Spontaneous Intracranial Hypotension: Long-Term Follow-Up

,


Background and Objectives
The recognition of spontaneous intracranial hypotension (SIH) has increased in recent years.Although the incidence is estimated at 5 per 100,000, it is likely to be more common since the disease is often underdiagnosed [1].SIH results from spinal cerebrospinal fluid (CSF) leaks, which mostly occur at the thoracic level or at the cervicothoracic junction [2].Spinal MRI scans showing so-called spinal longitudinal extradural fluid (SLEC) are suggestive of ventral dural tears (type 1 leak) often caused by a calcified ventral osteophyte, leaking nerve root sleeves are typically associated with meningeal diverticula (type 2 leaks), and the third cause is CSFvenous fistulas (type 3 leaks).Type 2 leaks are SLEC-positive if the tear is proximal and SLEC-negative if it is distal, and type 3 leaks are always SLEC-negative [3].SIH typically manifests as orthostatic headache that resolves shortly after patients return to a recumbent position.Magnetic resonance imaging (MRI) abnormalities are found in most, but not all, and include diffuse pachymeningeal gadolinium enhancement (PGE), brain sagging, subdural hygroma, venous distension signs, and pituitary engorgement [2,4].
Orthostatic headache may disappear with time, but after months/year, patients may present with a chronic headache (migraine, tension type, or new daily persistent headache) [2,5].A meta-analysis suggests that the following can occur in SIH: absence of orthostatic headache, normal imaging findings, or normal opening pressure at lumbar puncture (LP) in up to 25% of cases [2].Epidural blood patch (EBP) is considered the treatment of choice for patients unresponsive to conservative or pharmacologic treatment [6,7], but repeated, preferably targeted EBPs are sometimes needed and relapses are not uncommon [6,7].
As there are relatively few long-term follow-up studies on SIH patients, the purpose of this study was to describe our clinical experience and long-term outcomes in our patient cohort with SIH and in a subgroup of patients with clinical symptoms but not confirmed by MR or LP, probable SIH (pSIH).

Methods
In this case series, a retrospective review of the medical records of 50 patients suspected of having SIH was conducted from March 2007 to March 2022.All referred patients were evaluated at our neurological department and/or in the outpatient clinic at the Danish Headache Center (DHC), which is a tertiary headache center, connected to the Department of Neurology, Rigshospitalet-Glostrup, University of Copenhagen, where the diagnostic work-up and follow-up was conducted.The study was done in accordance with the Declaration of Helsinki with later revisions.The study was approved by the Neuroscience Center at Rigshospitalet-Glostrup as a quality assurance study and approved as such by the Regional Ethics Committee (H-18008942), which determined that participant consent was not required as the study collected information from patients undergoing usual treatment and evaluation.The study was conducted in accordance with Danish data laws.This is the primary analysis of these data.The patients were identified when given the diagnostic ICD-10 code (DG96.0)and/or when identified by the authors or colleagues at the Danish Headache Center.

Inclusion Criteria.
We included patients who were available for follow-up at least 6 months after establishing the diagnosis and whose medical records contained sufficient and relevant data for the study.Patients were included if medical charts contained data on headache characteristics at onset, frequency, and intensity; MR imaging; treatment with EBP; and sufficient data on headache characteristics at follow-up.
Patients with orthostatic headaches whose diagnosis was confirmed by positive MRI findings or low CSF pressure were included in the SIH group.
Patients with a clear history of orthostatic headache (with or without other symptoms) at the onset but with a negative brain MRI or normal CSF pressure were included in the probable SIH (pSIH) group.Because the primary purpose of this study was to follow-up patients in the long term, we thought it would be interesting to also review the outcomes of patients with a negative MRI.The inclusion of these MRI negative patients was based on the observation that an MRI may be negative 4 months after the onset of the condition [4].The SIH and pSIH groups were analyzed separately.

