Headache ISSN 0017-8748C© 2006 by American Headache Society doi: 10.1111/j.1526-4610.2006.00376.xPublished by Blackwell Publishing

Research Submission

Acetaminophen, Aspirin, and Caffeine in CombinationVersus Ibuprofen for Acute Migraine: Results From a

Multicenter, Double-Blind, Randomized, Parallel-Group,Single-Dose, Placebo-Controlled Study

Jerome Goldstein, MD; Stephen D. Silberstein, MD; Joel R. Saper, MD; Robert E. Ryan, Jr., MD;

Richard B. Lipton, MD

Objective.—Compare the effectiveness of a combination analgesic containing acetaminophen, aspirin, andcaffeine to that of ibuprofen in the treatment of migraine.

Methods.—Multicenter, double-blind, randomized, parallel-group, placebo-controlled, single-dose study. Atotal of 1555 migraineurs were included in the analysis. No patients were excluded solely because of severity ofsymptoms or degree of disability. A single 2-tablet dose for each of the 3 treatment groups: a combination productcontaining acetaminophen 250 mg, aspirin 250 mg, and caffeine 65 mg per tablet (AAC); ibuprofen 200 mg pertablet (IB); or matching placebo. The primary efficacy endpoint was the weighted sum of pain relief (PAR) scoresat 2 hours postdose (TOTPAR2) and an important secondary endpoint was the time to onset of meaningful relief.

Results.—There were 669 patients in the AAC group, 666 patients in the IB group, and 220 patients in theplacebo group. The 3 treatment groups had similar demographic profiles, migraine histories, and baseline symptomprofiles. While both active treatments were significantly better than placebo in relieving the pain and associatedsymptoms of migraine, AAC was superior to IB for TOTPAR2, as well as for PAR, time to onset of meaningfulPAR, pain intensity reduction, headache response, and pain free. The mean TOTPAR2 scores for AAC, IB, andplacebo were 2.7, 2.4, and 2.0, respectively (AAC vs. IB, P < .03). The median time to meaningful PAR for AACwas 20 minutes earlier than that of IB (P < .036).

Conclusion.—AAC and IB are safe, cost-effective treatments for migraine; AAC provides significantly superiorefficacy and speed of onset compared with IB.

Key words: acetaminophen, aspirin, caffeine, ibuprofen, migraine

(Headache 2006;46:444-453)

From the San Francisco Headache Clinic, San Francisco, CA (Dr. Goldstein); Jefferson Headache Center, Philadelphia, PA(Dr. Silberstein); Michigan Head Pain & Neurological Institute, Ann Arbor, MI (Dr. Saper); Ryan Headache Center, Chesterfield,MO (Dr. Ryan); and Department of Neurology, Albert Einstein College of Medicine, Bronx, NY (Dr. Lipton).

Address all correspondence to Dr. Jerome Goldstein, San Francisco Headache Clinic, 909 Hyde Street, Suite 322, San Francisco,CA 94109.

Accepted for publication December 14, 2005.

