Abstract
The incidence and clinical relevance of subdural haematoma (SDH) in patients with spontaneous intracranial hypotension (SIH) remain undetermined. We reviewed 40 consecutive SIH patients (18 female, 22 male) in a tertiary hospital. Eight (20%) of them had SDH and nine (23%), non-haemorrhagic subdural collections. The presence of SDH was associated with higher frequencies of male gender, recurrence of severe headache and neurological deficits. Outcomes were satisfactory after supportive care or epidural blood patches except for one SDH patient, who developed transtentorial herniation resulting in Duret haemorrhage and infarctions of bilateral posterior cerebral artery territories. In conclusion, subdural fluid collections were common in patients with SIH. SDH was associated with headache worsening or neurological deficits. Patients with SDH generally recovered well; however, serious sequela might occur.
Keywords
Introduction
Spontaneous intracranial hypotension (SIH) is considered to be related to spinal cerebrospinal fluid (CSF) leakage (1, 2). The core symptom is postural headache (3). Some patients may develop neurological deficits such as cranial nerve palsy or unconsciousness (1). This headache is often self-limited, although some patients may have complications. Patients with SIH may develop subdural fluid collections, and sometimes subdural haematoma (SDH) may occur. Diverse clinical manifestations occur in SIH patients with superimposed SDH, including postural or non-postural headaches (4), improvement of previous headache (5), exacerbation/recurrence (6, 7) or being asymptomatic (8). However, most are single case reports and very few series have systematically analysed the relevant clinical features (5). In addition, the treatment of SDH in patients with SIH remains undetermined (5, 6, 9). In this study, we reviewed 40 consecutive in-patients with SIH to investigate specifically the clinical significance of SDH in patients with SIH.
Methods
We retrospectively reviewed consecutive patients with SIH at Taipei Veterans General Hospital (Taipei-VGH), a tertiary medical centre in Taiwan, between January 1998 and September 2005. Since there is no effective ‘gatekeeper’ for the healthcare delivery system in Taiwan (10), most of our SIH patients were self-referred either via our emergency department or headache clinic. Each patient completed a structured headache intake form when first visiting our headache clinic or when admitted to our ward.
For study inclusion, all patients had postural headache and at least one of the following three criteria of ‘headache attributed to spontaneous low CSF pressure’ proposed by the International Classification of Headache Disorders, 2nd edition (code 7.2.3) (ICHD-2): evidence of low pressure on magnetic resonance imaging (MRI) such as pachymeningeal enhancement; evidence of CSF leak on conventional myelography, computed tomographic (CT) myelography or cisternography; and low CSF pressure <60 mmH2O (3). We excluded patients who had prior dural punctures, epi- or peridural anaesthesia, ventriculoperitoneal shunt, or other causes of CSF fistula.
Clinical profiles
One author (T-H.L.) reviewed the charts and headache intake forms. The severity of headache was classified from mild to severe. Recurrence of severe headache and any postural component were recorded. Any associated neurological deficits and diagnostic dural puncture after headache onset were also documented.
Neuroimaging findings
A neuroradiologist (J-F.L.) reviewed the imaging findings and documented the presence or absence of any subdural fluid collections. In this study, patients with subdural collections were classified into those with SDH (SDH group) and those without (i.e. non-haemorrhagic subdural collection group). SDH was diagnosed if the fluid showed high signal intensity on both T1- and T2-weighted images (Fig. 1), whereas non-haemorrhagic subdural collections showed low or iso-signal intensity on T1-weighted images and were typically thin, located over the cerebral convexities and caused no mass effects (Fig. 2). Previous studies have suggested that most of them were subdural hygromas (11, 12) with variable signal intensities, depending on fluid protein concentration (12). In this study, patients were classified as SDH if any neuroimaging studies showed SDH during the whole clinical course. The timing of the performance of the MRI studies was documented in relation to the symptom onset of SIH, i.e. the appearance of postural headache. Every patient had at least one brain MRI examination on a 1.5-T unit. On our service, some patients repeated MRI studies if they had either: (i) documented SDH by a previous MRI examination, or (ii) symptomatic worsening. Iter, i.e. the opening of cerebral aqueduct, and cerebellar tonsil displacements were measured on the sagittal T1-weighted MRI scan to determine brain descent, defined as either iter displacement of ≥1.8 mm or cerebellar tonsil displacement of ≥4.3 mm (13). Midline shift of ≥5 mm was also recorded (14).
