Abstract
Objective:
Most patients with spinal cerebrospinal fluid (CSF) leakage require an epidural blood patch (EBP); however, the response to treatment is varied. This study aimed to compare the MRI findings at follow-up between EBP effective and non-effective groups and to identify imaging findings that predict EBP treatment failure.
Methods:
We retrospectively reviewed 48 patients who received EBP treatment for spinal CSF leakage. These patients were stratified into two groups: EBP effective (n = 27) and EBP non-effective (n = 21) using the results of the 3 month MRI as the end point.
Results:
Compared to the EBP non-effective group, the patients in the EBP effective group had a lower spinal CSF leakage number (2.67 vs 12.48; p = 0.001), lower spinal epidural fluid accumulation levels (3.00 vs 7.48; p = 0.004), brain descend (11.11% vs 38.10%; p = 0.027), pituitary hyperemia (18.52% vs 57.14%; p = 0.007), and decreased likelihood of ≥three numbers of spinal CSF leakage (25.93% vs 90.48%; p = 0.001) in the post-EBP MRI. Clinical non-responsiveness (OR: 57.84; 95% CI: 3.47–972.54; p = 0.005) and ≥three numbers of spinal CSF leakage (OR: 15.13; 95% CI: 1.45–159.06; p = 0.023) were associated with EBP failure. Between these variables, ≥three numbers of spinal CSF leakage identified using the post-EBP MRI demonstrated greater sensitivity in predicting EBP failure compared to clinical non-responsiveness (90.48% vs 61.9%).
Conclusion:
The number of spinal CSF leakage identified using the post-EBP MRI with a cut-off value of three is an effective predictor of EBP failure.
Advances in knowledge:
Compared to clinical responsiveness, the post-EBP MRI provided a more objective approach to predict the effectiveness of EBP treatment in patients with spinal CSF leakage.
Introduction
An increase in the incidence of spinal cerebrospinal fluid (CSF) leakage leading to CSF volume depletion has been observed in the past decades.1 Orthostatic headache is the typical presentation of spinal CSF leakage, resulting from decreasing CSF volume-related pain-sensitive structure stretch during the standing position.2 Common etiologies of CSF leakage include post-lumbar puncture headache and spontaneous intracranial hypotension.3
A magnetic resonance myelography (MRM) is used to visualize the accumulation of spinal epidural fluid or the abnormal irregular signal lateral to the neural sleeve in order to confirm the diagnosis of spinal CSF leakage.4 In addition to these findings, brain and spinal MRI include diffuse pachymeningeal enhancement,5 dural sinus engorgement,6 brain descend, subdural hematoma, pituitary hyperemia,7 and spinal epidural venous plexus engorgement. The Monro-Kellie doctrine states that the decreased intracranial CSF volume led to increased brain tissue volume or engorged vascular structures which are observed in the MRI.8
Appropriate intravenous hydration is the standard treatment of newly onset spinal CSF leakage.9 Hydration can cause complete recovery from discomfort; however, symptomatic recurrence is commonly observed.10 Patients who experience treatment failure following intravenous hydration are managed by increasing hydration or performing epidural blood patch (EBP) treatment.11 Previous studies have found that predictors of the efficacy of EBP include an age <40 years,3 a short interval between symptom onset and EBP treatment, leaks from the spinal tap,12 and anterior epidural fluid accumulation length occupying less than eight vertebral segments as identified by MRI. However, these studies assessed the efficacy of EBP treatment based on symptomatic improvement, which may be diverse and subjective.
In our clinical practice, a post-EBP MRI within 1 week is usually performed to evaluate immediate and objective treatment response; furthermore, a repeat MRI is performed at the 3 month follow-up to evaluate the long-term outcomes. The effectiveness of this evaluation method in EBP outcome assessment in patients requiring salvage EBP remains unclear. Therefore, we aimed to investigate the post-EBP MRI findings to identify patients who are likely to fail the EBP treatment.
Methods
Patients
This retrospective study reviewed the medical records of patients with orthostatic headache who underwent a MRI to diagnose spinal CSF leakage from January 2010 to January 2021. Patients with the following characteristics were excluded from the study: patients who recovered using hydration treatment and those with incomplete data.
Initially, 130 patients with orthostatic headache were included. Among them, 75 patients were diagnosed with CSF leakage based on the results of the MRI. Furthermore, 23 patients with CSF leakage were excluded since they recovered completely following hydration therapy. A total of 52 patients had intolerable headache and were indicated for EBP treatment; however, four were excluded due to incomplete data. A total of 48 patients were included in the final analysis of the study (Figure 1). This study was approved by the local institutional review board. Requirement of informed consent was waived due to the retrospective nature of the study.
Figure 1.
Schematic of patient enrollment. CSF, cerebrospinal fluid; EBP, epidural blood patch; MRM, magnetic resonance myelography.
