Cohort characteristics
The IIH study cohort comprised 191 patients (Fig. 1). Seven patients (4%, n = 7/191) were diagnosed with IIH-WOP and 184 with IIH (96%, n = 184/191). Cohort characteristics at baseline are presented in Table 1. Patients were predominantly female (86.4%, n = 165/191, males: 13.6%, n = 26/191). Mean age at diagnosis was 35.5 ± 12.5 years (range: 14–75). Men were significantly older than women at the time of diagnosis (p = 0.001).
Baseline data on weight and height was available in 99% (n = 189/191) of patients. The mean BMI at baseline was 34.4 ± 8.4 kg/m2 (range: 20–80). Nineteen patients (10%) had a normal BMI (i.e. <25kg/m2) at baseline. Forty patients were overweight (21%, n = 40/189, BMI: 25–30 kg/m2). Out of 69% (130/189) obese patients (BMI > 30 kg/m2), 42 patients had a BMI > 40 kg/m2. Mean CSF-OP at diagnosis was 36.2 ± 8.1 cmCSF. BMI and CSF-OP did not significantly differ between women and men. Baseline BMI and CSF-OP showed a moderate correlation (Pearson correlation, r = 0.325, p = 0.002). Visual disturbance (89%) and headache (84%) were the most common symptoms at baseline. Headache was more frequently reported by women (p = 0.007). Median baseline headache intensity on the NRS was 6 (IQR: 0–8). CSF-OP did not differ between patients with or without headaches (p = 0.347).
Diplopia was reported by 39%, vertigo by 45% of patients and 9% had abducent nerve palsy, each symptom or sign similarly frequent between women and men (Table 1). Headache and visual disturbance were less frequent at follow-up (headache: 68%, p = 0.016, visual disturbance: 55%, p = < 0.001, Wilcoxon-test). Follow-up BMI was only available for 36 out of 90 patients, with the mean follow-up BMI being lower compared to baseline (p = 0.031). Mean CSF-OP at follow-up was lower than at baseline (p < 0.001, follow-up CSF-OP available in 42 of 90 patients).
Table 1. Baseline characteristics of the study cohort.
|
|
all patients
n = 191
|
females
n = 165
|
males
n = 26
|
p-value
|
age at diagnosis
|
35.5 ± 12.5
|
34.4 ± 11.6
|
42.8 ± 15.2
|
0.0011
|
BMI [kg/m2]
|
34.4 ± 8.4
|
34.7 ± 8.3
|
32.4 ± 8.8
|
0.21
|
CSF-OP [cmCSF]
|
36.2 ± 8.1
|
36.3 ± 8.4
|
34.8 ± 7.6
|
0.401
|
symptoms at baseline
headache
- visual disturbance
- abducens palsy
- vertigo
- diplopia
|
150/179 (84%)
155/174 (89%)
18 (9%)
32/71 (45%)
31/80 (39%)
|
137 (83%)
135/152 (89%)
16 (9.7%)
28/63 (44%)
27/72 (38%)
|
13 (50%)
20/22 (90%)
2 (7%)
4/8 (50%)
4/8 (50%)
|
0.0072
1.02
0.92
1.02
0.72
|
ophthalmologic findings
■ BCVA [logMAR]
■ Frisén grade
(median & IQR)
■ optic disc atrophy
|
0.17 ± 0.38
2 (1)
2
|
0.17 ± 0.41
2 (1)
2
|
0.15 ± 0.15
2 (1)
0
|
0.91
-
-
|
1two-sided p-values calculated by unpaired samples t-test.
2two-sided p-values obtained by Fisher’s exact test.
|
Figure 1. Flow chart of the inclusion/exclusion process.
Treatment
For the analysis of visual outcomes, follow-up data of 90 patients was available. During the follow-up period 17% were managed surgically by either implantation of a VP shunt (14%, 13/90) or by bariatric surgery (2%, 2/90). None of the patients were treated by sinus stenting. The group of surgically treated patients constituted of two men and 13 women. Surgical treatment was similarly frequent in men and women (men: 22%, n = 2/9, and women: 16%, 13/81; p = 0.712). Surgical treatment was more frequent in patients with moderate papilledema at baseline (22%) than in patients with mild (5%) and with severe papilledema (14%) at baseline.
Data on medication were available in 83/90 (92%) patients. Of those, 62% were treated pharmacologically (i.e. received either/or a combination of: acetazolamide, topiramate and furosemide). Forty-five percent (54%) of the pharmacologically treated patients received acetazolamide in a median daily dose (MDD) of 1000mg, 23% received topiramate (MDD: 100mg) and 8% furosemide (MDD: 40mg).
Ophthalmologic findings and visual outcome
At baseline 184 of 191 patients had papilledema. In the whole study cohort, median Frisén grade was 2 (range: 0–5, IQR: 2) and median decimal BCVA at baseline was 0.8 (IQR: 0.4). Two patients had optic disc atrophy at baseline.
Assessment of visual outcomes required the availability of ophthalmologic follow-up examinations at a minimum of 6 months after IIH diagnosis. Visual outcomes were available in 90 patients (i.e. the follow-up cohort), which represented 47% of the whole study cohort. Patients who were lost to follow up were older than patients in the follow-up cohort. Other baseline characteristics (sex, BMI, CSF-OP, BCVA, Frisén score) of the patients who were lost to follow-up did not significantly differ from those in the follow-up cohort (Supplemental Table 2).
