Chapter 21. Refractory Headaches

With appropriate diagnosis and treatment, the majority of headache sufferers can look forward to improved management (although not elimination) of their headaches. There remains, however, a group of patients whose headaches do not improve with treatment as expected. In addition to the pain of inadequately relieved headache, these patients are at risk of developing chronic pain syndrome, a condition characterized by significant disability, medication dependence and overuse, depression, and worsening headache.

Chronic pain syndrome often develops insidiously over time: as a result of the pain and incapacitation of frequent, poorly relieved headaches, patients miss substantial school or work time, and productivity is often reduced even when they are in attendance. Overuse or dependence on medication may result from desperate attempts to remain functional. This leads to secondary problems with rebound headache, altered sleep-wake cycles, or end-organ dysfunction, such as gastrointestinal hemorrhage related to overuse of anti-inflammatory medications, kidney damage from overuse of acetaminophen, sleep disruption related to sedative misuse, and so forth. Sufferers often limit social or leisure activities in order to devote themselves to essential work or school activities, leading to isolation from family or social contacts. Avoidance of physical exertion that may trigger headache not infrequently leads to profound physical deconditioning. In many patients, depression ensues or worsens.

Once underway, this complex downward spiral is difficult, if not impossible, to address using single-discipline treatment or medication. If the disability resulting from chronic disabling headache could be prevented, tremendous suffering would be avoided. This chapter discusses the magnitude of the problem of refractory headache, suggests ways to intercede in patients currently experiencing refractory headache, and considers how early intervention with appropriate, disease-specific measures might prevent the disaster of chronic headache from developing in many patients.

Data from the American Migraine Study suggest that 5% of American women and 3% of American men suffer from frequent, severe headaches, defined as headache more than 15 days a month.1 The vast majority can be assumed to have migraine or transformed migraine, with a small number likely experiencing cluster or other less common types of headache. Although tension-type headache in its chronic form can be debilitating, it is far less frequent than migraine among severely disabled patients.

The gender distribution of headache is not equal, with women more likely than men at all postpubertal ages to suffer from headache, and more likely than men to develop severe, disabling forms of the disorder. Epidemiologic data show that women require more bed rest per attack than do men,2 and are more likely to be moderately or severely disabled as a result of headache. Regardless of gender, the most disabled segment of the headache population accounts for very large percentages of the indirect costs attributed to migraine,3 suggesting that treatment targeted to this group of severely affected patients is highly cost-effective.2

The level of functioning of patients with chronic headache has been found to be worse in many areas than that of patients with conditions such as diabetes or arthritis. Among a group of patients with chronic illnesses, patients with chronic headache functioned at a level comparable to that of patients with congestive heart failure or recent myocardial infarction.3

Five factors, acting singly or in concert, can be responsible for a patient whose headaches have been refractory to treatment, and evidence of them should be vigorously sought when patients do not respond as expected to treatment. They are (1) inappropriate expectations of treatment benefit, (2) inaccurate diagnosis, (3) inadequate treatment, (4) psychiatric comorbidity, and (5) severe disease.

Headache is most appropriately viewed as a chronic illness of long duration, in which the goal is management, not cure. Chronic nonmalignant headache is currently understood as an inherited vulnerability to headache caused by abnormalities in central nervous system neurotransmitter systems. Changes in central pain processing may also play a role.4 Although progress has been made in understanding the genetic underpinnings of headache, particularly migraine,5 it is not possible to alter them, and treatment is aimed at trigger avoidance, improving abortive treatment of migraine, implementing prophylactic measures or medication to decrease headache frequency, and improving the ability of patients to cope with headache.

Patients, their families, or even treating clinicians may have an erroneous understanding of the nature of chronic headache and the ability of medical treatment to alter the natural history of the disease. They may expect that treatment will easily eliminate all headaches with a minimum of side effects, and may not be aware of the importance of active patient involvement in treatment. If uncorrected, these beliefs can lead to physician shopping, dissatisfaction, and poor treatment compliance (especially with prophylactic treatments that produce only gradual, incremental improvement and require relatively long use to become effective).

