In this single-centre observational study, ICP monitoring of patients with clinically suspected SIH identified a low ICP in 8%, while pathologically high ICP was observed in 16% of patients. These results demonstrate that a significant proportion of patients with orthostatic headache who have a negative diagnostic work-up and are non-responsive to epidural blood patches, have a CSF dynamics disorder that can be identified through ICP monitoring. As suggested in previously published case reports, this study provides evidence on the potential utility of ICP monitoring in patients with suspected SIH but a negative diagnostic work-up [10, 15].

Interestingly, these results also demonstrate that orthostatic headache can be misleading and may represent the paradoxical presentation for a high ICP state. It could be speculated that some of these patients were affected by idiopathic intracranial hypertension (IIH) without papilloedema and that they paradoxically presented with an orthostatic headache. Another possibility is that some of these patients were originally affected by IIH and then developed a CSF leak that could not be identified with the spinal imaging available at the time. This hypothesis is in part supported by the findings of Craven et al. who noted that a high ICP state can drive a persistent CSF leak in patients with previous spinal surgery [20]. A third possible explanation is that some of these patients were originally affected by a CSF leak but developed rebound intracranial hypertension after treatment with epidural blood patches, with persistence of the orthostatic headache due to central sensitisation. Alternatively, as highlighted in previous studies, some of these cases may in fact represent chronic CSF leaks with elevated CSF pressures as part of the physiological compensatory mechanisms.

The study by Kranz et al. investigated the lumbar puncture opening pressures of 106 patients with a diagnosis of SIH and reported that 5% of them had a CSF opening pressure above 20 cm H2O [8]. In addition, Hani et al. noted normalisation of lumbar baseline CSF pressure with increasing duration of symptoms, and all of the patients in this study had longstanding symptoms at the time of ICP monitoring [14].

The presence of audiovestibular disturbances such as aural fullness and/or muffled hearing was significantly more common in patients with abnormal CSF dynamics (high or low ICP) than patients with normal/borderline ICP results (Fisher’s exact test p = 0.015, Table 2), and may represent potential clinical markers for CSF pressure dysregulation as the cause of orthostatic headaches in imaging-negative patients with suspected SIH However, no reliable clinical predictors of low ICP results were identified in this study, although the study is not sufficiently powered to be able to draw definitive conclusion (Table 2).

There are limited and conflicting reports on the utility of ICP monitoring in patients with low CSF pressure/volume syndromes. Case reports by Jensen et al. and Zada et al. described the helpfulness of ICP monitoring in the diagnosis of SIH and for assessment of treatment response, whereas Fichtner et al. could not detect any ICP increase after epidural blood patch using a subdural ICP monitor and hypothesised that a rise in ICP may not be the therapeutic mechanism of epidural blood patches [10, 12, 15].

Eide et al. performed ICP monitoring in patients with low CSF pressure/volume syndrome (only four were SIH patients) and found that morphological changes of the ICP pulse waveform only occurred when ICP fell below − 15 mmHg [11]. The patients with low ICP in our study had median ICPs above this threshold and this may be the reason why we did not identify significant differences in pulse amplitude among the three ICP groups (Table 2). Furthermore, lumbar infusion studies have provided useful insights in the CSF dynamics of patients with confirmed CSF leaks and it is interesting to note how the abnormalities in CSF dynamics tend to become less evident in the chronic stages of the disease and may be relevant in our study given the chronicity of symptoms prior to ICP monitoring [13, 14].

