Probable medication-overuse headache

The International Classification of Headache Disorders, 2nd ed. (ICHD-II), introduced specific diagnostic criteria for medication-overuse headache (MOH).¹ A revised version of these diagnostic criteria has recently been published on behalf of the International Headache Society.² An important change in ICHD-II was a shift from a required minimum dose of drug per month to a required minimum monthly frequency of intake. The threshold for triptan overuse was supported by a recent study,³ whereas the thresholds for other medications were based on expert opinion rather than formal evidence. To be classified as MOH, headache must resolve or revert to its previous pattern within 2 months after discontinuation of the overused drug. Therefore, only a temporary diagnosis of probable MOH can be applied initially, until medication withdrawal is completed. Underlying the whole classification of MOH is the assumption that medication overuse by itself transforms and aggravates the primary headache and, conversely, that simply removing the overused medication should make the patients better. However, few if any headache clinics keep their patients drug-free for 2 months, and whereas several studies have reported on MOH, most studies have allowed prophylactic treatment before 2 months or have not provided information about drugs substituting for the overused medication. It remains uncertain whether or not patients get better by mere discontinuation of the overused drug and nothing else. The primary aim of the current study was therefore to describe the emerging profile of headache frequency in patients with medication overuse 2 months after discontinuation of the offending drug and without any other headache treatment. The secondary aim was to relate headache outcome to the primary headache diagnosis and the type of overused medication.

Methods.

The Danish Headache Center (DHC) is a tertiary outpatient referral headache center. It functions as the only national referral center for severely affected headache patients in Denmark (5 million inhabitants). At the time this study was conducted, only referrals made by neurologists were accepted.

All patients treated and discharged from DHC in 2002 and 2003 had prospectively filled out a diagnostic headache diary on a daily basis. Diagnoses, headache frequency, and medication use were determined from these prospective recordings and loaded into a Microsoft Access 2002 database together with information on socioeconomic status and previous pharmacologic and nonpharmacologic treatment. The patients were primarily diagnosed according to ICHD-I, but data allowed a subsequent reclassification according to ICHD-II.

Medication use was transformed to standard dosages and sorted into eight groups (table 1). Drugs that could not be sorted in one of the first eight groups were placed in the group “others” and registered as days per month of drug use. After establishment of a diagnosis of probable MOH, all acute headache medication was discontinued abruptly and patients were kept medication-free for 2 months as part of the general treatment program in DHC.⁴ In case of severe opioid overuse, Phenobarbital substitution was used for a short period to avoid abstinence syndromes. Overuse of barbiturates or benzodiazepines must be tapered slowly, but marked overuse of these substances was not documented among our patients. Levomepromazine or phenergan was allowed as the only rescue medication during withdrawal, primarily to be used for the first week. After withdrawal, headache frequency was reassessed and diagnoses were revised.

Results.

Among 1,326 patients treated and discharged from DHC in 2002 and 2003, we identified 337 (25%) with an initial diagnosis of probable MOH. We excluded 121 who did not stay medication-free for 2 months. Among the remaining 216 patients, there were 58 men (27%) and 158 women (73%) with a median age of 48 years (range 17 to 86 years). Twenty-one patients (10%) had migraine, 72 patients (33%) had TTH, 92 patients (43%) had MT, and 31 patients (14%) had other headache diagnoses. Median duration of primary headache was 17 years (range 1 to 76 years). The median headache frequency before withdrawal was 30 days/month (range 15 to 30 days/month) (table 2). Eight patients (4%) overused ergots, 43 patients (20%) overused triptans, 63 patients (29%) overused simple analgesics, 12 patients (6%) overused opioids, and 90 patients (42%) overused combination analgesics (table 3). There was no significant difference between excluded and enrolled patients with regard to age, sex, primary headache diagnoses, type of overuse, or initial headache frequency.

A large number of patients (n = 100) encountered 30 days/month with headache both before and after medication withdrawal. Based on a relative reduction in headache frequency greater than, equal to, or smaller than 0% following withdrawal, 45% of the patients had a reduction in headache frequency following medication withdrawal, 48% of the patients neither improved nor worsened, and only 7% had headache aggravation. The proportion of patients encountering headache aggravation was 9.5% in migraine, 6.9% in TTH, 6.5% in MT, and 9.7% in OD.

