The final study cohort comprised 97 patients with IIH. Baseline characteristics of the study cohort are shown in Table 1.
Forty-five patients (46.4%) were classified as IIH-MIG and 52 (53.6%) as IIH-noMIG including 11.3% with tension-type headache, 25.8% with unclassifiable headache and 16.5% without reported headache. Twenty-two patients (22.7%) had a preexisting diagnosis of migraine.
At baseline, the IIH-MIG group displayed significantly higher headache frequency than IIH-noMIG (22 vs. 15 headache days per month, p = 0.003) as well as higher headache severity (6.5/10 vs. 4.5/10; p < 0.001). However, these differences did not persist when removing patients without headache from the IIH-noMIG group (headache days per month: 21, headache severity: 6.5/10, p = 0.9, respectively).
IIH-MIG and IIH-noMIG did not differ regarding sex (93.3% vs. 84.6% female, p = 0.2), mean age (30.9 vs. 34.1 years; p = 0.2), median BMI (31.5 vs. 32.4, p = 0.9), median CSF opening pressure (31.5 vs. 30.5 cmH2O, p = 0.6) and baseline visual impairment (62.2% vs. 61.5%, p = 0.9), while the proportion of IIH-WOP was higher in the IIH-MIG group (5.1% vs. 1.0%, p = 0.093).
Median weight loss at follow-up amounted to 7kg (IQR 1–15) equaling 6.9% from baseline with no significant difference between IIH-MIG and IIH-noMIG (8kg vs. 7kg, p = 0.9). All patients in the study cohort received pharmacological treatment with acetazolamide (97/97 [100%]; median maximum dosage of 750mg, IQR 500–1000) or topiramate (13.4%, median maximum dosage 62.5mg, IQR 50–75). Neither rate nor dosage of pharmacological treatment significantly differed between IIH-MIG and IIH-noMIG. Overall, 12.4% received VP shunt. Again, there were no significant differences between IIH-MIG (13.3%) and IIH-noMIG (11.5% [p = 0.9]).
Resolution of papilledema was achieved in 60 of the 97 IIH patients (61.9%) with 64.4% of IIH-MIG patients compared to 59.6% of IIH-noMIG patients (p = 0.7).
|
|
(n = 97)
|
Females
|
n (%)
|
86 (88.7)
|
Age at diagnosis
|
mean (SD)
|
32.9 (11.1)
|
Diagnosis
|
|
|
IIH with papilledema
|
n (%)
|
91 (93.8)
|
IIH without papilledema
|
n (%)
|
6 (6.2)
|
BMI
|
median (range)
|
32.0 (17.3–60.9)
|
Overweight (BMI > 25)
|
n (%)
|
10 (89.7)
|
Obesity (BMI > 30)
|
n (%)
|
55 (56.7)
|
Headache
|
|
|
Migraine
|
n (%)
|
45 (46.4)
|
No migraine
|
n (%)
|
52 (53.6)
|
Tension-type headache
|
n (%)
|
11 (11.3)
|
Unclassifiable
|
n (%)
|
25 (25.8)
|
No headache
|
n (%)
|
16 (16.5)
|
Headache days per month
|
median (range)
|
18 (0–30)
|
Headache severity
|
median (range)
|
5.5/10 (0/10–10/10)
|
Visual impairment
|
n (%)
|
60 (61.9)
|
Visual acuity
|
|
|
Abnormal
|
n (%)
|
17 (17.5)
|
Worse eye (logMAR)3
|
median (range)
|
0 (-0.1–1.0)
|
Perimetry
|
|
|
Abnormal
|
n (%)
|
58 (59.8)
|
Mean deviation in perimetry of worse eye (dB)2
|
mean (SD)
|
-8.2 (10.2)
|
Frisén-Scale3
|
median (range)
|
3 (0–5)
|
Lumbar puncture opening pressure3
|
median (range)
|
31 (25–50)
|
Abbreviations: BMI: Body mass index. dB: decibels. IIH: idiopathic intracranial hypertension. logMAR: logarithm of the minimum angle of resolution
Headache outcome
In the whole cohort, headache improvement and freedom of headache were achieved in 78.4% and 27.8%, respectively.