Lumbar puncture opening pressure is often tested when SIH is suspected, but this investigation can be misleading. Whilst low CSF opening pressures are suggestive of SIH, normal or high CSF opening pressures do not allow exclusion of this diagnosis [2, 8]. As noted above, both Kranz and Hani have shown normal or high CSF opening pressure in patients with SIH [8, 14]. Mokri et al. also described the cases of patients with confirmed CSF leaks and consistently normal lumbar puncture opening pressures [21]. These studies emphasise the fact that a diagnosis of SIH cannot be based solely on the results of a lumbar puncture [21, 22]. This is also supported by the results of our study in which three of the patients who had normal CSF opening pressures, had abnormal ICP monitoring results (two low and one high ICP readings) as noted in Table 1. Lumbar punctures should be undertaken with caution bearing in mind that the sensitivity of this investigation is modest and there is a risk of worsening symptoms of CSF leak. The exception to this recommendation would be when performing a lateral decubitus myelographic study; low CSF opening pressure measurement would be highly suggestive of SIH and would help negate the potential need for ICP monitoring in the future.

The main limitations of this study include the relatively small cohort, its retrospective design, and its heterogeneity (e.g., variable and inconsistent management profiles prior to ICP monitoring). Despite these limitations, this case series remains the largest study to date describing ICP monitoring results in patients with suspected SIH. Given that ICP monitoring is an invasive procedure, it is reserved for highly disabled patients with a clinical phenotype suggestive of SIH, where less invasive work-up has been inconclusive. Therefore, the relatively small size of the cohort is reflective of the judicious use of invasive options at our centre.

Despite the retrospective design, the standardised screening process and the systematic structure of both clinical letters and ICP monitoring protocol, reduced the risk of selection, detection and reporting biases. As for the heterogeneity in patient characteristics, this was expected given the absence in evidence- or consensus-based guidelines for the management of SIH at the time of the study.

Another important limitation is the scarce knowledge on normal ICP ranges. Performing ICP monitoring on healthy volunteers to obtain normative data would be unethical. In this context, the systematic review and meta-analysis by Norager et al. currently provide the best evidence on normal ICP ranges [19]. The results of this review align with our clinical experience and partly rely on data published by our centre [17]. We have used this evidence to classify the ICP monitoring results in our study in an objective and repeatable manner, while being open to the possibility that some patients classified as having a normal ICP in our study may be better classified as having borderline high or low ICP results.

In more recent years, a better understanding of SIH has led to the discovery of CSF-venous fistulas (CVF) as the cause of some cases of SIH [23, 24]. They were first described in 2014 and are not typically associated with a SLEC and their identification with traditional spinal imaging techniques can be challenging [23]. Dynamic CT myelography or digital subtraction myelography (DSM) with lateral decubitus views can be useful in the identification of CSF-venous fistulas, though these techniques were not previously widely available. Many of the patients included in this study had been investigated for SIH prior to the recognition of the CVF as an entity, and/or the availability of specialised spinal imaging techniques required to make the diagnosis. Schievink et al. (2020) reported that 10% of patients with orthostatic headaches who had normal brain and spinal imaging, including CT myelography, were found to have a CVF on digital subtraction myelography [25]. Future studies should investigate for the presence of CVFs with dedicated spinal imaging before ICP monitoring is employed.

Based on this study, ICP monitoring can be valuable for the investigation of a subset of patients with orthostatic headache. However, its use requires careful patient selection rather than being seen as part of the routine work-up for SIH. In all cases, the potential benefits of ICP monitoring must outweigh the possible risks. The procedure should be conducted at a specialist centre with expertise in managing SIH and performing ICP monitoring. In addition, the following criteria for the employment of ICP monitoring are suggested:

i. strong clinical suspicion for SIH,

ii. normal brain imaging (brain MRI with contrast),

iii. normal spinal imaging dedicated to the identification of possible epidural fluid collections (e.g., whole spine MRI with contrast, and dynamic CTM or DSM),

iv. normal spinal imaging dedicated to the identification of CSF-venous fistulas (e.g., dynamic CTM or DSM with lateral decubitus views),

v. normal lumbar puncture opening pressures, if this investigation is performed in the context of myelography (or for other diagnostic reasons),

vi. exclusion of other potential causes of orthostatic headaches including PoTS (unless already optimally treated), cervicogenic headaches, primary new daily persistent headaches and craniocervical instability,

vii. failure to respond to large volume epidural blood patch.