Chapter 22. Chronic Daily Headache

Chronic daily headache relates to the daily or almost-daily occurrence of headache. However, not all daily or almost-daily headaches fall under this denominator, as is the case with the daily or almost-daily headaches of (chronic) cluster headache and paroxysmal hemicrania, as well as those of hypnic headache or nocturnal migraine. These conditions can be referred to as paroxysmal daily headaches in which the headaches occur in well-defined attack patterns. In cluster headache, the attack pattern is that of headaches occurring once or twice a day and lasting 1 to 2 hours, whereas in paroxysmal hemicrania, it is that of headaches occurring 5 to 15 times per day and lasting 10 to 30 minutes. In hypnic headache or nocturnal migraine, the headaches occur once a day, waking the patient from sleep at night, usually between 4 and 6 am, and lasting for a variable amount of time, also depending on the efficacy of treatment.

Of the nonparoxysmal daily headaches, hemicrania continua is a condition that does not fall under the denominator of chronic daily headache, either. However, it is discussed in this chapter because it is very difficult, if not impossible, to distinguish from chronic daily headache on the basis of presentation alone. It differs from chronic daily headache in having a somewhat more consistent and less variable intensity of the pain and in an absolute response to preventive treatment with indomethacin.

Chronic daily headache is not the same as chronic tension-type headache, as defined by the International Headache Society (IHS).1 The IHS defines chronic tension-type headache as headaches with an average frequency of 15 days per month (180 days per year) or more for at least 6 months. In addition, the headaches must have at least two of the following features: pressing quality, mild or moderate intensity, bilateral location, and no aggravation by routine physical activity. Both of the following have to apply: no vomiting and no more than one of the following symptoms: nausea, photophobia, or phonophobia. Finally, the following conditions must be met: the history as well as the physical and neurologic examination do not suggest the presence of another cause of headache; if the presence of such a cause is suggested, it is ruled out by appropriate investigations or, if such a cause is present, the headaches did not occur for the first time in close temporal relation to occurrence of the disorder.

Although it is stated in the IHS classification that sometimes migraine gradually transforms into chronic tension-type headache, in my opinion, that is incorrect. Migraine may, however, transform into chronic daily headache, as discussed later. The confusion stems from the fact that the IHS identifies chronic daily headache as a synonym for chronic tension-type headache, which is incorrect. Chronic daily headache is, rather, a replacement for the old term, mixed or combined headache. In the practice of medicine, the group of patients with this condition is also the majority of those with the diagnosis of chronic daily headache. In the general population, such patients may comprise only half of the sufferers, with the other half made up of those with chronic tension-type headache.

With regard to the prevalence of daily headache in the general population, the only reliable information comes from a study conducted in the Netherlands in 1975–1976.2 The study was conducted in two districts of Zoetermeer, a middle-size town near Leiden, and involved a random sample of 15,563 subjects. The sample size was 4,522 (29% of the population), and the response rate was 77%, generating 2,198 subjects who were 20 years of age or older. The respondents were asked to fill out a questionnaire that included the following question: “How often do you have headache?” One of the answer options was “daily”. In the study, 6% of the respondents aged 20 years or older (4% of the men, 8% of the women) acknowledged the daily occurrence of headache. The highest prevalence was found in the age groups 20 to 24 years (8%) and older than 64 years (8%), and the lowest prevalence in the age group 35 to 54 years (5%).

The prevalence of frequent headache, that is, headaches occurring at least 180 days per year, in the general population is known from two more recent studies.3,4 One of the studies was conducted in Baltimore County, Maryland. It involved 13,343 randomly selected subjects 18 to 65 years of age, comprising 77% of the total of 17,237 eligible subjects. Of the respondents, 40% were men and 60%, women, and their median age was 38 years. The 1-year prevalence of frequent headache was 4%: 3% in men and 5% in women. Using the IHS criteria for chronic tension-type headache and Silberstein’s modified criteria for transformed migraine,5 the investigators found a prevalence of 2.2% for chronic tension-type headache, 1.3% for frequent headache with migrainous features, and 0.6% for other frequent headaches. They found the prevalence of frequent headache to be highest in the age group 41 to 55 years and lowest in the age group 56 to 65 years.