Neuroimaging
Upon consultation, whole-spine MRM, and conventional brain and spine MRI were performed simultaneously in all the patients. In the patients who underwent EBP treatment, whole spine MRM and non-contrast brain MRI were performed within 1 week, defined as “post-EBP MRI”. To further evaluate the efficacy of the treatment, the initial imaging protocol at consultation was repeated at the 3 month follow-up as “3 month MRI”.
We used a 1.5 T MRI scanner (MAGNETOM Aera, Siemens Healthcare, Erlangen, Germany) for the imaging survey. Conventional brain MRI included axial spin-echo T1 weighted images (T1WI) (repetition time (TR)/ echo time (TE), 500/10), axial fast-spin echo T2 weighted images (T2WI) (TR/TE 3200/115), and Gadolinium (Gd)-enhanced spin-echo T1WI in the axial, sagittal, and coronal planes. Additionally, we evaluated epidural venous engorgement using three-dimensional VIBE Gd-enhanced T1WI which focused on the cervical-to-thoracic spine level. We used whole-spine MRM with three-dimensional sampling perfection and optimized contrast using different flip-angle evolution (3D-SPACE) sequences. The MRM parameters included: TR = 3000 ms, TE = 560 ms, isotropic voxel size = 0.9 mm3, matrix size = 320×320 pixels, and field of view (FOV) = 200 mm. Additionally, we used fat suppression and a generalized auto-calibrating partially parallel acquisition (GRAPPA) imaging reconstruction with an acceleration factor of two. We captured images volumetrically in the coronal plane of the cervical-to-thoracic and thoracic-to-lumbar regions of the spine.
Additionally, this study evaluated specific qualitative brain MRI parameters including pituitary hyperemia, dural sinus engorgement, and diffuse pachymeningeal enhancement. Pituitary hyperemia was defined as a convex-shaped pituitary gland. Dural sinus engorgement was identified as a convexity in the upper surface of the midline sinus confluence using the sagittal view of the post-contrast T1WI. Diffuse pachymeningeal enhancement was defined as persistently enhanced dura mater.
Spinal leak was evaluated using the number of abnormal CSF signals along each spinal neural sleeve as identified by MRM.13 For example, abnormal linear high signals at bilateral C2-3 and C3-4 on MRM were designated a score of four. The levels of spinal epidural fluid accumulation were defined as the levels of vertebral bodies where the abnormal accumulation of either anterior or posterior epidural fluid was identified on MRM. All imaging findings were interpreted by an experienced neuroradiologist. The neuroradiologist performed the window adjustment according to the signal intensity of CSF to detect small CSF leakage while interpreting the MRM.
Targeted EBP injection
The EBP was administered at one or two vertebral levels below the site where the greatest volume of abnormal CSF signals in the neural sleeves were identified on the MRM in all the patients. The procedure was performed by an experienced anesthesiologist with a 20-gauge epidural Tuohy needle via a midline approach with the patient in a lateral recumbent position. To identify the epidural space, the loss of resistance technique was used for the lumbar spine and the hanging-drop technique for the cervical or thoracic spine.14 The autologous blood injection was continued until the patients complained of headache, back pain, or other discomforts. The patients were asked to maintain a supine position for at least 2 h following the procedure.
Treatment response assessment
The symptomatic improvement following EBP treatment was categorized into excellent, good, fair, and stationary/worse. Treatment response was classified as excellent when the patients experienced complete resolution of headache after the treatment; good when >50% of the symptoms resolved; fair when <50% of the symptoms recovered; and stationary/worse when the symptoms persisted or worsened.3 Furthermore, patients were stratified according to the degree of response into either clinical responsiveness or clinical non-responsiveness. Clinical responsiveness to EBP treatment was defined as patients with excellent or good response. On the other hand, clinical non-responsiveness was defined as fair or stationary/worse response.
The effectiveness of EBP treatment was determined using the 3 month MRI. We defined the effectiveness of EBP treatment when both the post-EBP MRI and the 3 month MRI showed no CSF leakage. Furthermore, EBP was also considered effective when patients who had spinal CSF leakage as identified by the post-EBP MRI recovered spontaneously in the 3 month MRI without additional EBP. The non-effectiveness of EBP treatment was defined as patients receiving further EBP treatment before the 3 month MRI.
Statistical analyses
Data analysis was performed using SPSS (v. 21; SPSS Inc, Chicago, IL). The normal distribution of continuous variables was tested by the Kolmogorov–Smirnov method. A nonparametric Mann–Whitney test was used to examine non-normally distributed variables. The Fisher’s exact test and χ2 test were used to analyze nominal variables. Factors attributed to EBP effectiveness were analyzed by logistic regression and quantified as odds ratios (ORs) and their accompanying 95% confidence intervals (CIs). The receiver operating characteristic (ROC) curve was used to identify the cutoff numbers of spinal CSF leakage. The predictive factors of EBP failure were determined by a forward selection model of logistic regression. Statistical significance was set at p < 0.05.
