American Journal of Epidemiology
Copyright © 1998 by The Johns Hopkins University School of Hygiene and Public Health
All rights reserved
Vol. 148, No. 5
Printed in U.S.A.
Influence of Consanguinity and Maternal Education on Risk of Stillbirth and
Infant Death in Norway, 1967-1993
Camilla Stoltenberg,1 Per Magnus,1 Rolv Terje Lie,2 Anne Kjersti Daltveit,2 and Lorentz M. Irgens2
To analyze the influence of consanguinity and maternal education on stillbirth and infant death for children
born in Norway between 1967 and 1993, the authors studied 7,274 children of ethnic Pakistani origin and
1,431,055 children of Norwegian ethnic origin. Of these children, 31.0% of the Pakistani children and 0.1 % of
the Norwegian children had parents who were first cousins. Consanguinity increased the relative risk of
stillbirth (odds ratio (OR) = 1.3, 95% confidence interval (Cl) 1.0-1.6) and infant death (OR = 2.4, 95% Cl
2.0-3.0), after adjustment for maternal education, maternal age, parity, and year of birth. In 1985-1993, 29%
(95% Cl 13-45) of stillbirths and infant deaths among the Pakistani group were attributable to consanguinity.
In the Norwegian group, 17% (95% Cl 13-21) of the deaths were attributable to factors associated with low
maternal education, while in the Pakistani group, the corresponding estimate was nonsignificant. The risks of
stillbirth and infant death were similar for children with non-consanguineous parents in both populations. This
is an important observation considering the differences in socioeconomic status between the two groups. The
authors conclude that consanguinity influenced stillbirth and infant death independent of maternal education,
and that a large proportion of deaths could be attributed to consanguinity in the Pakistani group due to high
frequencies of consanguinity. Am J Epidemiol 1998; 148:452-9.
consanguinity; education; ethnic groups; fetal death; genetics; infant death; social class
Consanguinity, defined as marriage between close
relatives, increases the risk of stillbirth and infant
death (1-3), but the extent to which socioeconomic
conditions can account for this association is disputed
(2, 4, 5).
In 1987, a meta-analysis of studies on prereproductive death (before 20 years of age) and consanguinity
(2) concluded that the proportion of deaths attributable
to consanguinity was low in all the populations considered. Since the publication of that meta-analysis, it
has become clear that the attributable risk of early
death due to consanguinity seems to be increasing
globally. Consanguinity appears to be more prevalent
than was estimated about 10 years ago (6-18), and in
many populations where consanguinity is prevalent,
the frequency of consanguinity does not decline (6,
18). Several of the populations with high rates of
consanguineous marriages now experience decreasing
or low rates of mortality due to social and economic
development and migration to wealthier countries, and
it has been postulated that the relative influence of
consanguinity may be larger when mortality rates are
low (2). Consanguinity influences stillbirth and infant
mortality mainly by increasing the frequency of autosomal recessive diseases (1). In addition, a slight increase in the frequency of diseases with a polygenic
etiology is considered to play a role (1).
British and Norwegian studies have reported higher
mortality rates among children of Pakistani immigrants than in other ethnic groups (19-21). Parental
consanguinity has been found to be related to these
mortality differences (19, 21), but there has been little
attempt to control for the influence of social class,
even though it is established that in most populations
consanguinity is more prevalent among parents with
low socioeconomic status (8, 22-26).
The aims of this study were to estimate the risk of
stillbirth and infant death in Norway among children whose parents were ethnic Norwegians or ethnic
Pakistanis in order to: 1) estimate the relative risks and
population attributable risks associated with consanguinity and maternal education within each population
group, 2) analyze whether consanguinity is a risk
Received for publication March 24, 1997, and accepted for publication February 19, 1998.
Abbreviations: AR, attributable risk; Cl, confidence interval; F,
coefficient of inbreeding—proportion of autosomal loci predicted
from pedigree analysis to be homozygous though inheritance of
identical genes from common ancestors; OR, odds ratio.
1
National Institute of Public Health, Oslo, Norway.
2
Medical Birth Registry of Norway, University of Bergen, Bergen,
Norway.
Reprint requests to Dr. Camilla Stoltenberg, Section of Epidemiology, Department of Population Health Sciences, National Institute
of Public Health, P.O. Box 4404 Torshov, 0403 Oslo, Norway.
452
Stillbirth, Infant Death, and Parental Consanguinity
factor for stillbirth and infant death after socioeconomic factors are controlled for, and 3) explore
whether consanguinity and differences in socioeconomic factors, such as maternal education, explain risk
differences between the two populations.