Exclusion Criteria.
Patients with previous head or spinal trauma, lumbar punctures, and/or spinal anesthesia were excluded from the study.By neuroimaging modalities (MRI, CT/MR angiography), possible causes for an orthostatic headache such as colloid cyst of the third ventricle, intracranial neoplasm, sinusitis, and cerebral venous thrombosis were excluded.Overview of inclusion and exclusion criteria are shown in Table 1.

Data Collection.
The following data were collected and analyzed: demographic variables, comorbidities, and precipitating factors at onset.Headache frequency and intensity at onset and follow-up was collected at an interview and based on the patient headache diary.In addition, we obtained the neuroimaging descriptions (brain MRI, spine MRI, CT, and myelography), the number of EBPs performed, response to pharmacological treatment and EBPs, and relapse rate and time to relapse.In order to reflect the clinical daily practice, we obtained all the initial imaging descriptions by experienced neuroradiologists.In case of doubt or unclear descriptions, our neuroradiologist (ES) reassessed the images.
At our headache center, we do not routinely use LP to diagnose SIH, because LP pressures are often normal in SIH [7] and to avoid trauma to the connective tissue [8,9].
However, LP has been performed in several patients to rule out other diagnoses.Spinal MRI and CT myelography were not performed on all patients; however, they were used in those whose initial treatment with EBP was unsuccessful.
2.4.Follow-Up Data.All patients enrolled in the study were followed up for at least 6 months.They were included if they could provide sufficient data on at least 2 follow-up visits after the last treatment.In cases where EBP was ineffective and a new treatment and/or diagnostic plan were required, patients were instructed to self-refer to the outpatient clinic after 2-3 weeks.
2.5.Outcome.On follow-up, it was noted that either no headaches (or additional symptoms) were present or, if headaches were present, the frequency was reported as episodic (1-14 days per month) or chronic (15 days or more per month).Patients were asked to define their headache using the visual analog scale (VAS), and the intensity was then reported as mild (VAS 1-3), moderate (VAS 4-7), or severe (VAS 8-10) (Tables 2 and 3).
2.6.Treatment.Depending on the clinical data, patients were offered treatment (in most cases, nontargeted EBP in the lumbar region) followed by additional investigation if needed (Appendix).The EBP was performed as follows: the patient was placed lying in the left lateral position.A standard epidural kit and an 18-gauge or 20-gauge angiocatheter was used to draw autologous blood in a sterile fashion.The epidural space was identified in the standard fashion using loss-of-resistance to air or saline.Once the placement of the needle tip of the epidural needle was confirmed, approximately 20-40 mL of autologous blood was injected slowly (30 to 60 seconds) into the patient's epidural space to create the blood patch.
Patients were discharged from our clinic if they were symptom-free for at least 6 months after the last EBP treatment.Patients who experienced a relapse of symptoms were treated with repeat EBP and managed according to the above discharge criteria.
A successful outcome was defined as no headache after the last treatment at 6-month follow-up or mild headache in some cases associated with physical activity but not requiring additional treatment.Persistent chronic headache (≥15 days per month) of moderate or severe intensity was considered a poor outcome.These outcome measures are somewhat more stringent than those used in similar studies, in which a good response is defined as having achieved at least a 50% reduction in the VAS score within 48 hours of the intervention and persisting for at least 6 months [7].

Statistical Analysis.
This is an observational descriptive study in which data is presented as frequencies, percentages, means, and ranges.The SPSS statistics software was used to calculate descriptive data.

Demographics.
The study included 50 patients, 37 in SIH and 13 in the pSIH group.In the SIH group, 22/37 (59%) were women and in the pSIH 6/13 (46%).The average age of symptom onset was 44 years for the SIH group and 42 years for the pSIH group (range 19-75 years): in the SIH group, 41 years in men and 47 years in women (range  and in the pSIH group, 48 years in men and 32 years in women (range .The time from onset to diagnosis was 6.5 months (range 1-41).In 20/37 (54%) of the patients, the diagnosis was established within 3 months.
3.2.Follow-Up.The mean follow-up time was 60 months (range 6-178 months).The number of cases was low in 2007, at the beginning of our inclusion period.This was due to very limited focus on that disorder.In recent years, there have been much more interest in the condition, in line with better diagnostics and treatment possibilities, which has resulted in an increasing referral rate.
The clinical data, diagnostic work-up, treatment, and follow-up for both groups are presented in Tables 2 and 3.

