In our cohort, only 44% of the children with idiopathic intracranial hypertension had initial papilledema. Although the data on papilledema in pediatric idiopathic intracranial hypertension are not uniform, our results are in agreement with some previous retrospective publications on children with this diagnosis, where only 50% [9, 29], 52% [30], or 66% [10] had initial papilledema, respectively. A further small study on secondary intracranial hypertension in children with cryopyrin-associated periodic syndrome found papilledema only in 1/6 (16.7%) patients despite intracranial pressure values between 28 cmH20 and 45 cmH20 [31]. Although optical coherence tomography is used more frequently to aid in the diagnosis of papilledema in children [32], it requires relatively high levels of equipment, expertise, and cooperation from the child and cannot be performed at the bedside.

The pathophysiology of idiopathic intracranial hypertension appears to be multifactorial and more complex in children compared to adults [2], where predominantly young, overweight women are affected, with a male to female ratio of 1:8.5 [33]. Our cohort had a male to female ratio of 2.3:1, similar to the results of Masri et al. with a male to female ratio of 2.1:1 [10]. Another study described different sex ratios in prepubertal (male to female ratio 8:5) vs. pubertal children (male to female ratio 5:9) [34].

Our results, together with the existing literature, suggest that in pediatric idiopathic intracranial hypertension, more males may be affected than in adults and, secondly, papilledema—as an indicator of increased intracranial pressure—seems to be present in about 50% of patients. Therefore, other more reliable non-invasive methods for diagnosis and follow-up are desirable.

Symptoms of idiopathic intracranial hypertension in children are often non-specific [4] or associated with more common conditions such as hydrocephalus. In addition, the classic diagnostic setup is based on MRI, invasive measurement of intracranial pressure [8], and fundoscopy, the latter of which has the problem of potentially low sensitivity [9, 10, 29,30,31]. Therefore, further diagnostic screening tools, that can quickly, reliably, and non-invasively guide suspicion towards elevated intracranial pressure and exclude hydrocephalus, are needed and useful in suspected pediatric idiopathic intracranial hypertension.

US measurement of the optic nerve sheath diameter is known to provide a non-invasive assessment of elevated intracranial pressure, as optic nerve sheath diameter correlates well with intracranial pressure values and thresholds for potentially elevated pressure are defined [15, 35, 36]. The use of US optic nerve sheath diameter in idiopathic intracranial hypertension has already been described in both adults [37] and children [19]. Its reliability in children has been demonstrated in a comparative MRI-based study [38].

Transtemporal US measurement of the third ventricle diameter is a well-established method [39], originally described in patients with neurodegenerative diseases [40]. Since the third ventricle diameter in childhood hydrocephalus reflects the lateral ventricular width (excluding rare hydrocephalus forms above the third ventricle due to foramen Monro blockage), it can be examined by proxy [21]. The reliability of this method in children has also been demonstrated in a comparative MRI-based study [23].

In our patient cohort, the mean initial optic nerve sheath diameter was 6.45±0.65 mm. This corresponds to a high probability of an intracranial pressure >27 cmH2O compared to a published US optic nerve sheath diameter cut-off value of 5.75 mm [35]. Due to the very narrow third ventricle diameter (1.69±0.65 mm), the working diagnosis after US screening was “significant intracranial pressure elevation without underlying hydrocephalus,” which was ultimately confirmed by subsequent diagnostics.

The herein reported US optic nerve sheath diameter values confirm findings of smaller cohort studies in children with idiopathic intracranial hypertension. One prospective study on 13 children found US optic nerve sheath diameter values between 5.3 and 8.2 mm at intracranial pressures >25/30 cmH20 [19]. Aslan et al. described mean US optic nerve sheath diameter values of 6.7 mm in a prospective study on seven children [41]. Another study on eight children found initial US optic nerve sheath diameter values of 5.94±0.46 mm [20].

Irazuzta et al. described US optic nerve sheath diameter values between 3.75 and 4.9 mm in children with intracranial pressure <25 cmH20 [19]; other studies published optic nerve sheath diameter cut-off values of <5.2/5.3 mm for intracranial pressure <10 mmHg [35, 42], or for a healthy control group [41], respectively. Correspondingly, our control group had a mean optic nerve sheath diameter of 4.96±0.32 mm.

Regarding the third ventricle diameter, only Sari et al. have provided data on a normal third ventricle diameter in children (maximum width of third ventricle diameter in girls 4.98±0.34 mm, 5.54±0.35 mm in boys) [43]. Our finding of a subnormal, significantly narrower third ventricle diameter in idiopathic intracranial hypertension raises the question of whether subnormal ventricle width in children is pathognomonic for the disease. Similar pediatric studies on the topic do not yet exist. Studies in adults with idiopathic intracranial hypertension have described narrow lateral ventricles in a small percentage (3.3%), but a general association between this diagnosis and very small ventricles has not been proven yet [44]. However, according to the Kelly-Monroe doctrine [45], compression of the ventricles in the context of idiopathic intracranial hypertension with increasing intracranial pressure due to venous hypertension and thus increasing periventricular cerebral blood volume is quite conceivable as an underlying mechanism, resulting in a subnormal ventricular width.

