Family, Twin, and Adoption Studies of Bipolar Disorder
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
American Journal of Medical Genetics Part C (Semin. Med. Genet.) 123C:48 – 58 (2003)
A R T I C L E
Family, Twin, and Adoption Studies
of Bipolar Disorder
JORDAN W. SMOLLER* AND CHRISTINE T. FINN
Family, twin, and adoption studies have been essential in defining the genetic epidemiology of bipolar disorder
over the past several decades. Family studies have documented that first-degree relatives of affected individuals
have an excess risk of the disorder, while twin studies (and to a lesser extent, adoption studies) suggest that genes
are largely responsible for this familial aggregation. We review these studies, including the magnitude of familial
risk and heritability estimates, efforts to identify familial subtypes of bipolar disorder, and the implications of
family/genetic data for validating nosologic boundaries. Taken together, these studies indicate that bipolar
disorder is phenotypically and genetically complex. ß 2003 Wiley-Liss, Inc.
KEY WORDS: family studies; twin studies; adoption studies; bipolar disorder; genetic epidemiology
INTRODUCTION point. Despite differences in study also led to the elaboration of standar-
methodology, sample ascertainment, dized diagnostic criteria for clinical prac-
Family, twin, and adoption studies
and diagnostic conventions, available tice with the publication of Diagnostic
represent different approaches to esti-
family, twin, and adoption studies have and Statistical Manual of Mental Dis-
mating the effect of familial and genetic
been consistent overall in documenting orders (DSM)-III in 1980 [American
influences on a disorder. Studies exam-
that bipolar disorder runs in families and Psychiatric Association, 1980] and cul-
ining the contribution of these influ-
that this familial aggregation is strongly minating most recently in the DSM-IV
ences to bipolar disorder have set the
influenced by genes. criteria [American Psychiatric Associa-
stage for the intensive efforts currently
A key consideration in interpreting tion, 1994]. The latter includes a further
under way to identify specific bipolar
genetic studies (whether they are epide- evolution of the bipolar phenotype: the
disorder susceptibility genes. In addition
miologic or molecular) is the impact of distinction between bipolar I disorder
to defining the genetic epidemiology
phenotype definition. The nature and (requiring an episode of mania) and
of the disorder, these studies provide
strength of genetic influences observ- bipolar II disorder (in which hypo-
information that may be of great interest
ed may vary substantially depending manic episodes occur along with depres-
to clinicians, patients, and families. For
on how precisely or how broadly the sive episodes) [Dunner et al., 1976].
example, family studies can provide em-
phenotype is defined. Prior to about Throughout this historical develop-
piric estimates of recurrence risk (the
1960, the definition of manic depres- ment, family, twin, and adoption studies
probability that a relative of an affected
sive illness generally followed Emil have played a key role in testing the
individual will develop the disorder).
Kraepelin’s conception, which included validity of nosologic distinctions by
Bipolar disorder has been among the
syndromes of both mania and depressive examining whether these diagnostic
most extensively studied psychiatric
episodes as well as recurrent depression categories are themselves familial and
disorders from a familial/genetic stand-
alone (see Angst and Marneros [2001] heritable.
and Baldessarini [2000] for recent re- In this article, we will summarize
views of the evolution of the nosology of highlights of the extensive literature on
Jordan W. Smoller is Assistant Professor of bipolar disorder). The modern emphasis the genetic epidemiology of bipolar
Psychiatry at Harvard Medical School and on distinguishing bipolar from unipolar disorder. The interested reader should
Director of the Psychiatric Genetics Program
in Mood and Anxiety Disorders in the Out-
affective illness is often attributed to also be aware of previous reviews, in-
patient Division of the Massachusetts Gen- Leonhard [1959] and has been incorpo- cluding Tsuang and Faraone’s [1990]
eral Hospital Department of Psychiatry. rated into most studies appearing after comprehensive review of studies pub-
Christine T. Finn is a psychiatrist at
Massachusetts General Hospital and a clin-
1960. The 1970s saw the develop- lished prior to the 1990s. We will focus
ical fellow in genetics in the Harvard Medical ment of standardized diagnostic criteria on studies that have used modern diag-
School genetics training program. [Feighner et al., 1972; Spitzer et al., nostic definitions, and study methods
*Correspondence to: Jordan W. Smoller,
15 Parkman St., WAC-812, Boston, MA
1978], enhancing the reliability and and review evidence addressing the
02114. E-mail: jsmoller@hms.harvard.edu comparability of research diagnoses as- magnitude of familial and genetic influ-
DOI 10.1002/ajmg.c.20013 signed in different studies. These efforts ences on bipolar disorder and putative
ß 2003 Wiley-Liss, Inc.ARTICLE AMERICAN JOURNAL OF MEDICAL GENETICS (SEMIN. MED. GENET.) 49
subtypes. We will also examine the im- polar and bipolar disorders when study- because they are methodologically rig-
plications of family, twin, and adoption ing risks of affective illness in families orous (e.g., predominant use of direct,
data for defining the boundaries of the semistructured interviews and diagnoses
bipolar phenotype, and briefly consider of relatives blind to proband affection
the clinical implications of familial Family studies published status). The inclusion of a control group
risk estimates for genetic counseling of can be important to the calibration and
prior to 1960 did not
patients and families. interpretation of results for relatives of
distinguish between unipolar affected probands. For example, the risk
and bipolar disorders of unipolar disorder among relatives of
FAMILY STUDIES
controls in the studies described below
Family studies attempt to answer the when studying risks of affective varies up to 10-fold. Thus, while
question of whether a disorder of interest illness in families. Gershon et al. [1975] found a lower risk
aggregates in families. To do this, studies of unipolar disorder among relatives of
typically compare the prevalence of the bipolar probands (8.7%) than did Tsuang
disorder among first-degree relatives of [Tsuang and Faraone, 1990]. Additional et al. [1980] (12.4%), the risk relative to
affected probands (cases) to the preva- methodological limitations included the controls is substantially higher in the
lence in the population or among re- lack of control groups, structured assess- former study (12-fold) than in the latter
latives of unaffected probands (controls). ments, standardized diagnostic criteria, (2-fold) because of a 10-fold difference
(Because younger relatives may not have and age correction of risk estimates. in the base rate of unipolar depression
passed through the period of risk for the Despite these shortcomings, as reviewed among control relatives in the two
disorder, prevalence estimates are typi- by Tsuang and Faraone [1990], 13 studies.
