CLINICAL REVIEW CLINICIAN’S CORNER Spontaneous Spinal Cerebrospinal Fluid Leaks and Intracranial Hypotension Wouter I. Schievink, MD Context Spontaneous intracranial hypotension is caused by spontaneous spinal ce- PATIENT PRESENTS WITH A rebrospinal fluid (CSF) leaks and is known for causing orthostatic headaches. It is an new headache that occurs important cause of new headaches in young and middle-aged individuals, but initial shortly after assuming an up- misdiagnosis is common. right position and is re- Objective To summarize existing evidence regarding the epidemiology, pathophysi- Alieved by lying down. Although such a ology, diagnosis, and management of spontaneous spinal CSF leaks and intracranial positional headache pattern is well- hypotension. known following a diagnostic lumbar Evidence Acquisition MEDLINE (1966-2005) and OLDMEDLINE (1950-1965) were puncture, the spontaneous onset of an searched using the terms intracranial hypotension, CSF leak, low pressure headache, orthostatic headache is not well recog- and CSF hypovolemia. Reference lists of these articles and ongoing investigations in nized and the patient may be diag- this area were used as well. nosed with migraine, tension head- Evidence Synthesis Spontaneous intracranial hypotension is caused by single or ache, viral meningitis, or malingering. multiple spinal CSF leaks. The incidence has been estimated at 5 per 100 000 per year, This has been a typical scenario for with a peak around age 40 years. Women are affected more commonly than men. Mechanical factors combine with an underlying connective tissue disorder to cause many patients experiencing spontane- 1 the CSF leaks. An orthostatic headache is the prototypical manifestation but other head- ous intracranial hypotension. The ache patterns occur as well, and associated symptoms are common. Typical magnetic spontaneous form of intracranial hy- resonance imaging findings include subdural fluid collections, enhancement of the pachy- potension was first described in 1938,2 meninges, engorgement of venous structures, pituitary hyperemia, and sagging of the and much has been learned about this brain (mnemonic: SEEPS). Myelography is the study of choice to identify the spinal syndrome, particularly since the early CSF leak. Treatments include bed rest, epidural blood patching, percutaneous place- 1990s,3-15 but an initial misdiagnosis re- ment of fibrin sealant, and surgical CSF leak repair, but outcomes have been poorly mains the norm. Unfamiliarity with studied and no management strategies have been studied in properly controlled ran- spontaneous intracranial hypotension domized trials. among physicians in general and the Conclusions Spontaneous intracranial hypotension is not rare but it remains under- unusually varied spectrum of clinical diagnosed. The spectrum of clinical and radiographic manifestations is varied, with di- and radiographic manifestations may all agnosis largely based on clinical suspicion, cranial magnetic resonance imaging, and myelography. Numerous treatment options are available, but much remains to be learned contribute to a delay in diagnosis that about this disorder. often is measured in months or even JAMA. 2006;295:2286-2296 www.jama.com years and decades.1 EVIDENCE ACQUISITION also were used. Clinical trials were not spontaneous intracranial hypotension The material covered in this review is available, and prospective studies were is not that rare and has to be consid- based on a systematic review of journal selected over retrospective studies. Se- ered an important cause of new daily articles in MEDLINE (1966-2005) and lected articles were largely those pub- persistent headaches, particularly OLDMEDLINE (1950-1965) using the lished within the past 10 years and hav- among young and middle-aged indi- terms intracranial hypotension, CSF leak, ing adequate documentation and relevant viduals. In the past, our knowledge re- low pressure headache, and CSF hypovo- clinical information, but older articles were also included if they were com- Author Affiliation: Maxine Dunitz Neurosurgical In- lemia. Reference lists of these articles stitute, Cedars-Sinai Medical Center, Los Angeles, Calif. and ongoing investigations in this area monly referenced and highly regarded. Corresponding Author: Wouter I. Schievink, MD, Max- ine Dunitz Neurosurgical Institute, Cedars-Sinai Medi- EVIDENCE SYNTHESIS cal Center, 8631 W Third St, Suite 800E, Los Ange- See also Patient Page. les, CA 90048 ([email protected]). Epidemiology Clinical Review Section Editor: Michael S. Lauer, MD. CME available online at Once considered an exceedingly rare We encourage authors to submit papers for consid- www.jama.com eration as a Clinical Review. Please contact Michael disorder, recent evidence suggests that S. Lauer, MD, at [email protected]. 