Axial T1-weighted (a) and T2-weighted (b) magnetic resonance images exhibiting bilateral subdural haematoma, more on the left side, with a slight mass effect in a patient with spontaneous intracranial hypotension. Note the hyperintensity in both T1- and T2-weighted images (arrows).
The study defined non-haemorrhagic subdural collections as iso-intensity in T1- (a) and hyperintensity in T2-weighted (b) images (arrows).
Treatment
Patients with SIH were given conservative treatments, such as hydration, and medications including caffeine, theophylline, fludrocortisone and/or analgesics. If conservative treatments failed, epidural blood patches (EBP) were performed. None of the patients received surgical repair of any CSF leakage. A surgical intervention decision for SDH was made only on consensus among the in-charge neurologist, the neurosurgeon and the patient/family.
Short- and long-term follow-up
Short-term outcome, based on chart review, was defined at 3 months post treatment. For patients with a disease duration of ≥1 year, long-term follow-up was either by in-person or telephone interview (October to December 2005). The Migraine Disability Assessment Score (MIDAS) questionnaire was used to evaluate headache disability (15).
Statistics
SPSS Version 11.0 for Windows (SPSS Inc., Chicago, IL, USA) was used for statistical analysis. For continuous measures, one-way analysis of variance (
Results
Patients
During the study period, 40 consecutive patients (18 female, 22 male, age range 24–78 years) with SIH were reviewed. Of these, 32 received one brain MRI study, five received two studies and three patients received three studies. The indications for those who received more than one MRI study were clinical worsening (n = 3) and follow-up of SDH (n = 7). In this study, eight patients (20%) had SDH in at least one series of neuroimaging studies and nine patients had non-haemorrhagic subdural collections (23%). Of the eight patients with SDH, six were bilateral. The remaining two patients had non-haemorrhagic subdural collections on the contralateral side. Two patients had non-haemorrhagic subdural collections prior to SDH development. The timing of the MRI examinations was 69 days from disease onset for patients with SDH, 24 days for those with non-haemorrhagic subdural collections and 12 days for those without any subdural fluid collections.
The demographics, clinical and radiological characteristics of the three groups of patients are listed in Table 1. Headache intensity was moderate or severe in all patients. Recurrence of severe headache was noted in four patients. The mean interval between remission and recurrence was 53 days (35 days in a patient without subdural fluid collection and 50, 61, 66 days in three patients with SDH). Three of them had severe headache even in lying positions; however, they still maintained a postural component. Associated neurological deficits were noted in five patients: consciousness change (n = 3), abducens nerve palsy (n = 1) and numbness of both hands (n = 1). All these patients had SDH except the one with abducens nerve palsy. A high percentage of patients with SDH had had a prior diagnostic lumbar puncture (50%); however, the difference did not reach statistical significance in comparison with the other two groups of patients. In general, the clinical features were similar between those with non-haemorrhagic subdural collections and those without subdural fluid collections. When compared with those without SDH, the SDH patient group was more likely to be male, have recurrence of severe headaches and experience neurological deficits. A midline shift of ≥5 mm (n = 3) was identified only in the SDH group.
Demographics, headache characteristics and imaging findings among spontaneous intracranial hypotension patients with subdural haematoma, with non-haemorrhagic subdural collections and without any subdural fluid collections
SDH, Subdural haematoma; NHSC, non-haemorrhagic subdural collections; NA, not applicable.
By Fisher's exact, or one-way analysis of variance tests.
Not all lumbar punctures recorded opening pressure levels.
Treatment
In our series, the majority of patients received conservative treatment and/or EBP. Ten (25%) patients received EBP, which showed no difference between groups (P = 0.88). Only two of the eight patients with SDH received surgical intervention. Both of them presented with recurrence of severe headache after recovery, which prompted the second CT or MRI, and SDH was identified. One of the two patients complained of acute numbness of both hands and received haematoma evacuation, which resolved the numbness and his headache improved gradually. The other patient was found comatose (E1/V1/M2 by Glasgow coma scale) due to transtentorial herniation and received emergency EBP, haematoma evacuation and suboccipital decompression. Duret haemorrhage and infarctions of bilateral posterior cerebral artery territories were shown in the follow-up brain MRI.
