Factors to predict recurrence after epidural blood patch in patients with spontaneous intracranial hypotension

• Published in: Headache Vol. 64; no. 4; pp. 380 - 389
• Main Authors: Lee, Seung Hyun, Lee, Jooyoung, Kim, Da-Woon, Kim, Dong Hyun, Ahn, Sung Jae, Choi, Moon Gwan, Jo, Sungyang, Suh, Chong Hyun, Chung, Sun J
• Format: Journal Article
• Language: English
• Published: United States Wiley Subscription Services, Inc 01.04.2024
• Subjects: Blood; Brain; Cerebrospinal fluid; Epidural; Health care facilities; Health services; Hypotension; Machine learning; Magnetic resonance; Magnetic resonance imaging; Medical imaging; Mesencephalon; Neuroimaging; Patients; Pons; Volumetric analysis; epidural blood patch; orthostatic headache; volumetric analysis; spontaneous intracranial hypotension
• ISSN: 0017-8748; 1526-4610
• DOI: 10.1111/head.14703

This study aimed to identify predictors for the recurrence of spontaneous intracranial hypotension (SIH) after epidural blood patch (EBP). Epidural blood patch is the main treatment option for SIH; however, the characteristics of patients who experience relapse after successful EBP treatment for SIH remain understudied. In this exploratory, retrospective, case-control study, we included 19 patients with SIH recurrence after EBP and 36 age- and sex-matched patients without recurrence from a single tertiary medical institution. We analyzed clinical characteristics, neuroimaging findings, and volume changes in intracranial structures after EBP treatment. Machine learning methods were utilized to predict the recurrence of SIH after EBP treatment. There were no significant differences in clinical features between the recurrence and no-recurrence groups. Among brain magnetic resonance imaging signs, diffuse pachymeningeal enhancement and cerebral venous dilatation were more prominent in the recurrence group than no-recurrence group after EBP (14/19 [73%] vs. eight of 36 [22%] patients, p = 0.001; 11/19 [57%] vs. seven of 36 [19%] patients, p = 0.010, respectively). The midbrain-pons angle decreased in the recurrence group compared to the no-recurrence group after EBP, at a mean (standard deviation [SD]) of -12.0 [16.7] vs. +1.8[18.3]° (p = 0.048). In volumetric analysis, volume changes after EBP were smaller in the recurrence group than in the no-recurrence group in intracranial cerebrospinal fluid (mean [SD] -11.6 [15.3] vs. +4.8 [17.1] mL, p = 0.001) and ventricles (mean [SD] +1.0 [2.0] vs. +2.0 [2.5] mL, p = 0.003). Notably, the random forest classifier indicated that the model constructed with brain volumetry was more accurate in discriminating SIH recurrence (area under the curve = 0.80 vs. 0.52). Our study suggests that volumetric analysis of intracranial structures may aid in predicting recurrence after EBP treatment in patients with SIH.