Neurologic Imaging

What Imaging Is Next for Chronic Headache When Intracranial Hypotension Is Suspected but Initial Scans Are Negative?

·
What Imaging Is Next for Chronic Headache When Intracranial Hypotension Is Suspected but Initial Scans Are Negative?

A 45-year-old patient presents to your clinic with a debilitating, months-long daily headache. The pain is consistently worse when they are upright and improves upon lying down, a classic orthostatic feature. You suspect spontaneous intracranial hypotension (SIH) from a cerebrospinal fluid (CSF) leak, but a recent non-contrast brain MRI and routine spine MRI were interpreted as normal. The clinical picture strongly points to a leak, but the initial, non-invasive imaging has failed to provide an answer. You are now faced with a critical decision: what is the next, more definitive imaging study to order to localize the occult CSF leak? This article details the ACR-guided workflow for this specific, challenging scenario. For this presentation, the American College of Radiology rates Radiographic myelography digital subtraction complete spine as Usually Appropriate, providing a pathway to diagnosis when standard imaging falls short.

Who Fits This Clinical Scenario for Suspected Intracranial Hypotension?

This guidance is specifically for an adult patient presenting with a chronic daily headache where there is a high clinical suspicion for intracranial hypotension, but initial imaging has been unrevealing. The key inclusion criteria are:

  • Adult patient with a chronic, often daily, headache.
  • Strong clinical features of SIH, most notably an orthostatic component (headache worsens upon standing or sitting and improves when supine). Other associated symptoms may include neck stiffness, tinnitus, muffled hearing, or cognitive changes.
  • Negative or non-diagnostic initial imaging. This is the crucial differentiator. The patient has already undergone standard, non-invasive studies like a non-contrast brain MRI and/or routine spine MRI, and these have not shown the classic signs of SIH (e.g., pachymeningeal enhancement, sagging brainstem, subdural fluid collections) or a clear site of CSF leakage.

Conversely, this workflow does not apply to several similar-appearing presentations. If a patient has a clear iatrogenic cause for a CSF leak, such as a headache developing within 72 hours of a dural puncture or spinal surgery, they fall into a different diagnostic category. Similarly, patients who present with obtundation or have obvious, classic findings of SIH on their initial brain MRI follow a distinct workup, as the primary question is no longer if a leak exists, but where it is. This article is for the diagnostically challenging cases where the leak is occult on first-pass imaging.

What Diagnoses Are You Working Up in This Scenario?

When initial imaging is negative despite a compelling clinical story, the differential diagnosis narrows, and the goal of the next study is to definitively locate a subtle cause of CSF volume loss. The primary considerations include:

Spontaneous Intracranial Hypotension (SIH) from an Occult CSF Leak
This remains the leading diagnosis. The absence of findings on standard MRI does not rule out a leak; it simply means the leak is not causing the typical secondary intracranial changes or is not large enough to be seen without intrathecal contrast. The leak may be intermittent, small, or located at a spinal nerve root sleeve, which is difficult to visualize on routine sequences. The purpose of advanced myelography is to make the CSF itself visible with contrast and directly observe its extravasation.

CSF-Venous Fistula
A less common but critical diagnosis to consider is a CSF-venous fistula. In this condition, CSF drains directly from the thecal sac into an adjacent spinal epidural vein. This pathophysiology often results in SIH without significant epidural fluid collection, making it nearly invisible on standard MRI and even challenging to see on conventional CT myelography. The rapid shunting of CSF requires high-temporal-resolution imaging, like digital subtraction myelography, to diagnose.

Dural Ectasia or Underlying Connective Tissue Disorder
While not a diagnosis in itself, the imaging may reveal predisposing factors. Conditions like Marfan syndrome, Ehlers-Danlos syndrome, or polycystic kidney disease can be associated with dural weakness (dural ectasia) and an increased propensity for developing spontaneous dural tears or diverticula that leak CSF. Identifying these features can support the diagnosis of SIH and inform long-term management.