IIH-MIG compared to IIH-noMIG showed significantly lower rates for headache improvement (66.7% vs 88.5%, p=0.009) and freedom of headache (11.1% vs 42.3%, p=0.006) (Figure 1 A-B). These differences remained significant when removing patients without headache (n=16) from the IIH-noMIG group for both headache improvement (66.7% vs 88.9%, p=0.019) and freedom of headache (11.1% vs 33.3%, p=0.015).
Resolution of papilledema was not significantly associated with headache outcome. When only analyzing patients with resolved papilledema at follow-up (n=60), IIH-MIG had less headache improvement than IIH-noMIG (62.1% vs 93.6%, p=0.003) and the difference in freedom of headache was even more pronounced (17.2% vs 61.3%, p<0.001) (Figure 1 C-D).
Weight loss was positively correlated with headache improvement (Spearman rho 0.231, p<0.001) and freedom of headache (Spearman rho 0.197, p=0.097), while neither rate or dosage of pharmacological treatment nor VP shunt displayed univariate correlation with headache outcome below the pre-defined threshold of p<0.2.
In multivariate analyses, IIH-MIG was associated with a significantly lower likelihood of headache improvement (odds ratio [OR] 0.57, p<0.001) and freedom of headache (OR 0.28, p<0.001) explaining 22.1% and 30.1% of variance (Table 3).
Neither age, BMI, IIH-WOP, CSF opening pressure and baseline headache frequency or severity, nor resolution of papilledema were significantly associated with headache outcome. Weight loss remained associated with increased probability of achieving headache improvement (OR 1.97 per 5% weight reduction, p<0.001) but not with freedom of headache.
Sensitivity analyses removing IIH-WOP patients (n=6), patients with preexisting diagnosis of migraine (n=22), and patients without headache (n=16) from the regression models did not significantly change the overall result or impact of single variables.
Table 2. Multivariable regression models regarding headache outcome 12 months after IIH diagnosis.
|
Headache improvement
|
Freedom of headache
|
|
ORa
|
95% CI
|
p-value
|
Change in R2
|
ORa
|
95% CI
|
p-value
|
Change in R2
|
Age (per 5 years increase)
|
0.96
|
0.84 – 1.09
|
0.488
|
.043
|
0.90
|
0.79 – 1.03
|
0.137
|
.051
|
BMI (per point)
|
0.87
|
0.62 – 1.21
|
0.403
|
.046
|
0.87
|
0.62 – 1.21
|
0.403
|
.046
|
CSF opening pressure (per 5cmH2O)
|
1.18
|
0.85 – 1.64
|
0.345
|
.103
|
1.20
|
0.83 – 1.71
|
0.426
|
.091
|
Baseline headache frequency (per 5 headache days/month)
|
1.21
|
0.87 – 1.62
|
0.321
|
.084
|
0.82
|
0.59 – 1.16
|
0.331
|
.079
|
Baseline headache severity (per VAS point)
|
1.12
|
0.79 – 1.37
|
0.492
|
.047
|
1.04
|
0.63 – 1.27
|
0.631
|
.024
|
Migrainous headache (IIH-MIG)
|
0.57
|
0.31 – 0.82
|
<0.001
|
.221
|
0.28
|
0.11 – 0.48
|
<0.001
|
.301
|
Resolution of papilledemab
|
1.49
|
0.92 – 2.23
|
0.103
|
.104
|
1.36
|
0.84 – 2.01
|
0.221
|
.083
|
Weight loss (per 5%)
|
1.97
|
1.21 – 3.01
|
<0.001
|
.123
|
1.41
|
0.96 – 2.11
|
0.107
|
.104
|
|
R2 overall: 0.771; p<0.001
|
R2 overall: 0.779; p<0.001
|
BMI: body mass index. OR: odds ratio. VAS: visual analogue scale. 95% CI: 95% confidence interval.
