)

0.564

Posterior globe flattening

5.20 (0.64–42.4)

0.124

5.14 (0.62–42.5)

0.129

1.18 (0.39–3.54)

0.775

1.18 (0.39–3.56)

0.770

Transverse sinus stenosis c

2.47 (0.70–8.75)

0.162

2.55 (0.69–9.45)

0.160

0.42 (0.15–1.21)

0.109

0.43 (0.16–1.22)

0.111

≥ 1 MRI feature

1.00 (0.25–4.05)

0.999

1.35 (0.31–5.83)

0.687

0.90 (0.28–2.91)

0.863

0.94 (0.29–3.09)

0.918

≥ 3 MRI features

0.71 (0.19–2.62)

0.606

0.94 (0.24–3.69)

0.925

0.42 (0.15–1.21)

0.109

0.38 (0.13–1.15)

0.086

  1. BMI Body mass index, CSF Cerebrospinal fluid, MRI Magnetic resonance imaging, N.a not applicable, OR Odds ratio, VAS Visual analogue scale, 95% CI 95% confidence interval. aCalculated by Firth's bias-reduced multivariable binary logistic regression models with headache improvement/ freedom of headache as dependent variable and MRI features as independent variables (not present [reference category] vs. present). Corrected Akaike information criterion (AICc) used to select the most parsimonious models, i.e. which variables were retained. bValues above/below 1 indicate higher/lower probability of headache improvement/ freedom of headache. conly available for 70 patients