Abstract
Background
Based on expert opinion, abducens nerve palsy and a neuroimaging criterion (≥3 neuroimaging signs suggestive of elevated intracranial pressure) were added to the diagnostic criteria for idiopathic intracranial hypertension. Our objective was to validate this.
Methods
This prospective study included patients with new-onset idiopathic intracranial hypertension for a standardized work-up: interview, neuro-ophthalmological exam, lumbar puncture, neuroimaging. Neuroimaging was evaluated by a blinded neuroradiologist.
Results
We included 157 patients classified as idiopathic intracranial hypertension (56.7%), probable idiopathic intracranial hypertension (1.9%), idiopathic intracranial hypertension without papilledema (idiopathic intracranial hypertension-without papill edema; 0%), suggested idiopathic intracranial hypertension-without papill edema (4.5%), or non-idiopathic intracranial hypertension (36.9%). Moderate suprasellar herniation was more common in idiopathic intracranial hypertension than non-idiopathic intracranial hypertension (71.4% versus 47.4%, p < 0.01), as was perioptic nerve sheath distension (69.8% versus 29.3%, p < 0.001), flattening of the globe (67.1% versus 11.1%, p < 0.001) and transverse sinus stenosis (60.2% versus 18.9%, p < 0.001). Abducens nerve palsy was of no diagnostic significance. Sensitivity of ≥3 neuroimaging signs was 59.5% and specificity was 93.5%.
Conclusion
Moderate suprasellar herniation, distension of the perioptic nerve sheath, flattening of the globe and transverse sinus stenosis were associated with idiopathic intracranial hypertension. We propose that idiopathic intracranial hypertension can be defined by two out of three objective findings (papilledema, opening pressure ≥25 cm cerebrospinal fluid and ≥3 neuroimaging signs).
Introduction
Idiopathic intracranial hypertension (IIH) is characterized by an unexplained increase in intracranial pressure (ICP) (1). The condition is linked to female sex and obesity, and the incidence is rapidly increasing in parallel with pandemic obesity (2–4). Since the first systematic characterization of the condition in 1937 (5) the diagnostic criteria have been revised several times (6–8). The present diagnostic criteria (7) are not data driven, but based on expert opinion, and their complexity and the lack of standardization makes them less operational in clinical practice. All versions have included the classic, diagnostic hallmarks of the condition, papilledema and elevated ICP (5–8), while diagnosis of IIH without papilledema (IIH-WOP) continues to spark controversy. In the latest diagnostic criteria (7), definite IIH is still defined by papilledema and elevated ICP, but two new objective findings – abducens nerve palsy and a neuroimaging criterion – are added, specifically for diagnosis of IIH-WOP and suggested IIH-WOP, respectively (Figure 1) (7). In all cases cerebro-spinal fluid (CSF) constituents must be normal, and neuroimaging must be without other structural lesions. Neurological exam is normal aside from any cranial nerve abnormalities (7).
Inclusion of Patients (Revised Friedman Criteria for IIH).
The suggested neuroimaging criterion requires the presence of at least three neuroimaging signs suggestive of ICP-elevation (empty sella, flattening of the posterior aspect of the globe, distension of the perioptic subarachnoid space, transverse venous sinus stenosis) (7). A retrospective study has found that this is in fact highly specific for IIH (9). However, IIH-WOP is rare and widespread use of MRI in the clinical diagnosis of IIH calls for validation. Further, the specific radiological method to evaluate the neuroimaging criterion varies and is not standardized (10). Prospective diagnostic studies including MRI, cerebral venography, and neuro-ophthalmological evaluation have been requested (9,10).
As symptoms of IIH are difficult to distinguish from primary headache disorders, this increased focus on objective findings is necessary. Nevertheless, considerable diagnostic challenges remain as illustrated by the fact that overdiagnosis occurs in 40% of patients diagnosed outside centers of tertiary care (11). The objective of our study was to address this unmet need, and validate the current diagnostic criteria in a large, prospective, diagnostic study.
