Longitudinal optical coherence tomography indices in idiopathic intracranial hypertension

Idiopathic intracranial hypertension (IIH) may result in optic nerve fiber loss and even atrophy. The timing of the optical coherence tomography (OCT) indices reaching the lowest point (nadir) and the factors that predict the patient's anatomical outcome are not known. We aimed to determine the timing and the factors that affect nadir retinal nerve fiber layer (RNFL) thickness. The medical records of 99 IIH patients who were treated from December 2009 to January 2020 were retrospectively reviewed. The mean RNFL thickness at presentation was 263.5 ± 106.4 µm. The mean time to nadir was 7.9 ± 6.3 months. The average RNFL and ganglion cell complex (GCC) thickness at the nadir were 92.6 ± 14.5 µm (47% showed thinning) and 77.9 ± 27.8 µm (70% showed thinning), respectively. The Frisén disc edema stage and average RNFL thickness at baseline correlated with a longer time to nadir, (r = 0.28 P = 0.003 and r = 0.24, P = 0.012, respectively). The nadir average RNFL thickness and the nadir average GCC thickness (r = 0.32, P = 0.001, r = 0.29, P = 0.002, respectively) correlated with the baseline visual field mean deviation. The final anatomical outcome of IIH episodes in this study resulted in RNFL and GCC thinning. The time to RNFL nadir and its values correlated with IIH severity at presentation.


Statistical analysis
The Kruskal-Wallis test and the Mann-Whitney test were used to compare continuous variables, while Spearman's rank correlation coefficient was used to study the association between continuous variables.The cumulative incidence of disease recurrence was described by 1 minus Kaplan-Meier curves.The length of follow-up was estimated with the reverse sundering method.SPSS software was used (IBM SPSS Statistics, version 25, Armnok, New York, USA 2017).

Demographics
Most (89.9%) of the 99 included patients were women, and the cohort's average age at presentation was 29.3 ± 9.2 years (range 18-58 years).A medical history relevant to IIH was noted for 30 patients (30.3%) (Table 1).Five patients (5.1%) have been taking drugs related to secondary IIH (Table 1

Management of patients
All patients were recommended to follow a weight loss and activity program at presentation.All patients were treated with acetazolamide (Diamox), and the initial recommended dose at presentation was between 500 and 1000 mg daily in 91 (91.9%) cases; the rest received a higher dosage of up to 2 g daily.In addition, 9 patients (9.1%) were treated with furosemide (Fusid) between 20 to 40 mg, and 4 patients (4.0%) were treated with topiramate (Topamax) between 25 and 100 mg.Six patients (6.1%) required optic nerve fenestration in the worse eye as determined by the worst baseline MD or the higher average RNFL thickness value at presentation in cases where both eyes had similar MD values.Four patients (4.0%) required a ventriculoperitoneal shunt, 7 patients (7.1%) underwent diagnostic angiography, and 6 patients (6.1%) underwent venous stenting.

Nadir
The

OCT
The mean gap between the baseline OCT and the nadir OCT finding was 7.9 ± 6.3 months.The mean time between the three consecutive OCT scans was 7.4 ± 2.2 months.The average RNFL and GCC thickness at the time of the nadir result were 92.6 ± 14.5 µm (range 57-129) and 77.1 ± 9.7 µm (range 37-93), respectively.At the time of nadir documentation, 46 (46.5%) of the patients had a thinner average RNFL thickness compared to the Cirrus HD-OCT normative database.Additionally, 69 (69.7%) patients had a thinner average GCC than the normative database.The average ratio of RNFL at nadir detection and the RNFL at presentation (RNFL ratio) was 0.42 ± 0.18.

Correlation between Nadir OCT outcomes, presenting parameters and disease course
The time to nadir correlated with the Frisén disc edema stage at presentation, r = 0.29 (p = 0.003) and with the baseline RNFL thickness, r = 0.62 (p < 0.001).In addition, a higher baseline LP OP correlated with the average baseline RNFL thickness (r = 0.22, p < 0.023) and with a lower nadir RNFL ratio (r = − 0.28, p = 0.004), (Table 2).

