ESQUIZOFRENIA REVIEW

7/28/2019 ESQUIZOFRENIA REVIEW

1/25

VOL. 19, NO. 2, 1993 The Genetics ofSchizophrenia: A Current,Genetic-EpidemiologicPerspective

261

by Kenneth S. Kendler andScott R. DlehlAbstractIn the "Special Report onSchizophrenia" published in theSchizophrenia Bulletin in 1987,the genetic basis of schizo-phrenia was reviewed. Here, weprovide our perspective on thecurrent status of this area of in-vestigation, focusing largely butnot exclusively on recent find-ings. Methodologically rigorousfamily studies have now clearlyshown that schizophrenia sub-stantially aggregates in families.Familial factors that predisposeto schizophrenia also increase therisk for certain schizophrenia-related personality disorders andprobably for some forms of non-schizophrenic nonaffective psy-chosis. Results from one newtwin study and updates fromtwo ongoing adoption studiescontinue to support the hypothe-sis that genetic factors play amajor role in the etiology ofschizophrenia. Little is knownabout how genetic liability toschizophrenia is transmitted, al-though statistical models suggestthat transmission is probably notdue solely to a single majorgene. Schizophrenia is clearly acomplex disorder in that genecarriers need not manifest theillness (incomplete penetrance),affected individuals need nothave the gene (environmentalforms or phenocopies), diagnosticuncertainties cannot be avoided,and different families may carrydifferent susceptibility genes (ge-netic heterogeneity). Therefore,segregation or linkage analysesare far more difficult to performwith schizophrenia than withMendelian genetic disorders.Given this complexity, it is nottoo surprising that no replicatedpositive evidence for linkage to

schizophrenia has been reportedto date. However, just as linkageanalysis of schizophrenia shouldnot be excessively embraced asthe only form of viable geneticresearch in schizophrenia, it alsoshouldn't be prematurelyspurned. If one or several genesof major effect exist for schizo-phrenia, large samples using newstatistical and laboratory meth-odologies have a good chance ofdetecting them. The authors thusrecommend a balanced researchapproach to the genetics ofschizophrenia that includes tradi-tional methods of family, twin,and adoption studies as well asa major effort in large-samplelinkage studies.

The status of the genetics ofschizophrenia was thoroughly re-viewed in 1987 in the "Special Re-port on Schizophrenia" (Gottesmanet al. 1987). In this review, weprovide our perspective on thecurrent status of this researchfield, focusing largely but not ex-clusively on recent findings. (Forother recent perspectives on thissubject, see Weeks et al. 1990;Gottesman 1991.)It has been an eventful time forthe field. Center stage has shiftedfrom traditional family, twin, andadoption studies to linkage analy-ses of "high-density" families. Thefocus of scientific inquiry hasmoved from determining the mag-nitude of familial aggregation, dis-entangling the effects of geneticand shared environmental factors,and understanding the boundaries

Reprint requests should be sent toDr. K.S. Kendler, Dept. of Psychiatry,Box 710, Medical College of Virginia,Richmond, VA 23298-0710.


2/25

262 SCHIZOPHRENIA BULLETIN

of the schizophrenia spectrum to-ward attempts to locate specificsusceptibility genes.This is not a new pursuit. Inestablishing the first modem psy-chiatric research institute inMunich early in this century,Kraepelin recruited the youngpsychiatrist/geneticist Ernst Rudinwith the specific goal of uncover-ing the Mendelian basis of schizo-phrenia. When one of us (KSK)visited with Manfred Bleuler in1986, Bleuler, then in his earlyeighties, appeared at the door withan old display case of butterflies.His high school science projecthad been performing simplecrosses and backcrosses in but-terflies to uncover the Mendelianbasis of wing patterns. He saidthen that it had become his life-long dream to discover the Men-delian basis of schizophrenia.

The psychiatric genetics com-munity has responded in threeways to the rapidly increasingpower of human gene-mappingmethods and their astounding suc-cesses with classic Mendelian dis-orders. Some have concluded thatthese techniques are irrelevant, asthe genetic and etiologic complex-ity of schizophrenia renders such asimple approach useless. At theopposite extreme, others havegrasped the world of human genemapping with unbounded enthusi-asm. Overnight, traditional psychi-atric genetic methods such as fam-ily, twin, and adoption studieshave been dismissed as antiquated,no longer "cutting edge," andready for the dustbin of history.Avoiding either extreme, a thirdresponse has been that, while veryexciting and holding out thepromise of a true revolution inour understanding of the etiologyof schizophrenia, gene-mappingapproaches continue to represent

but one among many potentiallyuseful techniques in psychiatricgenetics.Our long-held view, reflected inthis article, is the third response(Kendler 1987). Given the complex-ity of psychiatric disorders and theslow nature of progress to date,impatience is understandable. Thepotential power of human genemapping is therefore seductive, en-couraging investigators to focus ex-clusively on these glamorous newtechnologies. However, as we hopeto demonstrate, the progress wecontinue to make in our under-standing of the etiology of schizo-phrenia with the more traditionalpsychiatric genetic methods is sub-stantial, even if it is not revolu-tionary. Thus, a balanced researchplan for this devastating illnessshould include both traditional andgene-mapping approaches.

Family StudiesThe first question in the geneticepidemiology of schizophrenia isthe degree to which the disorderaggregates (or "runs") in families.Historically, studies of this key is-sue can be divided approximatelyinto two periods. In the firstperiod, as summarized by Zerbin-Rudin (1967), studies were nearlyall performed nonblind, withoutcontrol groups, and without theuse of structured psychiatric as-sessments or operationalized diag-nostic criteria. In many of thesestudies, it is unclear how manyrelatives were seen individuallyrather than evaluated from second-ary sources (e.g., hospital or parishrecords, reports from relatives).Although many of these studieswere performed by thorough andcareful researchers, their findingsthat schizophrenia consistently and

substantially aggregates infamilieswere open to method-ological criticism.In the early 1980s, two researchgroups (Pope et al. 1982; Abramsand Taylor 1983) questioned thevalidity of these studies and ar-gued that the previous evidencefor the familial aggregation ofschizophrenia arose from meth-odological deficiencies, particularlyan overly broad diagnostic ap-proach to the disorder. Whenschizophrenia was "narrowly" di-agnosed, these groups argued, evi-dence of familial aggregationmight be weak or absent.We are, in 1993, in a good posi-tion to evaluate these claims. Since1980, a substantial number ofsecond-generation family studies ofschizophrenia have been com-pleted. These share three keymethodological features: (1) a nor-mal control group, (2) structuredpsychiatric assessment and opera-tionalized diagnostic criteria, and(3) blind assessment and diagnosis.We judged seven studies as meet-ing these criteria (table 1). Threeother valuable studies (Scharfetterand Niisperli 1980; Guze et al.1983; Onstad et al. 1991a) wereeliminated because the only controlgroup consisted of nonschizo-phrenia psychiatric patients; afourth (Tsuang et al. 1980) waseliminated because it used a con-sensus diagnostic procedure notbased on a single system of opera-tionalized diagnostic criteria.These seven selected studieswere performed in the UnitedStates (Baron et al. 1985; Kendleret al. 1985a; Coryell and Zimmer-man 1988; Gershon et al. 1988),Greece (Frangos et al. 1985), Ger-many (Maier et al. 1990), and Ire-land (Kendler et al., in press a),with the total number of assessedrelatives ranging from 232 (Coryell

at

ot

ca

UvesdadCatocadeC

eo

Octobe

9,0

ttp://sc

op

eabuet.o

odjou

as.og/

ow

oaded

o
http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/


3/25

Te1Smmayresohamiaaeooszoeanmaore

amiysueo

szoeahnu

acogops

nevewwhreavabnda

so

reav

Frsahy B

oea

(1 Fa

e

a1

K

e

(1 Cya

Zmmema

(1 Ges

e

a1

Maeea

(1 K

ee

a (npea

Da

c

cea

R DM-

DM-

DM-

R R coc

sz

ea

R DMR

Cos

Nmas

Nmas

Sesgcpes

N vues

Vue

cos

N maccos

U

e

maccos

Frsde

reavo

szoea

po Szoea

B

n

3

1

4

2

7

2

7

1

9

3

4

2

2

1

M % 58 54 37 14 31 50 65B 3 5 9 1 3 2 4

Frsde

reavo

co

po Szoea

n 2 6 2 0 2 1 2

MR % 06 11 02 0 06 03 05

P00 00 8

8 N00 00 00

Ceao

oay

0300

0200

0300

021

0200

0300

0400

NoeBob

eoaemevesosM=mbdsS=sadeoR=RcDa

cCeaSzea1

DMJDM-RoDa

caSascM

oMaDsdsAmcPacA

ao11

N=nsgc

'nudneapoepbec

2 Rsnueavwhoyhaedb

ohognsubT

ea1

O O ro u ro C

at Pontificia Universidad Catolica de Chile on October 29, 2012http://schizophreniabulletin.oxfordjournals.org/om
http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/


4/25


and Zimmerman 1988) to 1,634(Kendler et al. 1985a). Studies usedeither Research Diagnostic Criteria(RDC; Spitzer et al. 1978) (Coryelland Zimmerman 1988; Maier et al.1990), modified RDC (Baron et al.1985; Gershon et al. 1988), DSM-III (American Psychiatric Associa-tion 1980) (Frangos et al. 1985;Kendler et al. 1985


5/25

VOL. 19, NO. 2, 1993 26 5

were all performed using DSM-IH-R operationalized criteria. Thestudy also has three methodologi-cal weaknesses. First, the samplesize, which included only 52 same-sex twin pairs, was small. Second,interviews and diagnoses wereperformed nonblind. Third, zy-gosities were based only on self-report measures.Probandwise concordance forschizophrenia in monozygotic (MZ)

and same-sex dizygotic (DZ) twinswas 15/31 (48.4%) and 1/28(3.6%), respectively. Assuming apopulation risk for schizophreniaof 0.8 percent, the correlation in li-ability in MZ and DZ twins canbe calculated at +0.87 0.08 and+0.24 0.16, respectively. Becausethe correlation in MZ twins sub-stantially exceeds twice that in DZtwins, the best estimate of broadheritability of liability from thissample is probably the MZ cor-relation itself. This result is towardthe upper end of the estimates ob-tained from previous twin studiesof schizophrenia (Kendler 1983).Overall, the results from this studyare reassuringly similar to thoseobtained from previous studies(Gottesman and Shields 1966;Kendler 1983) in suggesting thatgenetic factors play a major role inthe etiology of schizophrenia. Thisreport is also consistent with pre-vious twin studies in suggestingthat familial environment makeslittle or no contribution to the lia-bility for schizophrenia (Kendler1983).