Methods
Participants
This is a prospective diagnostic study. Patients were consecutively recruited between January 2018 and August 2020 at two specialized Danish centers with multi-disciplinary expertise in IIH (Odense University Hospital (OUH) and Danish Headache Center (DHC) Rigshospitalet-Glostrup) covering a population of 3.9 million people. Inclusion criteria were 18 years of age and clinical suspicion of new-onset IIH. Clinical suspicion of IIH was determined by a neurologist or ophthalmologist based on symptoms, phenotype, and any initial objective findings (e.g. abducens nerve palsy or papilledema). Exclusion criteria were pregnancy, breastfeeding, missing data, previous IIH and/or secondary intracranial hypertension (IH). Exclusion due to missing data was done only in cases where missing data made it impossible to verify or exclude the diagnosis (e.g. patient declined lumbar puncture) or if neuroimaging could not be extracted for blinded review. Previous IIH was excluded to ensure that the diagnosis was based on a standardized, expert work-up, and to include patients at the same disease stage before any effects of treatment on objective findings such as papilledema.
IIH was diagnosed after a standardized diagnostic work-up according to the current diagnostic criteria as recommended in international guidelines (1,7). This included: a semi-structured medical interview, neurological exam, neuro-ophthalmological exam, lumbar puncture to measure opening pressure (OP), routine analysis of blood and CSF and neuroimaging. Measurement of OP was standardized, with the relaxed patient placed in left lateral decubitus position, legs and neck extended. A standard issue manometer was used (cm CSF). Neuro-ophthalmological exam was performed by an experienced consultant (LDM) (fundus exam, optical coherence tomography (OCT), visual fields, visual acuity, Ishihara color plates). Pseudo-papilledema was differentiated from papilledema based on the morphological appearance (e.g. optic disc drusen, small disc, which lacks a physiologic cup and which has no pathologic alteration in the vessels) supplemented by autofluorescence imaging and B-scan ultrasonography. Optic disc swelling due to other diseases was differentiated from papilledema based on disease specific characteristics.
Patients were divided into two groups:
The study was approved by the regional ethical committee (Region of Southern Denmark, ID: S-20170058) and followed the Helsinki Declaration and Danish data laws. Written informed consent was obtained from all participants before inclusion.
Evaluation of neuroimaging
Cerebral MRI without contrast (1.5T or 3T), with or without thin-cut, fat-suppressed orbital sequences, was used to evaluate pituitary and orbital morphology. Choice of modality for cerebral venography depended on contraindications and availability (contrast enhanced CT or MRI time-of-flight without contrast).
An experienced neuroradiologist (LH of OUH), blinded to results of the diagnostic work-up and patient identity, systematically evaluated the four neuroimaging signs described in the diagnostic criteria (7). In case of suboptimal imaging quality for specific signs, e.g., inability to assess the optic nerve due to eye movement, the neuroradiologist assessed the sign as “could not be evaluated”. Predefined standards for each sign are described below.
Empty sella (I)
Pituitary morphology was evaluated in the sagittal plane. The diagnostic criteria (7) mention “empty sella” as neuroimaging sign, and refer to Yuh et al. (12), who describe pituitary morphology in five categories, depending on the degree of suprasellar herniation of CSF into the sella (7,12). I = no herniation, II = mild herniation (<1/3 of the sella height), III = moderate herniation (1/3–2/3 of the sella height), IV = severe herniation (>2/3 of the sella height) and V = empty sella defined as enlarged sella turcica without observable pituitary parenchyma (12). In consensus with two experts (neuroradiologist LH of OUH and endocrinologist of OUH), we specified “no observable pituitary tissue” as <1 mm of visible pituitary gland tissue. Yuh et al. suggest grade III = moderate suprasellar herniation as the cut-off point for pituitary abnormality associated with IIH (12). Therefore, we used grade III as cut-off point (Figure 1, Table 1, Table 2, Table 3) but we also evaluated grade V = empty sella (Table 2, supplementary analysis Table 3).
Patient characteristics.
†No answer excluded. BMI (non-IIH, n = 1), dizziness (IIH-FC, n = 2 and non-IIH, n = 4), pulsatile tinnitus (IIH-FC, n = 4 and non-IIH, n = 10), double vision (IIH-FC, n = 4 and none-IIH, n = 8), opening pressure (non-IIH, n = 2), specific papilledema degree (IIH-FC, n = 1).
