British Journal of Psychiatry (1990), 157, 273—275 Brief Reports

Is Parental Age Related to the Risk of Alzheimer'sDisease?

ALBERTHOFMAN, CORNELIAM. van DUIJN, WILLEMSCHULTE,THEUNISA. TANJA,
ROBERT HAAXMA, ARIE J. LAMERIS and ROLF J. SAAN

Advancedmaternalandpaternalagewereinvestigated asputativeriskfactorsforADin 198

clinicallydiagnosedAlzheimerpatientsandin 198 randomlyselectedhealthycontrols.No
significantdifferencesinaverageageof fathersandof mothersat birthof the subjectwere
observed. The risk of AD was not significantly different across categories of maternal and
paternalage.Theassociation
with parentalagewas notdifferentfor sporadicandfamilial
AD. These findingsdo not supportthe view of a maternalor paternalage effect on AD.

Advanced parental age has been suggested as a usingthe clinicalhistory, a neurologicalexamination,and

putative risk factor for Alzheimer's disease (AD)
because of its clinical and pathogenetic links with
Down's syndrome. However, studies of this subject
have yielded equivocal results. In four studies the
offspring of old mothers had a higher risk of
developing the disease than those of young mothers
(e.g. Cohen et a!, 1982; Amaducci et a!, 1986) but
this was not the case in eight others (e.g. De
Braekeleer Ct a!, 1988; Jouan-Flahault et a!, 1989).
In the largest study, with 116 clinically diagnosed
Alzheimer's patients and 97 population controls,
maternal age over 40 years was associated with a
nearly five-times increased risk of AD (Amaducci et
a!, 1986).A recentadditionalanalysisof that study
showed that the increase in risk was restricted to
sporadic Alzheimer patients (those with no family
history of AD); in cases with familial AD the relative
risk was not elevated (Schoenberg et a!, 1988). We
report a study of the putative association of AD and
parental age in 198 Dutch Alzheimer patients who
were compared with 198 population controls.
neuropsychological and laboratory tests (McKhann et a!,
1984).The diagnosisof AD wasmade in patientsfulfilling
the following criteria:(a) slow progressivedecline of
intellectual function; (b) Clinical Dementia Rating scale
score of >0.5 (Hughes et a!, 1982); (c) Short Portable
Mental Status Questionnaire (SPMSQ) score of <20
(Pfeiffer, 1975); (d) Hachinsid scale score of @7
(Hachinski
et a!, 1974);and(e)noevidence
for abnormalities
other
than cerebral atrophy, on a computerised tomography (CT)
and no evidence for focal dysfunctionon an electro
encephalograph (EEG). Ofthe 278 patients brought to our
attention, 201 satisfied these criteria, and in 198, data on
putativeriskfactorsfor AD wereobtained.
Controls
Foreachpatienta referencesubjectwasrandomlyselected
from the municipalpopulation register, matched for age
(within 5-year age-groups) and gender, and living in the
same municipality as the patient at the time of diagnosis.
All controls were apparently healthy, with a normal
cognitive function and an SPMSQ score of @20.

Data collection and analysis

Method
The age of the parentsat birthof the Alzheimerpatient
Theagesof themotherandfatheratbirthof 198Alzheimer and the birth order of the patient and his/h
patients and 198age-and sex-matchedhealthypopulation obtained from the next of kin (in generalthe wife or
controls were assessed,as part of a case-controlstudy of husband, or a child), as part of a structured interview on
Alzheimer's disease in the Netherlands.
risk factors for AD. The parental This study
age in the control subjectshas been
described in detail elsewhere (Hofman eta!, 1989). In the was also obtained in the same way. Reliable information
patients the diagnosisof AD was made before the age of on maternal age and birth order was obtained in 184patients
70 years. and controls, and on paternal age in 160
controls. The data were checked either by using the
Patients municipal population register or informa
family member. Full information was collected on dementia
Allpatientsin whomtheclinicaldiagnosisof AD wasmade in the family.Onlythose withat leastone first-degreerela
in theperiodof January1980to July 1987,whowerenot tivewithdementiawereconsideredto havea familialAD.
yet 70 years at the time of diagnosis, and who lived in the Thestrengthof theassociationof maternalandpaternal
fournorthernprovincesof theNetherlandsorin theregion age and AD was assessedby computationof the relative
of the city of Rotterdam,wereeligiblefor thisstudy.All risk (odds ratio), with ‘¿ parents
younger than 25 years at
patientsintheseareaswereseenbythreeof theauthors(WS, birth of the patient or control' as the reference category.
TAT and RH). Dementiasother than AD were excluded The analysis was done for all cases and then separately

