OBJECTIVES: To analyze and compare differences between epidemiological and clinical aspects, as well as radiologic findings and treatment, in a series of adult patients with traumatic intracranial hypotension (TIH) and spontaneous intracranial hypotension (SIH) treated at our institution in order to identify predictors of recurrence. BACKGROUND: Cerebrospinal fluid hypotension headache (CSF-HH) is often caused by orthostasis and relieved by recumbency. Etiology can be either traumatic or spontaneous. Indirect signs of CSF hypotension are often observed on brain MRI. The most common therapeutic approach is conservative management and, when necessary, the use of an epidural blood patch. METHODS: Medical history and brain MRI of adult patients consulting our institution with a diagnosis of CSF-HH between January 2010 and March 2019, were retrospectively reviewed. Clinical criteria as per the International Classification of Headache Disorders, 3rd edition, were applied. Presence of typical MRI findings were assessed by two experienced neuroradiologists, previously informed of patients' clinical characteristics. Patients were divided into two different groups, namely: Group A: Spontaneous Intracranial Hypotension (SIH) and Group B: Traumatic Intracranial Hypotension (TIH). Recurrence was defined as return of symptoms after one month of remission. In order to find predictors of recurrent intracranial hypotension the patients were divided into three groups: Recurrent Orthostatic headache (ROH); (33 cases; 25%); Non-Recurrent Orthostatic headache (NROH) (84; 61%) and Patients missing follow-up (20; 15%). The latter were excluded from the regression analysis. RESULTS: 137 patients with CSF-HH were identified: 80 traumatic (54 women, age 33.8 ±â€¯10.4 years) and 57 spontaneous (31 women, age 43.9 ±â€¯15.2 years). Median follow-up was 35 months (range: 8 months-9 years). Compared with TIH, patients with SIH showed lower frequency of orthostatic headache and higher frequency of aural fullness. Also, in patients with SIH, brain MRI showed a higher frequency of pachymeningeal enhancement, ventricular collapse, cisternal obliteration, posterior fossa crowding, brainstem distortion, and a more likely presence of subdural collections. Patients with SIH required an epidural blood patch treatment more often, showed higher recurrence rates, and were more prone to develop subdural hematomas. RECURRENCE: As more days elapsed between headache onset and clinical consultation, the presence of spontaneous intracranial hypotension, tinnitus, and thoraco-lumbo-sacral pain were all more common in patients with recurrence. Findings on brain MRI suggesting higher recurrence rates in patients included: ventricular collapse, brainstem distortion, and posterior fossa crowding. Patients treated with invasive therapy (epidural blood patch) presented a higher recurrence rate. In the multivariate regression analysis, the only independent predictor of recurrence after adjusting for age, sex and traumatic vs spontaneous cause of IH, was brainstem distortion diagnosed on MRI (OR 5.13, 95% CI: 1.2-21.7; p = 0.026). CONCLUSIONS: SIH and TIH can no longer be likened, since there is considerable variability in clinical presentation, imaging findings, response to treatment and recurrence rates. Anatomical abnormalities underlying SIH leaks are often complex and not simply a disruption of normal structures as encountered in TIH, which could explain why treatment success is poor and recurrence rates remain high.