cally adjusted to yield more informative studies appearing before 1960 were In the first of the controlled studies,
age-corrected morbid risks.) A higher consistent with an increased risk of Gershon et al. [1975] reported a 19-fold
morbid risk among relatives of affected major mood disorder among relatives greater risk of bipolar disorder (3.8%)
probands indicates that the disorder can of affected probands compared to avail- and 12.4-fold greater risk of unipolar
be familial, but it does not necessarily able estimates of mood disorder risk in disorder risks (8.7%) among first-degree
mean that genes are involved; a disorder the general population. relatives of 54 bipolar probands than
may run in families for nongenetic Table I summarizes family studies among relatives of controls (0.2%) in an
reasons (i.e., because of shared environ- published after 1960 that examined the Israeli sample. Subsequently, Tsuang
ment). Some family studies compare transmission of bipolar disorder and et al. [1980] reported a family study of
familial risks in relatives of those affected unipolar disorder separately. Studies that probands with schizophrenia, affective
with one disorder (e.g., bipolar disorder) relied only on the family history method disorders and controls that included
to relatives of those affected with another are not included. Despite methodologic numerous methodological strengths
disorder (e.g., unipolar disorder); these differences among them, these studies (direct interviews; blinded, structured
studies may clarify whether disorders consistently reported excess risks of bi- assessments; and a best-estimate diag-
coaggregate in families and may also bear polar disorder among first-degree rela- nostic procedure). They observed a
on the validity of diagnostic distinctions tives of bipolar probands. Most of the nearly 18-fold risk of bipolar disorder
between disorders. An important meth- estimates included in Table I are age- among relatives of bipolar probands
odologic distinction is that between corrected morbid risks. For the uncon- compared to controls. The risk of bi-
studies using the family history method trolled studies, the magnitude of risk can polar disorder was also significantly
(in which relative diagnoses are based on be compared to available epidemiologic elevated among relatives of unipolar
indirect reports of probands or other estimates of population risk (on the probands (10-fold) compared to con-
family members) and direct-interview order of 0.5–1.0% for bipolar I disorder trols. The risk of unipolar disorder was
family studies (in which diagnoses are [Tsuang and Faraone, 1990; Weissman nonsignificantly elevated in relatives of
based primarily on direct reports from et al., 1996; Kessler et al., 1997]). For the bipolar probands (12.4% vs. 7.5% for
interviewed family members). The controlled studies, the risks can be controls) and significantly elevated in
family history method tends to be less directly compared to risks among rela- relatives of unipolar probands (15.2% vs.
sensitive than the family study method tives of controls. Table I also includes risk 7.5%).
and thus may underestimate the preval- estimates, when available, for probands Somewhat different results were
ence of disorders in families [Andreasen and relatives with unipolar disorder. observed in a subsequent large, metho-
et al., 1986]. In the following review, we As we discuss later, these estimates are dologically rigorous family study by
focus on studies that have relied primar- relevant to the familial/genetic rela- Gershon et al. [1982] of five proband
ily on direct-interview assessments of tionship between bipolar and unipolar groups: bipolar I, bipolar II, schizoaffec-
family members. disorder. tive, and unipolar disorders and normal
Family studies published prior to The controlled family studies listed controls. First-degree relatives of pro-
1960 did not distinguish between uni- in Table I are particularly informative bands with bipolar I disorder had similar50 AMERICAN JOURNAL OF MEDICAL GENETICS (SEMIN. MED. GENET.) ARTICLE
TABLE I. Morbid Risk Estimates of Bipolar and Unipolar Disorder From Family Studies of Bipolar Disorder
Morbid risk, FDR (%)
Relatives at risk:
Reference Proband group bipolar/unipolar Bipolar disorder Unipolar disorder
Uncontrolled Studies
Perris [1966] Bipolar 574 10.1 0.5
Unipolar 684 0.3 6.4
Mendlewicz and Rainer [1974] Bipolar 606/544 17.7 22.4
Helzer and Winokur [1974] Bipolar 151 4.6 10.6
James and Chapman [1975] Bipolar 265 6.4 13.2
Smeraldi et al. [1977] Bipolar 173 5.7 6.9
Unipolar 185 1.0 8.1
Johnson and Leeman [1977] Bipolar 180 15.5 19.8
Petterson [1977] Bipolar 472 4.4 —
Trzebiatowska-Trzeciak [1977] Bipolar 289 11.4 0
Unipolar 256 0.3 7.41
Abrams and Taylor [1980] Bipolar 47 8.5 6.4
Unipolar 107 4.7 7.5
Andreasen et al. [1987]a Bipolar I 569 3.9b (8.1)c 22.8
Bipolar II 267 1.1b (9.3)c 26.2
Unipolar 1171 0.6b (3.5)c 28.4
Pauls et al. [1992] Bipolar 408 8.7b (12.4)c 11.6
Grigoroiu-Serbanescu et al. [2001] Bipolar 867 5.3 —
Controlled Studies
Gershon et al. [1975] Bipolar (I and II) 341/264 3.5b (3.8)c 8.7
Unipolar 96/77 0b (2.1)c 14.2
Controls 518/411 0.2b (0.2)c 0.7
Tsuang et al. [1980] Bipolar 169 5.3 12.4
Unipolar 362 3.0 15.2
Controls 345 0.3 7.5
Gershon et al. [1982a] Bipolar I 441/422 4.5b (8.6)c 14.0
Bipolar II 157/150 2.6b (8.1)c 17.3
Unipolar 138/133 1.5b (3.0)c 16.6
Controls 217/208 0b (0.5)c 5.8
Weissman et al. [1984] Unipolard 287 0.8b (2.0)c 18.4
Controls 521 0.2b (1.3)c 5.9
Maier et al. [1993] Bipolar 389 7.0 21.9
Unipolar 697 1.8 21.6
Unscreened controls 419 1.8 10.6
Heun and Maier [1993] Screened controls 221 1.0 7.7
Weighted summary estimatee Bipolar 6365/4861 8.7 14.1
Unipolar 3983/3959 2.2 17.9
Controls 1822/1706 0.7 5.2
a
Number at risk and rates not age-corrected.
b
Risk of bipolar I.
c
Risk of bipolar I þ bipolar II disorder combined.
d
Includes some probands from Gershon et al. [1982].