Adoption StudiesUpdated findings from two majoradoption studies of schizophreniahave appeared since the last re-view. Preliminary results of a rep-lication of the Copenhagen Adop-tion Study of Kety and colleagues

have been presented (Kety 1987).In the Copenhagen study, 34adoptees were identified who theinvestigators considered to havechronic, acute, or latent schizo-phrenia. Their biologic and adop-tive relatives were assessedthrough hospital records (Kety etal. 1968) and later by personal in-terview (Kety et al. 1975). Sincethen, the same research team ex-amined adoptees outside ofCopenhagen (the so-called Provin-cial sample) and identified 42other adoptees with chronic, acute,or latent schizophrenia. As withthe initial reports from theCopenhagen sample (Kety et al.1968), hospital records have beensearched, blinded, and diag-nostically reviewed for the biologicand adoptive relatives of these in-dex adoptees and of a matchedgroup of control adoptees. Bothchronic schizophrenia and latentand uncertain schizophrenia werefound to be significantly morecommon in the biologic relativesof index adoptees than in the bio-logic relatives of control adoptees.Furthermore, the rates of thesedisorders were low and not dif-ferent in the adoptive relatives ofboth groups of adoptees (Kety1987).Preliminary results are also nowavailable from a blind diagnosticreview by Kety and colleagues ofpersonal interviews with both theadoptees and the biologic andadoptive relatives from the Provin-cial sample (Kety et al., in prepa-ration). The researchers foundschizophrenia to be significantlymore common in the biologic rela-tives of all schizophrenia adoptees(4.1%) than in the biologic rela-tives of all control adoptees (0.5%)(p = 0.01 by one-tailed Fisher'sexact test). By contrast, latent anduncertain schizophrenia was not

found to be significantly morecommon in the biologic relativesof the schizophrenia adoptees thanin those of the control adoptees(6.5% vs. 5.5%, respectively). How-ever, if the schizophrenia adopteeswere screened to include onlythose with confirmed chronicschizophrenia and the controladoptees were screened to elimi-nate any cases with major mentalillness, latent schizophrenia wouldhave been significantly more com-mon in the biologic relatives ofthe former group than in those ofthe latter group (8.2% vs. 2.4%, p= 0.025). Rates of schizophreniaand related disorders were verylow and equal in the two groupsof adoptive relatives.The largest study of adopted-away offspring of schizophreniaand control mothers is being con-ducted in Finland under the direc-tion of Tienari (1991). As of April1991, 361 adoptive families con-taining the adopted-away offspringof a schizophrenia or controlmother had been contacted forfield study. Results were availablefor 144 offspring of schizophreniamothers and 178 offspring of con-trol mothers. To date, 15 psychoticadoptees have been ascertained, ofwhom 13 are offspring of schizo-phrenia mothers (13/144 = 9.1%)an d 2 are offspring of controlmothers (2/178 = 1.1%). This is ahighly significant difference (x2 =11.20, df = 1, p < 0.000). Of the13 psychotic adopted-away off-spring of schizophrenia mothers,by DSM-HI-R criteria, 7 hadschizophrenia, 2 had schizo-phreniform disorder, 2 had delu-sional disorder, and 2 had psy-chotic bipolar illness. Both of thepsychotic offspring of controlmothers had schizophrenia. There-fore, the prevalence of schizo-phrenia is also significantly greater


6/25

26 6 SCHIZOPHRENIA BULLETIN

in offspring of index mothers(7/144 = 4.9%) than in offspringof control mothers (2/178 = 1.1%)(X2 = 4.09, df = 1, p < 0.04).This adoption study has also ex-amined the role of familial-environmental factors in the etiol-ogy of schizophrenia and found asubstantial correlation between thefunctioning of the adoptive familyand the psychiatric outcome in theadoptee. Of particular interest, thisrelationship was greatest when theadoptee was an offspring of aschizophrenia mother. These resultsare consistent with a model inwhich schizophrenia emerges whengenetically susceptible individualsexperience disruptive family en-vironments. However, most of thefamilies were evaluated when theadoptees were well into adult life.Thus, this observed correlation be-tween adoptee and adoptive familyfunctioning could also be due todisturbed adoptees creating dis-turbed families. A prospectivestudy is now under way in a sub-set of this unique cohort with thehope of further clarifying this cen-tral issue.Summary of Family, Twin,and Adoption StudiesFamily studies suggest that schizo-phrenia strongly aggregates infamilies. Twin and adoption stud-ies continue to suggest that geneticfactors account for most of this fa-milial aggregation, although reportsfrom the Finnish adoption studycould suggest a significant role forenvironmental factors in vulnerableindividuals. Previous estimates ofthe heritability of liability toschizophrenia, based on eithertwin studies alone (Kendler 1983)or twin and family studies (not in-cluding any of the family studiesreviewed here [McGue et al.

1983]), range from 0.63 to 0.68,and the most recent twin study(Onstad et al. 1991b) suggests aheritability moderately higher thanthis. If, as indicated by nearly alltwin and adoption studies, most ofthe familial aggregation of schizo-phrenia is caused by genetic fac-tors, heritability should be abouttwice the correlation of liability infirst-degree relatives. The familystudy data agree closely with thisprediction. The correlation in lia-bility from the recent method-ologically strong family studies(around +0.35) is about half of theaggregate estimate of heritabilityobtained from twin studies. Theseconclusions, which apply to gen-eral populations of individualswith schizophrenia and no t to in-dividual cases, are not inconsistentwith the hypothesis that in someindividuals, schizophrenia islargely environmental in origin,while in others, the disorder iscaused largely by genetic factors.Boundaries of the"Schizophrenia Spectrum"An accurate definition of the af-fected phenotype is crucial in anygenetic investigation. A major areaof interest in the genetics ofschizophrenia has therefore been touse genetically informative designsto determine what psychiatric syn-dromes reflect the familial liabilityto schizophrenia. Such an en-deavor is also of interest becauseit can provide insights into the na-ture of that familial liability andhence into the etiology of schizo-phrenia itself. We have found ithelpful here to articulate four hy-potheses for the nature of thetransmitted liability to schizo-phrenia: (1) liability only to typicalschizophrenia, (2) liability toschizophrenia and schizophrenia-

like personality disorders, (3) lia-bility to all nonaffecrive psychosis,and (4) broad liability to al l psy-chiatric illness (Kendler 1988). Eachof these hypotheses predicts a dis-tinct pattern of illness in relativesof schizophrenia versus controlprobands.We will not review in this arti-cle the important attempts to de-fine the nature of the transmittedliability to schizophrenia usingnonclinical measures such as atten-tion, smooth-pursuit eye move-ments, or evoked potentials.Schizophrenia-Like PersonalityDiso rders. Many clinicians andresearchers over the last 100 yearshave noted that some close rela-tives of patients with schizo-phrenia, though never psychotic,had odd or eccentric personalitiesthat were clinically similar toschizophrenia (Kendler 1985). Thishas led to the second hypothesisarticulated abovethat the familialliability to schizophrenia is ex-pressed, at least in part, as apredisposition to a set of schizo-phrenia-like or schizotypal person-ality traits. Since this hypothesiswas first put to a rigorous test byKety and colleagues in the Danishadoption studies of schizophrenia(Kety et al. 1968, 1975), work inthis area has been strongly influ-enced by the development in theDSM-1I1 of operationalized diag-nostic criteria for schizotypal per-sonality disorder (SPD; Spitzer etal. 1979) and by its elaboration inth e DSM-III-R. Operationalizeddiagnostic criteria were alsoproposed for other putativeschizophrenia-related personalitydisorders, especially paranoid per-sonality disorder (PPD).

Seven family or adoption studiesusing personal interviews, DSM-IIIor DSM-III-R criteria, and blind


7/25

VOL. 19, NO. 2, 1993 26 7

diagnoses have examined the life-time prevalence or risk for SPDand/or PPD in first-degree biologicrelatives of schizophrenia andmatched control probands(table 2). Several observations onthe results of these studies are ofinterest.First, the rates of SPD/PPD inrelatives of patients with schizo-phrenia are widely variable andhighly significantly different acrossstudies (x2 = 101.31, df = 6, p


8/25

Te2Smmayresomaoamiyaa

osueunps

nevewoemin

pee

oDMaDMRszoyopadps

ydsdnrsde

reavoszoeaanmacopo

ro o

Frsahy

Lwnea

(1 K

eea

(1 Boea

(1 Fa

ea

(1 Cya

Zmmema

(1 Ges

ea

(1 K

eea

(npeb

Su Aeaw

osn

Booceav

oa

e

Nefamiysu

Nefamiysu

Nefamiysu

Nefamiysu

Nefamiysu

Ravo

szoeapo

Tan

3 1 3 5 7 1 3

So n 6 3 7 1 3 3 2P P

ee

1458

3015

2923

2707

4223

2816

8215

Ta 3 4 3 6 1 3 5

Ravo

copo

Sn

n 3 1 1 3 4 0 1P P

ee

7743

2121

4712

0603

2512

01705

P 02 00 padps

ydsdPee=empee

S=sadeo

'Daoyeeaceeshmbdsahhpee

ee

CO OI N o T 3D mzo c m z

at Pontificia Universidad Catolica de Chile on October 29, 2012http://schizophreniabulletin.oxfordjournals.org/om
http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/


9/25

VOL. 19, NO. 2, 1993 269

Table 3. Summ ary of the results of major recent familystudies examining the risk of nonschizophrenic nonaffectivepsychosis in relatives of schizophrenia versus controlprobands

Firstauthor/yearBaron et al. (1985)Frangos et al. (1985)Kendler (1985)Coryell and Zimmerman(1988)Gershon et al. (1988)Maier et al. (1990)Kendler et al. (in press d)

DisordersD D , APAPSF.SA.DD.APSASA+APSASF,SA,DD,AP

Risk Infirst-degree relativesSchizophreniaprobands

1.61.75.007.15.03.7

Controlprobands0.31.10.42.00.90.51.4

P0.06NS0.000NS0.000( 1 )

0.05Note.D D = delusional disorder AP atypical psychosis; SF = schlzophrenrform disorde rSA = schizoaffecttve disorder NS = not significant.'Cannot be calculated and not given In original report.

disorder or other nonaffective psy-chosis had significantly elevatedrisks for schizophrenia.