Clinical Findings.
All patients presented with a newonset orthostatic headache.Five patients in the SIH group and 3 in the pSIH group with a previous migraine noticed a new type of headache that was clearly distinguishable from their migraine attacks.Overall, 96% of patients complained of moderate-to-severe headaches when they were in the upright position.Over time, all patients developed a daily unremitting tension headache of mild-to-severe intensity, with an additional orthostatic feature.All patients had at least one associated symptom: cochleovestibular manifestations (tinnitus, phonophobia, and hearing loss) (62%), nausea (52%), neck pain (47%), dizziness (41%), interscapular pain (23%), paresthesia in the arms (21%), stabbing headache (12%), photophobia in patients without known migraine (8%), and cranial nerve palsies (abducens paresis) (1 patient).
The presenting symptoms did not differ between the SIH and pSIH groups.

Diagnostic Work-Up
3.4.1.SIH Group.Brain MRI showed signs of intracranial hypotension in all SIH patients, and all but 1 had diffuse PGE.Additional signs of intracranial hypotension on brain MRI were identified in 51% of patients: subdural hygroma (38%), brain sagging (16%), pituitary engorgement (14%), and venous distension sign (8%).Spine MRI was performed in 70% of the patients and revealed pathological findings in 73% of these.CT myelography revealed pathologic findings in 70% of the patients.

pSIH
Group.In all 13 patients, the brain MRI was normal, with no signs of intracranial hypotension.Spine MRI was performed in 31%, revealing a perineural cyst in 2 patients; one also had a herniated disc.Lumbar puncture was performed in 2 patients showing normal CSF pressure (Table 3).
Overall, CT myelography has been performed only in 10% of the patients, and this is due to 2 reasons: (1) this was a retrospective study with a primary aim to assess the long-term outcome of the patients with SIH; the majority of patients was included before our center began with the routine use of CT myelography; (2) if the patients had good outcome after the EBP, we did not proceed with further investigations.5 Acta Neurologica Scandinavica 3.5.Treatment 3.5.1.Conservative Treatment.All patients were instructed to have a high fluid intake including caffeine-containing drinks.All patients were offered simple analgesics, and those whose headache was lasting more than 1 month were offered preventive treatment.A total of 17/50 (34%) of patients had <50% effect on intensity after commencement of conservative treatment (simple analgesics in 4 patients, indomethacin in 5, amitriptyline in 3, and caffeine in 3 patients), while codeine achieved relief in one and cannabis tablets in another patient.The conservative treatment was stopped after 2-6 weeks if there was no response or in the presence of side effects or was continued until the patients received the EBP.As the conservative treatment was not satisfactory in most patients, they were offered EBP.
There was no medication overuse.
3.5.2.Epidural Blood Patch.In the SIH group, all but one patient was treated with EBP in a range from 1 to 8 (average 2.3) per patient.Only one patient received 8 EBPs with short-lasting but significant improvement in her clinical symptoms (severe headache and cognitive decline) and was eventually successfully treated surgically for a small thoracic CSF venous fistula.
Multilevel targeted EBP was performed in 3 patients, and the remaining patients underwent single-level nontargeted EBP in the lumbar region by injection of autologous venous blood.In the pSIH group, all but one patient were treated with EBP from 1 to 6 (average 1.5).