The other study was conducted in Camargo, Spain. The study involved 1,883 subjects older than 14 years of age, which was 84% of the randomly selected sample of 2,252. Of the respondents, 47% were men and 53% were women. Participants who indicated that they had headaches 10 days per month or more were requested to keep a headache diary for 1 month. On the basis of the diary, the prevalence of frequent headache, that is, headaches occurring 15 days per month or more, was determined to be 4.7%: 1.0% in men and 8.7% in women. The mean age of the subjects with frequent headache was 50 years; the mean age at onset of the frequent headaches was 38 years. Using the criteria presented earlier, the prevalence of chronic tension-type headache was determined to be 2.2% and that of transformed migraine, 2.4%. Overuse of abortive medications was found to be the case in (only) 19% of the patients with chronic tension-type headache and in 31% of those with transformed migraine.

With regard to frequent headache, Langemark and colleagues6 studied the clinical features of 148 patients with chronic tension headache. The patients had to have at least 10 days with headache per month and no more than one migraine attack. Ninety-three percent of them turned out to have at least 28 days with headache per month; that is, daily headaches (Figure 22-1). With regard to headache dynamics, this suggests that once headaches have increased to a frequency of 2 or 3 days per week, they rapidly progress to daily or almost-daily occurrence. The implication of this observation for the diagnostic criteria of chronic daily headache is that there is no need for an arbitrary number such as 15 (or 180). The diagnostic criterion for the condition with regard to frequency of headache could simply be “daily or near/almost daily,” and the same simplification could be made for chronic tension-type headache. Everything that falls short of this frequency criterion would be episodic; that is, episodic tension-type headache or (episodic) migraine. This observation also means that the prevalence of frequent headache can be equated with that of daily headache.

On the basis of the preceding epidemiologic studies, and taking the study by Langemark and colleagues into account, it can safely be stated that the prevalence of daily headache in the general population is approximately 5%. About half of these headaches are accounted for by chronic tension-type headache as defined by the IHS. With regard to age and gender characteristics of chronic daily headache, women are affected two times more often than are men, but age does not seem to have much of an effect on the prevalence of the condition.

The prevalence of chronic tension-type headache in the general population, as defined by the IHS, was also separately determined in Denmark.7 The study included 740 (76%) of 975 randomly selected subjects out of a total population of 325,621. The subjects were interviewed clinically, generating a prevalence number for chronic tension-type headache of 3%: 2% in men and 5% in women. It can, therefore, be safely stated that the prevalence of IHS-defined chronic tension-type headache in the general population is 2% to 3%, which accounts for about half of daily headaches.

In a study of chronic daily headache that my colleagues and I conducted,8–10 we defined the condition as headaches occurring at least 5 days per week for a period of 1 year or longer. We excluded only the patients with paroxysmal daily headaches, that is, cluster headache and paroxysmal hemicrania, in order to capture as much of the presentation, development, and outcome of chronic daily headache as possible. The study was conducted in 258 patients from my private headache practice, 19% men and 81% women, with an average age at consultation of 42 years. The distribution of the age of (any) headache onset for the men and women separately is shown in Figure 22-2. Seventy-seven percent of the patients (69% of the men and 79% of the women) experienced the onset of headache before the age of 30 years. The onset of headache occurred in the second decade of life in 36% of the women, compared with 24% of the men. The peak of headache onset in the second decade in women is consistent with the importance of the menstrual cycle in headache occurrence.

With regard to diurnal pattern, the daily headaches were present on awakening or occurred in the course of the morning in 79% of the patients, occurred in the afternoon or evening in 6%, and had a variable time of onset in 15% (Figure 22-3). In 25% of patients, the headaches were worst on awakening or in the course of the morning; in 53%, they were worst in the afternoon or evening, and in 22%, they were worst at a variable time of the day. The results agree with my clinical observation that daily headaches come in two distinct diurnal patterns. In the most common pattern, the headaches gradually increase in intensity as the day progresses, becoming worst in the afternoon or evening. According to the results of the study, this is the pattern in more than half of the patients with chronic daily headache. In the less common pattern, which I have referred to as reversed diurnal pattern, the headaches are worst on awakening in the morning and gradually improve as the day progresses. This was the case in one quarter of the patients, while in the remaining quarter, the diurnal course of the headaches was variable.

