Dementia Assessment and Cat Scans in Primary Care

To the Editor: We read with interest the recent editorial (Brummel-Smith, KV. Alzheimer's Disease and Managed Care: How Much Will It Cost? J Am Geriatr Soc 1998;46:780-781) and agree with most of its content. However, we must take issue with one of the author's points. We are not surprised at how “resistant primary care physicians are to the CAT scan.” They should, in fact, be somewhat circumspect. The Canadian Consensus Conference on the Assessment of Dementia in 1991 issued guidelines for primary care physicians suggesting that it was not necessary to do a CT scan in every patient in whom dementia was suspected.1 Several years later these guidelines were put to the test and found to be quite robust in the context of a tertiary care hospital memory clinic. If the CT scan was not helpful in the potentially reversible dementias, where the pretest probability of finding a CNS lesion is relatively high, we suggest that it will be even less useful in the primary care setting where the incidence of reversible dementias (especially caused by intracranial lesions) has been shown to be quite rare.2 These findings were consistent with a similar study,3 which assessed the guidelines of the American Academy of Neurology.4 Of 196 patients with dementia who had the complete work-up, including neuroimaging, only six (3.1%) were found on CT scan to have potentially reversible lesions. Four of these had normal pressure hydrocephalus, and two had a brain tumor. With neurosurgical treatment, the level of dementia declined in two patients, remained unchanged in two and continued to increase in two. None of these patients experienced complete resolution of the cause of their dysfunction. Of the six patients with the potentially reversible CT scan lesions, five would have met at least one of the CCCAD criteria for CT scanning. The patient who did not meet any of these criteria had a glioma, and mental status did not improve despite surgical intervention; he died in less than 1 year following diagnosis. In other words, by following these clinical guidelines the primary care physician can safely be quite conservative in choosing neuroradiological testing for his/her demented patients. Certainly, we support Dr. Brummel-Smith's suggestion that primary care physicians should be more attuned to the geriatric imperative. However, this goal will be reached more quickly through a sensitivity to the patient's history and background and a careful physical examination rather than through blind trust in tests (which certainly have their validity in certain circumstances) such as the Mini-Mental Status Examination or the CT scan. Editors note: The above letter was referred to the author of the original paper, and Dr. Brummel-Smith's reply follows. In reply: The point of Dr. Clarfield et al. regarding the lack of utility of a neuroimaging study in the evaluation of dementia is well taken. In fact, I agree with them. The point I made in my editorial was that primary care physicians are unlikely to perform such tests, including the use of a standardized mental assessment, in spite of a significant number of publications recommending their use. This “reality” has, in my mind, both positive and negative consequences. A number of publications have recommended the use of either CT or MRI in the workup of dementia.1–4 Even when it is not clearly recommended, authors often phrase the statement regarding use of neuroimaging as if to suggest that experts lean toward its use. For instance, in the Consensus Statement on Diagnosis and Treatment of Alzheimer's Disease and Related Disorders, Small et al. state, “Imaging studies are optional but recommended by many clinicians and experts.”5 On the other hand, the experience of many primary care physicians is exactly as Dr. Clarfield and colleagues found in their study. CT scanning rarely changes the diagnosis or provides guidance in treatment. I would certainly never recommend blind trust in any diagnostic tool, including the MMSE. However, I do have concerns that primary care physicians, operating in busy practices with 7 to 10 minutes allowed per visit, will be able to identify and diagnose older people with cognitive impairments correctly. The most frequent “mental status” test mentioned in primary care is “alert and oriented.” Yet with the low prevalence of dementia (approximately 3% for those aged 65 to 74) seen in primary care settings, a deficit in orientation has posttest probability of between 11 and 41% of predicting the presence of dementia.6 Indeed, Callahan et al. found that in cases where the patient had moderate to severe dementia, fewer than 25% of the subjects had been diagnosed with dementia.7 Primary care physicians are using proper discretion in limiting the use of neuroimaging studies. However, as I stated in the editorial, primary care physicians need a simple, quick, and valid screening tool that will help them decide which patients need more in-depth testing. I believe the evidence is clear that such in-depth testing should include a formal, standardized assessment tool (such as the MMSE). Better case finding and fewer unnecessary tests will better serve older patients.