Given that suffering from affec-tive illness represents a stress thatmight precipitate psychosis, oneprediction of this hypothesis isthat the probability of developingpsychosis given affective illnessshould be increased in relatives ofschizophrenia probands. Two stud-ies have confirmed this prediction(Kendler et al. 1985a, in press c) .However, other studies havebeen less consistent with this hy-pothesis. Three family studies donein the 1980s failed to find an in-creased rate for schizophrenia inrelatives of delusional disorderprobands (Watt 1980, 1985; Ken-dler et al. 1985b). Similarly, no in-creased risk for schizophreniaspectrum disorders was found inthe biologic relatives of the smallnumber of adoptees with schizo-phreniform disorder, delusionaldisorder, or atypical psychosis ina DSM-J/J-based rediagnosis of

the Copenhagen Danish AdoptionStudy (Kendler and Gruenberg1984).The preponderance of the evi-dence, however, supports the va-lidity of the third hypothesis thatthe familial vulnerability to schizo-phrenia may manifest itself, atleast in part, as a predisposition tomore broadly defined psychosis.Schizophrenia and Other Psychi-atric Disorders. The last hypothe-sis (that the familial vulnerabilityto schizophrenia is entirely non-specific) predicts that the risk forall major forms of psychopathol-ogy ought to be increased inrelatives of schizophrenia versuscontrol probands. We found ninerecent controlled studies that usedmodem operationalized criteria,blind diagnosis, and controlgroups to address this issue withrespect to three major categories ofnonschizophrenic psychopathology:affective illness, anxiety disorders,and alcoholism (table 4). All nine

studies examined the risk for af-fective illness in relatives ofschizophrenia and control pro-bands. The results are inconsistent.One study reports significantlyhigher rates of affective illness inthe relatives of the control pro-bands (Frangos et al. 1985); sixstudie s report no significant dif-ferences in the rates of illnessin the two groups of relatives(Kendler and Gruenberg 1984;Baron et al. 1985; Kendler et al.1985a; Coryell and Zimmerman1988; Maj et al. 1991; Ken dler etal., in press c); and two studiesreport significantly greater risks foraffective illness in the relatives ofthe schizophrenia probands(Gershon et al. 1988; Maier et al.1990). Overall, these resultsprovide some support for theKraepelinian dichotomy betweendementia praecox and manic-depressive insanity. However,these latter two studies (Maj et al.1991; Kendler et al., in press c)provide some evidence for a "uni-tary psychosis" theory most re-cently articulated by Crow (1990).In both of these studies (Gershonet al. 1988; Maier et al. 1990),when divided by polarity, onlyunipolar and not bipolar illnesswas significantly more common inrelatives of schizophrenia probandsthan in relatives of control pro-bands. This is inconsistent withmost unitary models for psychosis,which place bipolar illness closerto schizophrenia than unipolarillness.Table 4 also reviews resultsfrom studies that have examinedthe risk for anxiety disorders oralcoholism in relatives of schizo-phrenia versus control probands.For both these disorders, onestudy found an increased rate inrelatives of controls and the re-maining studies found no signifi-


10/25


(0

sf lc oE CO 3X COII D (0CO mN >

5

H o3 CQ CO

E"Sgsfc *8 -(0 CO*3 COO 1*

Q .


11/25

VOL. 19, NO. 2, 1993 271

Huntington's disease, musculardystrophy, or cystic fibrosis. Butalthough SML may exist forschizophrenia, it is now certainlyclear that schizophrenia differsfrom the classic Mendelian disor-ders in at least four crucial ways(McGue and Gottesman 1989;Risch 1990). First, most Mendeliandisorders are fully penetrant. Thatis, if one inherits a "bad gene,"one will always suffer from thedisorder if one lives through theperiod of risk. For schizophrenia,the extreme rarity of families withapparent Mendelian inheritance,the concordance rate in MZ twinsof well below 100 percent, and theevidence that offspring of unaf-fected MZ cotwins of schizophre-nia persons are at substantiallyincreased risk for schizophrenia(Gottesman and Bertelsen 1989) allsuggest that this disorder exhibitsreduced penetrance. That is, onecan have high genetic liability toschizophrenia and not manifest theillness. Second, in most Mendelianconditions, in all individuals whoexpress typical symptoms of thedisease, the symptoms are due todirect effects of the disease gene.For schizophrenia, however, this isnot true. Cases of schizophrenia-like symptoms produced by meta-bolic or neurologic conditions ordrug abuse clearly exist; thesecases are called phenocopies.Third, most Mendelian conditionsare rare and etiologically homoge-neous; that is, all cases arise fromabnormalities in the same gene.Different mutations in the samegene may cause variants of a dis-order, but this does not affectlinkage analysis because such al-lelic variants occur at the same ge-netic locus. Schizophrenia is, bycontrast, a relatively common con-dition and is likely to involve ge-netic heterogeneity. This means

that defects at several genes prob-ably influence the risk for formsof schizophrenia that are, at pres-ent, clinically indistinguishablethough etiologically distinct. Sucha perspective is comparable to themodern view of cancer as a collec-tion of disorders identified by acommon clinical manifestation (un-controlled cell division) but causedby very distinct genetic mecha-nisms (e.g., activation of growth-promoting oncogenes vs. loss oftumor-suppressing genes). Finally,in most Mendelian disorders thereis an obvious discontinuity be-tween affected and unaffected indi-viduals. As reviewed above, suchdiagnostic boundaries are lessclearly delineated for schizo-phrenia. Although there is goodevidence to consider an individualwith schizoaffective disorder orSPD in a high-density schizo-phrenia pedigree to be "affected,"family members with delusionaldisorder, atypical psychosis, orschizoid personality are moreproblematic.

Segregation AnalysisOne approach in genetic epidemi-ology to evaluating whether SMLexist is complex segregation anal-ysis. This method determines adisease's mode of transmission byexamining the pattern of diseasein a systematically collected sampleof nuclear families or extendedpedigrees. In contrast to linkageanalysis, this approach examinesonly disease phenotypes and notgenetic marker information.The rather inconclusive resultsfor schizophrenia obtained to datewith this method are reviewedelsewhere (Kendler 1988; Kendlerand Diehl, in press). One applica-tion of this method that was re-ported recently (Vogler et al. 1990)

used the large Swedish familystudy of Iindelius (1970). A multi-factorial model with high herit-ability (> 80%) provided a sub-stantially better fit to the datathan did any SML models. Techni-cal problems prevented the evalua-tion of a "mixed model" contain-ing both multifactorial and SMLcomponents. Nevertheless, andconsistent with most previous anal-yses, it appears unlikely that thedistribution of schizophrenia is duesolely to the effect of a SML(McGue and Gottesman 1989).Although complex segregationanalyses can include incompletepenetrance and phenocopies aswell as diagnostic uncertainty(by using different diagnostic"classes"), this method has no realpower to examine genetic hetero-geneity. Rather, results provideonly an "aggregate" mode of

transmission averaged across fam-ilies. Therefore, lack of evidence ofSML for schizophrenia should notbe interpreted as strong evidenceagainst the possibility that such"major" genes exist in a subset offamilies.Linkage AnalysisThe most striking recent develop-ment for research into the geneticsof schizophrenia has been in link-age analysis, the general principlesand methods of which are de-scribed elsewhere (e.g., Ott 1991).Before reviewing this area, how-ever, we again emphasize the spe-cial challenges in applying thismethod to psychiatric disorders,which, as discussed above, arelikely to be complex and hetero-geneous. Linkage analysis in hu-mans was developed originally forsimple Mendelian traits. Schizo-phrenia is clearly not such a sim-ple disorder, and this must be


12/25


kept in mind when performingand interpreting linkage studies.Reduced penetrance, phenocopies,diagnostic error, and locus hetero-geneity clearly make linkage stud-ies both qualitatively and quan-titatively more difficult forschizophrenia than for simpleMendelian disordersqualitativelybecause statistical methods devel-oped for classic Mendelian disor-ders must be modified beforebeing applied to schizophrenia andquantitatively because necessarysample sizes are likely to be muchlarger than those sufficient forsimple Mendelian disorders.

We illustrate these points by anexample. Several studies (Cavalli-Sforza and King 1986; Chakravartiet al. 1987; Clerget-Darpoux et al.1987; Goldin and Gershon 1988;Chen et al. 1992) have examinedthe impact of genetic heterogeneityon the power of linkage analysis.Two studies (Martinez and Goldin1989; Chen et al. 1992) have ex-amined conditions similar to thoseconfronted by researchers studyingmultiplex families in schizophrenia,but both presented results acrossonly a limited number of param-eters. Following the methods ofone of these reports (Martinez andGoldin 1989), we conducted ourown power calculations (table 5)under a series of assumptions (de-tailed in the footnote). Althoughthese assumptions are unlikely tobe exactly correct, we believe theyare reasonable guesses as to whatmust be confronted in attemptingto undertake linkage studies ofschizophrenia.