We defined cut-off values for US optic nerve sheath and third ventricle diameter to detect or exclude idiopathic intracranial hypertension in children in first-line diagnostic screening. Both optic nerve sheath and third ventricle diameter provide a high diagnostic accuracy, with elevated optic nerve sheath diameter indicating elevated intracranial pressure and subnormal third ventricle diameter excluding hydrocephalus as the underlying cause. The combined use of both parameters revealed a high diagnostic accuracy. In contrast to US optic nerve sheath diameter alone, the combination allows a clear distinction between increased intracranial pressure due to hydrocephalus (a more common cause in children) and idiopathic intracranial hypertension (a rarer cause). Several studies in the literature have established US optic nerve sheath diameter cut-off values for idiopathic intracranial hypertension, especially in adults, with excellent diagnostic accuracy comparable to our results [37, 46]. However, US third ventricle diameter cut-off values and the combination with optic nerve sheath diameters for diagnosis have not previously been published.

This is the largest study with the longest follow-up using US techniques in pediatric idiopathic intracranial hypertension. A prospective study on eight children and US optic nerve sheath diameter-based treatment monitoring was published with a maximum follow-up period of 18 months [20]. Initial optic nerve sheath diameter values were 5.94±0.46 mm and decreased, very similar to our results, to 4.59±0.12 mm. Relapse due to interruption of therapy was evident in two patients by a renewed increase in optic nerve sheath diameter.

In our patient cohort, decreasing optic nerve sheath diameter values exquisitely indicated success of therapy as evidenced by decreasing clinical symptoms, papilledema or intracranial pressure, if measured. In contrast to the rather rapid improvement of the clinical condition within weeks to a few months, the optic nerve sheath diameter levels decreased immediately, but did not normalize until follow-up 2/follow-up 3, which corresponds to a time interval of 8–38 months after the start of therapy. It is unknown if the intracranial pressure, since it was not measured in all patients, also took such a long time to normalize. Delayed regression of the elastic optic nerve sheath to normal values after long-standing extension in the sense of a hysteresis behavior is a likely explanation for the delayed normalization of the optic nerve sheath diameter [47]. A recently published retrospective study of 17 children with idiopathic intracranial hypertension confirmed the utility of US optic nerve sheath diameter as an initial diagnostic tool, but its use for follow-up was found to be limited [48]. A cystic formation of the optic nerve sheath was seen in 16/17 children which did not show a significant decrease in diameter during the course of treatment. We have occasionally made similar observations, which is reflected in the standard deviations, but the majority of our patients did not have cystic and persistently wide optic nerve sheath diameters. Such phenomena should be investigated in larger cohorts and must definitely be considered as a cause of persistent optic nerve sheath diameter enlargement. However, even in such cases, a general, albeit non-significant trend towards smaller US optic nerve sheath diameters may be indicative of treatment success.

Our study is the first to systematically investigate third ventricular width at initial presentation and during follow-up in pediatric idiopathic intracranial hypertension. Interestingly, the initially below-average third ventricle diameter behaved divergently to the decreasing optic nerve sheath diameter during follow-up since a significant increase occurred. Relapse was indicated by a renewed reduction of the third ventricle diameter with repeat increase after successful restoration of therapy. Moreover, the third ventricle diameter reached a stable level more quickly after successful treatment than the optic nerve sheath diameter (0–13 months), so that possible hysteresis effects of the third ventricle walls seem to be less pronounced.

Regarding the pathophysiology of the disease, the increased venous pressure should result in a cerebrospinal fluid reabsorption impairment via increase in cerebrospinal fluid outflow resistance [49], leading to ventricular distension. However, venous outflow impairment also results in a higher cerebral blood volume, thus higher turgor of the parenchyma, which prevents ventricular dilation. Furthermore, the increase in blood volume is compensated for by reduction of the cerebrospinal fluid volume [45], also resulting in ventricular narrowing.

One limitation of this study is, despite its prospective nature, the still small number of subjects, especially during the long-term follow-up. Larger, ideally multicenter studies might allow to derive general recommendations for diagnosis and follow-up of pediatric idiopathic intracranial hypertension with US. Another limitation is that all US investigations were performed by one (experienced) examiner. Although numerous studies have shown high repeatability and inter- and intraobserver reliability for both US optic nerve sheath [14, 50] and third ventricle diameter [23, 40], studies with multiple examiners may further support a general recommendation to use these techniques in children with idiopathic intracranial hypertension.