Surgical Treatment.
Before the SIH diagnosis was established, 2 patients underwent surgery for cerebral subdural fluid collection, with symptoms temporarily worsening in one patient.Three patients underwent surgery for a dural defect in the spine, with 2 patients being successful and the third with partial improvement.
During the follow-up period, 6/37 (16%) of patients experienced a relapse of SIH after symptoms had previously resolved completely for at least 3 months.
After relapse, 4/6 patients were successfully treated with repeated EBP.One patient was left with chronic moderate headache after unsuccessful repeated EBPs twice.One patient had remission after surgery for a dural defect and suffered only occasional exertional headaches.Relapses occurred over a period of 3 to 41 months (mean, 16 months).
Overall, the effect of EBP was good, even when patients were treated several months or even years after headache onset.A single patient had not received EBP because she had a good effect of pharmacologic treatment (Table 2).
No difference in response to EBP was found between men and women with respect to age or the presence of MRI findings.
Two women became pregnant after being diagnosed with SIH.No change in symptom pattern was observed during pregnancy, and both women delivered healthy babies without further complications.

Discussion
Our study is among very few long-term follow-up studies that analyzed the clinical and neuroimaging findings, treatment, and outcomes in patients with SIH diagnosed according to the ICHD-3 criteria [10].However, this is the first study that analyzed long-term follow-up in a subgroup of patients with probable SIH based on symptomatology but with a normal brain MRI.Our study showed that overall, 80% of our SIH patients and over half of the pSIH patients had a good outcome at long-term follow-up after treatment with EBP.Although studies have shown that up to 25% of patients with suspected SIH can have a negative MRI, the long-term follow-up of these patients was unknown.Our study however suggests that this subgroup should be treated as patients with evidence of low intracranial volume on MRI and should be offered the same treatment as even 50% has a good outcome.As in most other cases, conservative therapy is unspecific and insufficient, so patients with suspected SIH should be offered an EBP [7,11,12], even after several months to years after onset.
Treatment with EBP usually needs to be repeated, as only 1/3 of the patients respond well after the first EBP and up to half of the patients who receive three or more EBPs respond well [7,12].The outcome after treatment with EBP in our series of patients showed that remission of symptoms occurred in one-third of patients after a single EBP.Compared with a similar study in which a good response was defined as achieving at least a 50% reduction in the VAS score within 48 hours of the procedure and persisting for at least 6 months, the relatively poorer outcome in our study may be explained by more stringent outcome measures of frequency reduction [7].
One of the most important findings of our study is that the beneficial effect of EBP was not related to the duration of the disease; thus, some patients were successfully treated months and years after the onset of the disease.This is encouraging and proves that EBP should be tried in all patients with suspected SIH, no matter how late after the onset of symptoms.
Furthermore, a recent systematic review and metaanalysis has shown, interestingly, that no differences in terms 6 Acta Neurologica Scandinavica of outcome have been reported concerning targeted versus nontargeted EBP [13].This finding has significant implications, as nontargeted EBP is much easier to perform and can be performed at most hospitals as an efficient treatment, thus sparing the time and resources.As Kranz et al. point out, when diagnosing SIH, one should keep in mind that SIH is not always characterized by orthostatic headache, a negative MRI of the brain does not rule out SIH, the opening pressure may be normal, and an epidural blood patch does not immediately cure SIH [14].
In reality, the IHCD-III diagnostic criteria [9] for headaches attributed to low CSF volume are difficult to apply in clinical practice because the temporal relationship can be very difficult to determine, there is often no direct trauma or procedure to relate to, the opening pressure may be normal and thus misleading, and the clinical presentation may be vague over time.
There are few studies addressing long-term follow-up of patients with SIH, and most are case reports.In a large series of patients in which the median follow-up was 60 months, the overall response rate after EBP (injection of 10 mL of autologous venous blood fibrin glue and contrast agent) was 68.3% at the 6-month follow-up, with a reduction in pain scores of at least 50%.Symptoms recurred in 31.7%; of these patients who underwent a second procedure, the overall response rate at the last follow-up was 89.1% [7].
A series of 11 patients in whom 3-year follow-up was available showed that 83% were completely free of clinical symptoms and 8.3% had sporadic orthostatic headache [15].A retrospective analysis of 31 patients showed that EBP resulted in significant improvement of headache in 77% of patients [16].
Recurrent spontaneous spinal CSF leaks are not uncommon and can occur at any time after the initial CSF leak.Our results showed that recurrence occurred in 16% of patients over an average period of 16 months, two-thirds of whom were successfully treated with repeat EBP.Recurrence was also observed in up to one-third of patients in other longterm follow-up studies [7,17,18], which is a higher recurrence rate than in our study.
A retrospective analysis of 30 patients found that 50% achieved dramatic symptom relief after the first EBP and symptoms recurred in 6 patients [19].In another study, symptoms recurred in 32 patients (31.7%) after successful EBP.These patients underwent a second intervention, and the response rate at 6-month follow-up was 78.1%.Seven patients did not improve after a third procedure and remained symptomatic [7].