Re-evaluation of Primary Headache Disorders
If advanced imaging is also negative, it forces a re-evaluation of primary headache diagnoses like New Daily Persistent Headache (NDPH) or chronic migraine, which can sometimes mimic SIH. However, given the strong orthostatic component in this scenario, exhausting the search for a structural cause is paramount before settling on a primary headache diagnosis.

Why Is Radiographic Myelography the Next Step When Initial Imaging Is Negative?

When non-invasive imaging fails to identify a CSF leak, a more direct and dynamic study is required. The ACR designates Radiographic myelography digital subtraction complete spine as a Usually Appropriate next step. This procedure involves injecting iodinated contrast directly into the thecal sac via lumbar puncture and then using fluoroscopy with digital subtraction techniques to watch the contrast flow in real-time. Its strength lies in its ability to detect active, often fast-flowing, CSF leaks that would be missed on static images. The subtraction technique removes the overlying bone, providing an unobscured view of contrast extravasation, which is particularly effective for identifying elusive CSF-venous fistulas.

The ACR also rates CT myelography dynamic complete spine as Usually Appropriate. This study also involves intrathecal contrast but uses rapid, sequential CT acquisitions to visualize the leak. It provides excellent anatomic detail of the spine and can clearly delineate the path of extravasated CSF into the epidural space. The choice between digital subtraction myelography (DSM) and dynamic CT myelography (dCTM) often depends on institutional expertise and the specific clinical question. DSM is often favored for its superior temporal resolution in detecting very fast leaks or CSF-venous fistulas.

Alternative studies are rated lower for this specific scenario:

  • MR myelography complete spine is rated May be appropriate. This non-invasive technique uses heavily T2-weighted sequences to visualize the thecal sac. While it avoids radiation and intrathecal injection, its sensitivity for detecting active leaks is substantially lower than contrast-based myelography. In a patient whose initial standard MRI was already negative, it is less likely to provide a definitive answer.
  • DTPA cisternography is also rated May be appropriate. This nuclear medicine study involves injecting a radiotracer into the CSF and imaging over several hours to look for tracer accumulation outside the thecal sac. While sensitive for detecting the presence of a leak, its spatial resolution is poor, making it suboptimal for localizing the precise leak site for targeted therapy.

The recommended studies, DSM and dCTM, involve significant radiation (ACR RRL ☢☢☢☢ and ☢☢☢☢☢, respectively) and the risks of a dural puncture. However, this trade-off is justified in the context of a debilitating condition where less invasive methods have failed to yield a diagnosis, and localization is required for effective treatment.

What Is the Downstream Workflow After a Digital Subtraction Myelogram?

The results of the myelogram directly guide the subsequent clinical pathway. The goal shifts from diagnosis to targeted intervention.

If the study is positive and localizes a leak: The next step is targeted treatment. For most simple dural tears or leaking nerve root diverticula, the first-line therapy is a targeted epidural blood patch (EBP) or a fibrin glue patch performed under imaging guidance at the precise spinal level identified. This is significantly more effective than a “blind” or non-targeted patch. If a CSF-venous fistula is identified, treatment is more complex and may involve transvenous embolization of the draining vein or direct surgical ligation. Referral to a center with neurointerventional and neurosurgical expertise in these procedures is essential.

If the study is negative: A negative high-resolution myelogram in a patient with persistent, classic symptoms presents a true diagnostic challenge. If clinical suspicion remains exceptionally high, options include repeating the study, sometimes with provocative maneuvers (e.g., imaging in an upright position if possible) or considering a different advanced modality (e.g., dCTM if DSM was done first, or vice versa). Alternatively, some clinicians may proceed with a trial of non-targeted, multi-level epidural blood patching, which serves as both a diagnostic and therapeutic maneuver. If the headache resolves, it strongly supports the diagnosis of an occult leak. If it does not, a thorough re-evaluation for alternative diagnoses like NDPH is warranted.

If the study is indeterminate: Ambiguous findings, such as a possible nerve root diverticulum without definite contrast extravasation, require multidisciplinary discussion among neurology, neuroradiology, and neurosurgery. A consensus decision may lead to a targeted EBP as a diagnostic trial or further observation.