e
Includes bipolar I and II probands independently for studies separating these proband groups. For relatives, morbid risk of bipolar I and II
have been combined for studies estimating risk of each subtype. For control estimates derived from the family study of Maier et al. [1993],
data from screened controls [Heun and Maier, 1993] rather than unscreened population controls [Maier et al., 1993] were used.ARTICLE AMERICAN JOURNAL OF MEDICAL GENETICS (SEMIN. MED. GENET.) 51
risks of bipolar I (4.5%) and bipolar II Finally, Maier et al. [1993] con- several studies that used age cutoff points
disorder (4.1%) compared to risks of 0% ducted a family study of psychotic and of 20 years [Pauls et al., 1992], 25 years
and 0.5%, respectively, for relatives of affective disorders in a German sample. [Grigoroiu-Serbanescu et al., 2001;
controls. Relatives of probands with Consistent with previous studies, they Somanath et al., 2002], 30 years [James,
observed a significantly increased risk of 1977; Taylor and Abrams, 1981], or even
bipolar disorder in relatives of bipolar 50 years old [Angst et al., 1980]. In
First-degree relatives probands but not relatives of unipolar addition, there is evidence that pediatric
probands, while the risk of unipolar bipolar disorder may represent a distinct
of probands with bipolar I disorder was increased in relatives of form of the disorder, and that it may be
disorder had similar both proband groups. This study is ad- genetically related to disruptive behavior
ditionally of interest because the general disorders, particularly attention-deficit/
risks of bipolar I and bipolar II population controls were not screened hyperactivity disorder [Spencer et al.,
disorder compared to risks of for psychiatric illness. Thus, the risk of
illness in these probands and their re-
0% and 0.5%, respectively,
latives should approximate that of the There is evidence that
for relatives of controls. underlying population. An estimate of
the recurrence risk ratio (sometimes
pediatric bipolar disorder may
referred to as l1)—the ratio comparing represent a distinct form
bipolar II disorder also had increased risk of disorder among first-degree of the disorder, and that it may
risks of bipolar I (2.6%) and bipolar II relatives of affected individuals to the
(4.5%) disorder. Relatives of bipolar I population risk of the disorder—could be genetically related to
and bipolar II probands had elevated therefore be made (with precision lim- disruptive behavior disorders,
risks of unipolar disorder as well (14.0% ited by the sample size). The prevalence
and 17.3%, respectively) compared to of illness in control families suggests a particularly attention-deficit/
relatives of controls (5.8%). On the other l1 of 4 for bipolar disorder and 2 for hyperactivity disorder.
hand, risk of bipolar disorder was not unipolar disorder. Using the weighted
significantly elevated among relatives average of risk estimates from the con-
of unipolar probands. (In an expanded trolled studies in Table I, relatives of 2001; Todd, 2002]. In general, early-
report, adding probands with unipolar bipolar probands have a 10-fold excess onset disorder may represent a more
depression and controls ascertained at risk of bipolar disorder compared to severe subtype with stronger genetic
Yale, Weissman et al. [1984] also found relatives of controls and a 2.96-fold risk loading [Schurhoff et al., 2000]. In a
stronger aggregation of bipolar I disorder of unipolar disorder. study of adolescent bipolar probands,
among relatives of bipolar probands than Overall, a summary estimate of Strober et al. [1988] found that those
among those with depression.) Gershon morbid risk based on the studies listed with prepubertal onset of psycho-
et al. [1982] also estimated the risk of in Table I, weighting studies by the pathology had a poorer response to
affective illness (including bipolar, uni- number of relatives at risk, indicates that lithium treatment and their first-degree
polar, or schizoaffective disorder) among the recurrence risk of bipolar disorder relatives had fourfold greater risk
offspring and found a significantly for first-degree relatives of bipolar pro- (29.4%) of bipolar disorder than relatives
higher risk if two parents have affective bands is 8.7%, while the risk for unipolar of probands with adolescent-onset bipo-
illness (74%) than if only one parent is depression is 14.1%. lar disorder (7.4%). In the large NIMH
affected (27%). Finally, these authors Collaborative Program on the Psycho-
report age-corrected risks of illness in biology of Depression, Rice et al. [1987]
CLINICAL MARKERS OF
second-degree relatives (aunts, uncles, also observed an increase in familial risk
FAMILIAL RISK
and grandparents) of affected indivi- with early-onset disorder, even control-
duals. Although the quality of the data Several studies have examined clinical ling for a cohort effect in which more
for second-degree relatives was not as features that might be associated with recent birth cohorts were associated
strong as for first-degree relatives, these greater familiality of bipolar disorder. with an earlier age of onset. In a segre-
estimates suggest that risks are compar- Overall, family studies have not found gation analysis, they found evidence
able to population risks, i.e., 0.4–1.1% evidence that risk of mood disorder in supporting the possibility of a single
risk of bipolar I or II disorder and 3.6– relatives of bipolar probands varies with major locus when an age of onset effect
5.4% risk of unipolar disorder among sex of the proband or relative [Gershon was included. This is consistent with
relatives of bipolar probands. The et al., 1982; Faraone et al., 1987; Rice findings from a recent study of Roma-
absence of elevated risks of these dis- et al., 1987; Pauls et al., 1992]. On the nian families in which the risk of bi-
orders among second-degree relatives other hand, earlier-onset bipolar disor- polar disorder in first-degree relatives
was also observed in an earlier study der in probands has been associated with of probands with early-onset disorder
[Smeraldi et al., 1977]. greater familial risk of mood disorder in (age 25) was significantly greater52 AMERICAN JOURNAL OF MEDICAL GENETICS (SEMIN. MED. GENET.) ARTICLE
(9.4%) than the risk to relatives of later- et al., 1997, 2002]. Another recent ana- nents: additive genetic influences, shared
onset probands (5.5%), and segregation lysis [Jones and Craddock, 2002] sug- familial environment (e.g., social class
analysis suggested different modes of ill- gests that puerperal manic or hypomanic during childhood, parents’rearing style),
ness transmission in the two subtypes, episodes may confer increased familial and individual-specific environment
with evidence for a major gene effect in risk of bipolar disorder. The effort to (e.g., stressful life events). The herit-
the early-onset families. identify familial subtypes is important ability of the disorder is an estimate of the
A number of investigators have for molecular genetic studies of bipolar proportion of phenotypic variance that
examined the possibility that response disorder. Given the complexity and can be attributed to genetic influences.
to antimanic medication, specifically heterogeneity of the bipolar disorder Heritability refers to the strength of
lithium, might provide a basis for iden- phenotype, delineation of more gene- genetic influences in a population—not
tifying familial subtypes of bipolar tically homogeneous subtypes might a particular individual—and heritability
disorder. Studies examining the rela- significantly enhance the power to estimates may differ depending on the
tionship between familial loading for detect susceptibility loci relevant to the population studied.