One study showed that recurrent spinal CSF leak developed in 5 of 13 patients who underwent surgical repair, compared with none of the five patients who had been treated nonsurgically.The recurrent leak occurred between 10 and 77 months after the initial spinal CSF leak, but in all patients, it occurred within 2 or 3 months after successful surgical repair of the leak [18].
Of our SIH patients, 38% had subdural cerebrospinal fluid collection on brain MRI, two were exposed to surgery, and one experienced transient clinical deterioration.Surgical removal of subdural cerebrospinal fluid collections should be avoided because patients do not experience lasting neurologic improvement and some even deteriorate [20].More than half of these patients experienced recurrence of hematomas due to neglected treatment of the spinal CSF leak [21].
The onset of SIH headache is usually sudden and may appear in the form of a thunderclap headache [22] without obvious trauma.Thereafter, the main clinical feature of SIH is an orthostatic headache that occurs immediately after a change to an upright position and resolves within minutes after a change to a recumbent position.In cases where the diagnosis is delayed, the clinical presentation of headache often lacks the orthostatic character.Over time, most patients develop a continuous headache that is no longer orthostatic or a delayed response to postural changes.The headache worsens after several hours in an upright position but does not necessarily improve after the patient has been in a recumbent position for an extended period.This has led to the term "second-half-of-the-day headache," which is probably a manifestation of a slow-flowing CSF leak [23,24].
In our group, 26% of patients (SIH+pSIH) had a chronic headache that was no longer orthostatic but could worsen during the day.Most patients presenting to specialized headache clinics have chronic headaches and may not remember the characteristics of the headache (i.e., orthostatic) at baseline.Therefore, they may be misclassified as chronic migraine or new daily persistent headache.Headaches in the second half of the day have low sensitivity and specificity because they may also resemble a TTH.In cases where the headache starts abruptly, leading to a definite orthostatic headache that changes over time to a chronic headache that no longer has orthostatic features but worsens in the afternoon, this observation should raise the suspicion of underlying SIH during history taking.
Therefore, to raise awareness of SIH, it is essential to specifically ask patients about the onset of headache and any orthostatic components and to offer them a blood patch in suspected cases.In some cases, SIH may mimic benign exertional headache [25].In our study, 27% of patients had exertional headache as the sole sequela of a previous SIH at follow-up.
In all our patients, we found symptoms other than orthostatic headache, which are likely due to traction of the meninges and other brain structures.This is consistent with observations from other clinical studies [2,26,27].Therefore, it is important to ask about cochleovestibular manifestations, neck pain, nausea, interscapular pain, and upper extremity paresthesia in the history.In our study, almost 1/3 of patients developed SIH after a specific activity.Therefore, the history should include information about events that may be associated with SIH, such as mild trauma, sneezing, airplane travel, sudden body flexion, another type of physical activity, or even activities such as yoga [28].We did not find any reports of pregnant women diagnosed with SIH in the literature.The two patients who were pregnant during follow-up in our study experienced no changes in clinical symptoms during pregnancy and gave birth to healthy infants.Although it is not possible to draw definitive conclusions based on only two patients, our observation is 7 Acta Neurologica Scandinavica that pregnancy may not pose risks in women previously diagnosed with SIH.
There is no clear diagnostic algorithm for the diagnostic approach in a patient with suspected SIH.Clinical features and neurologic imaging findings remain the most important criteria for diagnosis [2,[27][28][29][30][31].However, a recent metaanalysis showed that brain MRI findings were normal in 19% of patients [2].When we combined our SIH and pSIH groups, 26% of patients had normal MRI of the brain.As 54% of pSIH patients had a good outcome after treatment with EBP, we suggest that patients with a clear history of orthostatic headache should be offered standard treatment, even if there is no evidence of low intracranial pressure on neuroimaging.
A neuroimaging study has shown that a significantly shorter latency between the time of MRI examination and the time of headache onset was observed in patients with a negative initial MRI examination compared with patients with a positive PGE.It is estimated that PGE becomes apparent approximately 7-10 days after headache onset in most patients and may resolve as early as 25 days after headache onset [5].
The presence of an underlying cause of SIH can sometimes be difficult to detect; these include these included perineural cysts, osteophytes, meningeal diverticula, and cerebrospinal fluid venous fistula [2,27,[29][30][31][32][33].Various neuroimaging techniques have been proposed [26].Most investigators suggest standard and heavily T2-weighted MR imaging and/or a CT myelogram [4,29,34].The advantage of CT myelography is that it can detect both disc osteophyte spurs and CSF venous fistulas [29,35].Depending on the neuroimaging method chosen, leaks due to high flow/high volume and slow flow or intermittent leak can be captured [28], but it can be time-consuming and repeated examinations are often required.We suggest that spinal MRI and CT myelography should be part of the work-up in patients with probable SIH as recent studies have shown an improved sensitivity with improved MR technique [2].This would be of clear advantage in the pSIH group, which might be underdiagnosed regarding the spinal CSF leak, since they all improved by the treatment.

