Content uploaded by Richard Lipton
Author content
All content in this area was uploaded by Richard Lipton on Dec 18, 2017
Content may be subject to copyright.
2000;54;308- Neurology
N. Breslau, L. R. Schultz, W. F. Stewart, R. B. Lipton, V. C. Lucia and K. M. A. Welch
Headache and major depression: Is the association specific to migraine?
This information is current as of June 5, 2006
http://www.neurology.org/cgi/content/full/54/2/308
located on the World Wide Web at:
The online version of this article, along with updated information and services, is
Print ISSN: 0028-3878. Online ISSN: 1526-632X.
published continuously since 1951. Copyright © 2000 by AAN Enterprises, Inc. All rights reserved.
Neurology is the official journal of AAN Enterprises, Inc. A bi-monthly publication, it has been
at ALBERT EINSTEIN COLL MED on June 5, 2006 www.neurology.orgDownloaded from
Headache and major depression
Is the association specific to migraine?
N. Breslau, PhD; L.R. Schultz, PhD; W.F. Stewart, PhD, MPH; R.B. Lipton, MD; V.C. Lucia, MA;
and K.M.A. Welch, MD
Article abstract—Objective: To examine the relationship between migraine and major depression, by estimating the risk
for first-onset major depression associated with prior migraine and the risk for first migraine associated with prior major
depression. We also examined the extent to which comorbidity with major depression is specific to migraine or is observed
in other severe headaches. Methods: Representative samples of persons 25 to 55 years of age with migraine or other severe
headaches (i.e., disabling headaches without migraine features) and controls with no history of severe headaches were
identified by a telephone survey and later interviewed in person to ascertain history of common psychiatric disorders.
Results: Lifetime prevalence of major depression was approximately three times higher in persons with migraine and in
persons with severe headaches compared with controls. Significant bidirectional relationships were observed between major
edepression and migraine, with migraine predicting first-onset depression and depression predicting first-onset migraine.
In contrast, persons with severe headaches had a higher incidence of first-onset major depression (hazard ratio 53.6), but
major depression did not predict a significantly increased incidence of other severe headaches (hazard ratio 51.6).
Conclusions: The contrasting results regarding the relationship of major depression with migraine versus other severe
headaches suggest that different causes may underlie the co-occurrence of major depression in persons with migraine
compared with persons with other severe headaches. Key words: Migraine—Comorbidity—Major depression—
Epidemiology.
NEUROLOGY 2000;54:308–313
Epidemiologic studies report an association between
migraine headache and major depression.
1-8
The as-
sociation might be noncausal, reflecting shared
genetic or environmental causes. Alternatively, mi-
graine might cause major depression or might be
caused by it. The hypothesis that depression in per-
sons with migraine reflects a psychological response
to the stress of recurrent severe headaches would
predict an influence only from migraine to depres-
sion but not from depression to migraine. In con-
trast, the hypothesis of shared causes would predict
that each disorder would increase the risk of the
first-time occurrence of the other. We previously re-
ported, based on prospective data, bidirectional influ-
ences between migraine and major depression, with
each disorder increasing the risk of first onset of the
other.
5
These findings dampen the plausibility that
the migraine–depression association is caused by the
demoralizing experience of recurrent and disabling
headaches and suggest the shared-causes hypothesis.
Depression has been linked to other pain
conditions.
9-14
Reciprocal psychological effects of
pain and depressive symptoms have been postu-
lated: pain precipitates worry and pessimism, and
distress impairs the ability to cope with pain.
15,16
These findings call into question the specificity of
the migraine–major depression association. Is non-
migraine severe headache also associated with ma-
jor depression? Are there bidirectional influences in
relation to nonmigraine severe headache? A previ-
ous epidemiologic survey reported a higher preva-
lence of major depression in persons with migraine
headache but not tension-type headache.
7
However,
as the authors noted, the difference in the rates of
comorbid major depression between these headache
types might be explained by differences in headache
severity.
In this study, we address the migraine–major de-
pression comorbidity in a large-scale epidemiologic
study, the Detroit Area Study of Headache. We first
estimate the lifetime prevalence of major depression
in persons with migraine, persons with other severe
headaches of comparable pain intensity and disabil-
ity, and persons with no history of severe headaches.