The reversed diurnal pattern is, in my experience, particularly associated with the overuse of analgesics or vasoconstrictors, or both, for headache. Overuse in this context is defined as medication intake that is detrimental rather than beneficial to headache. In the reversed diurnal pattern, the severe headaches on awakening in the morning are caused by the withdrawal of medication overnight, and the gradual improvement during the day results from the resumption of medication intake. This scenario also is associated with the most frequent nighttime awakenings with headache. In the study, nocturnal awakening by headache occurred at least once a week in 36% of the patients. Of those patients who were awakened by headache at least once a week, 48% experienced the worst headache on awakening or in the course of the morning, compared with 22% of the patients who were awakened by headache less than once a week.

Ninety-four percent of the patients experienced severe headaches in addition to the daily headaches. The distribution of the frequency of the severe headaches in days per month is shown in Figure 22-4. Twenty-six percent of the patients experienced severe headaches more than 15 days per month. Otherwise, they experienced severe headaches mostly 10 days per month or less (63%). The results suggest that the majority of the patients with chronic daily headache who seek specialty care for their headaches do not have chronic tension-type headache. Instead, they have chronic tension-type headache combined with migraine or, as I prefer to call it, tension-type vascular headache.

In the development of chronic daily headache, medication intake is considered to play an important role, in particular the intake of analgesics and vasoconstrictors. A widely used vasoconstrictor for the abortive treatment of headache is caffeine, which is contained in beverages, especially coffee, but also in prescription and nonprescription medications. We determined the caffeine intake in our patients with chronic daily headache by looking at their coffee and medication intake. A cup of coffee was considered to contain 100 mg of caffeine. We found that 43% of the patients used less than 100 mg of caffeine per day, 35% used between 100 and 300 mg, and 22% used more than 300 mg. The average caffeine intake was 170 mg per day, which is approximately the equivalent of two cups of coffee.

With regard to analgesic use, we considered only the non-opioid medications, because opioids were hardly used by the patients for the treatment of their headaches. Also, of the barbiturate-containing medications, we did not take into account the barbiturate component because it is not strictly an analgesic. With these limitations, we found that 26% of the patients used less than 500 mg of aspirin-equivalents per day and 48% less than 1,500 mg. The average analgesic intake was 1,860 mg of aspirin-equivalents per day.

Of the 230 patients in the study with known onset of daily headaches, 22% experienced daily headaches from the onset. This could be called primary chronic daily headache, in the same way that we speak of primary and secondary chronic cluster headache. The remaining 78% initially experienced intermittent headaches, that is, had secondary chronic daily headache. The distribution of the age of onset of the (daily) headaches in the patients with daily headaches from the onset, or primary chronic daily headache, is shown in Figure 22-5. Sixty-six percent of the patients experienced onset of daily headaches between the ages of 10 and 39 years.

Of the patients with daily headaches but who initially had intermittent headaches, that is, of those with secondary chronic daily headache, 19% experienced an abrupt onset of daily headaches and 81%, a gradual onset. The distribution of the age of onset of daily headaches in the patients with abrupt-onset secondary chronic daily headache was similar to that of the patients with primary chronic daily headache shown in Figure 22-5.

The circumstances related to the onset of daily headaches in the patients with primary chronic daily headache and in those with abrupt-onset secondary chronic daily headache are shown in Table 22-1. The table also shows the circumstances of daily headache onset for the two groups combined, as there was no difference in distribution of the circumstances between the two groups. The most common circumstance of daily headache onset in the two groups combined was head, neck, or back injury, caused by a motor vehicle accident in 61%. This was followed by flulike illness or sinusitis and medical illness or surgical procedure as causes of daily headache onset. Examples of medical illness associated with the (abrupt) onset of chronic daily headache are colitis, fibromyalgia, vertigo, encephalitis, and meningitis.

There were also no differences between the patients with primary chronic daily headache and those with abrupt-onset secondary chronic daily headache with regard to the following features: gender distribution, time of daily headache occurrence, worst headache time daily, nocturnal headache awakening, laterality of the daily headaches, occurrence and frequency of severe headaches, laterality of the severe headaches, and parental occurrence of headache. The only difference between the two groups was the association of the daily and severe headaches with nausea. Nausea was more common in the patients with abrupt-onset secondary chronic daily headache than in those with primary chronic daily headache. The difference probably results from the fact that 57% of the patients in the abrupt-onset group had a prior history of severe headaches, which tend to be associated with gastrointestinal symptoms.