bipolar disorder and lithium responsive- disorder. Linkage and association ana- As was the case with family studies,
ness have had mixed results with some lyses using this strategy have focused on the interpretation of early twin studies
finding better response among patients lithium-responsive bipolar disorder and of bipolar disorder is complicated by
with a positive family history [Mendle- are illustrated by a report linking this methodological shortcomings (includ-
wicz et al., 1973; Smeraldi et al., 1984; phenotype to a locus on chromosome ing lack of blinded and structured
Grof et al., 1994] while others have 15q [Turecki et al., 2001]. The subtyping assessments and lack of specificity
found no significant relationship or even of bipolar disorder on the basis of comor- in diagnostic procedures). These early
an inverse relationship [Engstrom et al., bid panic disorder has also had utility in studies were comprehensively reviewed
1997; Coryell et al., 2000]. However, molecular genetic studies [MacKinnon by Tsuang and Faraone [1990], who
using a stringent definition requiring et al., 1998; Rotondo et al., 2002]. summarized the concordance rates and
years of no bipolar disorder episodes in derived heritability estimates where
patients who had multiple episodes prior possible from the data provided in these
TWIN STUDIES
to lithium prophylaxis, Alda et al. [1994, studies. Because many of these studies
1997] have observed familial clustering As we have mentioned, while evidence did not distinguish bipolar from unipolar
of lithium responsiveness and report se- of familiality supports the possibility that mood disorders, the summary measures
gregation analyses consistent with single genes influence a disorder, family studies reported by Tsuang and Faraone [1990]
major gene transmission [Grof et al., cannot establish the role of genes or refer to the composite phenotype of
2002]. estimate the magnitude of their influ- mood disorder. Combining the results of
The existence of other putative ence. Twin studies, by essentially com- 11 twin studies published between 1928
familial subtypes of bipolar disorder have paring groups of twin pairs matched for and 1986, comprising 195 MZ pairs and
been supported by recent family studies. shared environment but differing in 255 same-sex DZ pairs, they report a
In an analysis of multiplex families with degree of genetic relatedness, can help proband-wise concordance rate of 78%
bipolar spectrum disorders, Potash et al. parse genetic and environmental con- for MZ twins and 29% for DZ pairs. The
[2001, 2003] found that a history of tributions. Twin studies typically com- summary estimate of heritability was
psychotic bipolar disorder (with hallu- pare the concordance rates of a disorder 63%. In the largest and most informative
cinations and delusions) in probands between monozygotic (MZ) twins (who of these studies, Bertelsen et al. [1977]
conferred increased risk of bipolar dis- are essentially genetically identical) and ascertained, from the Danish Psychiatric
order among relatives; furthermore, dizygotic (DZ) twins (who share on Twin Register, 55 MZ and 52 DZ twin
psychotic bipolar disorder itself clustered average half of their genes). Assuming pairs in which the proband had been
in families of probands with psychotic that shared environmental influences on psychiatrically hospitalized and was
bipolar disorder, suggesting that this MZ twins are not different from envir- diagnosed with manic depressive disor-
phenotype may breed true to some ex- onmental influences on DZ twins (the der. Although diagnostic interviews
tent. Degree of psychosis also appeared equal environments assumption), signif- were unstructured and not blinded, the
to be familial in another sample of icantly higher concordance rates in MZ diagnosis of manic depressive disorder
families ascertained for linkage studies twins reflect the action of genes. Never- involved recurrent, episodic disorders of
of bipolar disorder [Omahony et al., theless, an MZ concordance rate that is mood (including both bipolar and uni-
2002]. less than 100% means that environmen- polar cases). The proband-wise concor-
A subtype of bipolar disorder tal factors influence the phenotype. dance rate for MZ twins (67%) was
comorbid with panic disorder has also Twin studies can also be used to estimate significantly greater than that observed
been proposed; data from independent the contribution of genetic and environ- for DZ twins (20%). Similar differences
family sets suggest that risk for panic mental factors to the variance in liability were observed when the analysis was
disorder in families segregating bipolar to the disorder [Kendler, 2001]. These restricted to bipolar/bipolar twin pairs
disorder is a familial trait [MacKinnon are often partitioned into three compo- (62% vs. 8%).ARTICLE AMERICAN JOURNAL OF MEDICAL GENETICS (SEMIN. MED. GENET.) 53
Kendler et al. [1993] reported a (87%) and similar to that for mania alone ever, MZ concordance rates and herit-
study of 486 twin pairs ascertained (84%). Table II summarizes data from the ability estimates are less than 100%,
through either the Swedish Psychiatric three largest and most methodologically demonstrating that environmental fac-
Twin Register (probands hospitalized rigorous twin studies that used modern tors are also influential.
for affective illness) or the population- definitions of bipolar disorder and in-
based Swedish Twin Registry. DSM- cluded more than 20 pairs with a bipolar
ADOPTION STUDIES
III-R diagnoses were made using a proband.
self-report questionnaire containing sec- Despite differences in ascertain- Adoption studies can help distinguish
tions of the Structured Clinical Inter- ment, assessment, and diagnostic meth- genetic and environmental influences
view for DSM-III-R for mania and ods, twin studies of bipolar disorder are on family resemblance by comparing
major depression. Bipolar disorder was generally consistent in observing greater rates of a disorder in biological family
diagnosed in five of the 13 co-twins concordance among MZ twins than DZ members to those in adoptive family
of a bipolar twin proband (concor- twins. This provides compelling evid- members. If genes influence the risk
dance ¼ 38.5%), but only one of 22 DZ ence for the hypothesis that suscept- of a disorder, biological (genetically
pairs (4.5%). Model fitting indicated that ibility genes contribute to the familiality related) family members should re-
the heritability of bipolar illness was of bipolar disorder. In fact, the results semble each other more than do adop-
79%, with a residual 21% of the variance tive (environmentally related) family
attributable to individual-specific envir- members. Unfortunately, because they
onment [Kendler et al., 1995]. In this Despite differences in are logistically difficult to conduct and
best-fitting model, there was no signi- ascertainment, assessment, and subject to a number of potential con-
ficant contribution of shared family founds [Kendler, 1993], the availabi-
environment to the liability to bipolar diagnostic methods, twin lity and interpretability of adoption
disorder. studies of bipolar disorder are studies of mood disorders has been
Finally, a study of 224 twin pairs limited.
ascertained from the Maudsley Twin generally consistent in Although there have been several
Register reported significantly higher observing greater concordance adoption studies addressing genetic
concordance rates for mania plus hypo- and environmental contributions to
among MZ twins than DZ
mania, defined by Research Diagnostic mood disorders, only two have exam-
Criteria, in MZ twins (44%) than DZ twins. This provides compelling ined the inheritance of bipolar disorder
twins (9.1%) [Cardno et al., 1999]. On evidence for the hypothesis using a modern definition of the phe-
the other hand, the difference in MZ notype [Mendlewicz and Rainer, 1977;
versus DZ concordance rates for mania that susceptibility genes Wender et al., 1986]. Mendlewicz and
alone (36.4% vs. 7.4%) did not reach contribute to the familiality of Rainer [1977] compared rates of illness
statistical significance, perhaps due to in adoptive and biological parents of
lower power for this comparison. As
bipolar disorder. 29 bipolar adoptees and those of 22
observed in the Swedish twin sample, unaffected adoptees; they also studied
the best-fitting biometrical model in- of biometrical modeling suggest that two additional sets of controls: parents
cluded no significant contribution from familial aggregation is due predomi- of 31 bipolar patients who were not
common environment. The heritability nantly to genetic factors, with herit- adopted, and parents of 20 individuals
for mania plus hypomania, which may ability estimates in the range of 60–85% who had contracted polio during child-
correspond to a combination of bipolar and little evidence that shared family hood or adolescence. The last group was
I and II disorder cases, was substantial environment plays a major role. How- intended to control for factors involved
TABLE II. Largest Recent Twin Studies of Bipolar Disorder
MZ twins DZ twins
a
Reference Number of pairs Concordance Number of pairsa Concordance Heritability
Bertelsen et al. [1977] 34 62% 37 8% 59%
Kendler et al. [1993, 1995] 13 38.5% 22 4.5% 79%
Cardno et al. [1999]b 25 44% 33 9.1% 87%
a
Number of pairs with affected bipolar proband.