Thorough exploration of the beliefs of all relevant parties may reveal unrealistic ideas and expectations, providing an opportunity for education and reformulation of treatment goals. In general, realistic treatment goals for this group of severely affected patients are to reduce headache frequency (ideally to two or fewer headaches a week); to find abortive treatment that works quickly, consistently, and well; and to involve the patients in lifestyle alterations that improve their ability to cope with headache, such as regular exercise, good sleep habits, and the use of relaxation or other nonpharmacologic pain control methods. When these goals can be accomplished, most patients remain able to participate in work, school, or social activities despite frequent headache.

Failure of a patient to respond as expected to headache treatment, or a significant change in headache pattern or characteristics, should prompt reconsideration of the diagnosis. An initial inaccurate diagnosis of a primary headache disorder may have been made in a patient who really has a secondary headache, or a secondary cause of headache may have developed in a patient since the original headache diagnosis was made.

In any case, a thorough review of the headache history and repeat physical and neurologic examination, as well as testing, should be considered. Neurologic and physical findings are generally normal in patients with primary headache disorders, but special attention should be devoted to palpation of the pericranial musculature, funduscopic examination, including the presence of venous pulsations, and mental status. It is important to remember that some secondary causes of headache, such as Arnold-Chiari malformations or high-pressure headaches without papilledema, can present in subtle ways. It is always worth carefully reviewing a patient’s medications as well, because many frequently used medications can cause or exacerbate headaches.

There is considerable but not complete overlap in treatment options that are helpful for the various primary headache disorders. This incomplete overlap means that even if the diagnosis of a primary headache disorder has been made correctly, incorrect assignment to the various subtypes of primary headache can have treatment implications. For example, a patient with cluster headache who is considered to have migraine (not an infrequent occurrence) is not likely to benefit from treatment with beta blockers or tricyclic antidepressants, because these drugs are not useful in cluster headache. If such a patient has a partial response to sumatriptan (effective in both migraine and cluster headache), used in the erroneous assumption that the headache in question is migraine, the clinician may be reinforced in the inaccurate diagnosis and recognition of the correct diagnosis delayed. Thus, response to a particular treatment modality should not be considered to have clear diagnostic significance.

Evaluation of a patient referred for refractory headaches must include a thorough review of previous courses of treatment. A limited number of medications are useful in the management of headache disorders; it is essential that options for treatment not be discarded as unhelpful on the basis of inadequate previous experience. Detailed information about previous treatment is often difficult to obtain. A complete list may require acquisition of previous medical or pharmacy records. On review, it is clear that many patients referred to specialty clinics with a diagnosis of refractory headache have not had the benefit of aggressive, high-dose trials of treatment. Scrupulous review of previous treatment attempts by an expert in the pharmacotherapy of difficult headache disorders can often produce important suggestions for future treatment trials.

Some patients with frequent or severe headaches have been maintained only on abortive agents, or have never been given the opportunity to try the newer, disease-specific abortive agents such as the triptans, perhaps out of exaggerated fears about the safety of the drugs. Not uncommonly patients report that they are “allergic” or have had an adverse reaction, to an important class of therapeutic drugs for headache. If possible, this claim should be investigated. The author has had many patients report so-called allergic reactions to sumatriptan or dihydroergotamine, for example, which on review represented a dystonic reaction to a phenothiazine antinauseant given concomitantly.

Another common problem is that appropriate medication classes have been tried, but the doses used have not been adequate or the medication has not been tried for an adequate length of time. In general, the maximal benefit of most prophylactic agents requires use for 6 to 8 weeks at the highest safe or tolerated dose before maximal benefit can be assessed.6 Worsening or improvement in headaches observed over a shorter period of time may be the result of the waxing and waning nature of chronic headaches. Noncompliance may also be a factor in apparently ineffective treatment, and drugs with tolerable side effects and once-daily regimens will likely improve treatment outcome in this chronic disorder.7

Treatment attempts can be compromised by concomitant overuse of drugs that cause rebound headache. These are usually short-acting, abortive medications such as combination analgesics, ergotamine, or caffeine-containing drugs. Frequent use of triptans has also been suggested as a cause of rebound headaches.8 Overuse of these classes of medications has been implicated as the single most important factor leading to the transformation of intermittent migraine to chronic daily headache. In many cases, patients are reluctant to disclose the true extent of medication overuse, or do not report overuse of nonprescription medications.