Correlation between OCT Nadir parameters and visual functions
The average nadir RNFL thickness highly correlated with the baseline MD (r = 0.32, p = 0.001).The RNFL ratio also correlated with the baseline VA (r = − 0.19, p = 0.047) and with the baseline MD (r = 0.37, p < 0.001).Additionally, the average nadir GCC thickness correlated with the baseline VA, CV, and VF MD (r = − 0.26, p = 0.006, r = 0.19, p = 0.045, and r = 0.29, p = 0.002, respectively), (Table 3).The nadir OCT RNFL thickness measurements that were within the normal range correlated with significantly higher HRR scores (p = 0.043), and the nadir OCT RNFL values that were within the normal range correlated with higher baseline MD values (p < 0.05).Furthermore, the nadir OCT RNFL thickness values that were thickest as well as those that were within the normal range correlated with higher average nadir GCC thickness values (p = 0.003 and p = 0.041, respectively).There was a significant positive correlation between the nadir average GCC thickness and the baseline average GCC thickness (r = 0.47, p < 0.0001).The RNFL ratio and the nadir average GCC thickness correlated highly with the initial dose of acetazolamide treatment (r = − 0.26, p = 0.007 and r = − 0.28, p = 0.004, respectively).www.nature.com/scientificreports/

Follow-up and recurrence episodes of IIH
The median follow-up duration of IIH patients was 40.4 ± 2.4 months, with an interquartile range of 20.7-65 months.Thirteen patients (13.1%) sustained a recurrent episode during this period (Fig. 2).