Results
Clinical characteristics
Among the 48 patients, 27 (56.25%) were included in the EBP-effective group, while 21 (43.75%) in the non-effective group according to MRI findings following the first targeted EBP treatment. A longer onset-diagnosis (23.87 vs 9.33 d; p = 0.042) and onset-EBP (26.59 vs 11.9 d; p = 0.040) interval were observed in the EBP effective group compared to the EBP non-effective group. Importantly, patients in the EBP effective group had a higher likelihood of clinical responsiveness compared to the EBP non-effective group (96.30% vs 38.10%; p = 0.001). Furthermore, the EBP effective group had fewer numbers of the spinal CSF leakage than the EBP non-effective group (mean: 10.56 vs 13.86) at the initial MRI, although the difference was not statistically significant (p = 0.063). No significant difference was observed in relation to other demographic and imaging characteristics between the two groups at diagnosis (Table 1). Notably, 21 patients required additional EBP; among them, 14, 5, and 2 patients required two, three, and four additional EBP treatments within 3 months, respectively.
Table 1.
Comparison of the clinical and initial MRI parameters regarding to effectiveness of EBP
| Total (n = 48) |
EBP effective (n = 27) |
EBP non-effective (n = 21) | p-value | |
|---|---|---|---|---|
| Age, years | 39.19 ± 9.46 | 39.59 ± 10.69 | 38.67 ± 7.84 | 0.942a |
| Gender, n (%) | 0.770b | |||
| Male | 19 (39.58%) | 10 (37.04%) | 9 (42.86%) | |
| Female | 29 (60.42%) | 17 (62.96%) | 12 (57.14%) | |
| Etiology | 1.000c | |||
| Post-lumbar puncture headache | 3 | 2 (7.41%) | 1 (4.76%) | |
| Spontaneous intracranial hypotension | 45 | 25 (92.59%) | 20 (95.24%) | |
| Headache score | 7.24 ± 1.88 | 6.98 ± 1.90 | 7.57 ± 1.83 | 0.363a |
| Onset-diagnosis interval, days | 17.52 ± 27.67 | 23.89 ± 35.47 | 9.33 ± 6.28 | 0.042a |
| Onset-EBP interval, days | 29.17 ± 27.75 | 26.59 ± 35.43 | 11.9 ± 7.11 | 0.040a |
| Days between EBP and follow up MRI | 3.46 ± 2.25 | 3.33 ± 2.39 | 3.62 ± 2.11 | 0.611a |
| Hospitalization duration | 15.67 ± 8.46 | 11.85 ± 4.29 | 20.57 ± 9.97 | 0.001a |
| Numbers of spinal CSF leakage | 12.00 ± 7.26 | 10.56 ± 7.45 | 13.86 ± 6.72 | 0.063a |
| Levels of spinal epidural fluid accumulation | 6.94 ± 5.11 | 5.81 ± 5.00 | 8.38 ± 5.01 | 0.083a |
| Spine epidural venous engorgement, n (%) | 0.082b | |||
| No | 21 (43.75%) | 15 (55.56%) | 6 (28.57%) | |
| Yes | 27 (56.25%) | 12 (44.44%) | 15 (71.43%) | |
| Diffuse pachymeningeal enhancement, n (%) | 1.000b | |||
| No | 12 (25.00%) | 7 (25.93%) | 5 (23.81%) | |
| Yes | 36 (75.00%) | 20 (74.07%) | 16 (76.19%) | |
| Brain descend, n (%) | 0.555c | |||
| No | 32 (66.67%) | 19 (70.37%) | 13 (61.90%) | |
| Yes | 16 (33.33%) | 8 (29.63%) | 8 (38.10%) | |
| Dural sinus engorgement, n (%) | 0.485b | |||
| No | 9 (18.75%) | 6 (22.22%) | 3 (14.29%) | |
| Yes | 39 (81.25%) | 21 (77.78%) | 18 (85.71%) | |
| Subdural hematoma, n (%) | 0.897b | |||
| No | 37 (77.08%) | 21 (77.78%) | 16 (76.19%) | |
| Yes | 11 (22.92%) | 6 (22.22%) | 5 (23.81%) | |
| Pituitary hyperemia, n (%) | 0.153b | |||
| No | 22 (45.83%) | 15 (55.56%) | 7 (33.33%) | |
| Yes | 26 (54.17%) | 12 (44.44%) | 14 (66.67%) | |
| Clinical response to EBP, n (%) | 0.001b | |||
| Responsive | 34 (70.83%) | 26 (96.30%) | 8 (38.10%) | |
| Non-responsive | 14 (29.17%) | 1 (3.70%) | 13 (61.90%) |
CSF, cerebrospinal fluid; EBP, epidural blood patch.
Numerical data are presented as mean ± standard deviation.
Mann–Whitney test.
χ2 test.
Fisher’s Exact test.