The distribution of the age of onset of the daily headaches in the patients with gradual-onset secondary chronic daily headache is shown in Figure 22-6 (solid line). Seventy-eight percent of the patients experienced the onset of daily headaches between the ages of 20 and 49 years. The distribution of the age of onset of the initial, intermittent headaches in these patients is also shown in Figure 22-6 (interrupted line). The average duration of the transition of the headaches from intermittent to daily was 11 years, which is reflected in the figure by the separation of the two distributions by approximately a decade.

With regard to parental occurrence, headache in the father or mother, or both, was more common in the patients with gradual-onset secondary chronic daily headache than in the combined group of those with primary chronic daily headache and abrupt-onset secondary chronic daily headache (69% versus 45%). This finding is interesting because conditions that develop abruptly generally have lesser genetic involvement than do those that develop gradually. On the basis of the information gathered on parental headache occurrence, this also seems to be the case in chronic daily headache.

With regard to the intensity of the initial headaches, in the 145 patients with gradual-onset secondary chronic daily headache, the headaches were mild in 33% and severe in 67% (Table 22-2). The mild headaches were associated with nausea in 25% and vomiting in 0% compared with the severe headaches, which were associated with nausea in 84% and vomiting in 72%. With regard to the frequency of the initial headaches, there was no difference between the mild and severe headaches. The mild headaches occurred less than twice per week in 88% and the severe headaches in 91%.

The features of the daily headaches that these patients ultimately developed were the same, whether the initial headaches were mild or severe in intensity. They were the same with regard to age of onset of the (initial) headaches, gender distribution, diurnal headache pattern, nocturnal headache awakening, associated symptoms and laterality of the daily headaches, occurrence of severe headaches, and the frequency, associated symptoms, and laterality of the severe headaches.

From a classification perspective, does it make sense to distinguish between primary and secondary chronic daily headache as we did and, within the latter group, between abrupt- and gradual onset? Judging from the age of onset of the daily headaches, gender distribution, headache presentation, circumstances of headache onset, and parental headache occurrence, there does not seem to be a reason for the distinction between primary chronic daily headache and secondary chronic daily headache with abrupt onset. The two groups probably should be considered as having the same chronic daily headache condition, which could be referred to as abrupt-onset chronic daily headache, representing 37% of our study group. However, this group should probably be distinguished from the group having chronic daily headache with gradual onset because of the very different development of the headaches and the difference in parental headache occurrence. The latter group could be referred to as gradual-onset chronic daily headache, and future studies are needed to determine whether this distinction is meaningful in terms of predicting treatment or outcome, or both.

The above mentioned findings with regard to the development of chronic daily headache support the proposed headache continuum, shown schematically in Figure 22-7.11 The headache continuum includes episodic and chronic tension-type headache, migraine, and tension-type vascular headache. Tension-type headache and migraine are the two most common headache conditions. Episodic tension-type headache is experienced almost universally, whereas migraine affects 10% to 15% of the population. The major distinction between the two, in my opinion, is the intensity of the headaches. Headache intensity is traditionally divided into three categories—mild, moderate, and severe—depending on the extent to which the headache affects the ability to function. A mild headache does not affect the ability to function, a moderate headache affects the ability to function but does not necessitate bed rest, and a severe headache is incapacitating and requires bed rest. Tension-type headaches are mild or moderate in intensity, whereas migraine headaches are moderate or severe.

Related to the (peripheral) mechanisms involved in causing the pain, the migraine headache is localized whereas the tension-type headache is more diffuse in location. The migraine headache is localized not only to one side of the head but, within the side of the head, to areas such as the temple or eye. The pain tends to be throbbing or sharp, steady in nature, whereas the pain of tension-type headache is dull and steady. Also related to the mechanisms involved in causing the pain, the migraine headache is affected by movement and activity, which is not the case with tension-type headache. The migraine headache often develops during the night and is present on awakening in the morning or wakes the patient from sleep at night, usually between 4 am and 6 am. Episodic tension-type headache, on the other hand, generally develops during the day, often in the late afternoon, between 4 pm and 6 pm. The episodic tension-type headache lasts a couple of hours, whereas the migraine headache lasts from part of a day to several days. Related to the low intensity of the pain, tension-type headache has few, if any, symptoms associated with it, and when symptoms are present, they are mild in intensity. Migraine headache, on the other hand, has intense associated symptoms related to the high intensity of the pain. Almost universally present in migraine are photophobia and phonophobia; however, nausea also is common and with the most intense migraine headaches, vomiting occurs as well.