Withdrawal from such offending medications can be difficult, especially on an outpatient basis. Many patients are reluctant to discontinue their reliance on these medications until their headaches are better; unfortunately, the headaches will not improve until they discontinue these medications.9 Unproductive treatment stalemates can be avoided by educating the patient about the nature of rebound headache, and outlining a realistic withdrawal regimen and interim headache control plan. The use of repetitive intravenous dihydroergotamine10 or corticosteroids can be very helpful in weaning patients from these drugs. Patients and physicians need to realize that improvement in chronic headache after withdrawal of excess analgesic medication use will be gradual and can take up to 6 months. Hospitalization may be required if patients cannot successfully complete this step of treatment on an outpatient basis.

In still other cases, rational use of co-pharmacy, or combinations of prophylactic medications, has not been attempted. When possible, therapy with a single agent is preferred to minimize the possibility of drug interactions and adverse events. However, as with other chronic illnesses, the combination of two classes of drugs can prove effective. Often-used combinations include sodium valproate and selective serotonin reuptake inhibitors (SSRIs), or beta blockers and tricyclic antidepressants.

Underuse of nonpharmacologic methods of treatment is common, partly because of poor insurance reimbursement for this treatment. Often such treatments are tried late in the course of the disorder, when a sense of treatment failure is established and the disease well established. In this setting, nonpharmacologic treatment is less effective. Findings from randomized controlled trials support the use of biofeedback (which appeared similar in efficacy to preventive pharmacotherapy); relaxation strategies, such as progressive muscle relaxation, breathing exercises, and imagery; and cognitive-behavioral therapy.11 Given these data, and considering the substantial side-effect penalties of pharmacologic prophylaxis for migraine, it is disappointing that most insurance and managed care organizations still refuse to pay for these therapies.

Sequential treatment is often all that is possible on an outpatient basis, where formidable bureaucratic barriers and time pressures make it difficult to arrange concomitant treatment with several different modalities. Nonetheless, attempts should be made to capitalize on the synergistic improvements possible with a combination of treatments. For example, many patients obtain better relief with the combination of physical therapy, biofeedback, and pharmacologic treatments when these methods are employed simultaneously, rather than singly. In explaining this to patients, it is often helpful to employ the analogy of a car stuck in a snowbank. Several people pushing together are more likely to meet with success than if each tries alone to move the car. Similarly, if single attempts at headache treatment have not proved effective, it does not necessarily follow that they should be abandoned. Rather, consideration should be given to the use of additional treatment modalities.

Patients with chronic headache are more likely than controls in the general population to suffer from a spectrum of affective disorders, including depression and anxiety. In specialty headache and pain clinics, personality disorders and a history of childhood trauma are also frequently found. Depression and anxiety disorders are comorbid with migraine, and it has been proposed that the predisposition to develop both is based on inherited disturbances in central nervous system serotonergic systems.12 Patients suffering from migraine with aura may also be at heightened risk for suicide.13 It is simplistic to view one as causing the other, and equally ingenuous to assume that treatment of one disorder will somehow cure the other. Adequate treatment of depression can render the burden of headache easier to cope with, and improvement in headache can lead to improvements in depression, but both disorders require specific attention and treatment.

There is no evidence of shared etiologic mechanisms for chronic headache and personality disorders; more likely, the overrepresentation of these patients among refractory populations reflects the fact that when a personality disorder exists in a person who also has a severe headache problem, the personality disorder complicates treatment. The presence of a personality disorder may lead to treatment noncompliance, difficulty establishing a working relationship with caregivers, and less effective coping with the social and occupational burdens of headache. All of this fosters the development of maladaptive behaviors and decreases the success of long-term treatment.