Comparison of the post-EBP MRI findings between the EBP effective and non-effective groups
Compared to the EBP non-effective group, the EBP effective group had fewer numbers of spinal CSF leakage (2.67 vs 12.48; p = 0.001), fewer levels of spinal epidural fluid accumulation (3.00 vs 7.48; p = 0.004), brain descend (11.11% vs 38.10%; p = 0.027), dural sinus engorgement (33.33% vs 71.43%; p = 0.008), pituitary hyperemia (18.52% vs 57.14%; p = 0.007), and a decreased likelihood of obtaining ≥three numbers of spinal CSF leakage (25.93% vs 90.48%; p = 0.001) (Table 2).
Table 2.
Comparison of the post-EBP MRI parameters regarding to effectiveness of EBP
| Total (n = 48) |
EBP effective (n = 27) |
EBP non-effective (n = 21) | p-value | |
|---|---|---|---|---|
| Numbers of spinal CSF leakage | 6.96 ± 7.25 | 2.67 ± 3.80 | 12.48 ± 6.93 | 0.001 |
| Levels of spinal epidural fluid accumulation | 4.96 ± 5.28 | 3.00 ± 4.62 | 7.48 ± 5.10 | 0.004 |
| Brain descend, n (%) | 0.027b | |||
| No | 37 (77.08%) | 24 (88.89%) | 13 (61.90%) | |
| Yes | 11 (22.92%) | 3 (11.11%) | 8 (38.10%) | |
| Dural sinus engorgement, n (%) | 0.008c | |||
| No | 25 (52.08%) | 19 (66.67%) | 6 (28.57%) | |
| Yes | 23 (47.92%) | 8 (33.33%) | 15 (71.43%) | |
| Subdural hematoma, n (%) | 0.741c | |||
| No | 36 (75.00%) | 21 (77.78%) | 15 (71.43%) | |
| Yes | 12 (25.00%) | 6 (22.22%) | 6 (28.57%) | |
| Pituitary hyperemia, n (%) | 0.007c | |||
| No | 31 (64.58%) | 22 (81.48%) | 9 (42.86%) | |
| Yes | 17 (35.42%) | 5 (18.52%) | 12 (57.14%) | |
| Levels of spinal CSF leakage ≥3, n (%) | 0.001c | |||
| No | 22 (45.83%) | 20 (74.07%) | 2 (9.52%) | |
| Yes | 26 (54.17%) | 7 (25.93%) | 19 (90.48%) | |
| Clinical response, n (%) | 0.001c | |||
| Responsive | 34 (70.83%) | 26 (96.30%) | 8 (38.10%) | |
| Non-responsive | 14 (29.17%) | 1 (3.70%) | 13 (61.90%) |
CSF, cerebrospinal fluid; EBP, epidural blood patch.
Numerical data are presented as mean ± standard deviation.
Mann–Whitney test.
Fisher’s Exact test.
χ2 test.
Post-EBP MRI findings associated with failure of targeted EBP therapy
The univariate analysis showed that ≥three numbers of spinal CSF leakage (OR: 27.14; 95% CI: 5.00–147.43; p = 0.000), greater levels of spinal epidural fluid accumulation (OR: 1.2; 95% CI: 1.05–1.37; p = 0.006), pituitary hyperemia (OR: 5.87; 95% CI: 1.39–16.20; p = 0.013), brain descend (OR: 4.92; 95% CI: 1.11–21.82; p = 0.036), dural sinus engorgement (OR: 27.14; 95% CI: 1.69–20.86; p = 0.005), and clinical non-responsiveness (OR: 42.25; 95% CI: 4.76–374.83; p = 0.001) were significantly associated with non-effective EBP treatment (Table 3).
Table 3.
Risk factors for “EBP non-effective”
| Univariate analysis | Multivariate analysis | |||||
|---|---|---|---|---|---|---|
| OR | 95 % CI | p -value | OR | 95 % CI | p -value | |
| Onset-diagnosis interval | 0.928 | 0.84–1.04 | 0.063 | 0.286 | ||
| Onset-EBP interval | 0.778 | 0.388 | 0.377 | |||
| Numbers of spinal CSF leakage (≥3 vs.<3) | 27.14 | 5.00–147.43 | 0.000 | 15.13 | 1.45–159.06 | 0.023 |
| Levels of spinal epidural fluid accumulation | 1.20 | 1.05–1.37 | 0.006 | 0.707 | ||
| Pituitary hyperemia (Yes vs No) | 5.87 | 1.60–21.53 | 0.008 | 0.954 | ||
| Brain descend (Yes vs No) | 4.92 | 1.11–21.82 | 0.036 | 0.986 | ||
| Dural sinus engorgement (Yes vs No) | 5.94 | 1.69–20.86 | 0.005 | 0.554 | ||
| Clinical response (non-responsive vs responsive) | 42.25 | 4.76–374.83 | 0.001 | 57.84 | 3.47–972.54 | 0.005 |
CI, confidence interval; CSF, cerebrospinal fluid; EBP, epidural blood patch; OR, odds ratio.
Among these factors, further multivariate analysis validated both ≥three numbers of spinal CSF leakage (OR: 15.13; 95% CI: 1.45–159.06; p = 0.023) and clinical non-responsiveness (OR: 57.84; 95% CI: 3.47–972.54; p = 0.005) as independent factors that predict the failure of EBP treatment (Table 3).