2286 JAMA, May 17, 2006—Vol 295, No. 19 (Reprinted) ©2006 American Medical Association. All rights reserved. Downloaded From: http://jama.jamanetwork.com/ by a STANFORD Univ Med Center User on 04/24/2015 SPONTANEOUS SPINAL CEREBROSPINAL FLUID LEAKS AND INTRACRANIAL HYPOTENSION garding spontaneous intracranial hy- Comprehensive population-based epi- symptoms typically is in the fourth or potension was derived from case reports demiologic studies, however, are not yet fifth decade of life, with a peak inci- only, and no epidemiologic data were available. In the past, spontaneous in- dence around age 40 years, but chil- available. In a community-based study tracranial hypotension was probably dren and elderly persons also may be conducted in 1994, the prevalence of more frequently underdiagnosed than affected.1-16 spontaneous intracranial hypotension it is now, and it is unlikely that there was estimated at 1 per 50 000.16 In a has been an actual increase in its inci- Etiology and Pathogenesis more recent emergency department– dence, although that possibility can- Spontaneous intracranial hypotension based study (2003-2004),17 spontane- not be entirely excluded. is caused by spontaneous spinal cere- ous intracranial hypotension was half Spontaneous intracranial hypoten- brospinal fluid (CSF) leaks. Because as common as spontaneous subarach- sion affects women more frequently spinal CSF leaks generally do not noid hemorrhage, for an estimated than men, with a female-male ratio cause any local symptoms, they annual incidence of 5 per 100 000. of approximately 2:1.1-16 Onset of remain undetected unless actively Figure 1. Spinal Cord Anatomy and Intraoperative Photograph With Corresponding Line Drawing of a Complex Meningeal Diverticulum Arising From a Thoracic Nerve Root Sleeve in a 27-Year-Old Woman A Spinal Cord Anatomy Epidural Venous Plexus Spinal Nerve Root (Covered in Dura) Transverse Spinous Process Process Dura Mater Spinal Meninges Arachnoid Subarachnoid Pia Mater Space T8 T9 T10 Dorsal Root Spinal Nerve Root Ventral Root B Complex Meningeal Diverticulum INTRAOPERATIVE VIEW TRANSVERSE SECTION POSTERIOR Multilobed Meningeal Dura Mater Multilobed Meningeal Diverticulum Diverticulum Arachnoid 0.5 cm DURA MATER LEFT RIGHT ANTERIOR Spinal Nerve Root (Covered in Dura) ©2006 American Medical Association. All rights reserved. (Reprinted) JAMA, May 17, 2006—Vol 295, No. 19 2287 Downloaded From: http://jama.jamanetwork.com/ by a STANFORD Univ Med Center User on 04/24/2015 SPONTANEOUS SPINAL CEREBROSPINAL FLUID LEAKS AND INTRACRANIAL HYPOTENSION ver to large amounts of CSF spontane- A distinct and uncommon cause of Box 1. Connective Tissue ously pouring out into the paraspinal spontaneous intracranial hypotension Disorders Associated soft tissues. not associated with a primary dural de- With Spontaneous Spinal There is good evidence to suggest fect is the presence of osseous spinal pa- Cerebrospinal Fluid Leaks that a generalized connective tissue thology. A congenital osseous spur,35 as and Intracranial Hypotension disorder plays a crucial role in the well as acquired degenerative disk dis- 36-39 Named Syndromes development of spontaneous spinal ease piercing the dura, has been CSF leaks. First reported in 1994,20 described. Marfan syndrome this association has been confirmed by Before precise imaging was able Ehlers-Danlos syndrome type II numerous subsequent studies.11,22-25 to detect the underlying spinal CSF Autosomal dominant polycystic Based on physical examination alone, leak, some authors speculated that kidney disease evidence for an underlying generalized spontaneous intracranial hypotension Unnamed Syndromes/ connective tissue disorder is found in resulted from decreased CSF secretion Associations about two thirds of patients.25 This or generalized CSF hyperabsorption, Isolated skeletal features of Marfan group of disorders is heterogeneous, but there are no data to support such syndrome possibly affecting different compo- alternate mechanisms. It has been pos- Isolated joint hypermobility nents of the dural extracellular matrix tulated that a decrease in CSF volume, (BOX 1). Isolated joint hypermobility rather than in CSF pressure, may be Joint hypermobility with fascial thinning is found in approximately two fifths the final common pathway in the of patients with spontaneous intracra- pathophysiology of spontaneous intra- Spontaneous retinal detachment nial hypotension and may be associ- cranial hypotension.40 Therefore, ated with attenuation of the dorsal “spontaneous CSF hypovolemia” has muscular fascia, complicating surgi- been introduced as an alternative looked for in a patient suspected of cal wound closure.25 Approximately one term.40 However, this is an oversimpli-