As expected, the higher the pen-etrance (the probability of havingschizophrenia if one has one ortwo copies of the predisposinggene), the smaller the number offamilies required to detect linkage.This makes sense because, as pen-

Table 5. Num ber of families needed to detect linkage toschizophrenia with 80 percent power, given varying levels ofgenetic heterogeneity

a1.00.90.80.70.60.50.40.30.20.1

0.54761769913719630052211064359

P e n e t r a n c e0.64356719312618027748310453824

0.740516484114160246434922

3550

0.83444567399140213370794

2994Note.a = proportion of famil ies In which the schizophrenia-predisposing gene is l inked tothe marker. The power calculations depicted in this table are based on the fol lowingassumptions. (1) nudear famil ies consisting of both parents and five offspring, in which bothparents were unaffected and the number of affected offspring was modeled to equal thatfound in our Ir ish study of high-density schizophrenia pedigrees (Dlehl and Kendler 1989; Su1991; Su et al . 1993), (2) schizophrenia modeled as a dominant trai t with Incompletepenetrance ranging from 50-80 percent, a phenocopy rate of 0.10 percent, and populationrisk of 1 percent In both the l inked and unl inked famil ies; (3) 6 (the recombination fractionbetween the marker and the disease gene) = 5 percent; (4) power = 80 percent; (5) a LODscore of 4.27 required as evidence for linkage (the level found by MacLean et al. [1993] toequal a LOD of 3.0, given that the LOD score Is maximized with respect to penetrance andphenotyplc defini t ion); and (6) PIC value (polymorphism Information content, a measure ofhow variable the marker Is In the population) of 0.70. We test for l inkage assuminghomogeneity because of evidence that, with family structures as found In this project, this isusual ly as powerful as or more powerful than testing for l inkage assuming heterogeneity(Martinez and Qoldin 1989; Risen 1989).etrance becomes higher, the clinicalphenotype becomes a more accu-rate means to determine the un-derlying genotype. More importantfor our discussion here is the im-pact of genetic heterogeneity onthe power to detect linkage. As-suming for a moment that thepenetrance of our putativedominant-like schizophrenia geneis 60 percent, if all high-densityfamilies were due to this gene, itwould require only 43 families(301 individuals) to provide an 80-percent chance of detecting link-age. Unfortunately, few would ex-pect schizophrenia to be soetiologically homogeneous. To us,it would seem optimistic to as-sume even the modest hetero-

geneity reflected in 70 percent ofhigh-density families being linkedto one gene, in which case the re-quired sample size would morethan double to 93 families (651 in-dividuals). More likely, the singlemost common gene responsible forschizophrenia is present in only 50or 30 percent of high-density fam-ilies, so that required sampleswould amount to 180 families(1,260 individuals) or 483 families(3381 individuals), respectively. Ifschizophrenia is highly hetero-geneous and the most commonsingle gene accounts for only 10or 20 percent of high-density fam-ilies, sample sizes required to de-tect linkage by current methodsmay be unobtainable with realistic


13/25

VOL 19, NO. 2, 1993 273

resources.Similar results are obtained if,instead of assuming that high-density families each have one andonly one schizophrenia gene inthem, one assumes that, in mostsuch families, there exist a modestnumber of genes (e.g., about five),each of which influences liabilityto schizophrenia. This model,termed "oligogenic" for postulatinga few important genes, also pre-dicts that sample sizes needed forlinkage analysis are many timesgreater than those needed for sim-ple Mendelian traits (Suarez et al,in press).The first linkage study of schizo-phrenia known to us was per-formed in Switzerland and pub-lished in 1958 (Constantinidis1958). McGuffin et al. (1983) re-viewed this and other early link-age studies of schizophrenia usingconventional protein-based markerssuch as human leukocyte antigens(HLA), blood groups, and enzymepolymorphisms. Most such studiessince that time, such as the workof Goldin et al. (1987), focused onthe HLA region, although onestudy examined blood-groupmarkers in 19 multiplex families(Andrew et al. 1987). No repli-cated positive evidence for linkageto schizophrenia emerged from theuse of these protein polymor-phisms.In principle, the nature of ge-netic markers used in linkagestudies is immaterial. Markers sim-ply distinguish which of the fourparental alleles at some chromo-somal location are transmitted toeach child in order to detect non-independent assortment withinfamilies of chromosomal segmentsand a disease. In practice, modernmolecular genetics has revolu-tionized linkage studies in man.Previously, informative markers

were available for only a verysmall fraction of the human ge-nome. Thus, rare positive findingsof linkage were of interest, butnegative findings meant little be-cause only a very small portion ofthe genome could be examined.However, modem molecular ge-netic techniques have providedboth large numbers of markers forvirtually all locations in the humangenome and great improvementsin marker informativeness (i.e.,variation among individuals). Todetermine whether a disease iscosegregating with a chromosomalregion, it is necessary to dis-tinguish the chromosomes of ma-ternal and paternal origin as wellas the two homologous chromo-somes of each parent. A markerwith only 2 variants (alleles) ismuch less likely to provide thesedistinctions than a marker with 10common alleles. One goal of theHuman Genome Project (alreadynearly attained) is the developmentof highly informative deoxyribo-nucleic acid (DNA)-based markers,spanning the entire human ge-nome, along with technologies fortheir accurate and efficient charac-terization (e.g., Ziegle et al. 1992).Certainly the most dramatic se-ries of events in the genetics ofschizophrenia in recent years be-gan with the observation(McGillivray et al. 1990) of ayoung man and his maternal un-cle, both of whom had schizo-phrenia and subtle dysmorphicracial features. Karyotype analysisrevealed that both affected individ-uals had a chromosome 5qll.2-ql3.3 trisomy, with the unaffectedmother having a balanced invertedinsertion. This isolated observationof two relatives with a partial tri-somy and schizophrenia couldeasily have been a chance occur-rence. Alternatively, these two rel-

atives could have had a very rarekind of schizophrenia etiologicallyunrelated to more common forms.However, this finding did repre-sent a potential clue to the loca-tion of schizophrenia genes, whichwas pursued very quickly.In late 1988, Sherrington et al.reported linkage in seven high-density pedigrees (five fromIceland and two from England)between schizophrenia-related

phenotypes and two DNA markersin the region of the 5q trisomy.This article has been extensivelyreviewed elsewhere (McGuffin etal. 1990; Owen et al. 1990; Wattand Edwards 1991), so our com-ments here will be limited to itsespecially striking features. First,the pedigrees themselves were re-markable for both their size andtheir very high density of schizo-phrenia. Such pedigrees haverarely been found by other inves-tigators. Second, evidence in favorof linkage of the 5qll-13 region toschizophrenia-related phenotypeswas very strong, with a maximumLOD score1 of 6.49. Despite the

'The LOD score (Logarithm of theOdds) is a ratio of the likelihood ofthe observed pattern of disease andmarker phenotypes in the families un-der study under two alternative hy-potheses. The null hypothesis in thedenominator is the hypothesis of nolinkage (recombination fraction = 0.5),and this is contrasted with a series ofalternative hypotheses ranging fromvery close linkage of the marker anddisease loci (recombination fraction =0.0) to various degrees of looser link-age (e.g., recombination fractions 0.01,0.05, 0.10 ... < 0.5). Traditionally(based on arguments appropriate onlyfor simple, nonheterogeneous dis-eases), a LOD score greater than 3.0(i.e., odds of 1,000:1) is consideredstatistically significant evidence insupport of linkage.


14/25


15/25

VOL 19, NO. 2, 1993 275

Table 6. Linkage studies of schizophrenia ContinuedStudy Locus/region 1 Exclusion or support of linkageKennedy et al.(19916)O'Neill et al. (1991)Hallmayer et al.(1992)Kennedy et al.(1991c)Wildenauer et al.

(1991)Vallada et al. (1992)Zatz et al. (1991)Wildenauer et al.(1991)d'Amato et al. (1992)Nonomura et al.(1991)Asherson et al.(1992)Collinge et al. (1991)DeLisi et al. (1991)McGuffin and Shirt(19862)Byerly et al. (1991)Polymeropoulos et al.(1991)Barr et al. (1991)

DR D4/11p15.5 Exclusion assum ing homogeneityHRA S/11p15.5 Exclusion assum ing homogeneityHTR 2/13q Exclusion assum ing homogeneityAD M/17q Exclusion assum ing homogeneityAD M/19 Exclusion assum ing homogeneityCYP 2D/22 Exclusion assum ing homogeneityDM D/Xp21 Weak suggestion of l inkage with muscular dystrophyADM /XYpter Exclusion assum ing homogeneityADM /XYpter Sugges tion of l inkageADM /XYpter Exclusion assum ing homogeneityADM /XYpter Exclusion assum ing homogeneityAD M/XY pter Suggestion of l inkageAD M/X q27-28 Exclusion assuming homogenei tyPP/20 loci Exclusion assum ing homogeneity or inconclusiveAD M/150 loci Exclusion assum ing homogeneity or inconclusiveADM /30 loci Exclusion assum ing homogeneity or inconclusiveAD M/150 loci Exclusion assum ing homogeneity or inconclusive

Note.C itation s in table are presented based on chromosomal location (chromosome 1 through xy, followed by the genomew lde sear ches ).CYP2D cytochrome P450 subfamily IID; DRD = dopamlne receptor D#; HRAS = H-ras oncogens (located In dose vicinity of DRD4); HTR2 =serotonln 2 receptor (5-hydroxytryptamlne receptor #); HLA human leukocyte antigen (major hlstocompatibillty com plex); DMD = Duchennemuscular dystrophy.'Locus: Name of locus for markers that are genes or proteins of known function, ADM If anonymous DNA marker, PP If miscellaneous proteinpolymorphisms such as Isozymes or blood group antigens, and CR If chromosome rearrangement; Region: Chromosomal location of markerloci.2 Review of several Independent studies

ing have been uniformly negative.Only the reports of Su et al. (Su1991; Su et al. 1991) formally ex-amined the important question ofexclusion of linkage from this can-didate region in the presence ofpossible heterogeneity among fam-ilies within a study. Using a rela-tively large sample of Irish high-density schizophrenia pedigrees,this study ruled out linkage to adisease gene present in any more

than 25 to 50 percent of all fam-ilies, depending on the geneticmodel assumed. Although abso-lutely no positive evidence wasfound, even this sample, which isconsiderably larger than all othersin table 6, could not eliminate thepossibility that a major suscep-tibility gene in this chromosomalregion could be present in up tohalf of the families studied. Thisexemplifies the crucial need for

large samples for studies of thiscomplex disorder. A review byMcGuffin et al. (1990) examinedthe question of heterogeneityamong studies and found strongevidence that the pedigrees studiedby Sherrington et al. (1988) differsignificantly from those studied bysubsequent investigators.The most recent addition to thiscontroversial story has been con-

tributed by the same investigative


16/25


team that reported the initial posi-tive finding (Mankoo et al. 1991).Their new results include both re-tests of the original seven ped-igrees using more informativemarkers and tests of additionalhigh-density pedigrees from bothIceland and England. In the origi-nal pedigrees, evidence for linkagehas diminished considerably al-though it has not disappeared en-tirely. Furthermore, the new fam-ilies show strong evidence againstlinkage for this region, and the en-tire combined sample excludeslinkage over the entire trisomicregion.