The median relative reduction in headache frequency following withdrawal varied considerably between the diagnostic groups, with a 67% (range −46 to 100%) reduction in migraine, 0% (range −56 to 100%) in TTH, 37% (range −37 to 95%) in MT, and 0% (range −56 to 100%) in OD. This difference is illustrated in figure 1 for migraine and TTH. Paired comparisons of frequencies within the diagnostic groups showed a reduction both overall and for migraine, TTH, and MT (p < 0.0001), but not for OD (p = 0.20) (table 2).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1. Relative reduction in headache frequency following a 2-month drug-free period in patients with medication overuse. Results are summarized in four groups of improvement within each diagnostic group

Patients with both migraine and TTH had a baseline TTH frequency approximately 3 times higher than migraine frequency. The relative reduction was fairly evenly distributed between migraine and TTH days (table 2).

Because the OD group was heterogeneous and did not respond to medication withdrawal, we excluded it from further analyses.

Comparing migraine, TTH, and MT revealed a difference in relative reduction between the diagnostic groups (p < 0.0001). Post hoc analysis showed a smaller reduction in TTH than in migraine (p < 0.001) and the combination of migraine and TTH (p < 0.01). The difference between migraine and MT was not significant (p = 0.18). Controlling for medication group, the overall difference between migraine, TTH, and MT remained significant (p < 0.05). The post hoc analysis remained significant only for the comparison between MT and TTH (p < 0.05), whereas migraine vs TTH and migraine vs MT yielded no significant differences, probably due to small group sizes.

The outcome following withdrawal was clearly dependent on the type of overuse. The proportion of patients with unchanged or more headaches following drug withdrawal was, when overusing ergots, 13%, triptans 28%, simple analgesics 67%, opioids 67%, and combination analgesics 61%. This is illustrated in figure 2, which shows the relative reduction in headache frequency within each diagnostic subgroup of MOH. Comparing triptan/ergot overuse vs all other medication overuse showed a difference in favor of a better outcome for overusers of triptan/ergots (p < 0.0001). Controlling for primary headache diagnosis using a stratum-adjusted Kruskal–Wallis test eliminated this difference (p = 0.14).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2. Relative reduction in headache frequency following a 2-month drug-free period in patients with medication overuse. Results are summarized in four groups of improvement within each type of overused medication

For the diagnostic group with pure TTH, the relative reduction in headache intensity was also analyzed by means of a 0-to-3 scale. Data were available for 47 patients, with a median reduction in pain intensity of 8% and mean of 21% (quartiles 0 to 48%) (p < 0.0001).

Discussion.

This prospective study shows that patients with frequent headaches and medication overuse get better by merely discontinuing the offending drug, even if no other treatment is given.

This study was prospective, but not blinded and not controlled. We investigated the emerging profile of headache frequency in patients with probable MOH, following a medication-free period of 2 months. Excluding patients who did not comply with this strict regime resulted in an expected 36% dropout rate. Even then, our sample size is large. Rescue medication in the form of levomepromazine/phenergan was allowed in the first 1 to 2 weeks of withdrawal, but otherwise no analgesics or acute migraine medications were allowed. Detailed information about the use of rescue medication was not available for all patients, but a prior study from DHC has shown that only 66% of the patients undergoing medication withdrawal used the rescue medication with an average amount of five standard dosages during these 2 months.⁴ We therefore find it reasonable to conclude that rescue medication did not interfere with our results.

Our patients had a long history of severe chronic headache with no effect of multiple previous treatments. It is therefore unlikely that the observed effect of drug withdrawal could be a simple time effect. Furthermore, different results in different diagnostic categories should be unbiased. The marked differences between the diagnostic groups and the completely absent effect in the group OD therefore suggest a specific effect of drug withdrawal. It is important to emphasize that this effect can only be used with caution for causal inferences. It does not prove with certainty that medication overuse is the cause of aggravation in patients who improve following drug withdrawal. Likewise, in the absence of improvement after withdrawal, it is possible that a previous medication overuse has caused enduring changes in the pain perception that persist after exposure has been ceased.

A strength of our study is the highly valid headache data from prospective recordings in headache diaries. In addition, there is no social bias because medical treatment is free for all patients in Denmark.