Prediction of EBP failure based on symptomatic response and post-EBP MRI findings
We evaluated the potential of ≥three numbers of spinal CSF leakage identified by the post-EBP MRI as a surrogate marker for predicting EBP failure. Additionally, the significant relationship between clinical non-responsiveness and EBP failure prompted the comparison of clinical non-responsiveness and ≥three numbers of spinal CSF leakage identified by the post-EBP MRI as predictors of EBP failure. The results showed that the sensitivity of ≥three numbers of spinal CSF leakage identified by the post-EBP MRI and clinical non-responsiveness to EBP was 90.48 and 61.90%, respectively; additionally, the specificity was 74.07 and 96.30%, respectively. These data suggest that ≥three numbers of spinal CSF leakage identified by the post-EBP MRI was a more sensitive predictor of EBP failure compared to clinical non-responsiveness (Table 4).
Table 4.
Prediction of EBP treatment failure based on clinical response and spinal CSF leakage numbers according to the post-EBP MRI
| Total (n) | EBP effective (n) | EBP non-effective (n) | Sensitivity | Specificity | PPV | NPV | Accuracy | ||
|---|---|---|---|---|---|---|---|---|---|
| Clinically non-responsive to EBP | Yes | 14 | 1 | 13 | 61.90% | 96.3% | 92.86% | 76.47% | 81.25% |
| No | 34 | 26 | 8 | ||||||
| Numbers of spinal CSF leakage ≥3 | Yes | 26 | 7 | 19 | 90.48% | 74.07% | 73.08% | 90.91% | 81.25% |
| No | 22 | 20 | 2 |
CSF, cerebrospinal fluid; EBP, epidural blood patch; NPV, negative predictive value; PPV, positive predictive value.
Case presentation
A 34-year-old female complained of orthostatic headache lasting for 1 week. The MRI and MRM images showed CSF leakage into the bilateral C5-T10 neural sleeves (Figure 2A) with epidural fluid accumulation over C4-L2 (Figure 2B and C). Additionally, the conventional MRI demonstrated abnormal spinal epidural venous and dural sinus engorgement. She underwent her first targeted EBP at the T3 level 13 days following the onset of symptoms due to failure to respond to initial hydration. After the first salvage EBP, significant improvement in the patient’s orthostatic headache was observed. However, the post-EBP MRI performed 3 days after the treatment revealed abnormal CSF signals along bilateral C7-T6 neural sleeves (Figure 2D) with epidural fluid accumulation over T1-L2. The orthostatic headache worsened five days after EBP treatment. After three EBP treatments, complete resolution of symptoms was observed. The 3 month MRI revealed absence of CSF leakage (Figure 2E).
Figure 2.
(A) 3D MIP MRM of C-T spine shows CSF leakage at multiple neural sleeves (arrow). (B) Sagittal T2WI MRI of C-T spine shows abnormal anterior and posterior epidural fluid accumulation (arrowhead). (C) Sagittal T1WI magnetic resonance imaging of C-T spine shows abnormal anterior and posterior epidural fluid accumulation (arrowhead). (D) The post-epidural blood patch 3D MIP MRM reveals CSF leakage at multiple neural sleeves (arrow) despite significant improvements in the patient’s symptoms. (E) The 3 month 3D MIP MRM reveals absence of CSF leakage. 3D MIP, three-dimensional maximum intensity projection; CSF, cerebrospinal fluid; EBP, epidural blood patch; MRM, magnetic resonance myelography.; T1WI, T1 weighted imaging
Discussion
The current study aimed to evaluate the response of spinal CSF leakage to EBP treatment using the findings from the post-EBP MRI. We found that ≥three numbers of spinal CSF leakage identified by the post-EBP MRI was significantly associated with EBP treatment failure. The similar result was also observed in patients without clinical responsiveness. However, the post-EBP MRI provided a higher sensitivity than clinical non-responsiveness.
The clinical course of spinal CSF leakage is dependent on the etiology.15 In particular, 86% of patients with CSF leakage due to lumbar puncture are asymptomatic and require only observation.16 Additionally, intensive hydration is sufficient to manage headaches due to dural puncture-induced CSF leakage.17 On the other hand, the clinical course of patients with CSF leakage due to spontaneous intracranial hypotension is different.18 In these patients, initial hydration is usually ineffective; therefore, salvage EBP is a critical management procedure.8,19 The indication for EBP in most patients in this study was spontaneous intracranial hypotension which may be due to the ineffectiveness of initial hydration.