444

Headache 445

Migraine is a chronic neurologic disorder charac-

terized by episodic attacks of head pain and associated

symptoms, including photophobia, phonophobia, nau-

sea, vomiting, and aura.1 In the United States, about 28

million people have migraine; two thirds are women.2,3

Recent population-based studies suggest that adults

with migraine average 1.8 attacks per month,4 but the

frequency and severity of episodes tend to vary from

person to person and within individuals over time. Like

other chronic disorders, migraine imposes a significant

economic and social burden on those with headache,

their families, and the society.5 In American Migraine

Study II, more than half (53%) of migraineurs re-

ported severe impairment of activity. Approximately

31% missed at least 1 day of work or school in the

3 months preceding interview because of migraine.2

The vast majority of people with migraine use

medication to treat their condition. About 41% use

prescription medications, either alone or combined

with over-the-counter (OTC) medications.3 Approx-

imately 57% self-treat with OTC medications to the

exclusion of prescription drugs.3 Widely used OTC

medications for migraine include single-ingredient ac-

etaminophen (APAP), aspirin (ASA), or ibuprofen

(IB), as well as a combination analgesic containing ac-

etaminophen 250 mg, aspirin 250 mg, and caffeine 65

mg per tablet (AAC).2,3,6-10

In 3 well-controlled trials, AAC (Excedrin� Mi-

graine) has shown significant superiority to placebo

in the treatment of migraine.11 Retrospective analy-

ses of these data have confirmed that AAC is signifi-

cantly more effective than placebo in treating subsets

of migraineurs with severe head pain and associated

symptoms12 and women with menstruation-associated

migraine.13 Another randomized, controlled clinical

study in the early treatment of migraine found that

AAC was significantly more effective than sumatrip-

tan, a leading prescription medication.14 Separate in-

vestigations have demonstrated the superiority of IB

to placebo: the liquigel15 (Advil� Migraine) formula-

tion of IB in the treatment of migraine and the tablet16

(Motrin� Migraine Pain) in the treatment of migraine

pain. A single study also found that APAP was supe-

rior to placebo for relief of migraine pain.17 Despite

the established efficacy and availability of several OTC

alternatives for migraine, no comparative clinical trial

for treatment of migraine with OTC medications has

yet been published. However, one recently published

study, which included exclusively typical OTC patients,

has shown that 2 tablets of a triple combination very

similar to AAC (250 mg ASA + 200 mg APAP + 50 mg

caffeine per tablet) were significantly superior to a

treatment with 2 tablets of 250 mg ASA + 200 mg

APAP, 2 tablets of 500 mg ASA, two tablets of

500 mg APAP, 2 tablets of 50 mg caffeine, and placebo

in patients with migraine and episodic tension-type

headache (TTH).18

Since AAC and IB are the approved OTC alter-

natives for migraine, we decided to compare their effi-

cacy in a head-to-head clinical trial. Herein, we report

the results of a multicenter, double-blind, randomized,

parallel-group, placebo-controlled study assessing the

effectiveness of AAC versus IB and placebo in treating

a single migraine attack.

Most previous OTC migraine studies11-17 enrolled

patients who did not usually require bed rest with their

headaches and patients who vomited with less than

20% of attacks. To assess the efficacy of these agents

in an unrestricted patient population, we included the

full range of those with migraine, without constraints

on the level of disability or frequency of vomiting.

SUBJECTS AND METHODSSubjects.—This study used a double-blind,

randomized, parallel-group, placebo-controlled de-

sign. Patients were recruited with population-based

(random-digit dialing) and traditional (eg, private

practice, referrals, and local advertising) methods. A

detailed clinical assessment, conducted by the investi-

gator, ensured that each subject’s headaches met In-

ternational Headache Society (IHS) diagnostic crite-

ria for migraine without aura (IHS 1.1) or migraine

with aura (IHS 1.2).1 In addition, each subject was

at least 18 years old, was in good general health,

and had experienced a migraine attack at least

once every 2 months—but no more than 6 times

monthly—during the prior 12 months. Untreated at-

tacks were of at least moderate pain intensity. Pa-

tients whose headache symptoms may have been

caused or aggravated by recent head or neck trauma

and patients with cluster headache, specific mi-

graine variants, or other serious nonmigraine causes

446 March 2006

of headache were excluded. Patients who reported

using analgesic drug products for headache on more

than 12 days per month were also excluded. Unlike

earlier studies of OTC medications for migraine,11-17

no patients were excluded based on the requirement

for bed rest or the presence of frequent vomiting.

Study Design—Protocol.—The study was con-

ducted in 4 phases: screening, selection, treatment,

and follow-up. Trained interviewers identified poten-

tial patients at the screening phase. At the selection

phase, patients provided written informed consent and

a complete medical history. Physical and neurologic

exams were also completed at this time. Female pa-

tients also underwent urine pregnancy testing. Qual-

ified patients were randomly assigned (3:3:1 ratio) to

receive double-blinded study medication containing

a single, four-tablet dose consisting of 2 unbranded

AAC tablets (APAP 250 mg, ASA 250 mg, and caffeine

65 mg) and 2 dummy IB tablets, 2 unbranded IB tablets

(200 mg IB), and 2 dummy AAC tablets or 4 dummy

tablets to treat the pain and associated symptoms of

a single acute migraine attack. If the headache symp-

tom profile met the criteria for migraine and was of

at least moderate intensity, patients were instructed

to take study medication. They were asked not to

take rescue medication for at least 2 hours, if possi-

ble. All treatment information remained blinded until

all queries were resolved and the database was locked.

The authors affirm that an institutional review board

at each investigative site reviewed and approved the

study protocol and the conduct of the research.

