Abstract
Background
Intracranial structural dislocation in spontaneous intracranial hypotension (SIH) can be measured by various intracranial angles and distances. We aimed to identify the clinical significance of structural dislocation in relation to treatment outcome in patients with SIH.
Methods
In this retrospective analysis, we identified patients with SIH who received an epidural blood patch (EBP) at Samsung Medical Center from January 2005 to March 2015. Structural dislocation in pretreatment MRIs of SIH patients was assessed by measuring tonsillar herniation, mamillopontine distance, the angle between the vein of Galen and straight sinus (vG/SS angle), the pontomesencephalic angle, and the lateral ventricular angle. After the first EBP, poor response was defined as the persistence of symptoms that prompted a repeat EBP.
Results
Out of the 95 patients included, 31 (32.6%) showed poor response. Among the radiological markers of structural dislocation, the vG/SS angle was associated with poor response (49.82 ± 16.40° vs 66.58 ± 26.08°, p = 0.002). Among clinical variables, premorbid migraine (p = 0.036) was related to poor response. In multivariate analysis, reduced vG/SS angle was independently associated with poor response (OR 1.04 [95% CI 1.01 – 1.07] per 1° decrease, p = 0.006). In 23 patients who underwent MRI after successful treatment, the vG/SS angle significantly increased after the EBP (p < 0.001, by paired t-test), while two patients with aggravation or recurrence showed a further reduction of their vG/SS angles.
Conclusions
Intracranial structural dislocation, measured by the vG/SS angle, is associated with poor response to the first EBP in patients with SIH. Successful treatment can reverse the structural dislocation.
Introduction
Spontaneous intracranial hypotension (SIH) is one of the important secondary headache disorders (1). SIH is characterized by orthostatic headache, low cerebrospinal fluid (CSF) pressure or volume, and radiological evidence of CSF leakage. SIH is benign in most incidences, but may be deleterious if untreated or inadequately treated (2,3). The epidural blood patch (EBP) is the mainstay of treatment (4,5). Targeted EBP treatment has been reported to be superior to blind EBP (6). However, the reported success rate of EBP is low, ranging from 36% to 70% (4–6). EBP failure might lead to prolonged discomfort in patients, longer hospital days, re-admission, and consequent socioeconomic loss. Thus, predicting treatment response might be important to clinicians for educating patients and establishing a treatment strategy.
In previous studies, the absence of radiological markers for intracranial compensatory changes was related to poor treatment outcome (6,7). These markers include pachymeningeal enhancement, subdural fluid collections, prominence of cerebral venous sinuses, and pituitary hyperemia (1,8). However, several new markers have been reported in SIH, such as mamillopontine distance, pontomesencephalic angle, and the angle between the vein of Galen and the straight sinus (vG/SS angle), which reflect intracranial structural dislocation (3,9). The clinical significance of these markers has not been investigated. We hypothesized that the structural alteration reflects a more decompensated state and might serve as a predictor of poor response to the treatment.
In this study, we aimed to investigate radiological markers of structural alteration to predict treatment response in patients with SIH.
Methods
Study design
In this retrospective study, we identified patients who were admitted with SIH to the Samsung Medical Center from January 2005 to March 2015. All the diagnoses were made based on the most updated International Headache Society criteria, and confirmed by investigators after reviewing electronic medical records on the headache characteristics and disease course. After the electronic medical record review, patients who received their first EBP as a first-line treatment in the Samsung Medical Center were included in this study. Thus, patients who refused EBP and received only conservative treatment, who received a Burr-hole trephination as a first-line treatment, or who received the first EBP in another hospital were not included. Patients who lacked a pretreatment MRI were excluded in this study. The Samsung Medical Center Institutional Review Board approved this study.
Clinical characteristics
Data for demographic factors such as sex, age, body mass index (BMI), and smoking were collected by chart review. Headache characteristics, including a history of preceding minor trauma (such as severe cough, vomiting, or a fall), were also reviewed. Information regarding time interval from symptom onset to diagnosis, time interval from diagnosis to treatment, type (blind vs. targeted) and the location of the EBP (cervical, thoracic, or lumbar), and the amount of injected blood volume, were also collected.