We then examine the relationship between migraine
and major depression, by estimating the risk for the
onset of major depression in persons with prior mi-
graine and the risk for the onset of migraine in per-
sons with prior major depression. We compare these
results with the results from a corresponding analysis
of the relationship between major depression and se-
vere headaches other than migraine.
From the Departments of Psychiatry (Dr. Breslau and V.C. Lucia) and Biostatistics and Research Epidemiology (Drs. Breslau and Shultz), Henry Ford
Health System, Detroit, MI; the Department of Psychiatry (Dr. Breslau), Case Western Reserve University School of Medicine, Cleveland, OH; the
Department of Psychiatry (Dr. Breslau), University of Michigan School of Medicine, Ann Arbor, MI; Innovative Medical Research (Drs. Stewart and Lipton),
Baltimore, MD; and the University of Kansas School of Medicine (Dr. Welch), Kansas, MS.
Supported by NIH Headache Research Center (Bethesda, MD) grant P50 NS32399.
Received May 24, 1999. Accepted in final form August 27, 1999.
Address correspondence and reprint requests to Dr. Naomi Breslau, Director of Research, Department of Psychiatry, Henry Ford Health System, One Ford
Place, 3A, Detroit, MI 48202-3450; e-mail: nbresla1@hfhs.org
308 Copyright © 2000 by the American Academy of Neurology at ALBERT EINSTEIN COLL MED on June 5, 2006 www.neurology.orgDownloaded from
Methods. Sample and data. The Detroit Area Study of
Headache was designed to identify representative samples
of the general population with migraine or other severe
headaches and controls with no history of headache, and to
ascertain the prevalence of major depression in each
group. The study was conducted in two phases. In the first
phase, a large-scale random-digit dialing
17
telephone sur-
vey was conducted to identify the three population groups.
In the second phase, face-to-face interviews were conducted
with subsets of persons identified in the telephone survey
(i.e., those with migraine, those with severe headaches, and
controls) to assess common psychiatric disorders.
Telephone survey. The target population was the tri-
county area of Wayne, Macomb, and Oakland, Michigan,
which contained 3.9 million of the total 4.3 million resi-
dents of the Detroit, Michigan, primary metropolitan
statistical area at the time of the 1990 census.
18
The
random-digit dialing methods are appropriate in this
geographic area, because virtually all households
(.97%) had a telephone according to the 1990 census.
18
Persons 25 to 55 years old were targeted, based on pre-
vious findings indicating that persons younger than 25
or older than 55 years of age have considerably lower
rates of migraine.
19
Screening was completed in 82.5% of
households, and the cooperation rate in eligible house-
holds was 70%. Data were gathered on 4,765 persons 25
to 55 years of age.
A computer-assisted telephone interview was used. Re-
spondents were asked if they had experienced at least one
headache in the last year not caused by a head injury,
hangover, or illness. The number of different types of head-
aches each respondent experienced in the last year was
ascertained. Detailed questions about the International
Headache Society (IHS)
20
defining features of migraine,
age at onset, and associated impairment were first asked
about the most severe type of headache experienced in the
last year. If more than one type of headache was reported,
the same series of questions was asked about the second
most severe headache type.
21
The validity of the diagnosis for migraine by computer-
assisted telephone interview has been assessed in
population-based studies in the United States and the
United Kingdom. In each country, a population sample of
persons who met telephone-screening criteria for migraine
and a sample that did not were independently evaluated
by clinicians, using a semistructured interview for head-
ache diagnosis that has been extensively used in clinical
trials. Clinicians were blind to the telephone-screening di-
agnosis when applying IHS criteria. In both studies, the
positive predictive value of the telephone interview for the
diagnosis of migraine was approximately 94%.
22,23
The household interview: ascertaining psychiatric his-
tory. Of the 4,765 persons interviewed in the random-
digit dialing telephone survey, 1,696 were selected for psy-
chiatric assessment via face-to-face interviews. The sample
included all those who met IHS criteria for migraine in the
survey (n 5683), all those who experienced headaches
other than migraine and also met study criteria for severe
headache (described subsequently) (n 5253), and a subset
of the remaining subjects serving as controls (n 5760).