How can we explain the initialstrong positive evidence of linkage,followed first by uniformly nega-tive results in all independentstudies and then by an inability ofthe original investigators to repli-cate their own work? There arethree plausible explanations(McGuffin et al. 1990). First, theoriginal finding may represent ahighly unlikely Type I error. Howunlikely is not entirely clear, asthe original conceptual frameworkfor interpreting LOD scores wasconstructed for Mendelian condi-tions in which the presence of agene was assured and the modeof transmission was known. How-ever, it is clear that, even aftercorrecting for these effects, theoriginal LOD score above 6.0, if afalse positive, was very unlikely.Second, there may be a gene inthe 5qll-13 region that influencesliability to schizophrenia, but eitherit is of only modest effect or it ispresent in only a very small sub-set of families. This hypothesisalso seems unlikely, however, be-cause it is difficult to understandwhy the gene should be detectedin all or most of the seven fam-ilies in the initial study but essen-tially in very few or none of the

many other families studied byother investigators. This explana-tion also fails to account for theloss of most support, even in theoriginal families, upon reanalysiswith additional, more highly infor-mative marker loci. Third, the pos-sibility of a systematic source oferror in the original report mustbe acknowledged. Although somelimitations in the report by Sher-rington et al. (1988) have beennoted (Watt and Edwards 1991),there is no evidence of systematicbias that might arise, for example,if psychiatric diagnoses were per-formed with knowledge of markergenotypes. Thus, we remain in theunsatisfactory position of havingno adequate explanation for thispuzzling and discouraging seriesof events. Our own view is that areasonably common gene of majoreffect for the liability to schizo-phrenia almost certainly does notexist in the 5qll-13 region. How-ever, we do not yet possess datawith sufficient statistical power toconfidently rule out the existence,within this chromosomal region orat any other location in the humangenome, of either a quite raregene of major effect or a morecommon gene of minor effect onliability to schizophrenia.In addition to the 5qll-13 re-gion, linkage to schizophrenia hasbeen examined for several othercandidate regions (table 6). Tomake table 6 as comprehensive aspossible, we included citations towork thus far presented only asabstracts for meetings. Some ofthese studies are quite small insize and/or are preliminary, andwe caution that final reports mayoffer different conclusions fromthose indicated in the table. Per-haps of greatest interest are candi-date loci associated with thedopamine (DA) system, long con-

sidered to be centrally involved inthe pathophysiology of schizo-phrenia. Three studies (Byerley etal . 1991; Kennedy et al. 1991a;Wildenauer et al. 1991) examinedlinkage between schizophrenia andthe Dj locus located at 5q34-35,six studies (Byerly et al. 1991;M acciardi et a l. 1991; Mo ises et al.1991; Owen et al. 1991; Wilden-auer et al. 1991; Su et al. 1993)evaluated the D 2 locus on l lq,one study examined the D 3 recep-tor, and three linkage studies(Kennedy et al. 1991b; Macciardi etal . 1991; O'N eill et al. 1991) evalu-ated the region of the newly iden-tified D 4 locus on lip. All studiesconsidered only exclusion of link-age assuming homogeneity, exceptthe s tud y of Su et al. (Su 1991; Suet al. 1991, 1993), in which linkageto the D 2 locus was excluded forat least 50 to 75 percent of fam-ilies, depending on assumptionsused for testing various geneticmodels. Very weak support oflinkage was found for the D 2locus in one study (Macciardi etal . 1991). Another study (Owen etal . 1991) found results suggestiveof linkage on chromosome Ilq23when evaluating an additive modelof disease gene transmission in aregion proximal to the location ofthe D 2 DA receptor gene, but thissame region was excluded formost families for a nearly identicalgenetic model in a study reportedby Su et al. (Su 1991; Su et a l.1991, 1993). Evidence suggestive oflinkage has been reported forchromosome 2q in the one studythus far reporting results for thisregion (Aschauer et al. 1991), andone report noted weak evidence ofcosegregation of schizophreniawith the muscular dystrophy genelocated on chromosome Xp21 (Zatzet al. 1991). Linkage betweenschizophrenia and the serotonin
http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/


17/25

VOL 19, NO. 2, 1993 277

receptor located on chro-mosome 13, a possible receptor forneuroleptics, has also been ruledout in a large Swedish pedigree(Hallmayer et al. 1992).Some investigators suggestedthat affected siblings are moreoften of the same sex, especiallywhen schizophrenia appears to bepaternally transmitted (Crow 1988).This finding would be consistentwith a susceptibility gene locatedin the pseudoautosomal regions ofthe X and Y chromosomes (table6). Two studies (Collinge et al.1991; d'Amato et al. 1992) reportedsuggestive evidence of linkage todistinctly different locations in thischromosomal region while otherstudies (Nonomu ra et al. 1991;Wildenauer et al. 1991; Ashersonet al. 1992) were negative.Three groups (Barr et al. 1991;Byerly et al. 1991; Polymeropouloset al. 1991) reported preliminaryresults of genomewide searchesusing a very limited number offamilies (table 6). Thus far, nostrong positive evidence of linkagehas been found, which might notbe surprising given the likely het-erogeneity of this disorder and theneed to evaluate many families, asdiscussed above.Finally, brief mention should bemade of studies that provide weakevidence for the cosegregation oftranslocations involving severalchrom osom es (e.g., [2, 18], [1, 11],[6, 11], [9, 11]) with various formsof psychosis (Genest et al. 1976;Holland and G osden 1990; St.Clair et al. 1990; Wagner et al.1990). These studies may providehelpful leads in defining furtherpotential candidate regions forlinkage studies.In summary, no replicated posi-tive findings have yet emergedfrom efforts to locate individualgenetic loci that influence the lia-

bility to schizophrenia. This is notsurprising, nor should it be toodiscouraging. The candidate locithat we have for schizophrenia arerelatively weak in that evidence oftheir involvement in the patho-physiology of schizophrenia (i.e., asite of action of antipsychoticdrugs) is indirect at best. Thenumber of genes expressed in thecentral nervous system almost cer-tainly reaches the tens of thou-sands, only a minute fraction ofwhich have been identified. A pri-ori, if major genes exist for schizo-phrenia, it is more likely that theywill be among the large majorityof previously unknown genes thanamong the small minority of genesalready identified. Furthermore,from the perspective of exclusionmapping and with the possible ex-ception of ongoing studies of asmall number of families (Barr etal. 1991; Byerley et al. 1991; Poly-meropoulos et al. 1991), the pro-portion of the genome excludedunless we assume nearly completegenetic homogeneity for schizo-phreniaremains quite small. Ifone or more major genes forschizophrenia exist in the humangenome, it is unlikely that theywould have been identified by allcombined research efforts to date.Given the plausible scenario thatseveral such genes exist, somewith major effects and others hav-ing more minor influences, studiesto date must still be consideredpreliminary.Association StudiesWhile most interest has focused onlinkage analysis, association studiesrepresent another viable approach.Linkage examines the cosegrega-tion of markers and disease withinfamilies. Association studies usu-ally compare the frequency of in-

dividual genes (more technicallyalleles) in unrelated affected indi-viduals with that in matched con-trols. Association studies have twomajor advantages over linkagestudies in schizophrenia. First,finding large numbers of schizo-phrenia patients and controls ismuch easier and less costly thanfinding large numbers of high-density families. Second, in somecircumstances, genetic hetero-geneity may impair the statisticalpower of association tests less thanit does the power of tests forlinkage.However, compared with linkagestudies, association studies havetwo critical disadvantages. First,linkage studies examine general lo-cations in the genome while asso-ciation studies examine individualalleles (i.e., specific copies of agene that reside at a particular lo-cation). Although too technical todescribe here fully, associationstudies are most powerful for dis-eases in which many affected indi-viduals have inherited the samedisease gene dating back manygenerations to a very rare muta-tion event. For a genetic diseasethat has arisen from very manyindependent mutational events, as-sociation tests lose much of theirpower. Given the high frequencyof schizophrenia in the populationand the low reproductive success(i.e., fitness) of individuals affectedby this disorder, genetic theorywould suggest that mutationalevents leading to schizophreniashould be fairly common. Second,except for the case in which thesame DNA variation that causesthe disease simultaneously causesthe difference among marker al-leles, association works throughlinkage disequilibriumthat is, thetendency for alleles at differentplaces very close together on chro-


18/25


mosomes to be correlated in thepopulation. This occurs becausethe alleles are so close that cross-ing over (recombination) betweenthe neighboring locations occursvery rarely. The further back intime the mutational event oc-curred, the more crossing over willoccur in subsequent generationsand the smaller the chromosomaldistance will be over which link-age disequilibrium still exists.Linkage works on an entirely dif-ferent principle: the cosegregationof marker and disease geneswithin families.

For practical purposes, linkagecan detect a disease gene within5-10 centimorgans (cM) of themarker (where 1 cM equals a1-percent probability of a recom-bination in any single meiosis).Association studies, by contrast,can rarely detect a disease genefurther aw ay tha n 1 cM if that far.Therefore, for genetic diseases aris-ing from one or a few mutationalevents, association studies can bequite powerful if a marker is veryclose to the mutation. Conse-quently, association studies arelimited to tests of "candidategenes," which are suspected a pri-ori of having a possible etiologicrole in the disease owing toknown biochemical properties.When searching for disease genesnot yet characterized, however,linkage studies are almost certainlya superior method.