Childhood trauma may cause permanent changes in the central nervous system that increase susceptibility to the development of chronic pain.14,15 Patients may be reluctant to disclose a history of physical or sexual abuse; tactful queries about these topics are an important part of the history of any patient with refractory headache or pain. If a history of such abuse is elicited, it can alert the clinician to be cautious about such “hands-on” treatment options as physical therapy or biofeedback, techniques which may be poorly tolerated by patients with a history of previous abuse.

Psychosocial factors can amplify or reinforce illness behavior in patients with legitimate headache disorders. For example, the secondary gain of increased attention from family, avoidance of unpleasant work or personal responsibilities, or school avoidance can reinforce the behavior of headache. Caregivers may inadvertently increase the patient’s sense of dependence on external factors by overreliance or emphasis on the use of pills, procedures, or passive therapies such as massage. Attention to the role these factors are playing in refractory headache cases is important. Very careful consideration should be given before providing a patient with medical sanction to avoid school or work attendance. Although well-intentioned, such allowances often produce isolating behaviors that prove self-perpetuating and difficult to reverse.

Expert psychiatric and psychological evaluation and ongoing support are invaluable in all cases of refractory headache. Mental health professionals who have experience treating patients with chronic pain or headache, or both, should be sought. Patients may resist such referral unless they understand it does not imply that their headaches are not “real.” Most recognize that the burden of living with headache can aggravate any tendency to depression and welcome the opportunity to discuss how to cope with chronic illness. Ideally, the referring clinician will work closely with the mental health professional to coordinate treatment. Some antidepressants can exacerbate headache, and the use of sedatives or tranquilizers may be counterproductive as well, by limiting the options for rescue therapy of severe headache episodes.

Although significant progress has been made in understanding and treating headache disorders, not all patients have disease that is amenable to medical therapy. For example, subcutaneous sumatriptan injections, the gold standard in acute migraine therapy, work in only 80% of patients under the best circumstances, and even then provide complete, consistent relief across attacks in an even smaller percentage of patients.16 Contemporary prophylactic drugs are widely recognized as having limited efficacy with a high side-effect burden, and the best provide improvement of headache in only 50% to 60% of patients.17 This limited response to treatment likely reflects the underlying genetic heterogeneity of migraine, as well as the spectrum of disease severity. And, as with every other medically based illness, some patients have disease that is simply beyond the reach of currently available medical interventions.

Other patients, over a long course of headache, have developed maladaptive behaviors that significantly limit possibilities for improvement. Entrenched patterns of medication use (or overuse), emphasis on short-term treatment at the expense of long-term improvement, altered family dynamics, expectations of treatment failure, and secondary gain are all extremely difficult to reverse once established. The likelihood of meaningful, successful intervention in a long-established severe headache disorder decreases over time, a strong argument for aggressive intervention at the earliest possible stage.

Specialty Programs

For patients whose headaches have proved refractory to good outpatient management, who are experiencing significant life disruption from their headaches, who require withdrawal from large amounts of problematic medications, or who have medical or psychiatric problems that render traditional, single-discipline outpatient treatment unsuccessful, hospitalization or involvement in a multidisciplinary outpatient pain program are important treatment options. Relatively few specialized inpatient units exist, but the most severely affected patients should be offered this option. Multidisciplinary outpatient programs are increasingly available. A list of treatment programs can be obtained from the Commission on the Accreditation of Rehabilitation Facilities or the American Pain Society. Where possible, these programs should provide access to a clinician skilled in the management of headache patients.

In referring a patient for a multidisciplinary pain management approach, it is important to communicate to the patient the change in treatment strategy and philosophy that such a referral represents. Most such programs emphasize improvement in function rather than pain control as the goal of treatment, and it is important that patients accept this change in emphasis. Rather than focusing on further, largely futile, attempts to find medication or other procedures that will “fix” the pain of headache, these programs attempt to minimize the impact of headache on patients’ function. They work to reverse physical deconditioning through physical therapy, teach pacing and organizational strategies to minimize the overdo-underdo cycle in which many patients are trapped, and introduce relaxation and behavioral techniques to manage pain. An emphasis on pain and passive treatments is discouraged, and patients are rewarded for increases in functional behavior and taught to take an active role in their own treatment. Medication reduction programs, in addition to withdrawal from problematic medications, are required for many patients. These programs have been shown to produce measurable and persistent improvements in patients’ perceived quality of life despite minimal changes in pain scores. Use of health care resources and emergency department usage also declines.