Control subjects were frequency-matched to the migraine
subset by sex, age (within 5 years), and race. (Eligibility
criteria for controls are described subsequently.) A total of
1,287 subjects (75.9%) completed face-to-face interviews.
Participation varied by age, race, and level of education; it
was higher in older persons, in white subjects, and in per-
sons with more education. Participation also varied across
the three diagnostic groups, with the highest rate of partic-
ipation reached in the migraine group and the lowest in
the severe headache group. However, these differences in
participation were small. Participation rate did not vary
between the sexes.
Household interviews of headache patients were con-
ducted between July 1996 and January 1998, usually
within 2 to 4 weeks of the telephone interview. Interviews
of controls began in October 1996. Trained lay interview-
ers conducted the household interviews, using a computer-
assisted personal interview. Assessment of psychiatric
disorders was by the Composit International Diagnostic
Interview, version 2.1,
24
a fully structured interview sched-
ule that generates psychiatric disorders according to the
Diagnostic and Statistical Manual of Mental Disorders,
4th ed (DSM-IV).
25
Studies of this interview in general-
population samples showed high reliability for most diag-
noses, including major depression.
26
Definition of the three subject groups and key variables.
The classification of respondents into the three diagnostic
groups—migraine, severe headache, and controls—was
based on information about the two worst headache types
experienced in the last year identified in the telephone
interview. Migraine was defined according to the IHS cri-
teria for migraine with and without aura.
20
The two mi-
graine types were mutually exclusive, and migraine
without aura was diagnosed only in persons with no mi-
graine with aura. The definition of severe headache re-
quired headache duration of 4 to 72 hours (i.e., the same
duration criterion used in the IHS to define migraine); no
history of nausea, vomiting, photophobia, or phonophobia
associated with a headache; and a specified minimum
headache impact score on the Headache Impact Question-
naire (HImQ), described subsequently. Of the pool of re-
spondents with nonmigraine headaches, only 11.6% met
these eligibility criteria.
Controls could not meet criteria for either migraine or
severe headache; could not report a history of nausea, vom-
iting, photophobia, or phonophobia associated with a head-
ache; and could not have ever experienced headaches with
a pain severity score greater than 2 on a scale from 0 to 10
(i.e., mild).
The exclusionary rule with respect to nausea, vomiting,
photophobia, and phonophobia used in the definition of the
severe headache and control groups was intended to ex-
clude persons who might fit the IHS category of “migrain-
ous disorder not fulfilling above [migraine] criteria.”
Although the diagnostic definition of this category states
that it applies if “all but one” of the migraine criteria are
met, a high index of suspicion for migraine is raised by the
presence of nausea and vomiting or photophobia and pho-
nophobia, which are the prototypic symptoms in migraine.
This rule ensured that migraine sufferers were excluded
from the severe headache and control groups.
The HImQ was used to define headache severity.
22,23
The HImQ is the sum of two component measures: average
pain intensity (10-point scale) and total lost time in each
January (2 of 2) 2000 NEUROLOGY 54 309
at ALBERT EINSTEIN COLL MED on June 5, 2006 www.neurology.orgDownloaded from
of three domains of activity, expressed as lost days. Lost
time was derived as the sum of actual missed days in work
for pay or school, in household work, and in nonwork ac-
tivities (i.e., family, social, or recreation) plus reduced-
effectiveness day equivalents. This is the average
percentage of reduced effectiveness resulting from a head-
ache multiplied by the number of headaches (adjusted for
missed days in each domain). For the severe headache
group, a minimum HImQ of 38.05 was required, to approx-
imate the HImQ distribution of this group to the HImQ
distribution of the migraine group.
Major depression is defined in the DSM-IV by the pres-
ence of at least five from a list of nine symptoms occurring
within the same 2-week period, with at least one of the
symptoms pertaining to a depressed mood or loss of inter-
est or pleasure. The symptom groups are dysphoria, loss of
interest or pleasure, weight loss or weight gain, insomnia
or hypersomnia, psychomotor agitation or retardation, fa-
tigue, feelings of worthlessness or excessive guilt, concen-
tration problems, and suicidal ideations or attempts.