Efficacy Measurements.—The prospectively de-

fined primary efficacy endpoint was the weighted

sum of pain relief (PAR) scores at 2 hours postdose

(TOTPAR2). Other endpoints included TOTPAR

at 4 hours (TOTPAR4), time to meaningful PAR,

pain intensity difference from baseline (PID), 4-hour

weighted sum of pain intensity differences from base-

line (SPID4), proportion of patients with pain reduced

to mild or none (headache response, HR), and propor-

tion of patients with pain reduced to none (pain-free,

PF). At baseline, patients rated pain intensity, func-

tional disability, nausea, vomiting, photophobia, and

phonophobia in the study diary. They also rated these

symptoms along with PAR at 15, 30, 45, 60, 90, 120,

180, and 240 minutes postdose. The time to meaning-

ful PAR (defined as that point in time after the subject

ingested the study medication when relief of headache

pain was considered meaningful by the subject) was

assessed with a stopwatch.

Patients rated PAR on a 5-point scale (0 = no re-

lief; 1 = a little relief; 2 = some relief; 3 = a lot of

relief; and 4 = complete relief). They rated pain inten-

sity by means of a 4-point scale (0 = no pain; 1 = mild

pain; 2 = moderate pain; and 3 = severe pain). They

rated functional disability from 0 to 4 (0 = none; 1 =usual activities require a little additional effort; 2 = re-

quire some additional effort; 3 = require a great deal

of additional effort; and 4 = unable to perform usual

activities). The associated symptoms of nausea, vomit-

ing, photophobia, and phonophobia were recorded as

absent (0) or present (1) for each symptom.

Safety Assessments.—Patients recorded adverse

experiences (AEs) in the diary, and investigators com-

pleted the AE evaluation during the follow-up phase.

The AE type, intensity, duration, seriousness, relation

to the study drug, and outcome were recorded. Clinical

laboratory data were not routinely collected.

Sample Size Calculation.—Sample size calcula-

tions were based on the efficacy parameter TOTPAR4.

The selected sample size of 665 patients per active

treatment group provided at least 90% power to detect

0.7 units difference in TOTPAR4 (using a standard de-

viation estimate of 3.87 units) between the 2 treatment

groups (2-sided, α = 0.05).

Statistical Analysis.—Treatment group compa-

rability was assessed using analysis of covariance

(ANCOVA) for quantitative variables (eg, age) and

chi-square tests for categorical variables (eg, sex). The

treatment groups were compared with respect to de-

mographics and baseline characteristics.

Data missing for any scheduled postdose time

point evaluation in otherwise evaluable patients (eg,

patient fell asleep) were interpolated. For example, if

the 30-minute observation was missing, it was replaced

by the average of the 15-minute and the 45-minute

observation value. For patients requiring rescue med-

ication, postrescue medication intensity scores for

pain intensity, functional ability, and nausea were as-

signed either the baseline or the last recorded value,

whichever was higher. Postrescue medication PAR

scores were assigned “no relief.”

Headache 447

The primary efficacy analysis data set was the

intent-to-treat subject patient population.

Time to meaningful PAR was analyzed using

Wilcoxon rank sums and nonparametric survival tech-

niques, stratified by investigator. The analyses of PID,

SPID, PAR, and TOTPAR were made at each time

point using ANCOVA of treatment group and in-

vestigator as main factors and baseline pain intensity

as the covariate. The Cochran–Mantel–Haenszel test,

stratified by investigator and baseline pain intensity,

was used to compare treatment groups with respect

to cumulative proportions of patients who achieved

meaningful PAR, PF, HR, and those who remedicated

by each scheduled time point. The Cochran–Mantel–

Eligible patients (N=1714)

Randomized to AAC (N=737)

Randomized to placebo (N=243)

Took AAC (N=669)

Lost to follow up=36

No HA=32

Took placebo (N=221)

Lost to follow up=15

No HA=7

Intent-to-treat (N=669)

Excluded: 0

Efficacy-evaluable (N=656)

Excluded: 13

HA not migraine=12 Other=1

Randomized to IB (N=734)

Took IB (N=669)

Lost to follow up=38

No HA=27

Intent-to-treat (N=666)

Excluded: 3

--No postbaseline=3

Intent-to-treat (N=220)

Excluded: 1

--No postbaseline=1

Efficacy-evaluable (N=653)

Excluded: 13

HA not migraine=9 Interfering med =1

Other=3

Efficacy-evaluable (N=215)

Excluded: 5

HA not migraine=4 Rescue <1 hour=1

Fig 1.—Summary of patient disposition.