Predicting the clinical response of patients with CSF leakage to treatment remains a challenge. In relation to the initial MRI findings, only anterior epidural fluid accumulation is a consistent predictive factor of treatment failure, specifically, the accumulation of eight vertebral segments of anterior epidural fluid as demonstrated by a previous study.19 However, this was inconsistent with our findings. In our study, no significant difference was observed in relation to the anterior and posterior epidural fluid accumulation levels identified by the initial MRI between the EBP effective and non-effective groups. Furthermore, the results of our study indicated that anterior and posterior epidural fluid accumulation as identified by the post-EBP MRI were observed in 12 and 25 patients, respectively. However, the total levels (summation of levels of anterior and posterior) of epidural fluid accumulation and the posterior epidural fluid accumulation levels identified by the post-EBP MRI were lower in the EBP effective group compared to the non-effective group (data not shown). Instead of analyzing the separate anterior and posterior epidural fluid accumulation levels, the effectiveness of the total levels of epidural fluid accumulation was evaluated to simplify clinical application. Although the difference was not substantial, the number of spinal CSF leakage identified by the initial MRI was lower in the EBP effective group than that in the EBP non-effective group. In order to reach statistical significance, increasing the size of the study cohort might be beneficial in future studies. Nevertheless, this finding suggests the superior effectiveness of post-EBP MRI in predicting EBP efficacy compared to an initial MRI.
The cut-off value of spinal CSF leakage numbers to predict targeted EBP failure remains controversial. In this study, we used a regression analysis to identify that ≥three numbers of spinal CSF leakage was a significant predictive factor of targeted EBP treatment failure. Importantly, our study revealed that the median CSF leakage number was six during the post-EBP MRI, suggesting that ≥three numbers of spinal CSF leakage is feasible in daily practice. In relation to the interval between symptom onset and EBP treatment, Kanno et al3 found better treatment outcomes in patients with a shorter onset–EBP treatment interval. However, this was inconsistent with our finding of shorter onset–diagnosis and onset-EBP intervals in the EBP non-effective group compared to the EBP effective group. Kanno et al3 had a longer onset–EBP interval compared to our study (median: 1.5 years vs mean: 29 days), which indicates the difference in stages between our samples and may be the reason for the data discrepancy. An increased severity of leakage is associated with increased number of CSF leakage identified by the post-EBP MRI, and a shorter onset–diagnosis or onset–EBP treatment interval. The shorter intervals urge patients to seek medical help after symptom onset and avail of medical services which results in prompt medical management. Therefore, the numbers of CSF leakage identified by the post-EBP MRI might be more specific for leakage severity and have better predictive values for EBP efficacy compared to onset-diagnosis or onset–EBP intervals. In summary, decreasing disease severity remains the key to improving treatment outcomes in spinal CSF leakage.
Unlike other studies that used symptomatic improvement in evaluating treatment efficacy,18,20,21 our study utilized the 3 month MRI to measure outcomes because MRI findings were more objective compared to symptomatic improvements. In previous studies, patients who underwent additional EBP treatment for persistent headaches were classified as EBP non-effective. Compared with the objective image evaluation, symptom persistency remained an effective predictor for non-effectiveness of EBP treatment in our study (accuracy: 81.25%). However, spontaneous recovery remains a possibility in patients with spinal CSF leakage.18 In our study, we used the findings of the 3 month MRI as the end point. The observation was the clinical decision of patients tolerating their symptoms despite the persistence of spinal CSF leakage in their post-EBP MRI. Notably, the 3 month MRI found spontaneous recovery among patients with CSF leakage, especially those with <three numbers of spinal CSF leakage identified by the post-EBP MRI. However, rapid recurrence after symptomatic improvement following EBP treatment remains problematic, especially in patients with a higher number of spinal CSF leakage. The presented case serves as an example, which showed immediate recurrence of the headache when the post-EBP MRI identified ≥three numbers of spinal CSF leakage. This result suggested that immediate spinal CSF leakage may be observed in patients with ≥three numbers of spinal CSF leakage identified by the post-EBP MRI, despite significant improvements in orthostatic headache after EBP treatment. Therefore, a closer follow-up is warranted in these patients compared to those with <three numbers of spinal CSF leakage.
Many imaging modalities have been proposed to accurately assess the CSF leakage site. Among these modalities, the intrathecal digital subtraction myelography,22 or CT myelogrpahy23 are the gold-standard because these approaches directly visualize the contrast leakage from the leakage site and are more sensitive to ventral CSF leakage and CSF-venous fistula. However, the procedures of intrathecal digital subtraction myelography and CT myelography are relatively invasive, which might deter their clinical application. In the current study, we used MRM to identify the site of spinal CSF leakage because the MRM is non-invasive and resulted in less exposure to radiation. Furthermore, studies have demonstrated that the MRM, CT myelography, and intrathecal gadolinium-enhanced MRI are comparable in their diagnostic ability to identify spinal CSF leakage.24–26 Importantly, MRM and CT myelography had a substantial overall level-by-level concordance for CSF leaks along the nerve roots.26 In our study, all the patients had abnormal spinal CSF signals along the neural sleeves. Among all the study individuals, 10 patients without epidural fluid accumulation had CSF leakage along the neural sleeves and these 10 patients recovered completely after the single EBP. This result suggested that our study population might not contain patients with CSF-venous fistula, because patients with CSF-venous fistula usually do not have epidural fluid accumulation but are refractory to EBP treatment.22 Taking into consideration these factors, digital subtraction myelography might be the most appropriate diagnostic method to detect CSF-venous fistula for patients suspected to have spinal CSF leakage without epidural fluid accumulation or abnormal spinal CSF leakage along the neural sleeves as identified by MRM.27
This study has several limitations. The small sample size and retrospective nature of the study may limit the generalizability of the results. Additionally, symptomatic improvement is a subjective method of assessment which may have interfered with our outcome evaluation. Furthermore, all the images were evaluated by the same neuroradiologist, which might result in a potential assessment bias. Therefore, further studies utilizing a prospective design with a larger sample size are warranted to validate the results.