Maintenance Opioids

Treatment of nonmalignant chronic pain, including headache, with maintenance opioid medications has become increasingly common and accepted in the medical community over the past decade. Several studies have reviewed the impact of such use in patients with chronic headache. The evidence to date suggests that a substantial minority (probably in the neighborhood of 25%) of such patients derive long-term benefit from opioid maintenance without the development of unacceptable adverse events or aberrant drug-related behaviors. The risk of addiction in appropriately selected patients is probably quite low.

If elected, opioid maintenance should be guided by the principles applied in other nonmalignant pain states. A written contract is helpful in clarifying treatment goals and expectations, and regular office visits to monitor progress are essential. It should be remembered that most patients do not derive long-term benefit from this treatment, and that in the majority of cases it simply buys some time to consider other treatment options or obtain a measure of respite from relentless headaches. Those are not inconsiderable benefits, however, and with these realistic goals in mind, opioid maintenance should be considered for every patient truly refractory to other headache therapies.

The danger, of course, is that with the increasing popularity of opioid maintenance therapy for a variety of conditions, and with the emphasis on costs of treatment, opioid therapy (which is relatively inexpensive) will be inappropriately tried before more effective and desirable treatments have been exhausted. This is to be discouraged, for several reasons. In addition to the fact that only a minority of patients ultimately derives long-term benefit from maintenance opioids, they are a nonspecific treatment that masks but does nothing to address the underlying pathophysiology of headache. In addition, the subtle but cumulative effects of mental apathy and cognitive impairment that might be acceptable in terminal cancer patients are much less defensible in generally young, otherwise healthy, and potentially productive patients.

Several general principles can be kept profitably in mind when treating patients with refractory headaches who have failed to obtain benefit from careful consideration of the previously mentioned factors. The first is that, despite the apparent failure of previous attempts at prophylaxis, most such patients should be maintained on a reasonable prophylactic regimen for headache—not with the expectation that this will produce improvement, but in the hope that it may delay progression of the underlying disorder. Evidence is slowly emerging that some prophylactics may have disease-modifying effects and delay progression to even more severe levels of headache disability.

The temptation for patients to overuse abortive headache medications is often overwhelming, and strict limits must be agreed on with the patient and enforced. In general, in these severe situations, it is appropriate to allow patients to use strong abortive medications, including opioids, sedatives, or parenteral dihydroergotamine or sumatriptan up to 2 or 3 days a week. Although the clinician may not be able to make these patients better, one certain way to make things worse is to allow too-frequent use of acute medications, with the eventual development of dependence, addiction, or end-organ dysfunction or damage.

All patients should be seen regularly in consultation by a mental health professional and routinely evaluated for depression and suicidal ideation. A regular aerobic exercise regimen and good sleep hygiene should be maintained. Overuse of nonspecific sedative and sleep medications is to be carefully avoided, because this can limit options for rescue treatment of severe headaches.

Finally, as with patients who have other severe, incurable illnesses, these patients benefit from a stable, long-term relationship with an interested clinician. Regular office visits should be geared toward review of the patient’s medication and coping strategies, with a goal of preventing further harm from the effects of the illness itself, attendant occupational or social disability, or the injudicious use of medication to treat the illness. Regularly scheduled visits afford the clinician an opportunity to review with patients the merits of herbal, surgical, or alternative medical treatments they may be considering. Ensuring continued psychosocial support and reinforcement of the patient’s positive behaviors and attempts to cope constructively with disability are invaluable. Such a long-term relationship usually becomes an important and appreciated resource for a person coping with a chronic illness, and should be viewed as an essential part of the pain physician’s practice—to comfort always, even when we cannot cure.