Associations of various polymor-phisms, including isozymes, blood-group antigens, serum proteins,and the HLA region, have gener-ally not bee n replicated (McGuffinand Sturt 1986; Saha et al. 1990),although sample sizes have notbeen very large for many of thesestudies. An association of bor-derline statistical significance wasfound with a DNA polymorphism

in the adenosine deaminase genein a small sample of both bipolarand schizophrenia patients com-pared with control subjects(Detera-Wadleigh et al. 1987). Re-cently, two studies (Sanders et al.1991; Diehl et al., submitted forpublication) found associationswith the porphobilinogen de-aminase (PBGD) gene located dis-tal to the D 2 DA receptor gene onchromosome llq. This gene is apotential candidate for schizo-phrenia because PBGD deficiencycauses acute intermittent porphyria,a rare disorder that, like schizo-phrenia, usually has a postpubertalonset and can present with psy-chotic symptoms. This disorder ismore common than expected inpsychiatric patients. One of thesestudies (Diehl et al., submitted forpublication) used the same sampleof schizophrenia and unaffectedfamily members to test for bothPBGD association and linkage tothis chromosomal region. Althoughmodestly significant evidence ofassociation was found, no positiveevidence of linkage was observed,and the PBGD map position wasexcluded for 60 to 80 percent ofthe families, depending on the ge-netic model parameters assumed.While the complexities of interpret-ing this example are beyond thescope of this review, it does sug-gest that, if the candidate gene ap-proach is viable, a combination ofboth linkage and associationmethods may be optimal for at-tempting to uncover the complex-ities of a disorder such asschizophrenia.

Conclusions and FutureDirectionsIn the more traditional areas ofpsychiatric genetics, our under-standing of the genetics of schizo-

phrenia has advanced modestly bysteady increments in the last sev-eral years. The question of thefamilial aggregation of narrowlydefined schizophrenia has beenrather definitively settled. Althoughit is possible that schizophrenia ismore familial in some populationsthan in others, this remains to befirmly established. The ongoingtwin and adoption studies continueto confirm, with ever-increasingconfidence, the major etiologic roleplayed by genetic factors in theetiology of schizophrenia. Ourknowledge of the boundary of theschizophrenia spectrum is cominginto sharper focus. However, link-age studies of schizophrenia haveyet to live up to their promise. In-stead, we have had to face thefrustration of nonreplication of ap-parently strong positive reports.

What might the future hold forthe genetics of schizophrenia?First, we caution against pre-maturely abandoning the moretried and true methods of psychi-atric genetic research and replacingthem entirely with linkage studies.Competition for funds and fornew investigators is inevitable in aworld of finite resources. It is truethat linkage studies provide thepossibility of great breakthroughsin our knowledge of the geneticsand etiology of schizophrenia atbasic biochemical and physiologicallevels that may never be ad-dressed by traditional methods.However, a complete understand-ing of schizophrenia, from DNA tophenotype, will undoubtedly re-quire a host of methods, includingthe traditional family, twin, andadoption studies, which many in-vestigators now consider to beoutdated. We reemphasize that, re-gardless of how sophisticated ourmolecular genetic and statisticalmethods become, chances of sue-
http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/http://schizophreniabulletin.oxfordjournals.org/


19/25

VOL 19, NO. 2, 1993 2 7 9

cess still ultimately depend on ourability to measure and classifypsychiatric phenotypes in a man-ner consistent with their underly-ing genetic basis.The best analogy might be thatof a stock portfolio. Family, twin,and adoption studies are low-risk,slow-growth, dependable invest-ments that will continue, at amodest speed, to provide increas-ing knowledge about the geneticsof schizophrenia. Linkage studiesare hot, new, high-risk investmentsthat might produce great break-throughs but also might stall oreven go bust. Individual investorswill hold different views as to theoptimal balance of these alternativestrategies, depending on their ob-jective assessment of the relativechances of success or failure andthe rewards associated with eachstrategy, as well as on their per-sonal comfort level in dealing withvarying levels of uncertainty. Mostinvestment counselors would sug-gest that a portfolio should includeat least some of both kinds of in-vestments. As a field, we woulddo well to follow such advice.

Second, it is critical that weavoid premature disillusionmentwith linkage studies of schizo-phrenia. The human brain is verycomplex and quite difficult to ac-cess, and schizophrenia is a com-mon and crippling condition. Oneof the very best hopes for ap-proaching a complete understand-ing of the pathophysiology of thisdisorder, which could lead totherapeutic options currently un-dreamt of, lies in the "positionalcloning" strategy (Collins 1992).For a complex disorder such asschizophrenia, this approach wouldmost likely begin with gene map-ping by linkage analysis. The ag-gregate results from twin andadoption studies allow us to con-

clude with some confidence thatgenes that influence liability toschizophrenia exist somewhere inthe human genome. The crucialquestions to which we do nothave answers are (1) How manysuch genes are there? (2) Howcommon are they? and (3) Howlarge are their individual effects? Ifthere are any relatively commongenes of moderate to large effect,we have a very good probabilityof detecting them reliably in moststudy populations if we perseverein our study of large samples andmaximize our statistical power todetect linkage under complexmodes of inheritance. If there arevery many genes, none of whichhas any more than a small effecton liability, current methods andprojected sample sizes are almostcertainly inadequate and will yieldnegative or unreplicated results.The path to replicated linkageresults for schizophrenia will prob-ably not be a smooth one. A pat-tern of tentative findings by onegroup not replicated by othergroups may be likely. It is impor-tant that we begin this undertak-ing well informed of the risks anddifficulties as well as of the possi-ble benefits. If immediate returnsare expected to come too easily,failure to fulfill this unrealistic ex-pectation may lead to withdrawalof support. Thus, as we previouslywrote:

Linkage analysis could join themany scientific approaches toschizophrenia which have beencharacterized by rapid andoverly enthusiastic endorsementby the psychiatric communityonly to be followed by disap-pointment and precipitous rejec-tion. [Kendler 1987, p. 31]To carry out a truly credible ex-ecution of the linkage strategy fora disease as complex and hetero-

geneous as schizophrenia, largenumbers of carefully diagnosedfamilies and highly informativemarkers are required. These re-sources are just now beginning tobe brought into action. While defi-nitely not offering a guaranteedsuccess, this approach, if allowedsufficient time to mature, couldyield truly unprecedented insightsinto the etiology of this disorder.

ReferencesAbrams, R., and Taylor, M.A. Thegenetics of schizophrenia: A reas-sessment using modern criteria.American Journal of Psychiatry,140:171-175, 1983.American Psychiatric Association.DSM -III: Diagno stic and StatisticalManual of Mental Disorders. 3rded. Washington, DC: The Associa-tion, 1980.American Psychiatric Association.DSM -III-R: Diagno stic and Statisti-cal Manual of Mental Disorders. 3rded., revised. Washington, DC: TheAssociation, 1987.Andrew, B.; Watt, D.C.; Gillespie,C; and Chapel, H. A study ofgenetic linkage in schizophrenia.Psychological Medicine, 17:363-370,1987.Aschauer, H.N.; Aschauer-Treiber,G.; Isenberg, K.E.; Todd, R.D.;Knesevich, M.A.; Garver, D.L.;Reich, T.; and Cloninger, C.R. Noevidence for linkage between chro-mosome 5 markers and schizo-phrenia. Human Heredity, 40:109-115, 1990.Aschauer, H.N.; Meszaros, K.;Aschauer-Treiber, G.; Willinger, U.;Fessl, D.; Chaudry, H.R.; Stompe,T.; Fuchs, K.; and Sieghart, W.RFLP-linkage studies of schizo-phrenia on chromosome 2. Psychi-atric Genetics, 2:12, 1991.


20/25


Asherson, P.; Parfitt, E.; Sargeant,M.; Tidmarsh, S.; Buckland, P.;Taylor, C; Clements, A.; Gill, M.;McGuffin, P.; and Owen, M. Noevidence for a pseudoautosomallocus for schizophrenia linkageanaly sis of m ultip ly affected fam-ilies. British Journal of Psychiatry,161:63-68, 1992.Baron, M.; Gruen, R.; Rainer, J.D.;Kane, J.; Asnis, L.; and Lord, A. Afamily study of schizophrenia andnormal control probands: Implica-tions for the spectrum concept ofschizophrenia. American Journal ofPsychiatry, 142:447-455, 1985.Barr, C.L.; Kennedy, J.L.; Pakstis,A.J.; Wetterberg, L.; Sjogren, B.;Gelernter, J.; Hallmayer, J.; Moises,H.W.; Cavalli-Sforza, L.L.; andKidd, K.K. Progress in genomescan for linkage in schizophrenia.[Poster] American Journal of HumanGenetics, 49(Suppl.):335, 1991.Byerly, W.; Plaetke, R.; Jensen, S.;Holik, J.; Hoff, M.; Myles-Worsley,M.; Wender, P.; Reimherr, F.; Lep-pert, M.; O'Connell, P.; Lalouel, J.;and White, R. Linkage analysis ofsix schizophrenia pedigrees with150 DNA markers. SchizophreniaResearch,4(Special Issue):274-275,1991.Cavalli-Sforza, L.L., and King,M.-C. Detecting linkage for genet-ically heterogeneous diseases anddetecting heterogeneity with link-age data. American Journal of Hu-man Genetics, 38:599-616, 1986.Chakravarti, A.; Badner, J.A.; andLi, C.C. Tests of linkage and het-erogeneity in Mendelian diseasesusing identity by descent scores.Genetic Epidemiology, 4:255, 1987.Chen, W.J.; Faraone, S.V.; andTsuang, M.T. Linkage studies ofschizophrenia: A simulation studyof statistical power. GeneticEpidemiology,9:123-139, 1992.