In conclusion, this study found that ≥three numbers of spinal CSF leakage identified using the post-EBP MRI was significantly related to EBP treatment failure and was a better predictor of EBP treatment outcomes compared to clinical non-responsiveness in patients with spinal CSF leakage who were non-responsive to primary hydration therapy.
Footnotes
Competing interests: Chieh-Lin Jerry Teng received an honorarium and consulting fees from Novartis, Roche, Takeda, Johnson & Johnson, Amgen, BMS Celgene, Kirin, AbbVie, and MSD.
Contributor Information
Hung-Chieh Chen, Email: hungchiehchen@gmail.com.
Jyh-wen Chai, Email: hubtchai@gmail.com.
Chih-Cheng Wu, Email: chihcheng.wu@gmail.com.
Po-Lin Chen, Email: boringtw@gmail.com.
Chieh-Lin Teng, Email: drteng@vghtc.gov.tw.
REFERENCES
- 1.Chen H-C, Chen P-L, Tsai Y-H, Chen C-H, Chen CC-C, Chai J-W. Quantitative measurement of CSF in patients with spontaneous intracranial hypotension. AJNR Am J Neuroradiol 2017; 38: 1061–7. doi: 10.3174/ajnr.A5134 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Kranz PG, Gray L, Malinzak MD, Amrhein TJ. Spontaneous intracranial hypotension: pathogenesis, diagnosis, and treatment. Neuroimaging Clin N Am 2019; 29: 581–94. doi: 10.1016/j.nic.2019.07.006 [DOI] [PubMed] [Google Scholar]
- 3.Kanno H, Yoshizumi T, Nakazato N, Shinonaga M. Predictors of the response to an epidural blood patch in patients with spinal leakage of cerebrospinal fluid. J Clin Neurol 2020; 16: 1–8. doi: 10.3988/jcn.2020.16.1.1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Yoo H-M, Kim SJ, Choi CG, Lee DH, Lee JH, Suh DC, et al. Detection of CSF leak in spinal CSF leak syndrome using MR myelography: correlation with radioisotope cisternography. AJNR Am J Neuroradiol 2008; 29: 649–54. doi: 10.3174/ajnr.A0920 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Pannullo SC, Reich JB, Krol G, Deck MD, Posner JB. MRI changes in intracranial hypotension. Neurology 1993; 43: 919–26. doi: 10.1212/WNL.43.5.919 [DOI] [PubMed] [Google Scholar]
- 6.Farb RI, Forghani R, Lee SK, Mikulis DJ, Agid R. The venous distension sign: a diagnostic sign of intracranial hypotension at MR imaging of the brain. AJNR Am J Neuroradiol 2007; 28: 1489–93. doi: 10.3174/ajnr.A0621 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Mokri B, Atkinson JL. False pituitary tumor in CSF leaks. Neurology 2000; 55: 573–5. doi: 10.1212/WNL.55.4.573 [DOI] [PubMed] [Google Scholar]
- 8.Mokri B. The Monro-Kellie hypothesis: applications in CSF volume depletion. Neurology 2001; 56: 1746–8. doi: 10.1212/WNL.56.12.1746 [DOI] [PubMed] [Google Scholar]
- 9.Harrington BE, Schmitt AM. Meningeal (postdural) puncture headache, unintentional dural puncture, and the epidural blood patch: a national survey of United States practice. Reg Anesth Pain Med 2009; 34: 430–7. doi: 10.1097/AAP.0b013e3181b493e9 [DOI] [PubMed] [Google Scholar]
- 10.van Kooten F, Oedit R, Bakker SLM, Dippel DWJ. Epidural blood patch in post dural puncture headache: a randomised, observer-blind, controlled clinical trial. J Neurol Neurosurg Psychiatry 2008; 79: 553–8. doi: 10.1136/jnnp.2007.122879 [DOI] [PubMed] [Google Scholar]
- 11.Mokri B. Spontaneous low pressure, low CSF volume headaches: spontaneous CSF leaks. Headache 2013; 53: 1034–53. doi: 10.1111/head.12149 [DOI] [PubMed] [Google Scholar]
- 12.Sencakova D, Mokri B, McClelland RL. The efficacy of epidural blood patch in spontaneous CSF leaks. Neurology 2001; 57: 1921–3. doi: 10.1212/WNL.57.10.1921 [DOI] [PubMed] [Google Scholar]
- 13.Chen C-H, Chen J-H, Chen H-C, Chai J-W, Chen P-L, Chen CC-C. Patterns of cerebrospinal fluid (CSF) distribution in patients with spontaneous intracranial hypotension: assessed with magnetic resonance myelography. J Chin Med Assoc 2017; 80: 109–16. doi: 10.1016/j.jcma.2016.02.013 [DOI] [PubMed] [Google Scholar]