Considering the selection bias found in tertiary headache centers, the mean amount of overused drugs per month was not excessive, ranging from 30 standard dosages for triptans, 60 standard dosages for combination analgesics, to 75 standard dosages per month for simple analgesics. Compared with other studies, our patients tended to use larger amount of simple analgesics, lower amounts of combination analgesics, and a similar amount of triptans.^3,5,6

The age was slightly older with less female preponderance compared with other studies.^3,6–9 Pure migraine was diagnosed in only 10% of the patients, whereas TTH accounted for 33% and the combination of migraine and TTH for 43%. This diagnostic distribution differs from that found in a recent study, which had a much larger proportion of pure migraine (71%) and lower proportions of TTH (14%) and their combination (15%).³ Consequently, the distribution of overuse also differs in the way of fewer triptan and ergot overusers in the current study. We found that 4% overused ergots, 20% triptans, and 71% analgesics or combination analgesics in contrast to 13% ergots, 39% triptans, and 48% analgesics in that previous study.³ These variations are probably due to either cultural differences or more likely to our systematic prospective use of diagnostic diaries. Episodic TTH is usually underdiagnosed in a diagnostic interview compared to a diagnostic diary, and nausea, photophobia, and phonophobia tend to be more pronounced at the clinical interview.¹⁰

By mere discontinuation of overused medication, 45% of the patients had a reduced headache frequency. Among the rest, 48% had no worsening and only 7% had aggravation of the headache. In prophylactic migraine trials, migraine frequency (days/month) is the most widely used efficacy measure, which is in accordance to the International Headache Society recommendation for migraine trials from 2000.¹¹ Days per month is also the gold standard in TTH trials according to the International Headache Society recommendation from 1995.¹² This recommendation has been questioned¹³ and contrasts the clinical experience that a modest reduction in headache intensity and duration, measured in hours per day, usually is considered highly relevant and a major improvement for patients with chronic TTH. Consequently, most clinical trials involving TTH make use of a headache index as efficacy measure, where intensity and duration are also taken into account.^14,15 Due to the relatively large proportion of patients with a primary diagnosis of TTH in the current study, our findings may reflect the difficulty using headache frequency as the primary outcome measure in TTH. We therefore analyzed headache intensity in the group with pure TTH and found a median reduction in intensity of 8%. The upper quartile for reduction in intensity was 48%, which shows that a substantial proportion of these patients have a significant reduction in intensity and that there is a large variation between patients. Though the reduction in TTH intensity was modest, the combined effect of a reduction in both frequency and intensity represents a clinically meaningful improvement. This is supported by our treatment evaluation questionnaire, where the group with pure TTH reported a median value of 2 (equal to good) on a 1-to-5 scale (excellent to miserable) when asked about treatment influence on headache and quality of life (unpublished data).

Our data show that with use of headache frequency as the sole criterion for improvement following medication withdrawal, the effect will be most pronounced in migraine patients, whereas TTH is more complex and probably should include intensity and duration in the efficacy measure.

The relative reduction in headache frequency varied considerably between the diagnostic groups and was in favor of a migraine diagnosis, with a 67% median reduction in pure migraine, 37% in the combination of migraine and TTH, and 0% in pure TTH and the group of other headaches. This is consistent with a previous study that reported a nonmigraine diagnosis in 70% of the patients that did not improve after medication withdrawal and a migraine combination diagnosis in the remaining 30%.⁷ A study containing mostly migraine patients found a 58% improvement in headache index (frequency × severity) after 12 weeks, by discontinuation of overused medications,⁶ whereas another study reported a 74% reduction in headache days 1 year after a successful and maintained detoxification in patients with transformed migraine.⁹ Yet these patients were also on prophylactics in contrast to our patients. These figures are nevertheless comparable with the 67% reduction we found and suggest that in migraine the full effect of withdrawal is seen already after 2 months. It therefore seems reasonable to evaluate the outcome of withdrawal after 2 months, as stated by ICHD-II in the diagnostic criteria for MOH.

Interestingly, the overall differences between the diagnostic groups remained significant when controlling for triptan/ergot overuse, whereas the difference in favor of a better outcome for triptan/ergot overusers was neutralized when controlling for primary headache diagnosis. These results should be interpreted with caution, as some of the group combinations in the stratified analysis become very small, entailing a risk of type 2 error. In any case, our data indicate an outcome effect related to the primary headache diagnosis irrespective of the type of overused medication.

It is important to emphasize that the reported effect of a 2-month drug-free period is the mere effect of a simple drug withdrawal and that any eventual benefits of prophylactic treatment have not been included. Assuming that medication overuse nullifies the effect of prophylactic agents,⁶ the possibility of a satisfactory long-term treatment outcome exists also in the patients who had no initial reduction in headache frequency.

Acknowledgment

The authors thank Mrs. Hanne Andresen for technical assistance during data collection.