Identification of CSF leakage
CSF leakage sites were identified by CT myelography, spinal MRI, or radioisotope (RI) cisternography. The leakage sites were categorized on the basis of the location in the cervical, thoracic, lumbar, and sacral segments. When two or more noncontiguous spinal segments were identified as primary CSF leakage sites, the patient was coded as having multiple leakage sites.
Radiological evaluations
The pretreatment brain MRI was analyzed. If patients had multiple pretreatment MRIs, the last MRI taken at the closest time before the first EBP was chosen. To assess compensatory changes, we evaluated subdural fluid collections, pachymeningeal enhancement, engorgement of venous structures, pituitary hyperemia, and subdural hematoma. Engorgement of venous structures was defined as prominent dural venous sinus enhancement with enlarged and rounded sinuses. The pituitary gland was considered enlarged when the height of the gland was more than 9 mm on a non-contrast midsagittal T1-weighted MR image.
To evaluate the presence and degree of structural alteration, we measured the sagging of the brain, cerebellar herniation, mamillopontine distance, and intracranial angles including the vG/SS angle, pontomesencephalic angle, and lateral ventricular angle. Tonsillar herniation was defined as the downward displacement of more than 5 mm below Chamberlain’s line (3,9–11). The vG/SS angle was measured on midsagittal T2- or T1-weighted MR images. In the case of severe curvature, the final 8 to 10 mm of vG was used to measure the vG/SS angle (Figure 1a) (3). To overcome the engorgement of vG and SS, the middle of the vein was tracked. The pontomesencephalic angle was defined as the angle between the line tangential to the anterior margin of the midbrain and the line tangential to the superior margin of the pons (Figure 1b) (9). The mamillopontine distance was measured on midsagittal T1-weighted MR from the inferior margin of the mammillary bodies to the top of the pons (Figure 1c) (9). The lateral ventricular angle was measured between the medial aspects of the right and left lateral ventricles (Figure 1d). This angle was measured on the coronal image at the level of the fornix, third ventricle, and pituitary infundibulum (9).
Measurement of intracranial structural dislocation: (a) The angle between the vein of Galen and straight sinus, (b) The pontomesencephalic angle, (c) The mamillopontine distance (asterisk denotes the mammillary body) and (d) The lateral ventricular angle. This angle was measured on the coronal image at the level of the fornices (arrowhead), third ventricle (straight arrow), and pituitary infundibulum (curved arrow).
The images were evaluated independently by two investigators (HC and MJL). Inter-rater reliability was good for the mamillopontine distance and vG/SS angle (Intraclass Correlation Coefficient [ICC] = 0.945 and 0.863 respectively, with p < 0.001 in both cases), and moderate for the pontomesencephalic angle (ICC = 0.739, p = 0.011). Kappa values of pachymeningeal enhancement and venous engorgement were moderate (K = 0.765, p < 0.001). In case of disagreement, the third investigator (JC) was consulted.
Intervention and analysis of outcomes
Targeted EBP was performed in all patients with determined sites of leakage. When the leakage site could not be identified, EBP was conducted in a blind manner. We defined a poor response to EBP as a persistent symptom or sign after the first EBP prompting a repeat procedure. A good response was defined as complete recovery or minimal symptoms after the first EBP (6).
Follow-up measurement
In patients who were followed up with a brain MRI, clinical outcome at the time of MRI (remission, recurrence, or aggravation) was determined by review of medical records. Recurrence was defined in patients whose symptoms initially remitted after the first EBP, but then recurred. Aggravation was identified in patients whose symptoms persisted and worsened despite the first EBP. The most significant variable at the baseline MRI was re-assessed by investigators who were blinded to the clinical status (HC and MJL).
Statistical analysis
Statistical analyses were performed using SPSS 18.0 software (SPSS Inc., Chicago, IL, USA). Student’s t-test or the Mann-Whitney test was used for comparisons of continuous variables and Fisher’s exact test or a χ2 test for categorical variables. The Receiver Operating Characteristic (ROC) curve was analysed to find a cutoff for predicting poor responders. Multivariate logistic regression analyses were performed to adjust known predictors of poor treatment outcome such as the absence of reactive brain changes and the mode of EBP (blind vs targeted). For follow-up analysis, a paired t-test was performed to assess post-treatment changes to study variables separately by different clinical outcomes. A two-tailed p-value of < 0.05 was considered significant.