Haenszel test, stratified by investigator, was used to an-

alyze baseline characteristics of the treated headache,

as well as the proportion of patients with nausea, pho-

tophobia, and phonophobia. Statistical significance

was declared when P was less than or equal to .05.

RESULTSPatient Population.—Of the 1714 randomized pa-

tients, 91.0% (1559) took study medication Figure 1.

Of the 1559 patients who took study medication, 4 pa-

tients completed only the baseline assessment in the

diary, leaving 99.7% (1555) of the patients in the pri-

mary efficacy analysis data set.

448 March 2006

Table 1.—Demographics and Migraine History ofIntent-to-Treat Patients

AAC IB Placebo Total(N = 669) (N = 666) (N = 220) (N = 1555)

Mean age (year) 38.3 38.4 38.3 38.3Sex (%)

Male 21.2 18.5 18.6 19.7Female 78.8 81.5 81.4 80.3

Race (%)White 74.3 76.6 73.6 75.2Black 20.2 18.0 20.0 19.2Asian 0.6 0.9 2.7 1.0Hispanic 3.9 4.2 3.6 4.0Other 1.0 0.3 0 0.6

Migraine type (%)Without aura 78.6 78.8 82.7 79.3With aura 21.4 21.2 17.3 20.7

Usual pain without treatment (%)None 0 0 0 0Mild 0 0.2 0 0.1Moderate 20.0 17.7 20.0 19.0Severe 80.0 82.1 80.0 80.9

Usual disability without treatment (%)None 0.1 0.3 0 0.2Mild 1.3 1.5 0.9 1.4Moderate 23.5 22.2 25.5 23.2Severe 41.4 43.2 40.0 42.0Incapacitating 33.6 32.7 33.6 33.2

Usual pharmacologic treatment (%)None 0.3 0.6 0.5 0.5Nonprescription 57.0 55.1 56.4 56.1onlyPrescription only 20.6 21.2 23.2 21.2Both 22.1 23.1 20.0 22.3

The 3 treatment groups had similar demographic

profiles and migraine histories (Table 1). The mean

age was 38.3 years, 80.3% were female, and 75.2%

were white. Untreated, usual migraine pain was mild in

0.1%, moderate in 19.0%, and severe in 80.9%. Usual

disability untreated was mild in 1.4%, moderate in

23.2%, severe in 42.0%, incapacitating in 33.2%, and

none in 0.2%. Migraine was usually treated with ex-

clusively OTC medications in 56.1% of patients, with

prescription medications only in 21.2%, with a combi-

nation of OTC and prescription medications in 22.3%,

and with no medication in 0.5%. Symptom profiles for

the treated headaches in the active and placebo groups

were comparable at baseline (Table 2).

Table 2.—Characteristics of Treated Attack

AAC IB Placebo Total(N = 669) (N = 666) (N = 220) (N = 1555)

Baseline pain (%)Moderate 57.2 56.3 62.3 57.6

Severe 42.8 43.7 37.7 42.4Functional disability (%)

None 0.7 0.9 0 0.7Mild 8.4 6.6 10.0 7.8Moderate 38.0 40.2 35.9 38.6Severe 38.6 38.4 42.3 39.0Incapacitating 14.3 13.8 11.8 13.8

Nausea (%)No 45.7 45.2 48.2 45.9Yes 54.3 54.8 51.8 54.1

Vomiting (%)No 99.0 97.7 99.1 98.5Yes 1.0 2.3 0.9 1.5

Photophobia (%)No 8.2 5.6 7.3 6.9Yes 91.8 94.4 92.7 93.1

Phonophobia (%)No 10.2 9.9 10.0 10.0Yes 89.8 90.1 90.0 90.0

Aura (%)No 80.9 81.2 83.6 81.4Yes 19.1 18.8 16.4 18.6

Unilateral (%)No 32.3 34.4 35.5 33.6Yes 67.7 65.6 64.5 66.4

Pulsating/throbbing (%)No 10.5 10.7 8.2 10.2Yes 89.5 89.3 91.8 89.8

Aggravated by physical activity (%)No 13.5 13.1 12.3 13.1Yes 86.5 86.8 87.7 86.8

Menstruating (%, females only)No 84.1 83.4 81.6 83.4Yes 15.4 16.4 18.4 16.3N/A 0.6 0.2 0 0.3

No statistically significant treatment-by-inves-

tigator interactions were detected for the primary ef-

ficacy variable.