Common Pitfalls in Working Up a Suspected Occult CSF Leak

Navigating the workup for an occult CSF leak requires persistence and attention to detail. Several common pitfalls can delay diagnosis:

  • Prematurely ending the workup: The most common error is accepting a negative non-contrast brain MRI as definitive evidence against SIH, especially when a classic orthostatic headache is present.
  • Suboptimal imaging technique: Performing a non-dynamic CT myelogram or failing to image the entire spine can easily miss an intermittent or high-flow leak. The protocol must be tailored to the clinical question.
  • Confounding post-puncture headache: The myelogram itself requires a dural puncture, which can cause a new iatrogenic low-pressure headache. It is crucial to distinguish this from the patient’s underlying chronic headache.
  • Overlooking CSF-venous fistulas: These require a high index of suspicion and specific imaging protocols (like DSM) to detect. Failure to consider this diagnosis can lead to a frustratingly negative workup.

If a definitive leak cannot be localized despite a thorough workup including advanced myelography, but the clinical syndrome is classic for SIH, escalation is appropriate. The next step is referral to a tertiary care center with a dedicated, multidisciplinary team specializing in CSF leaks and complex headache management.

Related ACR Topics and Tools

This article focuses on a single, complex clinical scenario. For a comprehensive overview of all clinical variants related to Imaging of Suspected Intracranial Hypotension, please consult our parent guide. You can also use the tools below to explore adjacent scenarios, review imaging protocols, and discuss radiation dose with your patients.

Frequently Asked Questions

Why not just perform an epidural blood patch empirically without trying to find the leak?

While a non-targeted (‘blind’) epidural blood patch is sometimes performed as a therapeutic trial, localizing the leak first is the standard of care. A targeted patch directed at the specific site of CSF egress is more effective and has a higher success rate than a blind patch. Furthermore, identifying the exact cause, such as a CSF-venous fistula or a dural tear from a bone spur, is critical for planning definitive treatment like surgery if patches fail.

What is the difference between digital subtraction myelography (DSM) and dynamic CT myelography (dCTM)?

Both are rated ‘Usually Appropriate’ for this scenario and are excellent at localizing leaks. DSM uses real-time X-ray (fluoroscopy) with bone subtraction to visualize contrast flow, giving it very high temporal resolution, which is ideal for seeing fast leaks and CSF-venous fistulas. Dynamic CTM uses rapid, sequential CT scans after contrast injection, providing superior cross-sectional anatomical detail of the leak site and surrounding structures. The best choice often depends on institutional expertise and the suspected type of leak.

If the initial brain MRI is negative, does that mean the intracranial pressure is normal?

Not necessarily. A significant percentage of patients with confirmed spontaneous intracranial hypotension have a normal initial brain MRI. The classic findings of pachymeningeal enhancement, brain sagging, and subdural fluid collections are specific but not perfectly sensitive. A compelling clinical history, especially a strong orthostatic headache, should prompt further investigation for a CSF leak even if the brain MRI is unremarkable.

What are the main risks associated with radiographic or CT myelography?

The primary risks are related to the lumbar puncture required to inject contrast. These include a post-lumbar puncture headache (which can confound the clinical picture), bleeding, infection, and nerve irritation. There is also exposure to ionizing radiation, which is higher with CT myelography than with digital subtraction myelography. Finally, there is a small risk of an allergic reaction to the iodinated contrast agent.

How long does a patient need to lie flat after a myelogram or an epidural blood patch?

Post-procedure protocols vary by institution. After a diagnostic myelogram, a period of bed rest (often a few hours) is typically recommended to reduce the risk of a post-dural puncture headache. After a therapeutic epidural blood patch, a longer period of strict bed rest, often 24 to 72 hours, is usually advised to allow the patch to set and seal the leak effectively. Patients are also typically instructed to avoid bending, lifting, or straining for several weeks.

Reviewed by Pouyan Golshani, MD, Interventional Radiologist — May 30, 2026