Clerget-Darpoux, F.; Babron, M.C.;and Bonaiti-Pellie, C.B. Power androbustness of the linkage homoge-neity test in genetic analysis ofcommon disorders. Journal of Psy-chiatric Research, 21:625-630, 1987.Collinge, J.; DeLisi, L.E.; Boccio,A.; Johnstone, E.C.; Lane, A.;Larkin, C; Leach, M.; Lofthouse,R.; Owen, F.; Poulter, M.; Shah, T.;Walsh, C ; and Crow , T.J. Evidencefor a pseudo-autosomal locus forschizophrenia using the method ofaffected sibling pairs. British Jour-nal of Psychiatry, 158:624-629, 1991.Collins, F.S. Positional cloning:Let's not call it reverse anymore.Nature Genetics, 1:3-6, 1992.Constantinidis, J.K. Les marqueursde chromosomes chez les schizo-phrenes et la recherche du linkageentre ces caracteres et la schizo-phrenie par la methode de Pen-rose. Journal de genetique humaine,7:189-242, 1958.Coryell, W., and Zimmerman, M.The heritability of schizophreniaand schizoaffective disorder: Afamily study. Archives of GeneralPsychiatry, 45:323-327, 1988.Crow, T.J. Sex chromosomesand psychosis: The case for apseudoautosomal locus. BritishJournal of Psychiatry, 153:675-683,1988.Crow, T.J. Nature of the geneticcontribution to psychotic illnessAcontinuum viewpoint. Ada Psychi-atrica Scandinavica, 81:401-408,1990.Crowe, R.R.; Black, D.W.; Wesner,R.; Andreasen, N.C.; Cookman, A.;and Roby, J. Lack of linkage tochromosome 5qll-ql3 markers insix schizophrenia pedigrees. Ar-chives of GeneralPsychiatry, 48:357-361, 1991.d'Amato, T.; Campion, D.; Gor-wo od, P.; Jay, M.; Sabate, O ; Petit,

C ; Abb ar, M.; Malafosse, A.;Leboyer, M.; Hillaire, D.; Clerget-Darpoux, F.; Feingold, J.; Waks-man, G.; and Mallet, J. Evidencefor a pseudoautosomal locus forschizophrenia. II: Replication of anon-random segregation of allelesat the DXYS14 locus. British Jour-nal of Psychiatry, 161:59-62, 1992.DeLisi, L.E.; Crow, T.J.; Davies,K.E.; Terwilliger, J.D.; Ott, J.; Ram,R.; Flint, T.; and Boccio, A. No ge-netic linkage detected for schizo-phrenia to Xq27-q28. British Jour-nal of Psychiatry, 158:630-634, 1991.Detera-Wadleigh, S.D.; de Miguel,C ; B errertini, W.H.; DeL isi, L.E.;Goldin, L.R.; and Gershon, E.S.Neuropeptide gene polymorphismsin affective disorder and schizo-phrenia. Journal of Psychiatric Re-search,21:581-587, 1987.Detera-Wadleigh, S.D.; Goldin,L.R.; Sherrington, R.; Encio, I.; deMiguel, C; Berrettini, W.; Gurling,H.; and Gershon, E.S. Exclusion oflinkage to 5qll-13 in families withschizophrenia and other psychiatricdisorders. Nature, 340:391-393,1989.Diehl, S.R., and Kendler, K.S.Strategies for linkage studies ofschizophrenia: Pedigrees, DNAmarkers, and statistical analyses.Schizophrenia Bulletin, 15:403-419,1989.Diehl, S.R.; Su, Y.; Murphy, B.;O'Neill, F.A.; Nie, L.; Burke, J.;Duke, B.W.; Kipps, B.R.; MacLean,C.J.; Walsh, D.; and Kendler, K.S."Association of PorphobilinogenDeaminase RFLPs With Schizo-phrenia Without Evidence of Link-age." Submitted for publication.Falconer, D.S. The inheritance ofliability to certain diseases, esti-mated from the incidence amongrelatives. Annals of Human Ge-netics, 29:51-76, 1965.

,

p

p

j

g


21/25

VOL 19, NO. 2, 1993 28 1

Frangos, E.; Athanassenas, G.;Tsitourides, S.; Katsanou, N.; andAlexandrakou, P. Prevalence ofDSM-HI schizophrenia among thefirst-degree relatives of schizo-phrenic probands. Ada PsychktricaScandinavica, 72:382-386, 1985.Genest, P.; Dumas, L.; and Genest,F.B. Translocation chromosomique t(2; 18) (q21; q23) chez un individu.schizophrene et sa fille. UnionMedicale du Canada, 105:1617-1681,1976.Gershon, E.S.; DeLisi, L.E.;Hamovit, J.; Nurnberger, J.I.;Maxwell, M.D.; Schreiber, J.;Dauphinais, D.; Dingman, C.W.;an d Guroff, J.J. A controlled familystudy of chronic psychoses. Ar-chives of GeneralPsychiatry, 45:328-336, 1988.Goldin, L.R.; DeLisi, L.E.; andGershon, E.S. Relationship of HLAto schizophrenia in 10 nuclearfamilies. Psychiatry Research,20:69-77 , 1987.Goldin, L.R., and Gershon, E.S.Power of the affected-sib-pairmethod for heterogeneous disor-ders. Genetic E pidemiology, 5:35-43,1988.Gottesman, I.I. Schizophrenia Gen-esis: The Origins of Madness. NewYork, NY: W.H. Freeman, 1991.Gottesman, 1.1., and Bertelsen, A.Confirming unexpressed genotypesfor schizophrenia: Risks in the off-spring of Fischer's Danish identicaland fraternal discordant twins. Ar-chives of GeneralPsychiatry, 46:867-872, 1989.Gottesman, I.I.; McGuffin, P.; andFarmer, A.E. Clinical genetics asclues to the "real" genetics ofschizophrenia: A decade of modestgains while playing for time.Schizophrenia Bulletin, 13:23-47,1987.

Gottesman, 1.1., and Shields, J.Contributions of twin studies toperspectives on schizophrenia. In:Maher, B.A., ed. Progress in Experi-mental Personality Research. Vol. 3.New York, NY: Academic Press,1966.Guze, S.B.; Cloninger, C.R.; Martin,R.L.; and Clayton, P.J. A follow-upand family study of schizophrenia.Archives of General Psychiatry,40:1273-1276, 1983.Hallmayer, J.; Kennedy, J.L.; Wet-terberg, L.; Sjogren, B.; Kidd, K.K.;and Cavalli-Sforza, L.L. Exclusionof linkage between the serotonin 2receptor and schizophrenia in alarge Swedish kindred. Archives ofGeneral Psychiatry, 49:216-219,1992.Holland, T., and Gosden, C. Abalanced chromosomal transloca-tion partially co-segregating withpsychotic illness in a family. Psy-chiatry Research,32:1-8, 1990.Kendler, K.S. Overview: A currentperspective on twin studies ofschizophrenia. American Journal ofPsychiatry, 140:1413-1425, 1983.Kendler, K.S. Diagnostic ap-proaches to schizotypal personalitydisorder A historical perspective.Schizophren ia Bulletin, 11:538-553,1985.Kendler, K.S. The feasibility oflinkage studies in schizophrenia.In: Helmchen, H., and Henn, F.A.,eds. Biological Perspectives ofSchizophrenia. Chichester, England:John Wiley & Sons, 1987. pp. 19-32.Kendler, K.S. The genetics ofschizophrenia: An overview. In:Nasrallah, H.A.; Tsuang, M.T.; andSimpson, J.C., eds. Handbook ofSchizophrenia: Nosology, Epidemiol-ogy, and Genetics. Vol. 3. Amster-dam, The Netherlands: Elsevier,1988. pp. 437-462.

Kendler, K.S., and Diehl, S.R. Thegenetics of schizophrenia. In:Kaplan, H.I., and Sadock, B.J., eds.Comprehensive Textbook of Psychia-try. Vol. 4. Baltimore, MD:Williams & Wilkins, in press.Kendler, K.S., and Gruenberg,A.M. An independent analysis ofthe Copenhagen sample of theDanish Adoption Study of Schizo-phren ia. VI: The pa ttern of psychi-atric illness as defined by DSM-IIIin adoptees and relatives. Archivesof GeneralPsychiatry, 41:555-564,1984.Kendler, K.S.; Gruenberg, A.M.;and Tsuang, M.T. Psychiatric ill-ness in first-degree relatives ofschizophrenic and surgical controlpatients: A family study usingDSM-HI criteria. Archives of Gen-eral Psychiatry, 42:770-779, 1985a.Kendler, KS.; Gruenberg, A.M.;and Tsuang, M.T. A DSM-IU fam-ily study of the non-schizophrenicpsychotic disorders. American Jour-nal of Psychiatry, 143:1098-1105,1986.Kendler, K.S.; Masterson, C; andDavis, K.L. Psychiatric illness infirst-degree relatives of patientswith paranoid psychosis, schizo-phrenia, and medical illness. Brit-ish Jou rnal of P sychiatry, 147:524-531, 1985b.Kendler, K.S.; McGuire, M.; Gruen-berg, A.M.; O'Hare, A.; Spellman,M.; and Walsh, D. The Roscom-mon Family Study. I: Methods, di-agnosis of probands and risk ofschizophrenia in relatives. Archivesof General Psychiatry, in press a.Kendler, K.S.; McGuire, M.; Gruen-berg, A.M.; O'Hare, A.; Spellman,M.; and Walsh, D. The Roscom-mon Family Study. HI: Schizo-phrenia-related personality disor-ders in relatives. Archives ofGeneral Psychiatry, in press b.


22/25


Kendler, K.S.; McGuire, M.; Gruen-berg, A.M.; O'Hare, A.; Spellman,M.; and Walsh, D. The Roscom-mon Family Study. IV: Affectiveillness, anxiety disorders and alco-holism in relatives. Archives ofGeneral Psychiatry, in press c.Kendler, K.S.; McGuire, M.; Gruen-berg, A.M.; Spellman, M.; O'Hare,A.; and Walsh, D. The Roscom-mon Family Study. II: The risk ofnonschizophrenic nonaffective psy-choses in relatives. Archives ofGeneral Psychiatry, in press d.Kennedy, J.L.; Bassett, A.S.; Siden-berg, D.G.; Van Tol, H.H.M.;Civelli, O.; Honer, W.G.; Collins,E.J.; Kamble, A.B.; and Barr, C.L.Linkage studies of the dopamineD 4 receptor gene and schizo-phrenia in eastern Canadian fam-ilies. [Poster] Psychiatric Genetics,2:81, 1991a.Kennedy, J.L.; Guiffra, L.A.;Moises, H.W.; Cavalli-Sforza, L.L.;Pakstis, A.J.; Kidd, J.R.; Castiglione,CM.; Sjogren, B.; Wetterberg, L.;and Kidd, K.K. Evidence againstlinkage of schizophrenia tomarkers on chromosome 5 in anorthern Swedish pedigree. Nature,336:167-170, 1988.Kennedy, J.L.; Honer, W.G.;Gelernter, J.; Wetterberg, L.;Kaufmann, C.A.; Niznik, H.B.;O'Dowd, B.F.; Civelli, O.; andKidd, K.K. Linkage studies of twonew genes of neuropsychiatric in-terest. [Poster] Schizophrenia Re-search, 4:280, 1991b.Kennedy, J.L.; Honer, W.G.; Mar-tignetti, J.; Macciardi, F.; Moreno,C ; Cavallini, M.C.; and Smeraldi,E. Antibody selection for candidategenes in schizophreniaLinkagestudies. Psychiatric Genetics, 2:31-32, 1991c.Kety, S.S. The significance of ge-netic factors in the etiology of