Statistical analysis. To estimate the risk for the first
occurrence of migraine associated with prior major depres-
sion and the risk for depression associated with prior mi-
graine, we used Cox proportional hazards models with
time-dependent covariates.
27-29
The hazard ratio calculated
in this model estimates the risk for first occurrence of the
outcome (e.g., migraine) in persons with the risk factor of
interest (e.g., major depression) compared with persons
without the risk factor. Time was defined as chronologic
age. Persons who had not experienced the outcome by the
time of the interview were censored. Similar Cox propor-
tional hazards models were used to estimate the risk for
the first occurrence of severe headache associated with
major depression and vice versa. Analyses that focus on
the migraine–major depression relationship excluded per-
sons in the severe headache group. Analyses that focused
on the severe headache–major depression relationship ex-
cluded the migraine group. Cases with onset of headache
and major depression in the same year were censored just
before the year in which these events occurred, because
information on chronologic order was not available.
Cox proportional hazards models, unlike logistic regres-
sion, take into account differences across individuals in the
period of risk for the outcome. The added advantage of
models with time-dependent covariates is that they allow
consideration of independent variables that change over
time. For example, a person’s status with respect to major
depression can change from no major depression to major
depression at any time until the age at first migraine at-
tack or censoring. The parameter estimate (b) in the Cox
proportional hazards model is a regression coefficient from
which the hazard ratio can be obtained by taking the anti-
logarithm of the parameter estimate.
To estimate the risk for first-onset major depression
associated with prior migraine with aura and migraine
without aura, we used Cox proportional hazards models
with two time-dependent covariates. To estimate the risks
for the first occurrence of migraine with aura and migraine
without aura associated with prior major depression, we
used competing risks analysis, in which the occurrence of
one type of outcome removes (censors) the subject from
risk for the other outcome.
30
x
2
analysis was used to test
the difference between the two coefficients. The SAS (Cary,
NC) procedure for proportional hazards regression was
used.
31
Results. Description of subsamples. On sex, race, and
age—the three variables on which controls were matched
to the migraine group—small differences were detected
between the two groups (table 1). Controls had a higher
proportion of college graduates than the two headache
groups. The small differences between the migraine group
and the controls, despite the matching scheme, result from
differences in participation rates in the household inter-
view. Of the total migraine group, 29.5% met criteria for
migraine with aura. The severe headache group comprised
persons with tension-type headache (58.8%), persons meet-
ing IHS criteria for migrainous disorder (i.e., meeting all
criteria except for nausea or vomiting and photophobia
plus phonophobia; 33%), and other undiagnosed headaches
(11.2%).
Lifetime association of major depression with migraine
and severe headache. Both the migraine and the severe
headache groups had markedly higher lifetime prevalences
of major depression than the controls (table 2). The sex-
adjusted odds ratios of major depression associated with
each of the headache types were greater than 3 and were
not significantly different from one another (Wald x
2
5
Table 1 Distribution of sex, race, age, and education in the three
diagnostic groups
Characteristic
Migraine,
n5536
Severe headache,*
n5162
Controls,
n5586
Female sex, % 83.0 64.8 76.8
Race, %
White 75.9 72.2 69.5
Black 19.4 23.5 24.7
Other 4.7 4.3 5.8
Age, y, mean (SD) 40.3 (8.3) 40.0 (8.1) 40.7 (8.9)
Education, %
,High school 5.8 5.6 5.7
High school 33.4 26.5 33.1
Some college 34.0 40.7 26.0
$College 26.8 27.2 35.2
* Other than migraine as defined by study criteria.
Table 2 Lifetime prevalence of major depression in migraine,
severe headache, and controls, and sex-adjusted odds ratios (OR)
(n 51,284)
Group
Major
depression, %
Sex-adjusted OR*
(95% CI)
Migraine, n 5536 40.7 3.51 (2.64, 4.64)
With aura, n 5158 49.4 4.90 (3.34, 7.19)
Without aura, n 5378 37.0 3.03 (2.23, 4.11)
Severe headache,† n 5162 35.8 3.18 (2.14, 4.73)
Controls, n 5586 16.0 1.00
* From multivariable logistic regression.
† Other than migraine as defined by study criteria.