Methodological Considerations.
The study covers a period of 15 years, and as the Capital Region has 1.5 million citizens, it would be expected to see 75 new patients alone per year.The reason why we only identified 50 patients in the current study are due to several factors: less awareness about the diagnosis years ago and insufficient data in the medical charts leading to patients being excluded.Since the study was retrospective, we included only patients with sufficient data for this study.In addition, there is a risk of recall bias and a certain number of patients were probably not identified.For the entire group of patients studied, there may also have been inclusion bias because they were enrolled from a tertiary center.
The diagnostic work-up was primarily based on clinical outcome, i.e., response to the first two EBPs, and therefore, not all patients underwent spinal MRI or CT myelography.Because of the relatively small number of patients included, we cannot draw definitive conclusions, especially in the pSIH group.However, because few published studies of SIH include long-term follow-up, our clinical observations may help physicians identify and treat patients with SIH for decision-making and prognostic evaluation.
One of the most important findings of our study is that the beneficial effect of EBP was not related to the duration of the disease, such that some patients were successfully treated months and years after the onset of the disease.Another finding is that 54% of pSIH patients responded positively to treatment with EBP, raising the question whether the current ICHD-3 criteria for SIH should be reevaluated and include this group of patients.4.2.Conclusions.The clinical picture of SIH is more typical in the early stages of the disease and is characterized by orthostatic headache.However, as this feature may disappear over time, a detailed history of disease onset is crucial.Signs of intracranial hypotension on MRI are common but absent in some cases.Even in cases where there is a significant delay in diagnosis, we recommend offering treatment with one or more EBPs even in pSIH patients.Long-term recurrences may occur, so follow-up is required.However, the long-term prognosis is good in most patients and more awareness of SIH is needed.

Table 1 :
Inclusion and exclusion criteria.

Table 2 :
Patients with spontaneous intracranial hypotension (SIH) and positive MR signs for intracranial hypotension.

Table 3 :
Patients with probable spontaneous intracranial hypotension (pSIH) but without MR signs for intracranial hypotension.