Pain Relief.—AAC and IB achieved significantly

greater TOTPAR scores than placebo at 2 (2.7, 2.4, and

2.0, respectively), 3 (5.1, 4.7, and 3.9, respectively), and

4 (7.8, 7.1, and 5.9, respectively) hours postdose (P <

.006). Only AAC was significantly better than placebo

for TOTPAR at all time points beginning as early as

1 hour postdose (P < .011). In addition, the mean

TOTPAR values for AAC were significantly superior

to IB at 2 hours (P < .03), 3 hours (P < .01), and 4 hours

Headache 449

0

0.5

1

1.5

2

2.5

3

0 15 30 45 60 90 120 180 240

Minutes Post Dose

Me

an

Pa

in R

eli

ef

Sc

ore

AAC Ibuprofen Placebo

* AAC vs. IB p<0.05

† AAC vs. IB p<0.10

*

†

†

*

*

*

Fig 2.—Pain relief.

(P < .007) after patients took study medication (see

PAR profile, Figure 2).

The onset of meaningful PAR was significantly

earlier for both AAC-treated and IB-treated patients

than patients who took placebo (placebo 167.1 min-

utes vs. AAC 128.4 minutes, P < .001, and placebo

167.1 minutes vs. IB 147.9 minutes, P = .049). Fur-

thermore, as shown in Figure 3, the median time to

meaningful PAR for the AAC treatment group was

significantly earlier (approximately 20 minutes) than

for the IB treatment group (P = .036).

AAC was significantly superior to placebo begin-

ning at 45 minutes postdose (P < .018), and it remained

significantly more effective than placebo at all subse-

quent time points (P < .002). IB provided significantly

higher mean PAR scores than placebo beginning at 90

minutes after dosing (P < .003), and it remained sig-

nificantly superior to placebo throughout the remain-

der of the study period (P < .005). AAC-treated pa-

tients had significantly higher mean PAR scores than

IB-treated patients at 45 minutes postdose (P < .022),

as well as at 2, 3, and 4 hours postdose (P < .022, .004,

and .011, respectively).

Pain Intensity Difference From Baseline.—AAC

and IB achieved significantly greater SPID scores than

placebo over 1 (0.4, 0.3, and 0.2, respectively), 2 (1.5,

1.4, and 1.1, respectively), 3 (3.0, 2.7, and 2.2, respec-

tively), and 4 (4.6, 4.2, and 3.3, respectively) hours post-

dose (P < .041). However, the mean SPID values for

AAC were also significantly superior to IB at 2 hours

(P < .045), 3 hours (P < .018), and 4 hours (P < .012)

after patients took study medication (see PID profile,

Figure 4).

AAC-treated patients had significantly higher

mean PID scores than placebo-treated patients be-

ginning at 45 minutes postdose (P < .002); they re-

mained significantly higher than placebo at all subse-

quent time points (P < .001). IB provided significantly

higher mean PID scores than placebo beginning at 60

minutes after dosing (P < .035), and it remained signif-

icantly superior to placebo throughout the remainder

of the study period (P < .007). However, AAC-treated

patients had significantly higher mean PID scores than

IB-treated patients at 45 minutes postdose (P < .049),

as well as at 2, 3, and 4 hours postdose (P < .046, .010,

and .013, respectively).

Effects on Other Headache Characteristics.—Figure 5 shows that the proportion of patients who

became PF was significantly higher for both AAC and

IB than for placebo at 3 and 4 hours postdose (P <

.04). However, the PF rates for AAC exceeded those

for IB at 3 and 4 hours postdose (P < .035). AAC

450 March 2006

0

10

20

30

40

50

60

70

80

0 15 30 45 60 90 120 180 240

Minutes Post Dose

Pro

po

rtio

n o

f S

ub

jec

ts

AAC Ibuprofen Placebo

Median Time (Minutes)

128.4 147.9 167.1

95% C.I. For Median Time

(120, 142) (135, 163) (146, 221)

AAC vs IB

0.036

AAC vs

Placebo 0.001

IB vs

Placebo 0.049

AAC IB Placebo

Fig 3.—Onset of meaningful pain relief.

also had a significantly higher PF rate than placebo at

2 hours postdose (P < .036).