310 NEUROLOGY 54 January (2 of 2) 2000 at ALBERT EINSTEIN COLL MED on June 5, 2006 www.neurology.orgDownloaded from
0.26, 1 df,p50.61). We tested the interaction between sex
and headache type with respect to the odds for major de-
pression, but detected no evidence of a significant interac-
tion (Wald x
2
53.20, 2 df,p50.20). A logistic regression
model that also included education as a covariate did not
alter the estimates. Within the severe headache group, the
lifetime prevalence of major depression was higher in per-
sons with migrainous disorders not fulfilling migraine cri-
teria than in persons with tension-type headache, 42.1%
versus 30.1%, respectively. We also found that in the ma-
jority of comorbid cases, the onset of headache preceded
the onset of major depression.
The lifetime prevalence of major depression was higher
in persons with migraine with aura than in persons with
migraine without aura (see table 2). The difference was
statistically significant (Wald x
2
56.198, 1 df,p50.013).
Migraine and major depression: examining the influence of
each condition on the first occurrence of the other. Previous
migraine and first onset of major depression. Using a Cox
proportional hazards model with migraine as a time-
dependent covariate, we estimated the hazard ratio of the
first onset of major depression associated with previous
migraine, controlling for sex. Persons with severe head-
ache were excluded. The sex-adjusted hazard ratio of the
first onset of major depression in persons with previous
migraine was 2.35 (95% CI 1.84 to 3.01). In a separate
model we tested the interaction between prior migraine
and sex but found no evidence for it (Wald x
2
50.455, 1 df,
p50.49). Including subjects’ education in the model did
not alter the results.
Previous major depression and first occurrence of mi-
graine. Estimated in a Cox proportional hazards model
with major depression as a time-dependent covariate and
controlling for sex, the hazard ratio for the first occurrence
of migraine in persons with prior major depression was
2.75 (95% CI 2.17 to 3.48). No significant sex interaction
was detected (Wald x
2
51.60, p50.20). Subjects’ educa-
tion had a significant inverse relationship with migraine
(Wald x
2
55.61, 1 df,p,0.02). However, the inclusion of
education in the model did not alter the hazard ratio of
migraine associated with prior major depression.
Migraine with aura and migraine without aura. The
sex-adjusted hazard ratio of the first onset of major depres-
sion associated with prior migraine with aura was 2.81
(95% CI 1.95 to 4.04), and with prior migraine without
aura, 2.18 (95% CI 1.65 to 2.88). The two estimates are not
significantly different (x
2
51.55, 1 df,p50.21). Estimat-
ing the influence in the reverse direction, from major de-
pression to each of the two migraine subtypes, we found
that the sex-adjusted hazard ratio of the first occurrence of
migraine with aura associated with prior major depression
was 3.98 (95% CI 2.70 to 5.86) and of the first migraine
without aura, 2.22 (95% CI 1.65 to 2.98), using competing
risks analysis
30
(x
2
for the difference between the two esti-
mates 55.56, 1 df,p50.018).
Severe headache and major depression: are there influ-
ences in both directions? Prior severe headache and first
onset of major depression. Using Cox proportional haz-
ards model with severe headache as a time-dependent co-
variate and controlling for sex, we estimated the hazard
ratio of first onset of major depression in persons with
prior severe headache relative to persons with no prior
severe headache. Persons with migraine were excluded.
The adjusted hazard ratio of first onset of major depression
associated with prior severe headache was 3.56 (95% CI
2.38 to 5.32). No evidence of sex interaction was detected
(Wald x
2
50.52, 1 df,p50.47). Education had no associ-
ation with major depression and its inclusion in the model
did not alter the results.
Prior major depression and first occurrence of severe
headache. Estimating the relationship in the reverse di-
rection, from prior major depression to first severe head-
ache, and applying the same statistical model, we found
that the sex-adjusted hazard ratio was 1.63 (95% CI 0.94
to 2.83; Wald x
2
53.08, 1 df,p50.08). No significant sex
interaction was detected (Wald x
2
51.10, 1 df,p50.29).
Education had an inverse relationship with severe head-
ache that did not reach statistical significance (Wald x
2
5
3.62, 1 df,p50.06). Controlling for education did not alter
the results.
A summary of the results of these analyses appears in
table 3.
