Background and objective Epidural blood patch (EBP) is a safe and effective treatment for spontaneous intracranial hypotension (SIH), but clinical and procedural variables that predict EBP efficacy remain nebulous.

Methods This study is an institutional review board-approved retrospective case series with dichotomized EBP efficacy defined at 3 months. The study included 202 patients receiving 604 EBPs; iatrogenic cerebrospinal fluid leaks were excluded.

Results Of the EBPs, 473 (78%) were single-level, 349 (58%) lumbar, 75 (12%) bilevel, and 56 (9%) multilevel (≥3 levels). Higher volume (OR 1.64; p<0.0001), bilevel (3.17, 1.91–5.27; p<0.0001), and multilevel (117.3, 28.04–490.67; p<0.0001) EBP strategies predicted greater efficacy. Only volume (1.64, 1.47–1.87; p<0.0001) remained significant in multivariate analysis. Site-directed patches were more effective than non-targeted patches (8.35, 0.97–72.1; p=0.033). Lower thoracic plus lumbar was the most successful bilevel strategy, lasting for a median of 74 (3–187) days.

Conclusions In this large cohort of EBP in SIH, volume, number of spinal levels injected, and site-directed strategies significantly correlated with greater likelihood of first EBP efficacy. Volume and leak site coverage likely explain the increased efficacy with bilevel and multilevel patches. In patients with cryptogenic leak site, and either moderate disability, negative prognostic brain MRI findings for successful EBP, or failed previous lumbar EBP, a low thoracic plus lumbar bilevel EBP strategy is recommended. Multilevel EBP incorporating transforaminal administration and fibrin glue should be considered in patients refractory to bilevel EBP. An algorithmic approach to treating SIH is proposed.