Recent Results in Acetaminophen Treatment for Migraine

William B. Young, MD, Neurologist, Jefferson Headache Center, Thomas Jefferson University Hospital, Philadelphia, Pennsylvania

The evidence-based medicine movement has reached the over-the-counter (OTC) migraine market. In 1998, the formulation of aspirin, acetaminophen, and caffeine was approved for use in migraine by the US Food and Drug Administration (FDA). Most OTC analgesics are not approved as treatments for migraine, however. Double-blind, placebo-controlled trials have been conducted on prescription-strength ketoprofen and ibuprofen, and aspirin and over-the-counter ibuprofen have been studied.

Recently, Lipton et al. (Arch Intern Med 2000;160:3486) published a large, two-center, double-blind, placebo-controlled trial of acetaminophen for the treatment of migraine headache. Using a new recruitment strategy based on random-digit dialing and using International Headache Society-based questions, they screened the community for eligible migraine sufferers. The screened patients were contacted, and those who were eligible were invited to make an appointment for a medical assessment.

Entrance criteria were typical for an acute migraine treatment trial. They included having between one migraine in 2 months to six migraines per month that were of at least moderate intensity if left untreated. Patients with very severe migraine (that is, severe headaches more than 50% of the time or vomiting more than 20% of the time) were excluded. Subjects treated one self-recognized migraine attack with either placebo or 1000 mg oral acetaminophen; they were asked to delay treatment until the pain reached at least moderate intensity. Subjects maintained a diary in which they recorded their pain intensity, pain relief, functional disability, nausea, photo/phonophobia, and vomiting in 30-minute increments to 2 hours after dosing and then in 1-hour increments to 6 hours after dosing. They were also asked to give an overall impression of the study drug.

A total of 351 people were enrolled; 176 were assigned to receive the drug and 175 to receive placebo. The demographics and headache characteristics of the two groups were similar, and 97% of subjects reported their treated headache as typical. Forty percent reported that their baseline pain before treatment was severe. Eighty-two percent took medicine and were included in an intent-to-treat analysis.

The primary end point, which has been universally used since the initial sumatriptan trial in 1991 (Cady et al. JAMA 1991;265:2831) and which has been known as the "Glaxo Criteria," was 2-hour headache improvement; that is, the percentage of patients who went from moderate or severe pain to mild or no pain. Improvement in the acetaminophen group was 57.8%, and in the placebo group, 38.7%; 22.4% of the acetaminophen group and 11.3% of the placebo group became pain-free 2 hours after dosing. The pain-free values increased at 6 hours after dosing to 46.3% in the acetaminophen group compared with 28.2% of the placebo group. All of these numbers were statistically significant at the .01 level or less. Acetaminophen was differentiated from placebo at one-half hour based on pain intensity differences from baseline and an adjusted pain relief score. The use of rescue medicine was significantly less in the acetaminophen group (15.0%) than in the control group (31.7%). Similarly, photophobia and phonophobia were reduced in the treated group. Nausea, however, was not reduced in the treatment group compared with the placebo group. Subjects' overall impressions of the study medication were superior to their impressions of the placebo.

Functional disability is an important attribute of migraine, with consequences both for the individual and for society. In this study, patients rated their ability to work or function as normal or mildly, moderately, or severely impaired. Slightly more than 96% of subjects reported functional disability before treatment; of these, 31.4% of acetaminophen-treated subjects reported no disability at 2 hours compared with 15.3% of control subjects. At 6 hours, these values were 55.7% for the active drug group and 34.3% for control subjects. These differences were statistically significant at both times (p = .002 and .001, respectively).

As expected, no serious adverse events were reported in this study. There were no significant differences between placebo- and acetaminophen-treated groups with respect to adverse events reported, whether these were due to study drug or not.

In contrast to many abortive prescription medication trials, and similar to OTC migraine medication trials, outcomes were not followed out to 24 hours. Therefore, this study could not evaluate headache recurrence rate, an important feature of acute treatment for many patients. In most studies, 24-hour recurrence rates are around one-third; however, different initial efficacy rates and headache recurrence definitions have rendered comparisons difficult. Nonetheless, recurrence data can be important in making the best medication selection, and the absence of this information is unfortunate.