Discussion
IIH is considered a benign condition with a favorable outcome 11 .It may, however, be complicated by optic nerve fiber loss and even optic nerve atrophy and may run a chronic and recurrent course of disease in 10-30% of cases 11,12 .The median follow-up duration in our study was 40.4 ± 2.4 months since IIH diagnosis.The relapse rate of our study patients was 13.8%.We demonstrated that a significant number of IIH patients sustained RNFL (47%) and GCC (70%) thinning after a mean of 8 ± 6 months.The time to RNFL nadir as well as the nadir OCT values correlated with IIH severity at presentation.The percentage of females in our study (89.9%) is similar to the findings in the literature 2,13 .The average age of our patients at presentation was relatively young (29.3 ± 9.2 years), similar to previous studies 2,12,13 .Their high mean BMI at presentation (32.9 ± 5.5 kg/m 2 ) was also consistent with earlier reports 2,13 .Furthermore, our patients shared the same common IIH symptoms as those observed by others 3,14 .
Our patients' mean baseline VA was − 0.01 ± 0.09 in logMAR (20/20 on the Snellen chart) and their mean HRR color test result was 5.2 ± 1.3 plates, even in the presence of disc edema.Other studies have similarly noted an intact VA in IIH 15 as well as good CV 11 .The final anatomical outcome of the IIH episode as defined by OCT resulted in RNFL and GCC thinning, with average RNFL and GCC thickness of 92.6 ± 14.5 µm and 77.9 ± 27.8 µm, respectively.A significant number of our patients (47%) showed RNFL thinning, and 70% showed GCC thinning.This important finding indicates that axonal loss might be more widespread than previously suspected (i.e., reportedly around 33%) 16 , and that it might occur even in apparently effectively treated IIH.The point at which RNFL loss appears on OCT is therefore difficult to define and highly individualized and it was reached after a mean of 8 months since baseline.
The percentage of patients presenting with symptoms was much lower at the appearance of a nadir OCT finding compared to baseline, and the BMI values decreased in 58% of our patients during that time period.There was a concurrent significant improvement in the CV and MD as well as an increase in the percentage of patients with VA of 20/20 or better (91%).The resolution of signs and symptoms correlated with the reduction of disc edema during the follow-up of patients in the IIHTT and the Iowa experience study 2,17 .
The time to RNFL nadir and its values were in correlation with the severity of the condition at presentation in our study.Higher LP OP at baseline was associated with a significantly lower ratio between the nadir RNFL average thickness and the baseline average RNFL thickness.Monteiro et al. 18 demonstrated that high LP OP values resulted in axonal loss and atrophy in chronic IIH patients.To the best of our knowledge, ours is the first study that demonstrates the quantitative relationship between the two.The results of the IIHTT showed a significant correlation between the average RNFL thickness at baseline and the Frisén grade as well as with the LP OP 19 .In addition, the results of that study found no correlation between the baseline VA or MD and the GCL + IPL thickness 19 .Our study results for the Frisén disc edema stage also correlated with a longer time to a nadir OCT finding.
The average GCC thickness at the time of a nadir OCT result and the RNFL ratio were found to be positively correlated with the VA at presentation.The average nadir GCC thickness was also positively correlated with the CV (HRR score) at presentation, and the nadir average RNFL thickness, the nadir average GCC thickness, and the RNFL ratio were positively correlated with the MD at presentation.In addition, thicker nadir OCT RNFL thickness values in comparison to the Cirrus HD-OCT normative database tended to correlate with higher baseline MD values.Furthermore, normal nadir OCT RNFL thicknesses were correlated with higher HRR scores.These results show that the final anatomical outcome of an IIH episode that results in RNFL and GCC thinning is correlated with the severity of the presentation as well as with the VA, CV, and VF outcomes.
We demonstrated a significant positive correlation between the average baseline GCC thickness and the average nadir GCC thickness.This is in contrast with Huang-Link et al. 's 20 study which employed the spectral domain OCT and found no change in the GCL-IPL layers irrespective of ICP during a 12-month follow-up period.This difference could be attributed to their small sample size (20 eyes).In our study a significant number of patients showed GCC thinning (69.7%) which correlated with the MD.This finding may suggest irreversible GCC loss from an IIH episode.
In contrast to previous studies emphasizing the importance of aggressive treatment of IIH 13 , in our study, the higher the initial dose of acetazolamide treatment, the lower was the average nadir GCC thickness.This finding could be explained by the greater disease severity in the IIH patients who required a higher initial dose of acetazolamide, as well as by their lower nadir values.
The relapse rate in our long-term study of 99 IIH patients was 13.1%.The relapse rate in our study was lower than the rate of 28% demonstrated by Yri et al. 10 and 26% by Tovia et al. 11 who defined relapse as the recurrence of either papilledema or symptoms.The mean observation period in their studies, however, was 74 months, compared to 40.4 months in our study.The relapse rate in our study was similar to that cited in the 10-year observational study by Shah et al. 21(15%) in which recurrence was defined as the return of symptoms and the return of previously resolved optic disc edema on clinical examination.We defined relapse as being shown not only by a clinical examination but also by OCT, which is a stricter criterion that increases the validity of establishing recurrence.Taken together, these findings serve to indicate that IIH patients warrant long-term follow-up in order to monitor the occurrence of IIH relapse.
The main limitation of the present study is its retrospective design.Additionally, the baseline GCC measurements were not accurate because of inaccurate segmentation in the edematous state.Moreover, in our study, 71 patients had enlarged optic nerve sheath complexes.All patients in our study had results from at least one brain imaging modality, yet not all patients have completed an MRI.Since CT cannot discriminate the optic nerve from the optic nerve sheaths (as MRI can), theoretically other causes of thick optic nerve sheet complexes cannot be ruled out, such as optic glioma, optic neuritis, or perineuritis.Yet, expansion of the optic complex and optic nerves without enhancement in the presence of an appropriate clinical presentation and normal CSF content rules out inflammatory or tumor involvement in typical cases.
The strength of this study lies in the strict inclusion criteria, the comparatively large sample size, the detailed reports of symptoms, signs, imaging findings, treatment, and the longitudinal follow-up.
In conclusion, this average 40-month follow-up study of 99 IIH patients showed a high rate of RNFL and GCC thinning in IIH patients that had been reached after a mean of 8 months.The time to a nadir RNFL finding and its values were in direct correlation with baseline disease severity.These findings indicate that long-term follow-up of IIH patients, including OCT imaging, is essential for optimal management.

Figure 2 .
Figure 2. Probability of the recurrence of edematous findings on ocular computerized tomography in idiopathic intracranial hypertension patients.
). Summary of all potentially suitable idiopathic intracranial hypertension patients retrieved from the database.

Table 1 .
Baseline characteristics of all potentially suitable patients.BMI body mass index, IIH idiopathic intracranial hypertension, SD standard deviation, min minimum, max maximum, PMH previous medical history.

Table 2 .
Spearman rank correlation coefficients for presentation parameters.Time to nadir time between the baseline OCT and the nadir OCT, RNFL retinal nerve fiber layer, GCC ganglion cell complex, Sig.2-tailed significance, BMI body mass index, LP OP lumbar puncture opening pressure.*Correlationissignificant at the 0.05 level(2-tailed).**Correlation is significant at the 0.01 level (2-tailed).Bold indicates significant values.