Calculated by multivariate binary logistic regression models with IIH-MIG as the dependent variable (with reference to IIH-noMIG) and headache improvement/ freedom of headache as the independent variable adjusted for age, body mass index (BMI), CSF opening pressure, baseline headache frequency, baseline headache severity, and resolution of papilledema as well as any other variable associated with headache outcome parameters at a p-value <0.2 in univariate analyses.
Contribution of variables of interest to explanation of variance was assessed by change in R2 through step-wise removal from the regression models.
aValues above/below 1 indicate higher/lower probability of headache improvement/ freedom of headache
bdefined as Frisén scale of 0 in fundoscopy 12 months after IIH diagnosis with reference to persistent papilledema defined as Frisén scale ≥1.
Visual outcome
Overall, persistent visual impairment occurred in 43.3% and visual worsening in 17.5%. There was no difference between the IIH-MIG and IIH-noMIG groups in the rate of persistent visual impairment (44.4% vs. 42.3%, p=0.8), whereas visual worsening was significantly less common in IIH-MIG (8.9% vs. 25.0%, p=0.037) (Figure 2A-B).
Removal of patients without headache (n=16) from the IIH-noMIG group did not change the distribution of persistent visual impairment (44.4% vs 41.7%, p=0.9) or visual worsening (8.9% vs. 22.2%, p=0.093).
None of the six IIH-WOP patients showed persistent visual impairment (vs. 46.2% in patients with papilledema, p=0.027) or visual worsening (vs. 21.0% in patients with papilledema, p=0.2). When removing IIH-WOP patients, there was still no significant difference between IIH-MIG and IIH-noMIG in persistent visual impairment (15/40 [37.5%] vs. 21/51 [41.2%], p=0.8), while visual worsening remained less frequent in IIH-MIG (4/40 [10.0%] vs 13/51 [25.5%]), although statistical significance was lost (p=0.060) (Figure 2C-D).
In univariate analyses, baseline visual impairment was significantly associated with higher persistent visual impairment (60% vs. 16.2%, p<0.001) but not with visual worsening (p=0.4). CSF opening pressure correlated with both persistent visual impairment (Spearman rho 0.292, p<0.001) and visual worsening (Spearman rho 0.365, p<0.001). Resolution of papilledema was not significantly associated with visual impairment (38.3% vs. 51.4%, p=0.2), but with less visual worsening (10.0% vs. 29.7%, p=0.013).
Weight loss was negatively correlated with persistent visual impairment and visual worsening (Spearman rho -0.208 and -0.297, p<0.001 respectively). Rate and dosage of pharmacological treatment was not correlated with visual outcome. Patients receiving a VP shunt had a significantly higher proportion of persistent visual impairment (91.7% vs. 36.4%, p<0.001) and visual worsening (75.0% vs. 9.4%, p<0.001).
Multivariate analysis revealed baseline visual impairment to be significantly associated with an increased probability of persistent visual impairment (OR 3.51, p<0.001, 32.2% of variance explained) (Table 3). VP-shunt also remained significantly associated with visual impairment at follow-up (explaining 9.3% of variance), while IIH-MIG as well as age, BMI, CSF opening pressure, resolution of papilledema and weight loss were not.
In the multivariable model regarding visual worsening, baseline visual impairment was significantly associated with a higher likelihood of visual worsening (OR 2.25, p<0.001, 24.7% of variance explained), while resolution of papilledema (OR 0.55, p=0.031, 10.1% of variance explained) and weight loss (OR 0.74, p=0.028, 9.6% of variance explained) decreased risk of visual worsening. The IIH-MIG group displayed a significantly lower probability of visual worsening (OR 0.39, p<0.001) independently explaining 11.3% of variance.