The proportion of patients whose pain intensity

was reduced to mild or none (HR) was significantly

higher for both AAC and IB than for placebo at 3 and

4 hours postdose (P < .04). Unlike IB, however, AAC

HR rates were also significantly greater than placebo

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

0 15 30 45 60 90 120 180 240

Minutes Post Dose

Me

an

PID

AAC Ibuprofen Placebo

* AAC vs. IB p<0.05

† AAC vs. IB p<0.10

*

*

†

†

*

*

Fig 4.—Pain intensity difference.

at 45, 90, and 120 minutes postdose (P < .035). Addi-

tionally, the HR rate for AAC was significantly higher

than IB at 2 hours postdose (AAC 67% vs. IB 62%,

P < .046).

The proportion of patients who were free of

the migraine-associated symptoms (functional disabil-

ity, nausea, photophobia, and phonophobia) were

Headache 451

0

10

20

30

40

50

60

0 15 30 45 60 90 120 180 240

Minutes Post Dose

Cu

mu

lati

ve

Pro

po

rtio

n o

f S

ub

jec

ts

AAC Ibuprofen Placebo

*

* AAC vs. IB p<0.05

*

Fig 5.—Pain free.

generally comparable for AAC and IB, and signifi-

cantly higher for both AAC and IB than for placebo

at most of the time points postdose.

Table 3.—Summary of Adverse Experiences by Body System

AAC% IB% Placebo%Adverse Experience (N = 669) (N = 669) (N = 221)

Body as a Whole 1.3 1.2 0.5Abdominal pain 0.6 0 0Headache 0.3 0.2 0Fever 0 0 0.5Viral infection 0 0.5 0

Cardiovascular 0.3 0 0.9Palpitation 0 0 0.9Tachycardia 0.3 0 0

Digestive System 3.4 0.9 1.8Dyspepsia 0.7 0.3 0.5Nausea 1.8 0.5 0.9Vomiting 0.4 0.2 0.5

Nervous System 5.7 2.2 3.6Confusion 0.1 0 0.5Dizziness 1.6 0.6 1.8Dry mouth 0.9 0.3 0Insomnia 0.3 0 0Nervousness 1.9 0.3 0.5Paresthesia 0.1 0 0.5Somnolence 0.3 0.8 0.5

Respiratory System 0.1 0.5 0Special Senses 0.7 0.5 0Urogenital System 0.1 0.2 0

Rescue Medication.—At 2 hours after treatment,

the proportion of patients who required rescue med-

ication was significantly higher in the IB (P = .025)

and placebo (P < .001) treatment groups than for the

AAC treatment group. While AAC maintained signif-

icant superiority versus placebo at 3 (P = .035) and 4

(P < .001) hours postdose, it did not sustain statistical

significance at 3 or 4 hours postdose compared with

IB.

Safety Results.—No serious AEs were reported,

and the incidence of adverse events was low: 9.7% in

the AAC group, 5.1% in the IB group, and 5.5% in

the placebo group (Table 3). AAC-treated patients re-

ported nervousness (1.9% vs. 0.3%) and nausea (1.6%

vs. 0.4%) more frequently than IB-treated patients, but

the IB treatment group reported somnolence more fre-

quently than AAC treatment group (0.7% vs. 0.3%).

Patients who received placebo had a higher incidence

of dizziness (1.8%) than the AAC (1.6%) and the IB

treatment groups (0.6%).

COMMENTSThis multicenter, double-blind, randomized,

parallel-group, placebo-controlled study shows that,

while both AAC and IB were more effective than

placebo in treating the full spectrum of migraine, AAC

was superior to IB. The superiority was exhibited by

452 March 2006

statistically significant and clinically greater effects

in PAR, pain intensity reduction, earlier onset of

meaningful PAR, as well as in PF and HR rates,

which are among the efficacy variables that are most

important to patients.19-21 In addition, significantly

more IB patients than AAC patients required rescue

medication, and significantly more AAC-treated

patients reported complete PAR.