schizophrenia: Results from the na-tional study of adoptees in Den-mark. Journal of Psychiatric Re-search, 21:423-129, 1987.Kety, S.S.; Rosenthal, D.; Wender,P.H.; and Schulsinger, F. The typesand prevalence of mental illness inthe biological and adoptive fam-ilies of adopted schizophrenics.Journal of Psychiatric Research,6:345-362, 1968.Kety, S.S.; Rosenthal, D.; Wender,P.H.; Schulsinger, F.; and Jacobsen,B. Mental illness in the biologicaland adoptive families of adoptedindividuals who have becomeschizophrenic: A preliminary reportbased on psychiatric interviews. In:Fieve, R.; Rosenthal, D.; and Brill,H., eds. Genetic R esearch in P sychi-atry. Baltimore, MD: JohnsHopkins Press, 1975. pp. 147-165.Kety, S.S.; Wender, P.; Jacobsen,B.; Ingraham, L.J.; Kinney, D.K.;Schulsinger, F.; Rosenthal, D.;Jansson, L.; and Faber, B. "MentalIllness in the Biological and Adop-tive Relatives of Adopted Individ-uals with Schizophrenia. Phase II:The National Sample Outside ofCopenhagen." In preparation.Lannfelt, L.; Sokoloff, P.; Martres,M.-P.; Pilon, C ; Giros, B.; andSchwartz, J.-C. The dopamine D 3receptor gene and schizophrenia.Psychiatric Genetics, 2:16, 1991.lindelius, R., ed. A study ofschizophrenia. A clinical, prognos-tic, and family investigation. ActaPsychiatrica Scandinavica,216(Suppl.):l-125, 1970.Lowing, P.A.; Mirsky, A.F.; andPereira, R. The inheritance ofschizophrenia spectrum disorders:A reanalysis of the DanishAdoptee Study data. AmericanJournal of Psychiatry, 140:1167-1171,1983.

Macciardi, F.; Cavallini, M.C.; Ken-nedy, J.L.; Marino, C; Grandy, D.;Van Tol, H.H.M.; Niznik, H.B.;and Smeraldi, E. Testing the hy-pothesis of genetic linkage betweenschizophrenia and dopamine recep-tor genes. Psychiatric Genetics,2:17-18, 1991.MacLean, C.J.; Bishop, C.T.; Sher-man, S.L.; and Diehl, S.R. Distribu-tion of LOD scores under uncer-tain mode of inheritance. AmericanJournal of H uman Genetics, 52:354361, 1993.Maier, W.; Hallmayer, J.; Minges,J.; and Iichtermann, D. Affectiveand schizoaffective disorders: Sim-ilarities and differences. In: Mar-neros, A., and Tsuang, M.T., eds.Morbid Risks in Relatives of Affec-tive, Schizoaffective, and Schizo-phrenic Patients Results of a Fam-ily Study. New York, NY:Springer-Verlag, 1990. pp. 201-207.Maj, J.M.J.; Starace, F.; and Pirozzi,R. A family study of DSM-III-Rschizoaffective disorder, depressivetype, compared with schizophreniaand psychotic and nonpsychoticmajor depression. American Journalof Psychiatry, 148:612-616, 1991.Mankoo, B.; Sherrington, R.; Bryn-jolfsson, J.; Kalsi, G.; Perursson, H.;Sigmundsson, T.; Read, T.; Mur-phy, P.; Curtis, D.; Melmer, G.;and Gurling, H. New microsatellitepolymorphisms provide a highlypolymorphic map of chromosome5 bands qll.2-ql3.3 for linkageanalysis of Icelandic and Englishfamilies affected by schizophrenia.[Abstract] Psychiatric Genetics, 2:17,1991.Martinez, M.M., and Goldin, L.R.The detection of Linkage and het-erogeneity in nuclear families forcomplex disorders: One versus twomarker loci. American Journal ofHuman Genetics, 44:552-559, 1989.

,

p

p

j

g


23/25

VOL. 19, NO. 2, 1993 28 3

McGillivray, B.C.; Bassett, AS.;Langlois, S.; Pantzar, T.; andWood, S. Familial 5qll.2-ql3.3 seg-mental duplication cosegregatingwith multiple anomalies, includingschizophrenia. American Journal ofMedical Genetics, 35:10-13, 1990.McGue, M., and Gottesman, I.I.Genetic linkage in schizophrenia:Perspectives from geneticepidemiology. Schizophrenia Bul-letin, 15:453-464, 1989.McGue, M.; Gottesman, I.I.; andRao, D.C. The transmission ofschizophrenia under a multifac-torial threshold model. AmericanJournal of Huma n Genetics, 35:1161-1178, 1983.McGuffin, P.; Festenstein, H.; andMurray, R. A family study ofHLA antigens and other geneticmarkers in schizophrenia. Psycho-logical Med icine, 13:31-43, 1983.McGuffin, P.; Sargeant, M.; Hetti,G.; Tidmarsh, S.; Whatley, S.; andMarchbanks, R.M. Exclusion of aschizophrenia susceptibility genefrom the chromosome 5qll-ql3 re-gion: New data and a reanalysisof previous reports. American Jour-nal of Human Genetics, 47:524-535,1990.McGuffin, P., and Sturt, E. Geneticmarkers in schizophrenia. HumanHeredity, 36:65-88, 1986.Moises, H.W.; Gelemter, J.; Giuffra,L.A.; Zarcone, V.; WetterbeTg, L.;Civelli, O.; Kidd, K.K.; Cavalli-Sforza, L.L.; Grandy, D.K.; Ken-nedy, J.L.; Vinogradov, S.; Mauer,J.; Litt, M.; and Sjogren, B. Nolinkage between D 2 dopamine re-ceptor gene region and schizo-phrenia. Archives of General Psychi-atry, 48:643-647, 1991.Nonomura, Y.; Yoneda, H.; Ina-yama, Y.; Kono, Y.; Ishida, T.; andSakai, T. A study of pseudoauto-somal region in sib-pairs with

schizophrenia. [Poster] PsychiatricGenetics, 2:89, 1991.O'Neill, F.A.; Hugher, A.; Diehl,S.R.; and Kendler, K.S. Exclusionof the H-ras region from linkageto schizophrenia in Irish families.[Poster] Psychiatric Genetics, 2:91-92, 1991.Onstad, S.; Skre, I.; Edvardsen, J.;Torgersen, S.; and Kringlen, E.Mental disorders in first-degreerelatives of schizophrenics. AdaPsychiatrica Scandinavica, 83:463-467, 1991a.Onstad, S.; Skre, I.; Torgersen, S.;and Kringlen, E. Twin concordancefor DSM-Ul-R schizophrenia. AdaPsychiatrica Scandinavica, 83:395-401, 1991b.Ott, J. Analysis of Human GeneticLinkage. Revised ed. Baltimore,MD: Johns Hopkins UniversityPress, 1991.Owen, M.; Craufurd, D.; and St.Clair, D. Localization of a suscep-tibility locus for schizophrenia onchromosome 5. British Journal ofPsychiatry, 157:123-127, 1990.Owen, M.; Parfirt, E.; Sargeant, M.;Asherson, P.; McGuffin, P.; Taylor,C; Collier, D.; Powell, J.; Murray,R.; and Gill, M. Linkage studies ofschizophrenia with markers ofchromosome llq. [Poster] AmericanJournal of Hu man Genetics, 49(Suppl.):354, 1991.Polymeropoulos, M.H.; Delisi, L.E.;Poulter, M.; Lofthouse, R.; Crow,T.J.; Xiao, H.; Boccio, A.; Weber,J.L.; and Merril, C.R. Genetic link-age studies in schizophrenia using(CA)n repeat polymorphisms. [Ab-stract] Schizophrenia Research,4:283,1991.Pope, H.G.; Jones, J.M.; Cohen,B.M.; and Iipinski, J.F. Failure tofind evidence of schizophrenia infirst-degree relatives of schizo-phrenic probands. American Journal

of Psychiatry, 139:826-828, 1982.Risch, N. Linkage detection testsunder heterogeneity. GeneticEpidemiology, 6:473-480, 1989.Risch, N. Genetic linkage andcomplex diseases, with special ref-erence to psychiatric disorders. Ge-netic Epidemiology, 7:3-16, 1990.Saha, N.; Tay, J.S.H.; Tsoi, W.F.;and Kua, E.H. Association ofDuffy blood group with schizo-phrenia in Chinese. GeneticEpidemiology, 7:303-305, 1990.Sanders, A.R; Hamilton, J.D.;Fann, W.E.; and Patel, P.I. Associa-tion of genetic variation at theporphobilinogen deaminase genewith schizophrenia. [Poster] Ameri-can Journal of Human Genetics, 49(Suppl.):358, 1991.Scharfetter, C, and Niisperli, M.The group of schizophrenias,schizoaffective psychoses, andaffective disorders. SchizophreniaBulletin, 6:586-591, 1980.Sherrington, R.; Brynjolfsson, B.;Petursson, R; Porter, M.; Dudle-ston, K.; Barradough, B.; Wasmuth,J.; Dobbs, M.; and Gurling, H. Lo-calization of a susceptibility locusfor schizophrenia on chromosome5. Nature, 336:164-167, 1988.Spitzer, R.L.; Endicott, J.; and Gib-bon, M. Crossing the border intoborderline personality and bor-derline schizophrenia. Archives ofGenera l Psychiatry, 36:17-24, 1979.Spitzer, R.L.; Endicott, J.; andRobins, E. Research Diagnostic Cri-teria: Rationale and reliabil

Date post:	03-Apr-2018
Category:	Documents
Upload:	sebastian-silva-soto
View:	222 times
Download:	0 times

ESQUIZOFRENIA REVIEW

Documents