Results
Patients
We identified 146 patients with SIH during the study period. After excluding 51 patients (38 who underwent a first-line treatment other than EBP (conservative management, n = 29; Burr hole trephination n = 3), eight with alleged treatment failure, and five without a pretreatment MRI), 95 patients (60 women and 35 men; mean age 41.4 years; range 25–81 years) who received their first EBP in the Samsung Medical Center were included in the analysis. Of the 95 patients, 31 (32.6%) were classified as having treatment failure.
Clinical characteristics
Demographic and clinical characteristics according to treatment outcome.
Values are presented as number (%) or median (interquartile ranges).
MRI = magnetic resonance imaging; CT = computerized tomography; RI = radioactive isotope; CSF = cerebrospinal fluid; EBP = Epidural blood patch.
Brain MRI findings
Brain MR findings in association to treatment outcome.
Values are presented as number (%) or mean ± SD.
vG/SS angle = the angle between the vein of Galen and the straight sinus.
Among the structural alterations, the vG/SS angle was significantly narrower in poor responders (49.82 ± 16.40° vs 66.58 ± 26.08°, p = 0.002; Table 2). Abnormally narrow mamillopontine distance was found more frequently in poor responders than in good responders (40% vs. 26%), although statistical significance was not reached. Other parameters, including sagging of the brain, pontomesencephalic angle, and lateral ventricular angle, did not differ between the two groups (Table 2). By ROC analysis, the vG/SS was a significant variable to discriminate poor responders (area under the curve = 0.698, 95% CI 0.589–0.806, p = 0.002), although the accuracy for classification was poor.
Multivariate analysis
Multivariate analysis of predictors of a poor treatment outcome.
OR = odds ratio, CI = confidence interval, EBP = epidural blood patch, vG/SS angle = the angle between the vein of Galen and the straight sinus.
The absence of reactive changes denotes a negative brain imaging of pachymeningeal enhancement, subdural fluid collection, subdural hemorrhage, engorgement of venous structures, and pituitary hyperemia.
Correlation analyses did not show any significant relationship between the vG/SS angle and other clinical and radiological variables including age, time interval from symptom onset to diagnosis, time interval from symptom onset to treatment, time interval from diagnosis to treatment, pituitary height, pontomesencephalic angle, mammillopontine distance, and lateral ventricular angle (Supplementary Table 1).
Follow-up measurement
Brain MRIs were followed up in 23 patients after complete remission of SIH, while only two patients with recurrence or aggravation underwent the follow-up MRIs (Figure 2). In 23 patients, the vG/SS angle was significantly increased (mean 60.7° [SD 23.72°] before EBP; mean 78.7° [SD 21.04°] after EBP, p < 0.001 by paired t-test) after a median of 6.5 months [interquartile range, 4.1–18.0] after remission (Figure 2a). Two patients with recurrence or aggravation showed further decrement of the vG/SS angles (−54.8° in the patient with recurrent SIH; −14.5° in the patient with symptom aggravation; Figure 2b). Representative cases are illustrated in Figure 3.
The angle between the vein of Galen and straight sinus (vG/SS angle) before and after treatment. (a) An increase in the vG/SS angle in 23 patients after complete remission and (b) A decrease in the vG/SS angle in two patients with aggravation (solid line) or recurrence (dashed line). Representative cases (a–b) A 52-year old male patient with spontaneous intracranial hypotension (SIH), showing (a) the angle between the vein of Galen and straight sinus (vG/SS angle) of 55.5°, (b) an increase in the vG/SS angle one month after successful treatment. (c–d) A 49-year old male patient with SIH, showing (c) a vG/SS angle of 59.8°, (d) a decrease in the vG/SS angle when the symptoms aggravated due to treatment failure.
Discussion
In this retrospective study, structural alteration, demonstrated by a narrower vG/SS angle, was significantly associated with treatment failure in patients with SIH. An increase in vG/SS angle on a post-EBP MRI correlated with treatment success. Radiological markers of reactive changes were not associated with treatment outcome.
Characteristic MRI findings of SIH can be classified into two categories. The first category includes compensatory changes in reduced intracranial volume explained by the Monro-Kellie hypothesis (12). Pachymeningeal enhancement, subdural fluid collections, prominence of cerebral venous sinuses, and pituitary hyperemia (1,8) are included in this category (13,14). The other category is the structural alteration caused by loss of CSF buoyancy (15). Downward displacement of brain structures is representative for this phenomenon (16). Recently, alteration of intracranial angles including mamillopontine distance, pontomesencephalic angle, and vG/SS angle has been reported as a new marker of structural alteration in patients with SIH (3,9).