ⱡIn case of immeasurably high ICP the highest readable level was used (IIH-FC, n = 20 and non-IIH, n = 2).
CSF, cerebrospinal fluid; IIH-FC, Idiopathic intracranial hypertension, any diagnostic sub-group, according to revised Friedman criteria; non-IIH, does not fulfill the revised Friedman criteria for Idiopathic Intracranial Hypertension; SD, standard deviation.
Significant p-values (≤0.05) are written as bold numbers.
Neuroimaging signs included in the revised Friedman criteria.
†Missing data excluded, could not be evaluated by neuroradiologist: Empty sella or herniation (n = 2), distension (n = 30), flattening of the globe (n = 42), sinus stenosis (n = 11), ≥3 neuroimaging signs (n = 25).
CSF, cerebrospinal fluid; IIH-FC, Idiopathic intracranial hypertension, any diagnostic sub-group, according to revised Friedman criteria; ITSS, index of transverse sinus stenosis; non-IIH, does not fulfill the revised Friedman criteria for Idiopathic Intracranial Hypertension; OP, opening pressure; TOF, time of flight.
Significant p-values (≤0.05) are written as bold numbers.
Sensitivity and specificity of neuroimaging signs with revised Friedman criteria as clinical reference standard.†
†The authors have excluded suggested IIH without papilledema from this part of the analyses, as presence of ≥3 neuroimaging signs is included as part of the diagnostic criteria for this sub-group in the clinical reference standard. The number of patients with suggested IIH without papilledema is n = 7. In the supplementary analysis an alternative cut-off point is used for pituitary morphology (grade V, empty sella) and in this case the number of excluded patients with suggested IIH-WOP is n = 1.
CI, confidence interval; IIH, idiopathic intracranial hypertension; IIH-FC with papilledema, idiopathic intracranial hypertension according to revised Friedman criteria excluding suggested IIH without papilledema; non-IIH, does not fulfill the revised Friedman criteria for Idiopathic Intracranial Hypertension.
Significant p-values (≤0.05) are written as bold numbers.
Flattening of the posterior aspect of the globe (II) and Distension of the perioptic subarachnoid space with or without a tortuous optic nerve (III)
Both were considered positive if present uni-or bilaterally. Flattening of the posterior aspect of the globe was a qualitative evaluation (axial plane, T2 weighted images). Distension of the perioptic subarachnoid space was defined as >2 mm distension in the coronal plane of T2 weighted images (10,13–15). Fat-suppressed, thin-cut orbital sequences were used when available. Presence or absence of optic nerve tortuosity has no diagnostic consequence in the diagnostic criteria and was not included (7).
Transverse venous sinus stenosis (IV)
Asymmetry of the transverse sinuses is a common normal variant whereas stenoses related to ICP elevation are bilateral (14,16–19). We employed a grading system designed to detect this (18). Stenosis was graded 1–4 on each side: 1 ≤ 33% stenosis, 2 = 33–66% stenosis, 3 ≥ 66% stenosis and 4 = hypoplasia or agenesia. Hypoplasia was defined as the full-length diameter being <1/3 of the superior sagittal sinus. A score (index of transverse sinus stenosis (ITSS)) was calculated by multiplying the two grades and a value of ≥4 was considered pathological (18).
Statistical analyses
Statistical analyses were performed in R. The Welch t test, Wilcoxon rank sum test or Fishers exact test were used as appropriate. In case of a missing answer the patient was excluded from the given analysis. Significance level was p ≤ 0.05. Standard issue manometers have an upper limit of 50 cm CSF, OP exceeded this in 20 IIH patients, and two non-IIH patients. For calculation of mean-OP the maximum readable value was used. As several diagnostic sub-groups in the current criteria, e.g. suggested IIH-WOP, are rare, occurring in less than 5% in previous studies, it was deemed necessary to include at least 100 patients with suspected IIH (20). Further power calculations were not performed as previous studies of diagnostic neuroimaging in IIH achieved statistical significance in considerably smaller cohorts (13).