In this study program, we used several recruit-

ment methods, including random-digit dialing, local

advertising, and clinical referrals to identify a broad

range of the migraine population, including individuals

who did and did not seek medical care. Unlike earlier

migraine research with OTC medications,11-17 this trial

did not exclude patients with the most severe migraine

symptoms (eg, those whose attacks usually cause them

to vomit or require bed rest). By demonstrating a ther-

apeutic response in an unrestricted migraine popula-

tion, both IB and AAC have shown that there is an

appropriate role for OTC medications in the majority

of patients with migraine. The epidemiologic profile

and patterns of medication in this study are similar to

results in population studies, suggesting that study pa-

tients in the United States are broadly representative

of those with migraine.2,3

A limitation of this trial is the lack of collected

data on headache recurrence or 24-hour PF rates.

In addition, there was a high placebo response rate

in this trial. Although placebo response in headache

research is traditionally high, ranging from 30% to

40%,22 methodological issues may have further ele-

vated levels in this study. For instance, several investi-

gators have shown that high randomization ratios (ie,

a high likelihood of getting active drug rather than

placebo) can raise placebo response rates,23 as pa-

tients are more likely to believe they are receiving ac-

tive medication. Future work may clarify issues with

placebo in studies of headache patients.

The per protocol primary efficacy variable was

TOTPAR2. The choice of primary endpoints in acute

migraine treatment trials is controversial; the value of

the most widely used endpoint, 2-hour HR, has been

widely and legitimately questioned.19,21,24 Although

HR at 2 hours postdose remains a frequently used

outcome measure, PAR and TOTPAR are the widely

used endpoints in analgesic trials.24 Our study indi-

cates that TOTPAR, which detected differences be-

tween the active treatment groups that HR failed to

identify, may represent a valid and sensitive endpoint

for detecting differences in pain response, particularly

when 2 active drugs are compared in future migraine

trials.

These findings have important implications for

medical professionals and their patients. For the first

time, rigorous data show that OTC medications are

safe and effective in an unrestricted migraine patient

population, and that there are significant differences

between the 2 approved OTC migraine therapies. Al-

though additional comparative trials of OTC agents

for migraine are recommended, these results show

that both AAC and IB are safe, cost-effective treat-

ments for a representative sample of people with mi-

graine and that AAC provides significantly superior

efficacy and speed of action compared with IB tablets.

This study verifies the evidence-based evaluations of

the U.S. Headache Consortium and the German Mi-

graine and Headache Society, who in their most recent

evidence-based therapeutic recommendations for the

treatment of migraine and TTH considered the combi-

nation of ASA, APAP, and caffeine a reasonable first-

line treatment for mild-to-moderate migraine attacks

and TTH.25,26

REFERENCES

1. Headache Classification Committee of the Interna-

tional Headache Society. Classification of the diag-

nostic criteria for headache disorders, cranial neural-

gias, and facial pain. Cephalalgia. 1988;8(suppl 7):1-

96.

2. Lipton RB, Stewart WF, Diamond S, Diamond ML,

Reed M. Prevalence and burden of migraine in the

United States: Data from the American Migraine

Study II. Headache. 2001;41(7):646-657.

3. Lipton RB, Diamond S, Reed M, Diamond ML, Stew-

art WF. Migraine diagnosis and treatment: Results

from the American Migraine Study II. Headache.2001;41(7):638-645.

4. Dahlof C, Linde M. One-year prevalence of mi-

graine in Sweden: A population-based study in adults.

Cephalalgia. 2001;21(6):664-671.

5. Lipton RB, Stewart WS, von Korff M. Burden of mi-

graine: Societal costs and therapeutic opportunities.

Neurology. 1997;48(suppl 3):S4-S9.

Headache 453

6. Lipton RB, Stewart WF. Migraine in the United

States: A review of epidemiology and health care use.

Neurology. 1993;43(suppl 3):S6-S10.

7. Lipton RB, Stewart WF, Celentano DD, Reed M.

Undiagnosed migraine headaches: A comparison of

symptom-based and reported physician diagnosis.

Arch Intern Med. 1992;152:1273-1278.

8. Sheftell FD. Role and impact of over-the-counter

medications in the management of headache. NeurolClin. 1997;15:187-198.

9. Lipton RB, Newman LC, Solomon S. Over-the

counter medication and the treatment of migraine.

Headache. 1994;34:547-548.