Discussion. The key findings of this study are as
follows:
1. The lifetime prevalence of major depression was
threefold higher in persons with migraine and in
persons with severe headache, compared with
persons with no history of severe headache.
2. The lifetime prevalence of major depression in
persons with migraine with aura was significantly
higher than in persons with migraine without
aura.
3. A bidirectional relationship between migraine and
depression was observed: migraine signaled an in-
creased risk for the first onset of major depres-
sion, and major depression signaled an increased
risk for the first time occurrence of migraine.
4. In contrast, severe headache signaled an in-
creased risk for major depression, but there was
no evidence of a significant influence in the re-
verse direction, from major depression to severe
headache.
Table 3 Estimates of the risk of major depression associated with headache and vice versa by headache type
Association All migraine With aura Without aura Severe headache*
Headache to major depression 2.35 (1.84, 3.01) 2.81 (1.95, 4.04) 2.18 (1.65, 2.88) 3.56 (2.38, 5.32)
Major depression to headache 2.75 (2.17, 3.48) 3.98 (2.70, 5.86) 2.22 (1.65, 2.98) 1.63 (.94, 2.83)
Values are sex-adjusted hazard ratios (95% CI). Sex-adjusted hazard ratios estimated in Cox proportional hazards models with time-
dependent covariates.
* Other than migraine as defined by study criteria.
January (2 of 2) 2000 NEUROLOGY 54 311
at ALBERT EINSTEIN COLL MED on June 5, 2006 www.neurology.orgDownloaded from
The findings confirm other reports on the life-
time association between migraine and major
depression.
1-4,6-8
They also replicate our previous
findings based on prospective data that showed sig-
nificant influences from migraine to the first onset of
major depression and from major depression to the
first occurrence of migraine.
5
They suggest that the
migraine–major depression association is unlikely to
be a psychological reaction to the demoralizing expe-
rience of recurrent headaches and suggest shared
causes, an explanation consistent with evidence that
similar neurochemical abnormalities might be impli-
cated in both conditions.
32
Our findings also confirm previous reports con-
cerning the status of migraine relative to other head-
ache types, with respect to headache severity and
associated impairment. They show that, although
only 17.3% of persons with nonmigraine headaches
scored above the impairment threshold used in this
study to define severe headaches, 70.5% of those
with migraine did so.
The evidence on the relationship between major
depression and severe headache, a category of non-
migraine headaches scoring above a specified level of
severity and disability, cannot be readily compared
with findings from other studies. The severe head-
ache category includes tension-type headache and
other nonmigraine headaches of considerable sever-
ity. Previous studies reported a higher prevalence of
major depression and higher scores on depression
scales in persons with undifferentiated recurrent
headaches compared with persons with no head-
aches.
9,15
Merikangas et al.
7
reported that the preva-
lence of major depression was not elevated in
persons with tension-type headaches, compared with
persons with no history of headaches. However, in
that study, severity of headache was not taken into
account, and persons with tension-type headache
whose headaches were milder than the typical mi-
graine headache were not excluded from the compar-
ison. Based on the large sample of our telephone
survey, most persons who met the IHS criteria for
tension-type headache (89%) scored below the 30th
percentile of the distribution of impairment scores of
the migraine group.
In contrast with the bidirectional association of
migraine and major depression, the association of
severe headache and major depression was found to
flow primarily in one direction, from severe headache
to major depression. This disparate pattern of re-
sults suggests the possibility that different causes
might account for the comorbidity of major depres-
sion in these two headache categories. The results
for migraine suggest shared causes, whereas those
for severe headache suggest a causal effect of head-
aches on major depression. To the extent that the
inclusion of persons with “partial” migraine in
the severe headache group might have influenced
the results, it could have only attenuated the differ-
ences between the two observed patterns.
The results suggest that migraine with aura
might have a stronger relationship with major de-
pression than migraine without aura: the lifetime
prevalence of major depression was higher in per-
sons with migraine with aura than without aura,
and major depression increased the risk for first oc-
currence of migraine with aura significantly more
than for migraine without aura. This finding is con-
sistent with previous reports.
3,7
The higher risk con-
ferred by major depression for migraine with aura
versus migraine without aura might account for the
observed difference in the lifetime association. The
scientific implications and clinical significance of
these differences are uncertain and require further
investigation in prospective studies.