There is ongoing debate with regard to the mechanisms involved in causing the pain of tension-type headache and migraine. It is my belief that peripheral mechanisms are important in both headache conditions, although the present thinking is oriented more toward central mechanisms. In tension-type headache, the peripheral mechanism is that of sustained contraction of the craniocervical muscles and in migraine, that of extracranial arterial vasodilation. The arterial vasodilation in migraine activates a secondary mechanism, known as neurogenic inflammation. Stretching of the nerve fibers involved in pain transmission (Aδ and C) surrounding the blood vessels causes the neurogenic inflammation. The stretching causes the nerve fibers to depolarize, which, on the one hand, generates the action potentials that are transmitted to the central nervous system. On the other hand, it causes the release of inflammatory chemicals into the peripheral tissues, such as substance P and calcitonin gene-related peptide. The inflammatory chemicals act to further dilate the arteries and also cause a lowering of the pain threshold locally in the peripheral tissues. Thus, a vicious cycle is created in which vasodilation causes inflammation, which, in turn, accentuates the vasodilation and renders it extremely painful.

When headaches occur regularly, they lead, through an involuntary reflex mechanism, to progressive tightening of the craniocervical muscles. The greater intensity of the pain makes this effect more pronounced in migraine than in tension-type headache. Hence, muscular symptoms are more prominent in migraine patients than in those with tension-type headache. Contributing factors to the process of tightening of the craniocervical muscles are treatment of tension-type headaches with analgesics (as opposed to efforts to relax the muscles) and lack of effective abortive treatment in migraine.

In tension-type headache, the progressive increase in tightness of the craniocervical muscles leads, over time, to an increase in frequency of the headaches. It also leads to a progressive earlier occurrence of the headaches during the day. Ultimately, a daily or almost-daily headache condition develops in which the headaches are present on awakening or begin shortly after arising. As long as the headaches remain mild or moderate in intensity, the condition can be referred to as chronic tension-type headache. However, once the headaches have taken up all available time, they tend to increase in intensity to create “migraine headaches.” The condition is then referred to as migraine with chronic tension-type headache. However, a better term would be tension-type vascular headache, emphasizing the existence of one single headache condition, rather than suggesting the presence of two separate conditions.

When the regular occurrence of migraine headaches leads to a progressive increase in tightness of the craniocervical muscles, a gradual increase in frequency of the headaches occurs as well as a progressive interposition of the migraine headaches with tension-type headaches. The increase in frequency of the migraine headaches stems from the fact that the muscle tightness itself becomes a trigger of migraine headaches. As the muscles become tighter, they begin to interfere mechanically with their own circulation, creating a stimulus for dilation of the feeding arteries to occur. One such feeding artery is the frontal branch of the superficial temporal artery, which overlies the powerful temporalis muscle. This artery is also involved preferentially in the process of migrainous vasodilation, causing the throbbing or sharp, steady pain in the temple so characteristic of migraine. Ultimately, the migraine and tension-type headaches merge into a condition of daily or almost-daily headaches with frequent migraine headaches. A condition is, thus, created that is identical to what has previously been described as migraine with chronic tension-type headache or tension-type vascular headache. However, the difference is that, in this case, the condition developed out of migraine, whereas the condition described earlier developed out of episodic tension-type headache.

Of the 145 patients in our study with gradual-onset (secondary) chronic daily headache, we were able to contact 91 (63%) for follow-up telephone interviews. Seven patients refused to participate in the follow-up interview, and 11 no longer remembered the nature of their initial headaches. One patient was excluded from the analysis because of the absence of headaches at the time of contact, and three patients because of missing data. The remaining 69 patients (77%) were able to provide adequate information to classify their initial and present headaches as tension-type headache or migraine according to IHS criteria.

