Imaging in Dizziness

Timothy C. Hain, MD • Page last modified: February 21, 2016

Imaging studies of the brain, and sometimes the cervical spine, are common in patients with Dizziness and Vertigo. The purpose of these studies is to identify a structural cause of dizziness such as a stroke, MS, tumor, or any number of a myriad of other disorders of the ear or central nervous system. This endeavor is driven by the difficulty in diagnosing dizziness, which can be caused by ear, medical, neurological, psychiatric sources. Additionally, there is a lot of undiagnosed dizziness -- often just nobody can figure it out.

The general consensus among those who pay for imaging, is that there is far too much imaging done. The data shows that most of the imaging does not identify the cause of dizziness. Often it seems that the second thing that physicians do when their patient says they are dizzy, after trying meclizine (first thing), is to treat their own and their patient's anxiety by getting an MRI scan done. Getting an image is also the safest thing to do for the physician, as it avoids the risk of being sued for malpractice should a tumor or similar lesion be identified subsequently. Thus there is certainly almost always value to imaging for both patient (less anxiety) and doctor (less malpractice risk), and medical institution (more money for the owner of the imaging system), but one could easily argue that tolerating some anxiety and eliminating malpractice suits based on adverse events given reasonable behavior, would be preferable to treating anxiety and malpractice risk with these expensive tools.

The biggest "trouble" areas, where the cost/benefit ratio is tilting more towards cost with imaging are MRI scans (which are still very expensive), CT angiograms (expensive and involve considerable radiation and dye), and CT scans (which are too easy to get). MRI scanners are now routine in almost any hospital. CT scanners are found in many outpatient offices, as they are smaller and cheaper. In rural settings, carotid dopplers are problematic as well.

The current system has a crude way of reducing unnecessary scans -- scans are "not approved" by many insurances in outpatient settings, but in the emergency department or hospital, anything goes. Ideally, one should have a way of estimating the benefit of imaging (or for that matter any diagnostic test), and adjusting the proportion that the patient vs. medicare or insurance company pays based on the probability of a useful outcome. The patient would need to be informed prior to the procedure of the risks/benefits of the scan, and their proportion of the actual bill (after contractual agreements of course). Thorny issues exist regarding the competence of the patient to make their own decisions regarding whether to proceed with an expensive procedure. The insurance provider would need to take on some of the risk of the scan -- if for example, an HMO plan "denied" doing an MRI on a patient with a brain tumor, then the HMO is should logically be the responsible party should irreversible damage occur. However, actuarial type decisions is what insurance companies are very good at. At this writing (2016), the data needed to do this is simply not there -- we would need to have an health care information system that gave payors data that would allow them to estimate risk/benefit. Perhaps in another 20 years we will have this.

CT and MRI scans in the Emergency Department (ED)

Brain imaging is frequent in the ED because it reduces risk of missing a diagnosis of a central cause. In addition, in the ED, there is no scrutiny from payors over what one can or cannot do, so "anything goes". Recently criticism has been leveled suggesting that imaging is obtained too frequently. According to Kerber, "the utilization of computerized tomography and magnetic resonance imaging (CT/MRI) increased 169% from 1995 to 2004, which was more than any other test". According to Saber et al (2013), the proportion of dizzy patients in the ED undergoing diagnostic imaging by computed tomography (CT), magnetic resonance imaging (MRI), or both in 2011 was estimated to be 39.9% (39.4% CT, 2.3% MRI). Neuroimaging was estimated to account for about 12% of the total costs for dizziness visits in 2011 (CT scans $360 million, MRI scans $110 million).

Lee et al (2013) reported that only 3.6% of patients referred for MRI's in the ED had stroke. The incidence of new central lesions in patients aged in their 40s, 50s, 60s, 70s and >/=80s was 0, 3.9%, 3.4%, 7.4% and 16.7%. Clearly then, MRI is more "productive" in older patients.

Lawhn-Heath (2013) reported that the diagnostic yield for head CT ordered in the ER for acute dizziness is low (2.2 %; 1.6 % for emergent findings), but MRI changes the diagnosis up to 16 % of the time, acutely in 8 % of cases. Ahsan et al (2013) reported even lower figures -- only 0.74% yielded clinically significant pathology requiring intervention. MRI scans were more productive -- 12.2% of MRI studies yielded discovery of significant abnormalities. These sorts of studies are intrinsically flawed because they are retrospective -- patients are likely selected for CT or MRI based on the clinicians judgement as to how much they are needed.

Nevertheless, the very low utility of ED CT scans for dizziness is unsurprising as the typical CT scan done in the ER is is generally "quick/dirty" -- suitable only to find large lesions. The CT cuts are often 5 mm (which is very wide -- the inner ear is only 10 mm in diameter), and contrast is often omitted to avoid risk of allergic reaction, and iatrogenic damage.

Tung, C., et al. (2013) found that only 10% of CT's done for urgent indications resulted in documentation of significant abnormality including acute cerebral infarction, intracranial hemorrhage, malignancy, infection, cerebral edema, or hydrocephalus. Applying the selection criteria as proposed by Rothrock (patient age >/=60 years, presence of new onset focal neurologic deficit, headache with vomiting, or altered mental status), allowed a very substantial reduction in CT frequency. These are a variant of the ABCD criteria.