The median time between baseline to follow-up was 3.4 years (range: 14 years, IQR: 3.5 years). Overall visual acuity was significantly better at follow-up than at baseline (p = 0.029). Follow-up perimetric MD was documented in 52% (47/90) of patients and exceeded 6dB in 40% (19/47). Perimetric MD was lower at follow-up, compared to baseline data, although not statistically significant (p = 0.155). Median papilledema grades at follow-up were lower than at baseline (p < 0.001). One patient in the follow-up cohort had optic disc atrophy at baseline, eleven patients (12%) had optic disc atrophy at follow-up. Ophthalmologic characteristics of the follow-up cohort are presented in Table 2.
Poor visual prognosis occurred in 36% (n = 32/90) of patients. This was defined as either 1) ophthalmologic worsening, which occurred in 22% and/or 2) persistent visual impairment, which occurred in 32% of patients. Worsening of BCVA occurred in 2 patients (3%) and worsening of papilledema occurred in 13 of 73 patients (i.e. 14%), who had follow-up documentation of papilledema grading.
In univariate group comparisons, neither sex, age, disease duration nor BMI baseline CSF-OP were significantly different between the visual outcome groups (Supplemental Table 1). Presence of headache at baseline (p = 0.8) and presence of abducens palsy (p = 1.0) were not associated with the visual outcome. Pharmacologic treatment was similarly frequent in the group of patients with poor and without poor outcome (p = 1.0, Suppl. Table 1) and was therefore not included into regression analysis.
In univariate regression analyses, the presence of moderate papilledema at baseline was associated with a lower likelihood of poor visual outcome (OR: 0.31, p = 0.043. Table 3). Here, surgical intervention was also strongly associated with a lower risk of poor visual outcome (OR: 0.21, p = 0.009, Table 3). Being a potential causal mediator of the papilledema-outcome relation, it was included into a multiple regression model as a sensitivity analysis, which yielded significant ORs for both moderate papilledema (OR: 0.13, p = 0.008) and for surgical intervention (OR: 14.5, p = 0.042), indicating partial mediation. Surgical intervention was therefore not included into the main multiple regression model in order to not bias the effect estimate.
Multivariable stepwise backward regression evaluated the influence of possible predictors to the visual outcome: The saturated model included the following independent variables: sex, baseline BMI, age at diagnosis, baseline CSF-OP, disease duration in months, and severity of papilledema at baseline. Due to listwise deletion of missing data, multivariable analysis included 71 patients. The final model included papilledema grades at baseline, sex and disease duration (Table 3, Nagelkerke R-square: 0.34). In multivariate analysis, the presence of moderate papilledema at baseline was associated with a lower likelihood of poor visual outcome compared to patients in the mild papilledema group (OR: 0.21, p = 0.016). The presence of severe papilledema at baseline showed no significant association with poor visual outcome compared to patients in the mild papilledema group (OR: 5.5, p = 0.157). Hosmer-Lemeshow-test indicated a good model fit (p = 0.537).
Table 2
Ophthalmologic findings at baseline and at follow-up in the follow-up cohort (90 patients).
| baseline | follow-up | p-value |
BCVA [logMAR]a (mean ± SD) | 0.13 ± 0.20 | 0.077 ± 0.13 | 0.0291 |
Frisén gradeb median (IQR) | 2 (2) | 0 (2) | < 0.0012 |
visual field MDc (mean ± SD) | 7.3 ± 7.4 | 4.8 ± 4.9 | 0.1551 |
optic disc atrophy n | 1 | 11 | - |
SD: standard deviation. IQR: interquartile range. 1two-sided p-values calculated by paired samples t-test. 2two-tailed p-values obtained by Wilcoxon-test. adocumented in 68 patients. bdocumented in 79 patients. cdocumented in 17 patients. |
Table 3
Regression analysis of the follow-up cohort (n = 90).
| poor visual outcome |
| univariable | multivariable |
| OR (95% CI) 1 | p-value | OR (95% CI)2 | p-value |
age at diagnosis | 1.02 (0.98–1.16) | 0.283 | not retained | - |
baseline BMI | 0.99 (0.93–1.04) | 0.600 | not retained | - |
baseline CSF-OP | 1.02 (0.97–1.08) | 0.391 | not retained | - |
disease duration | 1.01 (1.00–1.02) | 0.174 | 1.01 (0.99–1.03) | 0.075 |
male sex | 2.50 (0.62–10.07) | 0.198 | 7.17 (1.2–41.7) | 0.028 |
surgical therapy | 0.21 (0.06–0.68) | 0.009 | not included | |
baseline papilledema - moderate - severe | 0.31 (0.10–0.97) 6.67 (0.67–66.5) | 0.043 0.106 | 0.21 (0.06–0.75) 5.47 (0.52–57.3) | 0.016 0.157 |
CSF-OP: cerebrospinal fluid opening pressure, OR: odds ratio, 95%CI: 95% confidence interval. 1Obtained by univariable binary regression. 2Obtained by multivariable stepwise backward regression (likelihood ratio). Bold values indicate significant likelihood of the independent variables contributing to a poor visual prognosis. (Nagelkerke R-square: 0.34; Hosmer-Lemeshow test: p = 0.537) |