Twenty-three of the 69 patients (33%) still had daily headaches, whereas the remaining 46 (67%) again experienced intermittent headaches. Of the latter 46 patients, the initial headaches were classified as migraine in 39 (85%) and as tension-type headache in 7 (15%). Their present headaches were classified as migraine in 34 (74%) and as tension-type in 12 (26%). Thus, over time a shift had occurred from migraine to tension-type headache, accomplishing an improvement of the intermittent headaches for the group as a whole.

However, the question that we wanted to address was not whether the patients with intermittent headaches had improved in comparison with their initial headaches but whether patients with gradual-onset chronic daily headache revert back to their initial headache condition once the headaches become intermittent again. In the study, of the 39 patients whose initial headaches were classified as migraine, 30 (77%) also had migraine at follow-up and 9 (23%) tension-type headache. Of the 7 patients whose initial headaches were classified as tension-type headache, 3 (43%) had tension-type headache at follow-up and 4 (57%), migraine. Therefore, it seems that after experiencing daily headaches, migraine patients as a rule revert back to migraine, although some find their headaches improved to the extent that they are now classified as episodic tension-type headache. However, the situation is different for those patients who initially had episodic tension-type headache. They seem to be worse after having experienced daily headaches, with some patients experiencing headaches that have features and associated symptoms categorized as migraine according to the IHS classification.

The first step in the treatment of chronic daily headache is the accurate establishment of the use of analgesics and vasoconstrictors, both prescription and nonprescription. It is important to establish their use in terms of the number of tablets or capsules taken per day and the number of days of use per week or month. Patients tend to be notoriously vague about the intake of medications they use “as needed only”. They also often have to be reminded specifically to include nonprescription medications in their tally. Once the exact intake of analgesics and vasoconstrictors has been established, the clinician must determine whether overuse has occurred. As previously mentioned, overuse is defined as medication intake that is detrimental rather than beneficial to headache. It is use that promotes the occurrence of headache long term rather than providing headache relief. Analgesics and vasoconstrictors promote headache when they are taken for headache at time intervals shorter than their duration of action This dosing schedule allows them to accumulate in the system, with a return of headache when their effect wears off, a phenomenon known as rebound.

Rebound headache generally occurs when analgesics or vasoconstrictors are taken more often than 2 days per week on the average. This is particularly true for caffeine-containing medications because of the prolonged vasoconstrictor effect of caffeine, which can last for up to 60 hours. A higher frequency of intake of analgesics or vasoconstrictors can be permitted for simple analgesics and the shorter-acting triptans, such as sumatriptan. A lower frequency of intake should be considered with the longer-acting ergots, ergotamine and dihydroergotamine. However, it must be kept in mind that the rebound threshold has not been determined for any medication or group of medications. Furthermore, the diagnosis of rebound headache can only be made retrospectively, after withdrawal from analgesics and vasoconstrictors has been accomplished and improvement of headaches has occurred. A suspicion of rebound headache can be based not only on the frequency of medication intake, but also on increased medication usage over time with decreasing efficacy (Figure 22-8). The decreasing efficacy is often attributed to the development of tolerance but, in my opinion, it is more likely a manifestation of worsening of the headaches and an indication that use has become overuse.

If it is suspected that medication overuse and rebound headache are present, this situation needs to be addressed next. However, it can only be addressed after the patient has been given insight into the situation. With regard to vasoconstrictors, reference can be made to the vascular mechanism of headache. The vascular mechanism is antagonized by the vasoconstrictors, resulting in rebound vasodilation and headache recurrence when the vasoconstrictor effect wears off. Analgesics address only the pain of the headache and not the underlying mechanisms. Consequently—and as with symptomatic treatment in general—the underlying mechanisms deteriorate, resulting in worsening of headaches.

The withdrawal of analgesics or vasoconstrictors is generally best accomplished abruptly. However, whether that is possible also depends on the kind and quantity of the medications taken. When specific quantities of barbiturate-containing or opioid medications are involved, withdrawal may require hospitalization for close monitoring of withdrawal symptoms and intravenous administration of medications. The withdrawal of a significant amount of barbiturate-containing medication also requires a barbiturate taper to prevent seizure. The withdrawal of a significant amount of opioid medication requires expertise in addiction medicine and may have to be carried out in a detoxification center. Otherwise, it can often be accomplished on an outpatient basis, and several protocols have been developed to assist the patient with the withdrawal.