b
Data for phenotype of ‘‘mania plus hypomania.’’54 AMERICAN JOURNAL OF MEDICAL GENETICS (SEMIN. MED. GENET.) ARTICLE
with raising a disabled child. Bipolar mission of risk for mood disorders, Maier, 1993]. It seems likely that bipolar
patients were ascertained from review including bipolar disorder. II is a heterogeneous entity in which
of inpatient and outpatient medical some cases are more closely related to
records, and parents were assessed by bipolar I, some to unipolar depression,
FAMILIAL/GENETIC
semistructured interviews blind to clin- and others may represent a genetical-
BOUNDARIES OF
ical and adoptive status of the proband ly distinct disorder that breeds true
BIPOLAR DISORDER
offspring. The frequency of affective ill- [Blacker and Tsuang, 1992; Heun and
ness (comprising bipolar, unipolar, In addition to demonstrating the famili- Maier, 1993].
schizoaffective, and cyclothymic disor- ality of a disorder, family, twin, and Most of the family studies reviewed
ders) was significantly greater in the adoption studies can provide crucial earlier examined the familiality of both
biological parents (31%) than in the information about the etiologic rela- bipolar and unipolar disorder. The cross-
adoptive parents (12%) of bipolar pro- tionships and boundaries between dis- phenotype familial risks (i.e., the risk of
bands. The risk of affective illness was orders. For example, if two disorders, bipolar disorder among relatives of
similar in the biological parents of A and B, share familial determinants, unipolar probands and the risk of uni-
adopted and nonadopted bipolar pro- relatives of probands with either disorder polar disorder among relatives of bipolar
bands; lower risks, comparable to those A or B should show increased risks of probands) speak to the familial/genetic
in the adoptive parents of bipolar pro- both disorders. On the other hand, if relationship between the disorders. As
bands, were observed in biological and the disorders are distinct entities from summarized in Table I, there is some
adoptive parents of normal adoptees and a familial/genetic standpoint, relatives variability in risk estimates across studies.
biological parents of probands with will be at risk only for the disorder Focusing on the controlled family stu-
polio. These results implicate genetic affecting the proband (i.e., the disorders dies, one can compare illness risks in
factors in the familial aggregation of will breed true). Family/genetic data relatives of affected probands to risks in
affective illness, although small numbers have helped clarify the boundaries of relatives of controls. In these studies, the
precluded meaningful analyses of bipolar the bipolar disorder phenotype, al- risk of unipolar disorder among relatives
disorder alone. though ambiguities remain [Blacker of bipolar probands is comparable to the
Similar conclusions emerge from a and Tsuang, 1992]. risk among relatives of unipolar pro-
Danish adoption study reported by One distinction relevant to the bands and generally higher than the risk
Wender et al. [1986]. They compared nosology of bipolar disorder is the status among relatives of controls. On the
rates of illness in biological and adoptive of bipolar II disorder, characterized by other hand, the risk of bipolar disorder
parents of 71 adoptees who had been episodes of hypomania and depression. among relatives of unipolar probands
hospitalized with affective disorders Although the reliability of the bipolar II tends to be the same or intermediate
(including 27 with unipolar and 10 with diagnosis has been considered lower between the risk to bipolar relatives and
bipolar disorder), as well as 71 control than that of bipolar I, recent work sug- the risk to control relatives. The studies
adoptees matched for age, sex, time gests that good reliability can be achiev- differ somewhat with respect to this last
spent with biological mother, age at ed with direct interviews of relatives and point, however. For example, Tsuang
adoption, and socioeconomic status. careful diagnostic procedures [Simpson et al. [1980] found that relatives of uni-
They observed a significantly higher et al., 2002]. Family studies using direct polar probands had a statistically signifi-
frequency of major mood disorder interviews of first-degree relatives have cant 10-fold relative risk of bipolar
(including bipolar and unipolar disorder) provided some evidence that bipolar II disorder compared to relatives of con-
in biological parents of ill adoptees than disorder is a distinct entity. Risks of trols, whereas Maier et al. [1993] found
in biological relatives of controls. The bipolar II disorder have tended to be identical risks of bipolar disorder among
prevalence did not differ significantly highest among relatives of bipolar II relatives of unipolar or control probands.
between adoptive relatives of ill versus probands as opposed to those with bi- A controlled, longitudinal family study
control probands. Again, the small num- polar I or unipolar depression [Gershon from the multicenter Collaborative De-
ber of cases of bipolar disorder did not et al., 1982; Coryell et al., 1984; End- pression Study [Winokur et al., 1995]
permit meaningful statistical compari- icott et al., 1985; Andreasen et al., 1987; also found indistinguishable rates of
sons for this disorder per se. Also, be- Heun and Maier, 1993; Simpson et al., bipolar I disorder among relatives of
cause of differences in the demographic 1993]. However, the observation that unipolar or control probands. Overall,
characteristics of biological and adoptive familial risks of unipolar disorder are the available family studies support the
relatives, direct comparisons were not similar across these proband groups and conclusions that: 1) relatives of bipolar
made between these two groups. Over- (in some studies) that risk of bipolar I is probands have an elevated risk of both
all, then, the evidence from adoption also elevated in relatives of bipolar II bipolar disorder and unipolar disorder,
studies bearing directly on the herit- probands suggests that these affective and 2) relatives of unipolar probands are
ability of bipolar disorder has been disorders are not completely etiologi- at increased risk for unipolar disorder
limited; however, the available data are cally distinct [Gershon et al., 1982a; but do not appear to have a substantially
consistent with a role for genetic trans- Andreasen et al., 1987; Heun and elevated risk of bipolar disorder. TwinARTICLE AMERICAN JOURNAL OF MEDICAL GENETICS (SEMIN. MED. GENET.) 55
and Tsuang, 1992, 1993]. Analyses has been more difficult to define. This is
Overall, the available family aimed at clarifying these subgroups have perhaps not surprising given that the
not been able to identify robust indica- definition of schizoaffective disorder
studies support the
tors of bipolarity among unipolar rela- includes prominent mood symptoms.