Outcomes
The mean duration of follow-up was 42 ± 30 months (range 3–96 months). For short-term (3 month) outcome, one patient in the no subdural fluid collection group still had mild headaches due to persistent CSF leakage and one patient in the SDH group (see above) had severe disability (bed ridden and blind). For those whose disease duration was ≥1 year (long-term outcome) (n = 28), the MIDAS questionnaire was completed in 25 patients (89%). Only five (20%) of these still had some infrequent residual headaches and all their MIDAS scored 0–5, i.e. ‘minimal disability’ (15) and the others reported no headaches. Overall, both the short- and long-term outcomes were good in all three groups. Of note, none of the three SDH patients with midline shift ≥5 mm received surgery and all of them had total recovery.
Discussion
The incidence of subdural fluid collections in previous studies, including SDH and non-haemorrhagic subdural collections, in patients with SIH varied from 10% to 69% (16–18). However, a recent study has shown similar incidences to our findings (5), if SDH and non-haemorrhagic subdural collection were diagnosed separately. The wide variations in previous studies may also result from a discrepancy in the ability to detect small subdural effusions which ranged from subdural fluid collections with mass effects (17) to small effusions of 2 mm thickness (18).
Unlike SDH, patients with non-haemorrhagic subdural collections seemed to have similar clinical pictures to those without subdural fluid collections. Based on our results, three patients with SDH were associated with recurrence of severe headache. This phenomenon had been observed in some case reports (2, 4, 6–8, 19–21), but rarely systemically explored (5). Neurological deficits, especially consciousness changes, are also associated with SDH. Our group has reported one and reviewed six other SIH patients with consciousness changes (22). Six of these seven patients had SDH. In contrast, some authors have reported improvement of headache after SDH development and proposed that SDH could correct the low intracranial pressure (5). Although we cannot explain it, in our series male gender was also associated with SDH.
The true pathophysiology of SDH in patients with SIH is unknown and several mechanisms have been proposed. The first is rupture of bridging veins pulled away from the dura due to low intracranial pressure and brain descent (23). In addition, in light of meningeal biopsy results (24), some authors have suggested that the fragile and dilated new dural vessels in SIH patients might account for delayed SDH, even when normal pressure has returned and headache subsided (8). Another possible mechanism is that subdural fluid collections could predispose to SDH (6, 19): as trapped CSF gradually enlarges the subdural space, the bridging veins are increasingly stretched and in some cases may rupture (25). Similar to our series, Schievink et al. (5) reported that two of their eight SDH patients had prior subdural hygroma (similar to non-haemorrhagic subdural collection as defined in the current study). Our findings showed that the intervals of SDH occurrence spanned from 20 to 99 days, indicating the high variability of this complication. Therefore, the pathogenesis might differ among different patients. In this study, patients with SDH had a high percentage (50%) of prior diagnostic lumbar puncture when compared with the other two groups (22% and 33%). However, the difference was not significant. More studies are warranted to prove or disprove this association.
Our study proved that most patients with SIH run a benign course. However, like one of our SDH patients, several case reports have revealed severe sequelae (6, 26). De Noronha et al. reported one SIH patient with SDH, who had Duret haemorrhage and was dysarthric and ataxic 6 months after surgery (6). One Korean patient with SIH and SDH died of herniation (26). Therefore, patients with SIH and SDH are not completely without risk.
Our study has limitations. The incidence of SDH may be underestimated. First, the definition of SDH might miss some patients with chronic SDH with iso- to low signal intensity on T1-weighted images. Nevertheless, since our studied SIH patients were in their early stages, the possibility of misclassification might be low. Second, if patients had ‘asymptomatic’ SDH, their SDH might not be found (8). The latter assumption could also predispose to the false association between recurrence of symptoms and the development of SDH. In addition, since SIH is not a common disease, the case number of SDH in this study was small, although our study is one of the large series of SIH reported.
Competing interests
None declared.
Acknowledgements
This study was supported in part by grants from Taipei Veterans General Hospital (V94-295).