A recent protocol employs sumatriptan for withdrawal headache after abrupt discontinuation of the rebound-causing medications.12 Sumatriptan is given in a dose of 25 mg orally, three times per day for 10 days or until the patient is headache-free for 24 hours. Thereafter, it is used only on an as-needed basis for the abortive treatment of moderate or severe headache. Of the 35 sequentially selected patients to be enrolled in a study using this particular protocol, 9 left the clinic and 26 were actually treated. They had suffered from daily headaches for an average of 8.2 years. Of the 26 patients treated, 58% no longer experienced daily headaches after 1 month, and 69% had reverted back to intermittent headaches after 6 months.

An outpatient protocol that I have used successfully for the past 15 years to withdraw patients with daily headaches from daily or almost-daily use of abortive medications uses a short course of prednisone. Depending on the kind and quantity of the medications from which the patient has to be withdrawn, I give the prednisone for 3 or 6 days. The 3-day schedule consists of 15 mg prednisone four times per day for 1 day, 10 mg four times per day for 1 day, and 5 mg four times per day for 1 day; the days are doubled in the 6-day schedule. If the patient exhibits prominent muscular symptoms, that is, complains of tight or sore neck and shoulder muscles, I add diazepam to the schedule in a dose of 1 to 5 mg four times per day to help to relax the muscles.

I have used a similar schedule for patients admitted to the hospital when outpatient withdrawal was unsuccessful because of inability of the patient to tolerate the withdrawal headache or its association with severe nausea or vomiting. Under these circumstances, I add metoclopramide as an antinausea medication, given intravenously in a dose of 10 mg four times per day. It is important to start the metoclopramide immediately, that is, before the patient becomes sick, because once vomiting has developed, it is difficult to control even with intravenous administration of the medication. Instead of prednisone orally, dexamethasone can be given intravenously, in a dose of 4 mg four times per day, for several consecutive days. Diazepam can then be given every 6 hours, but only as needed for severe headache, and can also be given intravenously. An alternative to diazepam intravenously is lorazepam intramuscularly in a dose of 1 or 2 mg as needed every 6 hours.

An alternative to the preceding inpatient protocol with metoclopramide, dexamethasone, and diazepam or lorazepam is a regimen using metoclopramide and dihydroergotamine. In this protocol, both medications are given intravenously on a regular, generally 8-hour, schedule with the metoclopramide administered before the dihydroergotamine. This sequence is important to prevent the occurrence of nausea or vomiting as a result of the intravenous administration of dihydroergotamine. The dose of metoclopramide is usually 10 mg, and that of dihydroergotamine is gradually increased from 0.25 to 1 mg, depending on the ability of the patient to tolerate the medication, especially in terms of its gastrointestinal side effects. A long-term follow-up study of 50 consecutive patients with chronic daily headache treated with an intravenous dihydroergotamine protocol showed 44% to have good or excellent results after 3 months and 59%, after 2 years.13

After withdrawal from analgesics and vasoconstrictors, headaches may improve for up to 3 months. Often, preventive pharmacologic treatment is initiated immediately after the withdrawal but generally I do not do that. An exception is when the patient has problems sleeping at night, for which I will prescribe amitriptyline, doxepin, or trazodone. These are sedating tricyclic or tetracyclic antidepressants, of which the first two have also been shown to be effective in the preventive treatment of chronic tension headache. However, I do recommend patients to immediately begin nonpharmacologic treatments, such as using a heating pad daily on the neck and shoulders, to help decrease the muscle tightness that many patients have developed over time. At a later stage, I may prescribe more formal physical therapy, consisting of massage, ultrasound, stretching exercises, and so forth, or use trigger point injections to further relax the muscles.

For the daily headaches, I allow patients to use muscle relaxants, such as metaxalone or carisoprodol. For the severe headaches, promethazine, 50-mg suppositories, can be helpful as long as it is judged better for the patient not to use analgesics or vasoconstrictors. Once the headaches have become intermittent, I focus with my abortive treatment on the severe headaches, for which I try to find effective treatment, relying as much as possible on specific antimigraine medications. I define effective treatment as treatment that provides full relief of headache and associated symptoms within 2 hours of initiation. It is important for this treatment to be consistently effective as well, which would allow the patient to wait until the headache is severe before initiating it. This is the only way that patients can be prevented from falling back, over time, into the pattern of frequent intake of analgesics or vasoconstrictors.