MATERIALS AND METHODS
The study population consisted of the 7,494 children
who had two parents of ethnic Pakistani origin and the
1,448,766 children who had two parents of ethnic
Norwegian origin and who were born in Norway between 1967 and 1993. For simplicity, the two groups
will be referred to as "Pakistani" and "Norwegian,"
regardless of citizenship and that none of the children
were immigrants. Births missing information on
whether the child was still alive or not by December
31, 1993 (11 births in the Norwegian group), on parental consanguinity (13,153 births; 0.9 percent of the
Norwegians, 2.6 percent of the Pakistanis), and on
parity (5,773 births; 0.4 percent of the Norwegians and
0.6 percent of the Pakistanis) were excluded. The final
sample consisted of 7,274 Pakistani births and
1,431,055 Norwegian births. Information on survival
of the child, consanguinity between parents, maternal
age, parity, and year of birth was obtained from the
Medical Birth Registry of Norway (27), which has
collected data on all births from 1967 and onward.
Data for parental country of origin and maternal educational level obtained by 1990 were linked to the
453
Birth Registry by Statistics Norway. Maternal educational levels for the total population were obtained
from a census conducted in 1990. The categorization
and distributions for the variables are shown in table 1.
The outcomes were 1) stillbirth rate—number of
stillborn infants with gestational age S16 weeks per
1,000 births, and 2) infant mortality—number of
deaths in the first year of life per 1,000 live births.
Early mortality is the sum of stillbirths and infant
deaths per 1,000 births. Parental relations were categorized as "not consanguineous," "first cousins or
closer," or "other consanguineous relations." In the
Norwegian group, most of the parents assigned to the
category with "other consanguineous relations" were
related in more distant ways than first cousins, while,
in the Pakistani group, most of them were related in an
unspecified way (Appendix table). Consanguinity is
routinely recorded on a standardized form when the
mother undergoes her first pregnancy control, usually
before 12 weeks of gestation.
Because risk estimates were similar for the two
consanguineous categories in the Pakistani group, data
from first cousins and other relations were pooled for
some analyses. Mothers' education in years was used
as a measure of socioeconomic status. The proportion
of children with consanguineous parents was larger
among those whose mothers' educational level was
less than 10 years or missing, compared with those
with higher educational levels (table 2).
TABLE 1. Number of births, stillbirths, and infant deaths by parental consanguinity, maternal education, maternal age, parity,
and year of birth for children born in Norway, 1967-1993, with ethnic Norwegian or ethnic Pakistani parents
Births
variable
Norwegian
No. Of
infant deaths
%
No.
%
Norwegian
Pakistani
Norwegian
Pakistani
1,431,055
100.0
7,274
100.0
16,573
100
13,367
92
1,415,421
1,983
13,651
98.9
0.1
1.0
4,339
2,254
681
59.7
31.0
9.4
16,379
34
160
54
36
10
13,144
60
163
37
41
14
276,136
807,246
313,900
33,773
19.3
56.4
21.9
2.4
357
860
1,467
4,590
4.9
11.8
20.2
63.1
2,633
8,886
4,440
614
5
11
17
67
1,878
7,210
3,809
470
1
11
22
58
83,510
940,497
407,048
5.8
65.7
28.4
192
4,770
2,312
2.6
65.6
31.8
1,123
9,758
5,692
4
60
36
1.253
8,555
3,559
3
58
31
594,305
836,750
41.5
58.5
1.887
5,387
25.9
74.1
7,232
9,341
24
76
5,530
7,837
20
72
550,160
424,839
456,056
38.4
29.7
315
318
2,737
4,219
4.4
37.6
58.0
7,608
4,583
4,382
6
40
54
6,845
3.602
2.920
5
37
50
No.
Total
Parental consanguinity
Not consanguineous
First cousins or more closely related*
Other consanguineous relations!
Mother's education (years)
>12
10-12
<10
Missing
Maternal age (years)
<20
20-29
£30
No. of
stillbirths
Pakistani
Partfv
ramy
1
£2
Year of birth
1967-1975
1976-1984
1985-1993
* A total of 93 Norwegian and six Pakistani children had parents who were more closely related than first cousins.
t A total of 3,371 relatives in the Norwegian group and 344 in the Pakistani group were registered as unspecified relatives; the remaining relatives were more
distant than first cousins.
Am J Epidemiol
Vol. 148, No. 5, 1998
454
Stoltenberg et al.