10. Gilkey SJ, Ramadan NM. Use of over-the-counter

drugs in migraine. CNS Drugs. 1996;6:83-88.

11. Lipton RB, Stewart WF, Ryan RE Jr, Saper J,

Silberstein S, Sheftell F. Efficacy and safety of ac-

etaminophen, aspirin, and caffeine in alleviating mi-

graine headache pain. Three double-blind, random-

ized, placebo-controlled trials. Arch Neurol. 1998;55:

210-217.

12. Goldstein J, Hoffman HD, Armellino JJ, Battikha

JP, Hamelsky SW, Couch J, et al. Treatment of

severe, disabling migraine attacks in an over-the-

counter population of migraine sufferers: Results

from three randomized, placebo-controlled studies of

the combination of acetaminophen, aspirin, and caf-

feine. Cephalalgia. 1999;19:684-691.

13. Silberstein S, Stewart WF, Lipton RB. Treatment of

menstruation-associated migraine with the nonpre-

scription combination of acetaminophen, aspirin, and

caffeine: Results from three randomized, placebo-

controlled studies. Clin Ther. 1999;21(3):475-491.

14. Goldstein J, Silberstein SD, Saper JR, Elkind AH,

Smith TR, Gallagher RM, et al. Acetaminophen, as-

pirin, and caffeine versus sumatriptan succinate in the

early treatment of migraine: Results from the ASSET

trial. Headache. 2005;45(8):973-982.

15. Codispoti JR, Prior MJ, Fu M, Harte CM, Nelson

EB. Efficacy of nonprescription doses of ibuprofen

for treating migraine headache: A randomized con-

trolled trial. Headache. 2001;41:665-679.

16. Kellstein DE, Lipton RB, Geetha R, Koronkiewicz K,

Evans FT, Stewart WF, et al. Evaluation of a novel

solubilized formulation of ibuprofen in the treatment

of migraine headache: A randomized, double-blind,

placebo-controlled, dose-ranging study. Cephalalgia.2000;20:233-243.

17. Lipton RB, Baggish JS, Stewart WF, Codispoti JR, Fu

M. Efficacy and safety of acetaminophen in the treat-

ment of migraine: Results of a randomized, double-

blind, placebo-controlled, population-based study.

Arch Int Med. 2000;160(22):3486-3492.

18. Diener HC, Pfaffenrath V, Pageler L, Peil H, Aicher

B. The fixed combination of acetylsalicylic acid,

paracetamol and caffeine is more effective than single

substances and dual combination for the treatment of

headache: A multicenter, randomized, double-blind,

single dose, placebo-controlled parallel group study.

Cephalalgia. 2005;25:776-787.

19. Silberstein SD. Migraine symptoms: Results of

a survey of self-reported migraineurs. Headache.1995;35:387-396.

20. Dahlof C. How to assess patient preference of mi-

graine treatments. Cephalalgia. 1999;19(suppl 24):2-

5.

21. Lipton RB, Stewart WF. Acute migraine therapy: Do

doctors understand what patients with migraine want

from therapy? Headache. 1999;39(suppl 2):S20-S26.

22. Goldstein J. Migraine therapy: Putting clinical trials

to the test. Headache Q. 1999;10(suppl 1):15-17.

23. Max MB. The design of clinical trials for trea-

tment of pain. Available at: http://painconsor-

tium.nih.gov/symptomresearch/chapter 1/sec2/cmm-

s2pg1.htm. Accessed on October 14, 2005.

24. Marcus DA. Establishing a standard of speed for as-

sessing the efficacy of the serotonin 1B/1D agonists

(triptans). Arch Neurol. 2001;58:1056-1058.

25. Silberstein SD, for the U.S. Headache Consor-

tium. Practice parameter: Evidence-based guidelines

for migraine headache (an evidence-based review).

Report of the quality standards subcommittee of

the American Academy of Neurology. Neurology.2000;55:754-763.

26. Haag G, Evers S, May A, et al. Selbstmedika-

tion bei Migrane und Kopfschmerz vom Span-

nungstyp – Evidenz-basierte Empfehlungen der

Deutschen Migrane und Kopfschmerzgesellschaft

(Self-medication in migraine and tension-type

headache—Evidence-based treatment recommenda-

tions of the German Migraine and Headache Soci-

ety). Nervenheilkunde. 2004;23:415-430.