273
274 HOFMAN ET AL

TABLE I
Age ofparents at subject's birth and relative risk for AD in all patients and those with sporadic AD
Age of parent at
subject's birth:yearsMothersFathersNo.
Cf3patientscontrolspatientscontrolsAll
ofNo. ofRR'95% C!3No. ofNo. ofRR'95%

patients245246l.0@2721l.0@25—2954520.90.5—1.650420.90.5—1.930—3445450.90.5—1.643450.70.4—1.535—3918290.60.3—1.118320.40.2—1.040+15121

AD@242824l.0@1312l.0@25—2925270.80.4—1.725221.10.4—2.830—3424240.90.4—1.921201.00.4—2.635—3910120.70.3—1.910130
with sporadic

1. Relative risk (RR) adjusted for age, sex and area of residence.
2. Reference category.
3. 95!. confidence intervals.

TABLE II
Differences in mean (standarderror ofmean) age ofparents at birth ofpatients andcontro!s byfarnily history of dementia

Family historyof age:yearspatientscontrolsdifference95%

mothers'age:yearsMeon fathers'
dementiaMean
CI(s.e.m)(s.e.m)(s.e.m)(s.e.m)(s.e.m)(s.e.m)Yes28.4 CIpatientscontrolsdifference95'!.

3.6No28.8 (0.7)29.6(0.6)1.2 (0.9)—0.6,3.031.1 (0.9)32.5(0.8)1.4 (1.2)—

1.0,
3.8Total28.6 (0.7)29.1 (0.6)0.3(0.9)—
1.5,(0.7)32.0
2.130.4(0.8)1.6(1.1)—0.6,
(0.5)29.3
(0.5)0.7 (0.7)—0.7,2.130.7 3.1
(0.5)32.2 (0.6)1.5(0.8)—0.1,