Schievink et al. reported that a normal appearance on the brain MRI is suggestive of poor treatment outcome (7). They dichotomized MRI findings as a presence versus absence of compensatory changes (1,7,14). In our study, almost all patients showed at least one compensatory change in the brain MRI. However, their treatment outcome was not uniformly excellent. In our study, the treatment outcome was not associated with the absence of reactive changes. In contrast, the vG/SS angle, which is a marker of brainstem slumping, was associated with treatment outcome. Our finding suggests that evaluation of intracranial structural alteration requires as much focus as assessment of compensatory changes in the brain MRI of patients with SIH.
Our study results showed that the vG/SS angle was not a static marker; it was reversible, depending on the treatment outcome. After successful treatment, the vG/SS angle was significantly increased, while it decreased when the symptom recurred or aggravated. Savoiardo et al. reported that the vG/SS angle is associated with brain swelling (3). They suggested that a narrower vG/SS angle indicates more severe displacement of the brainstem and cerebellum, causing functional stenosis of venous drainage (3). Our study results provide evidence that the functional stenosis of vG/SS is a dynamic state that can be reversed after successful treatment.
Among clinical factors, premorbid migraine was more frequently reported in poor responders. A possible hypothesis is that patients with migraine had a sensitization in the pain-sensitive structures or central pain pathways. Consequently, their pain would be exaggerated even when the remaining CSF leakage after the EBP is minimal. In addition, migraine attack might be triggered in this situation. Further studies are needed to explore our hypotheses.
Our data are different from a previous study from our center (6). The previous study, which was performed by our neurosurgery team, recruited 56 patients from two distinct centers (Kangbuk Samsung Hospital and the Samsung Medical Center) between 1999 and 2009, while we recruited 95 patients from both neurological and neurosurgical wards in a single center (the Samsung Medical Center) between 2005 and 2015. Although some patients overlapped between the two studies, the diagnostic and therapeutic strategies were more homogenous in our study during the study period. Since 2005, CT myelography has been widely performed in our center because of its greater sensitivity and specificity to identify the leakage site (17). In terms of treatment, blind EBP was mostly performed before the year of 2005, regardless of CSF leakage sites. During the period of our study, targeted EBPs were preferred for all patients with determined sites of leakage. Consequently, the rate of blind EBP was low in the present study. Therefore, responder rates from the blind procedure could not be compared between the two studies.
Our study has several strengths. This is the first study to perform quantitative evaluation of structural imaging markers of SIH for clinical correlation. Second, the number of study participants was relatively large. Third, our patients underwent homogenous clinical and radiological evaluations at a single university hospital. However, our study also has some limitations. First, since the study was conducted in a single university hospital, the possibility of inclusion of more refractory cases was higher. However, in the Korean national insurance system, there is little prerequisite for patients to visit a tertiary hospital. Second, our data were from a single ethnic group. Thus, external validation with various ethnicities is required. Third, the rate of SDH was low in our study. This is attributed to our inclusion criteria, which led to an exclusion of patients who had Burr hole trephination before the EBP as their first-line treatment. Therefore our study might have a selection bias toward an exclusion of more severe patients. However, if SDH associated with SIH is undiagnosed and only treated by Burr hole trephination, the clinical outcome may be biased by the possible deterioration after the evacuation surgery (18).
In conclusion, intracranial structural alteration reflects poor treatment response in patients with SIH after EBP. Brainstem dislocation, measured by the vG/SS angle, was the most significant predictor of treatment failure. Successful treatment can reverse the structural dislocation. Further prospective studies are warranted to test our results to predict and monitor treatment response in patients with SIH.
Clinical implications
Structural alteration in spontaneous intracranial hypotension is associated with treatment outcome. A narrow angle between the vein of Galen and straight sinus is predictive of treatment failure, while the angle increases after successful treatment. Premorbid migraine is also associated with poor response to the first epidural blood patch treatment.
Footnotes
Declaration of conflicting interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was partly supported by Dong-A ST for data management.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