conclusions that: 1) relatives tives of bipolar probands [Blacker et al., Taken together, the available data argue
of bipolar probands have 1996]. for etiologic overlap in the familial/
Kraepelin’s fundamental diagnostic genetic basis of these disorders, particu-
an elevated risk of both bipolar distinction between manic depressive larly with the schizoaffective-bipolar or
disorder and unipolar illness and schizophrenia (dementia manic subtype. In several studies, the risk
praecox) was based in part on his con- of bipolar disorder among relatives of
disorder, and 2) relatives of clusion that these disorders have a schizoaffective probands has been com-
unipolar probands distinct hereditary basis. Nevertheless, parable to or greater than the risk among
are at increased risk for unipolar there is symptomatic overlap in that bipolar probands [Gershon et al.,
psychotic symptoms can be a feature of 1982; Andreasen et al., 1987; Rice
disorder but do not appear both disorders. The etiologic boundary et al., 1987; Maier et al., 1993]. On the
to have a substantially elevated between psychotic disorders and bipolar other hand, relatives of bipolar probands
disorder has been the subject of much have not been shown to have substan-
risk of bipolar disorder. attention, particularly in light of recent tially elevated risks of schizoaffective
observations that genetic loci influen- disorder [Angst et al., 1980; Gershon
studies have had limited power to cing these disorders may overlap [Badner et al., 1982; Pauls et al., 1992; Maier
examine the genetic relationship be- and Gershon, 2002]. While some studies et al., 1993].
tween bipolar and unipolar disorders; have suggested elevated risks of schizo- The overlap of genetic influences
cross-phenotype (e.g., unipolar/bipolar) phrenia among relatives of bipolar or on schizophrenic, schizoaffective, and
MZ twin pairs have been reported, but affective disorder probands [Tsuang manic disorders was also examined in
statistically meaningful comparisons to et al., 1980; Taylor et al., 1993; Valles a recent analysis of twins from the
DZ twin concordance rates have gen- et al., 2000], several controlled direct- Maudsley Twin Register [Cardno et al.,
erally not been possible [Bertelsen et al., interview family studies have not sup- 2002]. Cardno et al. [2002] applied
1977; Torgersen, 1986; Kendler et al., ported this [Gershon et al., 1975, 1982; nonhierarchical diagnostic definitions
1993]. Weissman et al., 1984; Maier et al., such that twins could be assigned life-
Variability in risks across studies 1993]. A similar picture emerges in time diagnoses for more than one of
may reflect, in part, differences in studies of schizophrenic probands and these disorders. They report significant
criteria applied to define cases. Tsuang their relatives; while one controlled correlations in genetic liability among
and Faraone [1990] pointed out that study [Taylor et al., 1993] found evi- the three syndromes; the genetic corre-
studies requiring recurrent depressive dence of increased familial risk of af- lations were 0.68 for schizophrenic and
episodes for the diagnosis of unipolar fective disorder (unipolar and bipolar manic syndromes and 0.88 for schizo-
disorder in probands tended not to find combined), most have not found an affective and manic syndromes. The
elevated rates of bipolar disorder among increase in bipolar disorder among genetic liability to schizoaffective dis-
their relatives. They suggest that pro- relatives of schizophrenic probands order was entirely shared with the two
bands with nonrecurrent depression may [Tsuang et al., 1980; Gershon et al., other syndromes. Kendler [2002] has
be more likely to represent latent cases 1988; Maier et al., 1993; Kendler and pointed out that interpretation of this
of bipolar disorder in which a manic Gardner, 1997]. Kendler and Gardner study is complicated because it is not
episode has not yet occurred. Consistent [1997] performed a meta-analysis of clear how often mania occurred in the
with this, a controlled study that in- three controlled, direct-interview stu- absence of schizophrenia in co-twins of
cluded probands with recurrent unipolar dies of schizophrenic probands (the schizophrenic probands. Nevertheless,
depression found no increased risk of Danish Adoption Study, the Iowa 500 this intriguing analysis supports the pos-
bipolar disorder in their families com- Non-500 Family Study, and the Ros- sibility that recent reports of overlapping
pared to families of unaffected controls common Family Study) and found that linkage findings for bipolar disorder and
[Heun and Maier, 1993]. On the other familial risk estimates of bipolar disorder schizophrenia represent shared genetic
hand, the risks of recurrent and non- were homogeneous across the studies, risk factors [Berrettini, 2001; Badner
recurrent depression were similar among with an nonsignificant common odds and Gershon, 2002].
relatives of bipolar probands. It is also ratio of 1.9 (95% confidence interval Overall, then, genetic epidemio-
possible that the excess risk of unipolar (CI) ¼ 0.7–5.2) for risk of bipolar dis- logic studies suggest that the bipolar
disorder among relatives of bipolar order in relatives of schizophrenic versus disorder phenotype is etiologically het-
probands may be overestimated if some control probands. erogeneous and that at least some pro-
apparently unipolar relatives are actually The boundary between schizo- portion of cases are genetically related
latent cases of bipolar disorder [Blacker affective disorder and bipolar disorder to unipolar depression and psychotic56 AMERICAN JOURNAL OF MEDICAL GENETICS (SEMIN. MED. GENET.) ARTICLE
disorders, particularly schizoaffective of risk, accompanied by the above
disorder. The identification of specific We would emphasize caveats.
susceptibility genes through molecular
here that some familiarity
genetic studies should help clarify the
with the results of SUMMARY
boundaries between bipolar disorder
and other psychiatric disorders. family/genetic studies is Family studies of bipolar disorder have
consistently demonstrated that the dis-
important for clinicians order can run in families. First-degree
CLINICAL IMPLICATIONS
working with patients relatives of affected individuals have
As we have seen, family, twin, and approximately a 10-fold increased risk
adoption studies have provided strong
and families affected with mood of the disorder compared to relatives of
evidence that familial and heritable fac- disorders; this includes unaffected controls. Twin studies (and to
tors contribute to the etiology of mood familiarity not only with a lesser extent adoption studies) have
disorders and, in particular, bipolar dis- further provided evidence that genes
order. In addition to providing moti- what is known but also with contribute strongly to familial transmis-
vation and justification for molecular what is not known. sion of the disorder. Earlier-onset bipo-
genetic efforts to identify susceptibility lar disorder appears to be associated
genes, these studies also have relevance with increased familial risk, and studies
to clinical practice. For example, dis- For example, family studies have pro- have identified other putative familial
cussing the contribution of genetic and vided useful empiric estimates of recur- subtypes, including lithium-responsive
biological risk factors can be an im- rence risk for first-degree relatives. At bipolar disorder and bipolar disorder
portant component of psychoeducation the same time, the precision of these with psychosis or comorbid panic dis-
with patients and families and may over- estimates is likely to be limited, as is clear order. Family and twin studies have
come misconceptions about the causes from the range of risks listed in Table I. In generally supported the validity of the
of bipolar disorder. As an illustration, addition, empiric risks from available bipolar disorder phenotype and its
although bipolar disorder can clearly run studies may not reflect the risks in in- separation from unipolar depression or
in families, the available evidence does dividual families; those risks may be far psychotic disorders. However, the data
not suggest that shared family environ- higher or lower depending on the mode also suggest that the familial/genetic
ment (e.g., parenting style) plays a sub- of transmission in a given family and determinants of these disorders are not
stantial role in the development of the the particular configuration of affected entirely distinct and/or that they are
disorder. family members. As an alternative ap- heterogeneous, with the existence of
Studies documenting the familiality proach, theoretical recurrence risks both distinct and overlapping forms of
and heritability of bipolar disorder are can be derived using estimates of cer- these phenotypes. Ultimately, molecular
also clearly relevant when patients or tain epidemiologic parameters. Moldin genetic studies will help clarify how
families have questions related to genetic [1997] calculated lifetime recurrence strong and how variably expressed the
counseling. For example, affected indi- risks for a child born into a five-member genetic influences on bipolar disorder
viduals planning to have children may family with varying configurations of may be.