With regard to preventive pharmacologic treatment, a particularly useful combination in those patients with frequent and severe headaches is that of a tricyclic and a beta blocker. The tricyclics that I prefer to use are amitriptyline, doxepin, and imipramine. I prescribe the first two when sedation is needed to help the patient fall asleep or sleep through the night. If sleep is not an issue, I prefer imipramine because it has fewer side effects, in particular a reduced likelihood of increased appetite and weight gain.

With regard to the beta blockers, six medications have been shown in randomized, double-blind, placebo-controlled studies to be effective in migraine prevention. These beta blockers are atenolol, bisoprolol, metoprolol, nadolol, propranolol, and timolol. In the patients with chronic daily headache, they are often effective in decreasing the intensity of headaches, whereas the tricyclics tend to have more of an effect on headache frequency. The calcium-entry blockers, in particular verapamil, I have found helpful if after analgesic or vasoconstrictor withdrawal, the headaches continue to awaken the patient regularly from sleep at night.

Kudrow determined the effect of analgesic withdrawal and preventive treatment with amitriptyline in 200 patients with chronic muscle-contraction headache who used analgesics daily, as documented by 1-month pretrial records.14 Patients were randomly divided into two groups and four subgroups. Half of the patients were prescribed amitriptyline (25 mg per day for 1 week and 50 mg per day thereafter). In each group, half of the patients were allowed to continue taking analgesics without restriction, while the other half were instructed to discontinue these medications. The percentage of headache improvement observed in the four groups is shown in Table 22-3. Analgesic withdrawal, by itself, resulted in a 43% improvement in headache as determined 1 month after initiation of treatment. The addition of amitriptyline to the analgesic withdrawal increased the headache improvement to 72%.

Our study suggests that with the previously outlined treatment approach, two thirds of patients who have daily headaches can be improved to intermittent headaches. Preventive treatment in these patients may reduce the symptoms of the intermittent headaches somewhat, as suggested by the shift from migraine to tension-type headache observed in our study, when the present headaches were compared with those that occurred initially.

What should be done with the patients who continue to have frequent and severe headaches, despite being taken off analgesics and vasoconstrictors and despite efforts at preventive treatment? This situation may be an indication for the use of long-acting opioids to relieve the pain sufficiently to allow these patients to function in their personal and professional lives. The long-acting opioids that I have used in such patients are fentanyl patch, oxycodone, and morphine sulfate. I have found these medications to be effective for a shorter period than manufacturers indicate and, therefore, use the fentanyl patch every 2, rather than 3, days and long-acting oxycodone every 6 or 8, rather than 12, hours. I gradually increase the dose of the medication until satisfactory pain control is achieved and the patient has returned to a relatively normal level of functioning. I prefer the use of long-acting to short-acting opioids because of the reduced likelihood that patients will develop tolerance and addiction. For some reason, rebound headache does not seem to develop with the long-acting opioids, although it is a routine consequence of short-acting analgesics, including opioids. The development of tolerance and rebound headache increases the use of opioid analgesics over time and makes it very difficult, if not impossible, to accomplish adequate pain control.

As far as its presentation is concerned, hemicrania continua can be looked on as a form of chronic daily headache. It is a nonparoxysmal daily headache, present continuously throughout the day, and limited to one side of the head. However, it is different from chronic daily headache in its treatment. Hemicrania continua is treated with indomethacin to which it has an absolute response, similar to paroxysmal hemicrania.15 The different treatment suggests a different etiology, which would preclude the condition from being grouped together with chronic daily headache.

It has been suggested that there is a form of hemicrania continua resistant to preventive treatment with indomethacin. Resistance to treatment with indomethacin, however, by definition means that it is not hemicrania continua. This does not mean that there are not numerous patients, who have continuous unilateral headaches with fixed lateralization that do not respond preventively to indomethacin. These patients have chronic daily headache and should be treated accordingly. The key to look for in the history is a response to aspirin, which seems to predict the responsiveness of the headaches to indomethacin. This is also the feature that ultimately led to the identification of both paroxysmal hemicrania and hemicrania continua as indomethacin-responsive headache syndromes.