In summary, this study showed that acetaminophen is superior to placebo along multiple parameters at 2 and 6 hours. The study was conducted using a population-based recruitment strategy. This strategy is relevant for the self-treated individuals who may not be under treatment and may not even be diagnosed. Patients who rated their headaches as severe still responded to acetaminophen. The generalizability of this study to the patient who visits the doctor's office for the chief complaint of severe recurrent headache is uncertain, since many of these patients self-select based on failure to respond to OTCs. These findings should be framed in the context of other placebo-controlled studies of over-the-counter and prescription medications.

Commentary: Acetaminophen and Migraine Trials

Stephen D. Silberstein, MD, FACP, Professor of Neurology, Jefferson Medical College; Director, Jefferson Headache Center, Thomas Jefferson University Hospital, Philadelphia, Pennsylvania

Douglas C. McCrory, MD, MHS, FACP, Assistant Professor of General Medicine, Duke Center for Clinical Health Policy Research, Durham, North Carolina

The type, severity, and frequency of a patient's migraine determine its treatment. Acute medication is indicated even if patients are using preventive medication. It is important to treat the headache as early as possible both to prevent its escalation and to increase the drug's effectiveness.

Often, migraine begins with pain of mild to moderate severity and is described as similar to a tension-type headache. As the headache increases in severity, it takes on more migraine-like features. Medication for an acute attack can be specific or nonspecific. Nonspecific medications are used to control the pain and associated symptoms of migraine as well as other pain disorders, whereas specific medications control the migraine attack but are not useful for non-headache pain disorders. Nonspecific acute headache medications include analgesics and combination analgesics, antiemetics, and opioids. Specific acute headache medications include ergotamine, dihydroergotamine (DHE), and the triptans. It is advisable to develop a treatment strategy for headaches of different severity, using one or more of these drug classes.

Headaches should be stratified primarily by severity and disability and treated with the medications most likely to be effective for that attack, taking into account the drug's efficacy, safety, and side effects. Individualized treatment of mild to moderate headache often begins with the use of analgesics. Most people achieve headache relief with a simple analgesic, such as aspirin or acetaminophen, either alone or in combination with caffeine or other adjuvants. Some physicians believe that these combinations are more likely to lead to overuse and dependence and may not offer enough additional pain control to justify their use. Acetaminophen is always an option to aspirin or the other nonsteroidal anti-inflammatory drugs (NSAIDs) in patients with gastritis or bleeding disorder and, because of the risk of Reye's syndrome, is preferable to aspirin in children younger than 15 years with nonspecific headache.

Until recently, the efficacy of acetaminophen had not been established in acute migraine treatment. Lipton et al. (Arch Intern Med 2000;160:3486) studied 289 patients in a placebo-controlled, double-blind randomized trial of acetaminophen 1000 mg versus placebo. They found that at 2 hours after dosing, significantly more subjects taking acetaminophen had headache relief (58% vs. 39%; p = .002) or were pain free (22% vs 11%; p = .01).

Previous Studies

These findings contrast with earlier placebo-controlled trials of acetaminophen for migraine. Diamond (Headache 1976;15:282) compared isometheptene in combination with acetaminophen and dichloralphenazone, acetaminophen alone, and placebo in a crossover study. Data on headache relief (graded by patients as complete, good, fair, or none at an unspecified time point) were reported for each treatment group. The investigator's analysis found that there was no significant difference between acetaminophen and placebo as far as headache relief was concerned.

Dexter et al. (Br J Clin Pract 1985;39:388) compared the combination of acetaminophen and metoclopramide (Paramax®, SmithKline Beecham) with placebo. The investigators found no significant difference between the two treatments as far as headache severity was concerned (severity was measured once, at the end of each attack). Of the 72 attacks treated with Paramax and included in the efficacy analysis, 29 (40%) were mild, 24 (33%) were severe, and 19 (27%) were incapacitating; of the 96 attacks treated with placebo and included in the analysis, 44 (44%) were mild, 27 (28%) were severe, and 25 (26%) were incapacitating. We were not able to calculate an odds ratio or effect size for this outcome on the basis of the data reported in the study.

Different Conclusions

Several differences between the studies may explain the differences in findings, for different reasons. These include differences in the study population, in the way the drugs were given, in the way the outcomes were measured, and in the study design (Table 1).