have questions about familial recurrence affected relatives using knowledge of
risks. In the absence of identified sus- disease prevalence and heritability and
ceptibility genes for bipolar disorder, assuming a polygenic mode of inheri- REFERENCES
family, twin, and adoption data represent tance. The risks for bipolar disorder vary Abrams R, Taylor MA. 1980. A comparison of
the primary basis for risk prediction. A from about 5–6% when there is only one unipolar and bipolar depressive illness. Am
J Psychiatry 137:1084–1087.
thorough discussion of the issues and affected first-degree relative to more Alda M, Grof P, Grof E, Zvolsky P, Walsh M.
techniques involved in genetic counsel- than 50% when there are four. He notes 1994. Mode of inheritance in families of
ing in psychiatry is beyond the scope of that, like empiric risks, these can serve as patients with lithium-responsive affective
disorders. Acta Psychiatr Scand 90:304–
this article, and the interested reader is a guide but do not represent individua- 310.
referred to recent work devoted to this lized risk estimates. In general, when Alda M, Grof E, Cavazzoni P, Duffy A, Martin R,
topic [Moldin, 1997; Hodgkinson et al., discussing recurrence risks with patients Ravindran L, Grof P. 1997. Autosomal
recessive inheritance of affective disorders
2001; Tsuang et al., 2001]. We would or their family members, it is important in families of responders to lithium prophy-
emphasize here, however, that some to highlight the limitations of the in- laxis? J Affect Disord 44:153–157.
familiarity with the results of family/ formation and to discuss how risks are American Psychiatric Association. 1980. Diagnos-
tic and statistical manual of mental disorders,
genetic studies is important for clinicians generated, the spectrum of diagnoses 3rd edition. Washington, DC: American
working with patients and families af- assessed in the supporting family studies, Psychiatric Association.
fected with mood disorders; this in- and the difference between empiric American Psychiatric Association. 1994. Diagnos-
tic and statistical manual of mental disorders,
cludes familiarity not only with what is and individualized risks. In our experi- 4th edition. Washington, DC: American
known but also with what is not known. ence, it is most helpful to provide a range Psychiatric Association.ARTICLE AMERICAN JOURNAL OF MEDICAL GENETICS (SEMIN. MED. GENET.) 57
Andreasen NC, Rice J, Endicott J, Reich T, Faraone SV, Lyons MJ, Tsuang MT. 1987. Sex Kendler KS, Gardner CO. 1997. The risk for
Coryell W. 1986. The family history ap- differences in affective disorder: genetic psychiatric disorders in relatives of schizo-
proach to diagnosis. How useful is it? Arch transmission. Genet Epidemiol 4:331–343. phrenic and control probands: a comparison
Gen Psychiatry 43:421–429. Feighner JP, Robins E, Guze SB, Woodruff RA Jr, of three independent studies. Psychol Med
Andreasen NC, Rice J, Endicott J, Coryell W, Winokur G, Munoz R. 1972. Diagnostic 27:411–419.
Grove WM, Reich T. 1987. Familial rates criteria for use in psychiatric research. Arch Kendler KS, Pedersen N, Johnson L, Neale MC,
of affective disorder. A report from the Gen Psychiatry 26:57–63. Mathe AA. 1993. A pilot Swedish twin
National Institute of Mental Health Colla- Gershon ES, Mark A, Cohen N, Belizon N, study of affective illness, including hospital-
borative Study. Arch Gen Psychiatry 44: Baron M, Knobe K. 1975. Transmitted fac- and population-ascertained subsamples.
461–469. tors in the morbid risk of affective disorders: Arch Gen Psychiatry 50:699–700.
Angst J, Marneros A. 2001. Bipolarity from a controlled study. J Psychiatr Res 12:283– Kendler KS, Pedersen NL, Neale MC, Mathe AA.
ancient to modern times: conception, birth 299. 1995. A pilot Swedish twin study of affective
and rebirth. J Affect Disord 67:3–19. Gershon ES, Hamovit J, Guroff JJ, Dibble E, illness including hospital- and population-
Angst J, Frey R, Lohmeyer B, Zerbin-Rudin E. Leckman JF, Sceery W, Targum SD, ascertained subsamples: results of model
1980. Bipolar manic-depressive psychoses: Nurnberger JI Jr, Goldin LR, Bunney WE fitting. Behav Genet 25:217–232.
results of a genetic investigation. Hum Jr. 1982. A family study of schizoaffective, Kessler RC, Rubinow DR, Holmes C, Abelson
Genet 55:237–254. bipolar I, bipolar II, unipolar, and normal JM, Zhao S. 1997. The epidemiology of
Badner JA, Gershon ES. 2002. Meta-analysis control probands. Arch Gen Psychiatry 39: DSM-III-R bipolar I disorder in a general
of whole-genome linkage scans of bipolar 1157–1167. population survey. Psychol Med 27:1079–
disorder and schizophrenia. Mol Psychiatry Gershon ES, Dehisi LE, Hamovit J, Nurnberger JI 1089.
7:405–411. Jr, Maxwell ME, Schrelber J, Davphinais D, Leonhard K. 1959. Aufteilung der endogenen
Baldessarini RJ. 2000. A plea for integrity of the Dingman CW 2nd, Guroff JJ. 1988. A Psychosen. Berlin: Akademie Verlag.
bipolar disorder concept. Bipolar Disord 2: controlled family study of chronic psy- MacKinnon DF, McMahon FJ, Simpson SG,
3–7. choses. Schizophrenia and schizoaffective McInnis MG, DePaulo JR. 1997. Panic
Berrettini WH. 2001. Molecular linkage studies disorder. Arch Gen Psychiatry 45:328–336. disorder with familial bipolar disorder. Biol
of bipolar disorders. Bipolar Disord 3:276– Grigoroiu-Serbanescu M, Martinez M, Nothen Psychiatry 42:90–95.