- 14.Hoffmann VL, Vercauteren MP, Vreugde JP, Hans GH, Coppejans HC, Adriaensen HA. Posterior epidural space depth: safety of the loss of resistance and hanging drop techniques. Br J Anaesth 1999; 83: 807–9. doi: 10.1093/bja/83.5.807 [DOI] [PubMed] [Google Scholar]
- 15.Inamasu J, Guiot BH. Intracranial hypotension with spinal pathology. Spine J 2006; 6: 591–9. doi: 10.1016/j.spinee.2005.12.026 [DOI] [PubMed] [Google Scholar]
- 16.Sakurai K, Matsukawa N, Okita K, Nishio M, Shimohira M, Ozawa Y, et al. Lumbar puncture-related cerebrospinal fluid leakage on magnetic resonance myelography: is it a clinically significant finding? BMC Anesthesiol 2013; 13: 35. doi: 10.1186/1471-2253-13-35 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Wang Y-F, Fuh J-L, Lirng J-F, Chen S-P, Hseu S-S, Wu J-C, et al. Cerebrospinal fluid leakage and headache after lumbar puncture: a prospective non-invasive imaging study. Brain 2015; 138(Pt 6): 1492–8. doi: 10.1093/brain/awv016 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Ferrante E, Trimboli M, Rubino F. Spontaneous intracranial hypotension: review and expert opinion. Acta Neurol Belg 2020; 120: 9–18. doi: 10.1007/s13760-019-01166-8 [DOI] [PubMed] [Google Scholar]
- 19.Wu J-W, Hseu S-S, Fuh J-L, Lirng J-F, Wang Y-F, Chen W-T, et al. Factors predicting response to the first epidural blood patch in spontaneous intracranial hypotension. Brain 2017; 140: 344–52. doi: 10.1093/brain/aww328 [DOI] [PubMed] [Google Scholar]
- 20.Berroir S, Loisel B, Ducros A, Boukobza M, Tzourio C, Valade D, et al. Early epidural blood patch in spontaneous intracranial hypotension. Neurology 2004; 63: 1950–1. doi: 10.1212/01.WNL.0000144339.34733.E9 [DOI] [PubMed] [Google Scholar]
- 21.Cho K-I, Moon H-S, Jeon H-J, Park K, Kong D-S. Spontaneous intracranial hypotension: efficacy of radiologic targeting vs blind blood patch. Neurology 2011; 76: 1139–44. doi: 10.1212/WNL.0b013e318212ab43 [DOI] [PubMed] [Google Scholar]
- 22.Luetzen N, Dovi-Akue P, Fung C, Beck J, Urbach H. Spontaneous intracranial hypotension: diagnostic and therapeutic workup. Neuroradiology 2021; 63: 1765–72. doi: 10.1007/s00234-021-02766-z [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Schievink WI. Spontaneous spinal cerebrospinal fluid leaks and intracranial hypotension. JAMA 2006; 295: 2286–96. doi: 10.1001/jama.295.19.2286 [DOI] [PubMed] [Google Scholar]
- 24.Dobrocky T, Winklehner A, Breiding PS, Grunder L, Peschi G, Häni L, et al. Spine MRI in spontaneous intracranial hypotension for CSF leak detection: Nonsuperiority of intrathecal gadolinium to heavily T2-weighted Fat-Saturated sequences. AJNR Am J Neuroradiol 2020; 41: 1309–15. doi: 10.3174/ajnr.A6592 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Tay AS-MS, Maya M, Moser FG, Nuño M, Schievink WI. Computed tomography vs heavily T2-weighted magnetic resonance myelography for the initial evaluation of patients with spontaneous intracranial hypotension. JAMA Neurol 2021; 78: 1275–6. doi: 10.1001/jamaneurol.2021.2868 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Wang Y-F, Lirng J-F, Fuh J-L, Hseu S-S, Wang S-J. Heavily T2-weighted MR myelography vs CT myelography in spontaneous intracranial hypotension. Neurology 2009; 73: 1892–8. doi: 10.1212/WNL.0b013e3181c3fd99 [DOI] [PubMed] [Google Scholar]
- 27.Farb RI, Nicholson PJ, Peng PW, Massicotte EM, Lay C, Krings T, et al. Spontaneous intracranial hypotension: a systematic imaging approach for CSF leak localization and management based on MRI and digital subtraction myelography. AJNR Am J Neuroradiol 2019; 40: 745–53. doi: 10.3174/ajnr.A6016 [DOI] [PMC free article] [PubMed] [Google Scholar]