Specificity and sensitivity of the neuroimaging criterion (≥3 neuroimaging signs) and 95% confidence intervals were calculated. Only patients where presence of ≥3 signs (yes/no) could be determined after blinded review (LH) were included. IIH with papilledema (definite IIH, probable IIH) – as described in the current diagnostic criteria – was defined as the clinical reference standard compared to patients in whom IIH was excluded (non-IIH) (1,7). The neuroimaging criterion, as suggested by the current criteria, was the index test (Figure 1). Patients with suggested IIH-WOP (Table 3, n = 7) were excluded in calculations of sensitivity and specificity as their diagnosis was partly based on neuroimaging findings and including them could falsely inflate sensitivity. Neuroimaging results were available to the assessor of the reference standard for exclusion of differential diagnoses.
Results
We evaluated 223 patients, 157 were included and 66 excluded due to: missing data (n = 16), previous IIH (n = 17), secondary IH (n = 29) and pregnancy (n = 4) (Figure 1).
In total 63.1% (n = 99) of the included patients fulfilled the revised Friedman criteria (7) (IIH-FC), distributed as IIH (56.7%; n = 89), probable IIH (1.9%; n = 3) and suggested IIH-WOP (4.5%; n = 7). Remarkably, we found no patients with IIH-WOP. The remaining 36.9% (n = 58; non-IIH) were mainly diagnosed with a primary headache disorder/medication overuse headache (53.4%; n = 31) or a neuro-ophthalmological condition including pseudo-papilledema and optic disc swelling (17.2%; n = 10, data not shown).
We compared patient characteristics in the IIH-FC group to the non-IIH group (Table 1). Both were predominantly female (96% versus 93.1%; p = 0.5), obese (BMI 36.5 versus 38.7; p = 0.1) and presented with headache (91.9% versus 96.6%; p = 0.3). Patients in the IIH-FC group were younger (29.1 versus 32.5 years; p = 0.03), suffered more often from pulsatile tinnitus (53.7% versus 22.9%; p < 0.001), transient visual obscurations (36.4% versus 10.3%, p < 0.001), blurry vision (66.7% versus 46.6%; p = 0.02) and subjective double vision (33% versus 17.6%; p = 0.05). Mean-OP was higher in IIH-FC (39.5 ± 10.2 cm CSF) than in the non-IIH group (26 cm CSF ± 8.4; p < 0.001). Remarkably, mean-OP in the non-IIH group was above the diagnostic limit (25 cm CSF). Presence of abducens nerve palsy was numerically more common in IIH-FC, but the difference was insignificant (8.1% versus 3.4%; p = 0.3). However, none of the patients presented with the combination of abducens nerve palsy and elevated OP, but without papilledema, and thus none fulfilled the criteria for IIH-WOP. We identified papilledema in 92.9% of IIH-FC patients (n = 92) and swollen optic discs in 5.2% of non-IIH patients (n = 3, p < 0.001). Non-IIH patients with swollen optic discs were diagnosed with other neuro-ophthalmological conditions (diabetic eye disease, non-arteritic anterior ischemic optic neuropathy). All non-IIH patients with swollen optic discs or abducens nerve palsy had OP < 25 cm CSF.
We evaluated a cerebral MRI in 100% of IIH-FC patients and 98.3% of non-IIH patients (p = 0.4). Cerebral venography (contrast-enhanced CT or time-of-flight MRI) was evaluated in 93.9% of IIH-FC and 91.4% of non-IIH patients (p = 0.5) (Table 2). Curiously, empty sella (grade V) was not found in IIH-FC patients but was present in 3.5% (p = 0.1) of non-IIH patients. In contrast, moderate suprasellar herniation (grade III or more) was more frequent in IIH-FC (71.4%) than in non-IIH (47.4%; p < 0.01), as was distension of the perioptic subarachnoid space (69.8% versus 29.3%; p < 0.001), flattening of the posterior aspect of the globe (67.1% versus 11.1%; p < 0.001), and bilateral sinus venous stenoses (60.2% versus 18.9%; p < 0.001). Mean-OP was higher in patients with moderate suprasellar herniation (35.7 versus 29.8 cm; p < 0.001), distension of the perioptic subarachnoid space (38.2 versus 30.2 cm; p < 0.001), flattening of the posterior aspect of the globe (39.2 versus 30.4 cm; p < 0.001), bilateral sinus venous stenoses (37.9 versus 33.1 cm; p = 0.01), and ≥3 neuroimaging signs (39.5 versus 31.1 cm; p < 0.001). Empty sella (grade V) was not associated with higher OP (32.5 versus 34.8 cm; p = 0.8) (Table 2).
