Diagnostic testing for the evaluation of headaches

Headaches are one of the most common symptoms that neurologists evaluate. Although most are caused by primary disorders, the list differential diagnoses is one of the longest in all of medicine, with over 300 different types and causes. The cause or type of most headaches can be determined by a careful history supplemented by a general and neurologic examination. Reasons for obtaining neuroimaging include medical indications as well as anxiety of patients and families and medico-legal concerns. In the era of managed care, concerns over deselection and negative capitation may dissuade the physician from ordering even a medically indicated scan. The yield of neuroimaging in the evaluation of patients with headache and a normal neurologic examination is quite low. Combining the results of multiple studies performed since 1977 for a total of 3026 scans reveals the overall percentages of various pathologies as: brain tumors, 0.8%; arteriovenous malformations, 0.2%; hydrocephalus, 0.3%; aneurysm, 0.1%; subdural hematoma, 0.2%; and strokes, including chronic ischemic processes, 1.2%. EEG is not useful in the routine evaluation of patients with headache. Similarly, the yield of neuroimaging in the evaluation of migraine is quite low. Combining the results of multiple studies performed since 1976 for a total of 1440 scans of patients with various types of migraine, the overall percentages of various pathologies are: brain tumor, 0.3%; arteriovenous malformation, 0.07%; and saccular aneurysm, 0.07%. WMA have been reported on MRI studies of patients with all types of migraine, with a range from 12% to 46%. The cause of WMA in migraine is not certain. Cerebral atrophy has been variably reported as more frequent and no more frequent in migraineurs compared with controls. The 'first or worst' headache has a long list of possible causes and always includes the possibility of acute subarachnoid hemorrhage. Headaches-especially the sentinel type caused by SAH-often are misdiagnosed. The probability of detecting an aneurysmal hemorrhage on CT scans performed at various intervals after the ictus is: day 0, 95%; day 3, 74%; 1 week, 50%; 2 weeks, 30%; and 3 weeks, almost nil. The location of a ruptured saccular aneurysm often is suggested by the predominant site of the SAH. The probability of detecting xanthochromia with spectrophotometry in the CSF at various times after a subarachnoid hemorrhage is: 12 hours, 100%; 1 week, 100%; 2 weeks, 100%; 3 weeks, more than 70%; and 4 weeks, more than 40%. The management of thunderclap headaches with normal CT scan and CSF examinations is controversial. Most patients have a benign course but an unruptured saccular aneurysm occasionally may be responsible for the headache. MR angiography may be a reasonable test to obtain instead of a cerebral arteriogram in many of these cases. About 30% to 90% of patients have headaches of various types and causes after mild head injury. Although most headaches are relatively benign, perhaps 1% to 3% of these patients have life-threatening pathology, including subdural and epidural hematomas, that are detected on CT and MRI scans. Headaches caused by subdural hematomas can be nonspecific. When new-onset headaches begin in patients over the age of 50 years, the physician always should consider whether it may be a secondary headache disorder requiring specific diagnostic testing and treatment. Up to 15% of patients 65 years and over who present to neurologists with new-onset headaches may have serious pathology such as stroke, TA, neoplasm, and subdural hematoma. Headaches are the most common symptom of TA, reported by 60% to 90%. The headaches of TA can be nonspecific and can occur in any location, with only 25% only over the temple. The diagnosis of TA is based on a high index of clinical suspicion that usually but not always is confirmed by laboratory testing. The erythrocyte sedimentation rate can be normal in 10% to 36% of patients with TA. A superficial temporal artery biopsy can give a false negative result in 5% to 44% of patients. As the era of managed care of medicine evolves, payers may increasingly question the benefit of diagnostic testing for the evaluation of headaches. A careful and complete history and examination enable the astute physician to make the proper diagnosis in most cases without testing. In some cases, however, the nonspecific nature of some headaches with a normal examination demand diagnostic testing even though the yield may be low. Headache diagnosis can be a field of unexpected land mines. Hopefully, the gatekeeper model will not prevent patients with troublesome headaches from obtaining appropriate diagnostic testing and neurologic consultation.