Table 1: Comparison of placebo-controlled studies of acetaminophen in migraine.

*We assessed the internal validity of individual trials using a scale devised by Jadad et al. (Controlled Clin Trials 1996;17:1). This scale evaluates methodological quality based on the following considerations: the use of random allocation; description of an adequate method of concealment of allocation; the use of double-blinding; description of an adequate method of blinding; and a description of drop-outs sufficient to determine the number of patients in each treatment group entering and completing the trial. Each trial could score between 0 and 5 points, with higher scores indicating higher quality in the conduct or reporting of the trial. Each of the items on this quality scale is an accepted criterion that has been empirically validated. The Jadad instrument is one of only a few such scales that has undergone a formal process of development and demonstrated good inter-rater reliability (Moher et al. Controlled Clin Trials 1995;16:62).

Incomplete reporting, especially in the case of Dexter et al. and Diamond, hampers precise characterization of the patients; however, there is enough information to conclude that the patients involved in the three trials were different. First, they were recruited from different settings: subjects in Lipton et al. were identified through random-digit dialing, and thus represent a population-based sample of community-dwelling migraine sufferers. In contrast, both Dexter et al. and Diamond recruited patients from among those attending a headache clinic. One might expect differences in the severity of headache symptoms between these populations.

The International Headache Society criteria were used to verify migraine diagnosis in Lipton et al.; however, the other two studies predated those criteria. The sex distribution was similar in the three studies, ranging from 71% to 79% women; the mean age was similar for Lipton et al. (37 years) and Dexter et al. (33 years) and somewhat higher for Diamond (49 years). Headache severity was described as "typically severe, lasting longer than 4 hours and frequently accompanied by nausea" by Diamond. Dexter et al. rated overall severity of migraine attacks (43% of attacks mild, 30% severe, 26% incapacitating) as well as headache duration and incidence and duration of nausea and vomiting. Headache pain severity as reported by Lipton et al. (59% moderate, 41% severe) was arguably not dissimilar from Dexter, assuming that pain severity is the main factor in rating overall migraine severity.

It is in symptoms of nausea and vomiting that the populations differ most. Lipton et al. excluded patients who usually required bedrest with their headaches or who vomited with more than 20% of attacks. As a result, the incidence of nausea (none 51%, mild 26%, moderate 20%, and severe 2%) and vomiting (yes 1.7%, no 98%) were very low compared to the incidence in Dexter et al., where 88% of patients reported nausea and vomiting. The difference in nausea and vomiting symptoms may explain the differences in findings if acetaminophen is effective at alleviating headache pain among patients with little nausea and vomiting (as found by Lipton et al.), but less so among patients with greater gastrointestinal symptoms.

Aside from the differences in the patient population, differences in methodology may have obscured assessment of acetaminophen's efficacy. Lipton et al. describe calculating the sample size required to provide 90% power to detect a difference of 20% in headache response and assumed a 35% placebo response rate. The actual placebo response rate (38.7%) and active treatment response rate (57.8%) were very similar to these estimates, and given that the suggested number of subjects were recruited, this study was appropriately statistically powered. In contrast, the small sample sizes in Diamond (N = 56) and Dexter et al. (N = 49) may have precluded finding a statistically significant effect of acetaminophen compared to placebo. In Dexter et al., no difference was observed in migraine severity (not even a trend); while in Diamond's study, acetaminophen-treated patients had higher headache relief scores that did not reach statistical significance. These findings are consistent with acetaminophen being effective but the study design being underpowered to detect a difference. However, these findings are also consistent with acetaminophen being effective for only a subpopulation (such as those without nausea and vomiting) of the study sample, and ineffective in the remainder (such as those with nausea and vomiting).

Additional differences in study design (crossover versus parallel group), number of attacks treated, dose of acetaminophen, or outcome variables measured are unlikely to be responsible for the differences in conclusions of these studies.

The Lipton et al. study is conclusive: In patients who rarely vomit during attacks, acetaminophen effectively relieves the pain of migraine headache and can help alleviate the phonophobia, photophobia, and functional disability that frequently occur with migraine attacks. However, the problem of insufficient statistical power in the Diamond and Dexter et al. studies prevents one from being certain whether acetaminophen is also effective for patients with more severe migraine symptoms including nausea and vomiting.