283. MM, Grinberg M, Sima D, Propping P, MacKinnon DF, Xu J, McMahon FJ, Simpson SG,
Bertelsen A, Harvald B, Hauge M. 1977. A Marinescu E, Hrestic M. 2001. Different Stine OC, McInnis MG, DePaulo JR. 1998.
Danish twin study of manic-depressive dis- familial transmission patterns in bipolar I Bipolar disorder and panic disorder in
orders. Br J Psychiatry 130:330–351. disorder with onset before and after age 25. families: an analysis of chromosome 18 data.
Blacker D, Tsuang MT. 1992. Contested bound- Am J Med Genet 105:765–773. Am J Psychiatry 155:829–831.
aries of bipolar disorder and the limits Grof P, Alda M, Grof E, Zvolsky P, Walsh M. MacKinnon DF, Zandi PP, Cooper J, Potash JB,
of categorical diagnosis in psychiatry. Am 1994. Lithium response and genetics Simpson SG, Gershon E, Nurnberger J,
J Psychiatry 149:1473–1483. of affective disorders. J Affect Disord 32: Reich T, DePaulo JR. 2002. Comorbid
Blacker D, Tsuang MT. 1993. Unipolar relatives in 85–95. bipolar disorder and panic disorder in
bipolar pedigrees: are they bipolar? Psychiatr Grof P, Duffy A, Cavazzoni P, Grof E, Garnham J, families with a high prevalence of bipolar
Genet 3:40–51. MacDougall M, O’Donovan C, Alda M. disorder. Am J Psychiatry 159:30–35.
Blacker D, Faraone SV, Rosen AE, Guroff JJ, 2002. Is response to prophylactic lithium a Maier W, Lichtermann D, Minges J, Hallmayer J,
Adams P, Weissman MM, Gershon ES. 1996. familial trait? J Clin Psychiatry 63:942–947. Heun R, Benkert O, Levinson DF. 1993.
Unipolar relatives in bipolar pedigrees: a Helzer JE, Winokur G. 1974. A family interview Continuity and discontinuity of affective
search for elusive indicators of underlying study of male manic depressives. Arch Gen disorders and schizophrenia. Results of a
bipolarity. Am J Med Genet 67:445–454. Psychiatry 31:73–77. controlled family study. Arch Gen Psychia-
Cardno AG, Marshall EJ, Coid B, Macdonald Heun R, Maier W. 1993. The distinction of try 50:871–883.
AM, Ribchester TR, Davies NJ, Venturi P, bipolar II disorder from bipolar I and Mendlewicz J, Rainer JD. 1974. Morbidity risk
Jones LA, Lewis SW, Sham PC, Gottesman recurrent unipolar depression: results of a and genetic transmission in manic-depres-
II, Farmer AE, McGuffin P, Reveley AM, controlled family study. Acta Psychiatr sive illness. Am J Hum Genet 26:692–701.
Murray RM. 1999. Heritability estimates Scand 87:279–284. Mendlewicz J, Rainer J. 1977. Adoption study
for psychotic disorders: the Maudsley twin Hodgkinson KA, Murphy J, O’Neill S, supporting genetic transmission in manic-
psychosis series. Arch Gen Psychiatry 56: Brzustowicz L, Bassett AS. 2001. Genetic depressive illness. Nature 268:326–329.
162–168. counselling for schizophrenia in the era of Mendlewicz J, Fieve RR, Stallone F. 1973.
Cardno AG, Rijsdijk FV, Sham PC, Murray RM, molecular genetics. Can J Psychiatry 46: Relationship between the effectiveness of
McGuffin P. 2002. A twin study of genetic 123–130. lithium therapy and family history. Am J
relationships between psychotic symptoms. James NM. 1977. Early- and late-onset bipolar Psychiatry 130:1011–1013.
Am J Psychiatry 159:539–545. affective disorder. A genetic study. Arch Gen Moldin SO. 1997. Psychiatric genetic counseling.
Coryell W, Endicott J, Reich T, Andreasen N, Psychiatry 34:715–717. In: Guze SB, editor. Washington University
Keller M. 1984. A family study of bipolar II James NM, Chapman CJ. 1975. A genetic study of adult psychiatry. St. Louis: Mosby. p 365–
disorder. Br J Psychiatry 145:49–54. bipolar affective disorder. Br J Psychiatry 381.
Coryell W, Akiskal H, Leon AC, Turvey C, 126:449–456. Omahony E, Corvin A, O’Connell R,
Solomon D, Endicott J. 2000. Family history Johnson GF, Leeman MM. 1977. Analysis of Comerford C, Larsen B, Jones R, McCand-
and symptom levels during treatment for familial factors in bipolar affective illness. less F, Kirov G, Cardno AG, Craddock N,
bipolar I affective disorder. Biol Psychiatry Arch Gen Psychiatry 34:1074–1083. Gill M. 2002. Sibling pairs with affective
47:1034–1042. Jones I, Craddock N. 2002. Do puerperal disorders: resemblance of demographic and
Dunner DL, Fleiss JL, Fieve RR. 1976. The psychotic episodes identify a more familial clinical features. Psychol Med 32:55–61.
course of development of mania in patients subtype of bipolar disorder? Results from a Pauls DL, Morton LA, Egeland JA. 1992. Risks of
with recurrent depression. Am J Psychiatry family history study. Psychiatr Genet 12: affective illness among first-degree relatives
133:905–908. 177–180. of bipolar I old-order Amish probands. Arch
Endicott J, Nee J, Andreasen N, Clayton P, Keller Kendler KS. 1993. Twin studies of psychiatric Gen Psychiatry 49:703–708.
M, Coryell W. 1985. Bipolar II. Combine or illness. Arch Gen Psychiatry 50:905–915. Perris C. 1966. A study of bipolar (manic-
keep separate? J Affect Disord 8:17–28. Kendler KS. 2001. Twin studies of psychiatric depressive) and unipolar recurrent depres-
Engstrom C, Astrom M, Nordqvist-Karlsson B, illness: an update. Arch Gen Psychiatry 58: sive psychoses. Introduction. Acta Psychiatr
Adolfsson R, Nylander PO. 1997. Relation- 1005–1014. Scand Suppl 194:9–14.
ship between prophylactic effect of lithium Kendler KS. 2002. Hierarchy and heritability: the Petterson U. 1977. Manic-depressive illness. A
therapy and family history of affective dis- role of diagnosis and modeling in psychiatric clinical, social and genetic study. Acta
orders. Biol Psychiatry 42:425–433. genetics. Am J Psychiatry 159:515–518. Psychiatr Scand Suppl 269:1–93.You can also read