We calculated specificity and sensitivity of ≥3 neuroimaging signs by comparing IIH-FC with papilledema (n = 92) to non-IIH (n = 58) (Table 3). In 59.5% (n = 47) of IIH-FC patients with papilledema ≥3 neuroimaging signs of elevated ICP were present as opposed to 6.5% of non-IIH patients (n = 3, p < 0.001). Sensitivity was 59.5% (95% CI: 47.9–70.4%) and specificity was 93.5% (95% CI: 82.1–98.6%). If the cut-off point for pituitary morphology was empty sella, grade V, specificity increased to 100% (95% CI: 93.4–100%) but sensitivity decreased to 28.2% (95% CI: 18.6–39.5%). Additionally, six of seven patients with suggested IIH-WOP would no longer fulfill the diagnostic criteria.
Discussion
There is an unmet demand for prospective studies that validate the diagnostic criteria for IIH – particularly requested are studies which include neuro-ophthalmological assessment and cerebral venography (9). Here, we present the first, large prospective diagnostic study of the current diagnostic criteria in a well-defined population with clinically suspected, new-onset IIH. We used a standardized method at two specialized centers and included OP, cerebral venography, neuro-ophthalmological assessment, and blinded evaluation of neuroimaging in patients with suspected new-onset IIH. This is novel and an important strength. As pointed out in a recent meta-analysis existing studies are either retrospective, include small sample sizes or use a case-control design (9,10,12,13,15,16,20,21). Furthermore, a primary diagnosis of IIH is wrong in 40% of cases (11). Therefore, we have conducted a prospective study based on interdisciplinary expert evaluation, which reduces bias and mimics the complicated, real-life diagnostic process. The non-IIH and the IIH-FC groups are similar in terms of sex, BMI, and headache status. This is a strength, as IIH is so intricately linked to female sex and obesity.
Remarkably, we did not diagnose any patients with IIH-WOP due to presence of abducens nerve palsy, suggesting that this finding is of limited diagnostic value. The frequency of suggested IIH-WOP did not exceed 4.5% confirming that this is rare (9,20).
In line with previous studies, we demonstrate that moderate suprasellar herniation, distension of the perioptic subarachnoid space, flattening of the posterior aspect of the globe and bilateral transverse sinus stenoses were highly associated with IIH-FC (13–15,19,21) and elevated OP, whereas empty sella was not. While the current diagnostic criteria (7) uses the term “empty sella”, a specific grade is not defined. In the literature moderate suprasellar herniation is widely accepted as the cut-off point for association with elevated ICP (9,10,12,21) and this is also strongly supported by our findings. Empty sella (grade V) is likely a long-term consequence of ICP elevation, which may explain why it is not found in new-onset IIH (12). It may also be a common incidental finding, especially in obese females 30–40 years old (22,23). We argue that “empty sella”, without further grading, is an inaccurate way to describe the pituitary morphology associated with IIH and elevated ICP. Moderate suprasellar herniation (grade III) is more precisely termed “partial empty sella”, as the pituitary gland is in fact still visible (12).
Adaption of the diagnostic criteria
Based on these results, we suggest an update of the current diagnostic criteria (7) (Table 4).
Suggested update to the revised Friedman criteria for diagnosis of idiopathic intracranial hypertension.
1IIH is probable in case of stand-alone papilledema if B–E are met. Consider new measurement of OP/ICP-monitoring.
2Pseudo-papilledema should be excluded by a neuro-ophthalmologist. Papilledema can be asymmetrical/unilateral.
3Thin-cut orbital MRI-sequences are highly recommended. Specificity is only sufficiently high if ≥3 neuroimaging signs are present. Neuroimaging signs are defined according to the following:
a. Minimum requirement for pituitary pathology is suprasellar herniation of CSF >1/3 into the sella turcica with or without compression of pituitary gland tissue (12).
b. Distension of the perioptic subarachnoid space: The space is >2 mm in the coronal/axial plane.
c. Flattening of the posterior aspect of the globe is a qualitative evaluation (neuroradiologist).
d. Bilateral sinus venous stenoses: ITSS score ≥4 (18). Stenosis was graded 1–4 on each side: 1 ≤33% stenosis, 2 =33–66% stenosis, 3 ≥66% stenosis and 4 =hypoplasia or agenesia. Hypoplasia was defined as the full-length diameter being <1/3 of the superior sagittal sinus. The ITSS score is calculated by multiplying the two grades.