for sporadiccases (no first-degreerelativeswith Alzheimer's significant differences in averageages for mothers and fathers
disease) only. The relative risk is presented with a 95% at birth of the subjectwereobserved(TableII). A separate
confidence interval (95% CI). The matching variables (age, analysis showed no significant differences in birth order
genderand area of residence)were taken into account between Alzheimer patients and controls. This applied both
by entering them into a model for logistic regression to those with and those without a family history of AD.
(Schlesselman, 1982). The reported relative risks are based
on the regression coefficient yielded by this model. A Discussion
separate matched-pair case-control analysis showed similar
results. In an additional analysis, the mean age of the There is much evidence to suggest that genetic factors
motherand fatherat birthof the subjectwas compared play a part in the aetiology ofAD. In epidemiological
for patients and controls. studies, strong familial aggregation of Alzheimer
patients has been observed, and recently, genetic
Resufts linkage between AD and marker loci on chromosome
21 has been reported (St George-Hyslop et a!, 1987).
The risk of AD was not significantly different across There is also evidence for clinical and neuro
categories of maternal and paternal age at birth of the
subject (Table I). The relative risk for maternal age of 40 pathological links between AD and Down's syndrome
years or over was 1.1 (95°/s CI 0.5—2.6). In sporadic (non (Anonymous, 1987). These observations have led to
familial) Alzheimer patients the risk was also not different the hypothesis that advanced parental age may cause
over strata of parental age (Table I). The relativerisk for abnormalities of chromosome-21 , which may lead
maternal age of 40 years or over was 0.9 (0.3—2.8). No to AD.
 PARENTAL AGE AND ALZHEIMER RISK 275
This study fails to confirm previous findings of References
a maternal age effect on the risk of AD. The average AMADUCCI, L. A., Fa&nouoM, L., Rocc@, W. A., el a! (1986) Risk
age of the parents at birth of the subjects was, if factors for clinically diagnosed Alzheimer's disease: a case
anything, slightly lower in the patients than in the control study of an Italianpopulation. Neurology, 36, 922-931.
controls. This study differs from others in that it is Ario@imous (1987) Alzheimer's disease, Down's syndrome, and
chromosome 21. Lancet, i, 1011—1012.
considerably larger, and employed randomly selected COHEN, D., EISDORFER,C. & LEVERENZ,J. (1982) Alzheimer's
population controls. disease and maternal age. Journal of the American Geriatric
Two methodological issues have to be discussed. Society, 30, 656—659.
Firstly to reduce misclassification of Alzheimer Da Ba@sxsiian, M., FRODA,S., GANTRIN, D., at a! (1988) Parental
age and birth order in Alzheimer's disease: a case-control study
patients as much as possible we appliedthe diagnostic in the Saguenay-Lac-St-Jeanarea (Quebec, Canada). Canadian
criteriasuggestedby McKhannet a! (1984) rigorously Journal of Neurology and Science, 15, 139—141.
and restricted our study to patients in whom HAcWNsKI, V. C., LASSEN,N. A. & MARSHELL,J. (1974) Multi
the diagnosis was made before the age of 70 infarct dementia: a cause of mental deterioration in the elderly.
Lancet, ii, 207—209.
years. However, we cannot rule out that some HOFMAN, A., SCHULTE,W., TANJA, 1. A., ci a! (1989) History of
misclassification has occurred and as in other dementia and Parkinson's diseasein first-degreerelativesof
studies this tends to dilute a true difference patients with Alzheimer's disease. Neurology, 39, 1589-1592.
in maternal age between the study groups. HUGHES, C. P., BERG, L., DOMZJGER, W. L., ci al (1982) A new
clinical scale for the staging of dementia. British Journal of
Secondly, this study was performed in early-onset Psychiatry, 140, 566—572.
Alzheimer patients. We do not know to what JOUAN-FLAHAULT, C., SEROUSSI, M. C. & COLVEZ, A. (1989)
extent the present findings apply to AD with a later Absence de liaison entre démence senile et age parental. Revue
onset. d'Epidémiologie cide SantePublique, 37,73—76.
McKii@, G., Dx,@cia.w4,D., FoLsmIN, M., ci a! (1984) Clinical
In conclusion, we suggest that parental age at diagnosis of Alzheimer's disease: report of a NINCDS-ADRDA
birth of the subject is not likely to be a risk factor Work Group. Neurology, 34, 939-944.
for AD. PFEIFFER, E. (1975) A short portable mental status questionnaire
for the assessment of organic brain deficit in elderly patients.
Journal oftheAmericanGeriatric Society, 23,433—441.
SCHLESSELMAN, J. J. (1982) Case-Control Studies. New York:
Acknowledgement Oxford University Press.
SCHOENBERG,B. S., RoccA, W. A., FRATIGLIONI,L., ci a! (1988)
The authors thank Verona Otten for her contribution to data Late maternal age as a risk factor for sporadic and familial
collection,
andDrsH.A.Valkenburg andH.W. ter Haarfortheir Alzheimer's disease. Neurology, 38, 311.
advice.
Thisstudy wassupported bytheSOOM foundation; itis ST GEORGE-HYSLOP, P. H., T@zi, R., POL1NSXY, R. J., ci a! (1987)
part of the Eurodem EC Concerted Action on the Epidemiology The genetic defect causing familial Alzheimer's disease maps on
of Dementia. chromosome 21. Science, 235, 885-890.

°A.Hofman, @,w,PhD,Professor of Epidemiology, Department of Epidemiology and BiostatLctics,Erasmus

University
Medical School, P0 Box 1738,
The Netherlands;
3000 Cornelia
DR Rotterdam,M. van Duijn,
MSc, Resident in Epidemiology; Willem Schulte, MD, Social Geriatrician; Teun A. Tanja, MD, General
Practitioner; Robert Haaxma, MD, PhD, Lecturer in Clinical Neurology; Arie J. Lameris, MD, General
Practitioner; Rolf J. Sun, MSc, Reader in Neuropsychology

•¿ Co@espondence

British Journal of Psychiatry (1990), 157, 275—278

The Effects of Chronic Lithium Treatment on

Psychomotor Performance Related to Driving
SIMON HATCHER,RUTH SIMS and DAVID THOMPSON

A groupof 16 psychiatricout-patientsin remission,who had beentaking lithiumcarbonate

as their sole medicationfor at least three months, were comparedwith a control group of