Several caveats are in order. First, although we
applied a statistical approach that takes into account
the temporal order between the onset of major de-
pression and the onset of headaches, the data are
based on retrospective accounts of age at onset and
are subject to recall errors. Clearly, prospective data
would provide more accurate information on the inci-
dence of major depression in persons with migraine
or severe headache, and vice versa. Nonetheless,
there is no reason to suspect that recall errors could
have biased the results. Furthermore, information
on history of migraine and other headaches was
gathered in the telephone interview, whereas infor-
mation on history of major depression was gathered
in the follow-up household interview. Respondents
were unaware of our interest in the relationship of
their headache history with major depression or the
chronologic sequence between various disorders. Sec-
ond, the subset of severe headache was relatively
small, yielding less reliable and precise estimates
than those of the migraine subset. However, we did
not merely rely on tests of statistical significance but
compared the actual estimates of the hazard ratios.
These comparisons showed plainly the asymmetry in
the severe headache–major depression relationship,
in contrast with the symmetry in the migraine–
major depression relationship. Third, the results on
the association between nonmigraine headache and
major depression are representative only of a small
fraction of the nonmigraine headache distribution—
that is, its severe tail end. The absence of a represen-
tative subset of most nonmigraine headaches—that
is, those that are less severe and less impairing—
limits our ability to test whether our findings on the
comorbidity with major depression characterizes the
entire nonmigraine headache category or is confined
to the severe tail end.
Important strengths also deserve mention. The
sample is representative of the general population
and thus is less prone to the bias that characterizes
studies of clinical samples. The grave limitations im-
posed by the bias in clinical samples for the under-
standing of psychiatric (and other) comorbidity
associated with headache has been emphasized.
33
Also, the few general population studies that have
focused on psychiatric comorbidity in migraine were
restricted to young adults. This study includes a
312 NEUROLOGY 54 January (2 of 2) 2000 at ALBERT EINSTEIN COLL MED on June 5, 2006 www.neurology.orgDownloaded from
broad age range, up to 55 years. We found no modi-
fying effects of age on the migraine–major depression
association, suggesting that the association applies
to a broad age range.
References
1. Merikangas KR, Risch NJ, Merikangas JR, Weissman MM,
Kidd KK. Migraine and depression: association and familial
transmission. J Psychiatr Res 1988;22:119–129.
2. Merikangas KR, Angst J, Isler H. Migraine and psychopathol-
ogy: results of the Zurich Cohort Study of young adults. Arch
Gen Psychiatry 1990;47:849–853.
3. Breslau N, Davis GC, Andreski P. Migraine, psychiatric disor-
ders and suicide attempts: an epidemiologic study in young
adults. Psychiatr Res 1991;37:11–23.
4. Breslau N, Davis GC. Migraine, physical health and psychiat-
ric disorder: a prospective epidemiologic study of young
adults. J Psychiatr Res 1993;27:211–221.
5. Breslau N, Davis GC, Schultz LR, Peterson EL. Migraine and
major depression: a longitudinal study. Headache 1994;34:
387–393.
6. Merikangas KK, Merikangas JR, Angst J. Headache syn-
dromes and psychiatric disorders: association and familial
transmission. J Psychiatr Res 1993;27:197–210.
7. Merikangas KR, Stevens DE, Angst J. Psychopathology and
headache syndromes in the community. Headache 1994;34:
S17–S26.
8. Breslau N, Merikangas K, Bowden CL. Comorbidity of mi-
graine and major affective disorders. Neurology 1994;44(suppl
7):S17–S22.
9. Von Korff M, Le Resche L, Dworkin SF. First onset of common
pain symptoms: a prospective study of depression as a risk
factor. Pain 1993;55:251–258.
10. Dworkin SF, Von Korff M, Le Resche L. Multiple pains and
psychiatric disturbance: an epidemiologic investigation. Arch
Gen Psychiatry 1990;47:239–244.
11. Fields H. Depression and pain: a neurobiological model. Neu-
ropsychiatry, Neuropsychol, Behav Neurol 1991;4:83–92.
12. Katon W, Kleinman A, Rosen G. Depression and somatization:
a review. Part I. Am J Med 1982;72:127–135.