Of note, IIH-WOP could not be included into the regression models because none of the 6 IIH-WOP patients showed persistent visual impairment or visual worsening. Sensitivity analyses removing IIH-WOP patients did not significantly change the overall results or the impact of single variables regarding persistent visual outcome. When excluding IIH-WOP patients from the visual worsening model, performance of the overall model did not significantly change (R2 0.749 vs. 0.722), but the impact of resolution of papilledema (OR 0.55 vs. OR 0.78) and IIH-MIG (OR 0.39 vs. OR 0.76) was reduced below the level of statistical significance (p=0.2 respectively) indicating that at least part of the effect of both papilledema resolution and migrainous headache phenotype on visual worsening is mediated by IIH-WOP.
Sensitivity analyses removing patients with preexisting diagnosis of migraine (n=22) and patients without headache (n=16) did neither significantly change results of any model of visual outcome nor single variables.
Table 3. Multivariable regression models regarding visual outcome 12 months after IIH diagnosis.
|
Impaired visual outcome
|
Visual worsening
|
|
ORa
|
95% CI
|
p-value
|
Change in R2
|
ORa
|
95% CI
|
p-value
|
Change in R2
|
Age (per 5 years increase)
|
0.97
|
0.93 – 1.02
|
0.236
|
.025
|
1.01
|
0.96 – 1.06
|
0.716
|
.003
|
BMI (per point)
|
0.99
|
0.93 – 1.04
|
0.482
|
.006
|
1.07
|
0.99 – 1.15
|
0.079
|
.083
|
CSF opening pressure (per 5cmH2O)
|
1.35
|
0.92 – 1.92
|
0.136
|
.129
|
1.39
|
0.94 – 2.11
|
0.106
|
.065
|
Visual impairment at baselineb
|
3.51
|
1.83 – 5.82
|
<0.001
|
.322
|
2.25
|
1.26 – 4.15
|
<0.001
|
.247
|
Migrainous headache (IIH-MIG)
|
0.84
|
0.33 – 2.14
|
0.713
|
.004
|
0.39
|
0.11 – 0.84
|
<0.001
|
.133
|
Resolution of papilledemac
|
0.54
|
0.21 – 1.38
|
0.196
|
.071
|
0.55
|
0.15 – 0.97
|
0.031
|
.101
|
Weight loss (per 5%)
|
0.79
|
0.56 – 1.12
|
0.099
|
.087
|
0.74
|
0.52 – 0.96
|
0.028
|
.096
|
Ventriculoperitoneal shunt
|
6.19
|
1.15 – 33.4
|
0.034
|
.093
|
1.24
|
0.20 – 7.64
|
0.814
|
.021
|
|
R2 overall: 0.737; p<0.001
|
R2 overall: 0.749; p<0.001
|
BMI: body mass index. IIH-WOP: idiopathic intracranial hypertension without papilledema. OR: odds ratio. 95% CI: 95% confidence interval.
Calculated by multivariate binary logistic regression models with IIH-MIG as the dependent variable (with reference to IIH-noMIG) and impaired visual outcome/ visual worsening as the independent variable adjusted for age, body mass index (BMI), CSF opening pressure, visual impairment at baseline, and resolution of papilledema as well as any other variable associated with headache outcome parameters at a p-value <0.2 in univariate analyses.
Idiopathic intracranial hypertension without papilledema (IIH-WOP) could not included into the model because none of the 6 IIH-WOP patients showed persistent visual impairment or visual worsening.
Contribution of variables of interest to explanation of variance of the dependent variables was assessed by change in R2 through stepwise removal from the regression models.
aValues above/below 1 indicate higher/lower probability of impaired visual outcome/ visual worsening.
bdefined as baseline visual acuity ≥0.1 logarithm of the minimum angle of resolution (Sloan charts) and/or <-2.0 mean deviation in decibels in static threshold perimetry (30-2 Swedish Interactive Threshold Algorithm) with reference to no visual impairment at baseline.
cdefined as Frisén scale of 0 in fundoscopy 12 months after IIH diagnosis with reference to persistent papilledema defined as Frisén scale ≥1.