4Secondary intracranial hypertension, pseudotumor cerebri syndrome, can be indistinguishable from IIH. IIH is rare in post-menopausal women, patients with BMI < 25 and in men. Secondary ICP-elevation should be ruled out by a specialist.
CSF, cerebrospinal fluid; CT, computed tomography; ICP, intracranial pressure; ITSS, index of transverse sinus stenosis; MRI, magnetic resonance imaging; OP, opening pressure.
Papilledema and elevated OP are the classic, objectifiable hallmarks of elevated ICP sufficient to diagnose IIH in the presence of normal CSF composition, neurological examination (except abducens nerve palsy) and neuroimaging (except the mentioned neuroimaging signs). Here, we present adequate prospective evidence of the high specificity of ≥3 neuroimaging signs (93.5%), confirming the findings of previous retrospective studies (9,10). We propose that the neuroimaging criterion (presence of ≥3 neuroimaging signs) is included in the diagnostic criteria for IIH on the same terms as papilledema and elevated OP, and suggest that the three objective findings of elevated ICP (papilledema, OP ≥ 25 cm CSF, neuroimaging criterion) are gathered in one criterion (Table 4, criterion A), which states that two of three must be present for diagnosis of IIH. This emphasizes the key principle in the current diagnostic criteria that at least one highly specific sign of elevated ICP (papilledema or neuroimaging) is mandatory, and it reduces the specific focus on rare sub-groups (IIH-WOP). Dedicated focus on this principle, rather than on rare sub-groups, could improve usability for clinicians.
We removed abducens nerve palsy as a separate diagnostic criterion, as it was rare and of no significant diagnostic impact. Additionally, we suggest that the term “partial empty sella” is used instead of “empty sella”. We further detail the exact cut-off point for pituitary abnormality (moderate suprasellar herniation, grade III). For the remaining three neuroimaging signs we also included detailed descriptions of the exact neuroradiological method to increase consistency (notes, Table 4). To ensure that all neuroimaging signs can be assessed reliably, we suggest that cerebral MRI with venography and thin-cut orbital MRI sequences are used. Adherence to these standardized principles is particularly important in cases where the diagnosis is suspected based on neuroimaging findings.
We find that this update of the diagnostic criteria is important as it continues the primary intention of the current criteria to reduce overdiagnosis. However, our suggested criteria still define IIH as a diagnosis of exclusion. Of particular importance neuroimaging should exclude structural lesions, sinus venous thrombosis and hydrocephalus and CSF analyses should exclude infection. IIH-WOP remains a controversial diagnosis, and we cannot exclude that IIH-WOP can be underdiagnosed in patients presenting with isolated elevation of OP. The mean-OP in the non-IIH group was above the diagnostic limit of 25 cm CSF, raising once again the question of the prevalence of IIH-WOP in obese patients with chronic headache. Nevertheless, IIH remains rare in comparison with other primary headache disorders. Migraine is highly prevalent, and obesity is a risk factor for migraine in young women (24–26). It is also well-recognized that OP may be moderately elevated in patients without IIH suffering from chronic headache or simple obesity (9,27,28), and the exact cut-off point is still debated with a so-called “grey zone” for OP between 25–30 cm CSF (1). Moreover, incidental finding of at least one of the four neuroimaging signs occurs in 49% of patients undergoing outpatient neuroimaging for other indications (29) and diagnosis of papilledema is challenging (11). This setting demands a high level of precision from the diagnostic criteria and precludes the use of isolated elevated OP or one to two neuroimaging signs as diagnostic for IIH-WOP. We argue that our non-IIH group remains the best control group for this type of study. The ability of the diagnostic criteria to separate healthy controls from patients with established IIH could also be tested. However, this would not mimic the actual setting in which the diagnostic criteria are expected to be used, and the criteria must be able to separate IIH-patients from other patients with a similar phenotype and symptoms.