13. Katon W, Kleinman A, Rosen G. Depression and somatization:
a review. Part II. Am J Med 1982;72:241–247.
14. Romano JM, Turner JA. Chronic pain and depression: does
the evidence support a relationship? Psychol Bull 1985;97:18–
34.
15. Von Korff M, Simon G. The relationship between pain and
depression. Br J Psychiatry 1996;30(suppl):101–108.
16. Ormel J, Von Korff M, Van den Brink W, Katon W, Brilman
E, Oldehinkel T. Depression, anxiety, and social disability
show synchrony of change in primary care patients. Am J Pub
Health 1993;83:385–390.
17. Potthoff RF. Telephone sampling in epidemiologic research: to
reap the benefits, avoid the pitfalls. Am J Epidemiol 1994;139:
967–978.
18. US Bureau of the Census. Census of Population and Housing,
1990 (United States): public use microdata sample: 1% sam-
ple. Washington, DC: Author, 1990.
19. Stewart WF, Lipton RB, Celentano DD, Reed ML. Prevalence
of migraine headache in the United States: relation to age,
income, race, and other sociodemographic factors. JAMA
1992;267:64–69.
20. Headache Classification Committee of the International
Headache Society. Classification and diagnostic criteria for
headache disorders, cranial neuralgias and facial pains. Ceph-
alalgia 1988;8(suppl 7):1–96.
21. Stewart WF, Lipton RB, Liberman J. Variation in migraine
prevalence by race. Neurology 1996;47:52–59.
22. Stewart WF, Lipton RB, Simon D, Von Korff M, Liberman J.
Reliability of an illness severity measure for headache in a
population sample of migraine sufferers. Cephalalgia 1998;18:
44–51.
23. Stewart WF, Lipton RB, Simon D, Liberman J, Von Korff M.
Validity of an illness severity measure for headache in a pop-
ulation sample of migraine sufferers. Pain 1999;79:291–301.
24. World Health Organization. Composite International Diag-
nostic Interview (CIDI), version 2.1. Geneva: World Health
Organization, 1997.
25. American Psychiatric Association. Diagnostic and statistical
manual of mental disorders, 4th ed. Washington, DC: Ameri-
can Psychiatric Association, 1994.
26. Wittchen H-U. Reliability and validity studies of the WHO-
Composite International Diagnostic Interview (CIDI): a criti-
cal review. J Psychiatr Res 1994;28:57–84.
27. Cox DR. Regression models and life tables. Series B. J R Stat
Soc 1972;34:187–220.
28. Cox DR. Partial likelihood. Biometrika 1975;62:269 –276.
29. Breslow N. Covariance analysis of censored survival data.
Biometrics 1974;30:89–99.
30. Allison PD. Survival analysis using SAS system: a practical
guide. Cary, NC: SAS Institute, 1995.
31. SAS Institute. SAS/STAT software: The PHREG procedure,
version 6.217. SAS technical report P-217. Cary, NC: SAS
Institute, 1991.
32. Haythornthwaite JA. Migraine headaches and psychopathol-
ogy: future directions. J Psychiatr Res 1993;27:183–186.
33. Lipton RB, Stewart WF. Migraine headaches: epidemiology
and comorbidity. Clin Neurosci 1998;5:2–9.
January (2 of 2) 2000 NEUROLOGY 54 313
at ALBERT EINSTEIN COLL MED on June 5, 2006 www.neurology.orgDownloaded from
2000;54;308- Neurology
N. Breslau, L. R. Schultz, W. F. Stewart, R. B. Lipton, V. C. Lucia and K. M. A. Welch
Headache and major depression: Is the association specific to migraine?
This information is current as of June 5, 2006
& Services
Updated Information http://www.neurology.org/cgi/content/full/54/2/308
including high-resolution figures, can be found at:
Permissions & Licensing
http://www.neurology.org/misc/Permissions.shtml
or in its entirety can be found online at:
Information about reproducing this article in parts (figures, tables)
Reprints http://www.neurology.org/misc/reprints.shtml
Information about ordering reprints can be found online:
at ALBERT EINSTEIN COLL MED on June 5, 2006 www.neurology.orgDownloaded from