We argue that the present update will increase usability for clinicians and reduce overdiagnosis, but it does still leave possibility for error, and additional studies as well as novel diagnostic biomarkers are needed. Nevertheless, the only existing alternative to the current diagnostic approach, as presented in this paper, is invasive, long-term ICP-monitoring to confirm the elevated OP. This is not feasible as it is limited by the discomfort for the patient, the possible complications, and the costs (30). There is an urgent need for improvement of diagnostic accuracy, and future studies focusing on further validation of the proposed diagnostic criteria are necessary. Researchers may also focus on evaluation of other neuroimaging signs associated with IIH, e.g. measurements of the optic disc, which are not included in this study.
Limitations
We could not confirm or exclude presence of ≥3 neuroimaging signs in a subset of patients with IIH and non-IIH (n = 13, 14.1% and n = 12, 20.7%, respectively). This was primarily due to suboptimal quality of orbital imaging. While inclusion at specialized centers guarantees a high-quality diagnostic work-up it may underestimate mild cases. Our cohort represents a mix of severity, and national guidelines do recommend referral of any suspected IIH-case to tertiary care, but we cannot exclude that severe cases are referred more consistently than mild ones.
Conclusion
Our study is the first prospective, diagnostic study to validate the current diagnostic criteria in a large cohort of patients with clinically suspected, new-onset IIH. We found that 63.1% of the included patients fulfilled the diagnostic criteria for IIH whereas 36.9% did not. Moderate suprasellar herniation (grade III, partial empty sella), distension of the perioptic subarachnoid space, flattening of the posterior aspect of the globe and bilateral transverse sinus stenoses were highly associated with IIH and elevated OP. Empty sella (grade V) was not. For diagnosis of IIH sensitivity of ≥3 neuroimaging signs was 59.5% and specificity was 93.5%. We have specified the neuroimaging criterion, and argue that it should be included in the diagnostic criteria for IIH alongside the traditional hallmarks of the disease, papilledema and elevated OP. Definite IIH-WOP was not found, and presence of abducens nerve palsy was rare. We propose that the criteria are revised according to the findings in this study, and hope that this will help clinicians navigate the pitfalls of diagnosing IIH.
Key findings
Based on the data presented in this prospective, diagnostic study we propose that IIH can be defined by two out of three objective findings (papilledema, opening pressure ≥25 cm CSF and ≥3 neuroimaging signs). We confirm the high specificity of the neuroimaging criterion for IIH and provide suggestions for standardization.
Footnotes
Acknowledgements
We thank medical students Sara Marie Pettersen and Helle Furuberg, University of Southern Denmark, who contributed considerably with entry of data collected during the study period.
We thank Prof. dr.med. Jes Olesen for valuable discussion regarding the suggested update to the diagnostic criteria.
Declaration of Conflicting Interests
The authors declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article:
JJK: Received funding from the Lundbeck Foundation.
RHJ: Gave lectures for Pfizer, Eli-Lilly, ATI, Merck, TEVA, Novartis, Lundbeck and Allergan. Investigator in clinical trials with ATI, Eli-Lilly, Novartis and Lundbeck. Director of Danish Headache Center, Lifting The Global Burden of Headache and Founder of Master of Headache Disorders at University of Copenhagen. Recived research funding from University of Copenhagen, Rigshospitalet, ATI, Lundbeck Foundation, The Medical Society in Copenhagen, NovoNordisk Foundation and Tryg Foundation.
LH: No conflicts of interest to disclose.
LDM: No conflicts of interest to disclose.
SMH: Received grants from Fight for Sight, Denmark and the Synoptik Foundation.
DB: Participated in clinical trials and served on advisory or speaker boards for Novartis, Eli Lilly, Lundbeck and Teva outside the submitted work (migraine). Received grants from Odense University Hospital and Rigshospitalet, University of Copenhagen during the conduct of the study.
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The study was funded by the Lundbeck Foundation (LBF R-276-2018-403-4), Candys Foundation (2015-146), Odense University Hospital and Rigshospitalet (25-A1320 and 69-A3346), Odense University Hospital (e-fond 177).