TABLE 2. Proportions (in percent*) of ethnic Norwegian and ethnic Pakistani births by parental consanguinity and maternal
education: births in Norway, 1967-1993
Parental
consanguinity
Not consanguineous
First cousins or more closely related
Other consanguineous relations
Total
Norwegian births,
by maternal education (years)
Pakistani births,
by maternal education (years)
>12
10-12
<10
Missing
>12
10-12
<10
Missing
99.3
0.1
0.7
100.0
98.9
0.1
0.9
100.0
98.5
0.3
1.2
100.0
98.6
0.4
1.0
100.0
77.0
17.1
5.9
100.0
70.3
21.9
7.8
100.0
53.0
36.4
10.6
100.0
58.4
32.0
9.5
100.0
Chi-square (Pearson)
Degrees of freedom
Significance
1,121
6
0.000
117
6
0.000
• Numbers are given in table 1.
Statistical methods
Multivariate logistic regression analyses were conducted (28, 29) to estimate relative risks of stillbirth,
infant death, or early death. Relative risks were calculated as odds ratios with 95 percent confidence intervals according to ethnic group, parental consanguinity,
and maternal educational levels. Odds ratios were adjusted for maternal age, parity, and year of birth. All
models were tested for significant two-way interactions (28, 29). Population attributable risks (AR) of
early death for consanguinity and low maternal education were calculated for the latest period, 1985—
1993, using the formula
AR = (p(RR - l))/(p(RR - 1) + 1),
where p is the prevalence of the risk factor (consanguinity or maternal education), and RR is the relative
risk associated with the factor (30). The relative risks
were approximated by odds ratios obtained from logistic regression models. We used a Monte Carlo
resampling technique to obtain confidence intervals
for the attributable risks. For a particular estimate of
the attributable risk, the estimated exposure prevalence (p) and the log odds ratio was used with their
estimated standard errors, assuming normal distributions, to produce 5,000 independent pairs of p and the
odds ratio. The formula above was then used to obtain
5,000 replicates of the attributable risk. The standard
error of the attributable risk obtained from this set was
then used to calculate 95 percent confidence intervals.
RESULTS
Consanguinity
A nonsignificant increased risk of stillbirth was observed in infants of consanguineous parents (odds ratio
(OR) = 1.3, 95 percent confidence interval (CI) 1.01.6). This pattern was similar for the Norwegian and
the Pakistani groups (table 3). The relative risk of
infant death for first cousin parents from both populations was 2.4 (95 percent CI 2.0-3.0). Again, the
pattern was similar for each of the two populations,
although the relative risk of infant death was nonsignificantly higher for Norwegian first cousin parents
than for the Pakistanis (table 3). In addition, an increased infant mortality was associated with other
consanguineous relations among the Pakistanis. The
effects of consanguinity were independent of maternal
education, maternal age, parity, and year of birth.
Maternal education
Among Norwegians, a significant negative association was found between maternal education and the
risk for stillbirth and infant death (table 3). For the
Pakistanis, there was a nonsignificant high risk of
infant mortality among infants whose mothers had low
or missing information on education. Effects due to
maternal age and parity slightly modified the influence
of maternal education on infant death in the Norwegian group.
Comparison between the Pakistani and the
Norwegian group
The Pakistani group as a whole experienced higher
mortality than the Norwegian group, with relative
risks of 1.3 for stillbirth and 1.7 for infant death after
adjustment for year of birth (table 4; model 1). There
were no significant risk differences between the Norwegian and Pakistani groups after adjusting for consanguinity (model 2). The results presented in table 3
indicate that, in the Norwegian group, maternal education had an effect on stillbirth and infant death. In
the Pakistani group, by contrast, maternal education
had no effect on stillbirth and only a possible effect on
infant death. However, adjustment for maternal education was made under the assumption that maternal
education had similar effects in the two populations
Am J Epidemiol
Vol. 148, No. 5, 1998
irental
Stillbirth, Infant Death, and Parental Consanguinity
O
S.
o>
(model 3). The difference in the risk of stillbirth disappeared after adjustment for maternal education, and
the difference in the risk of infant death was reduced.
Maternal age and parity were associated with the risk
of death, but the effects of these variables did not
explain risk differences between the two populations
(model 4). After adjusting for both consanguinity and
maternal education, no risk differences between the
groups were found (models 5 and 6).
co oo oo
o d d
.0"
in
a.
3
cp CE
2
°"
4orw egi
5
o
co
»-
I
o
»-
»- Tt
c\i oi
o ^r in cq
^ rt rt m
o>
oo»^- CM
n n in n
^ ^ ^
•S
Population attributable risk
eth
Infa
u
'£
o
During the study period, stillbirth and infant mortality decreased in both groups, and maternal educational levels increased considerably in the Norwegian
group. Population attributable risks were therefore
presented for 1985-1993 to illustrate the present situation better. In this period, the relative risk of early
death due to consanguinity was approximately 2.0 in
both populations (table 5). However, 29 percent of
early deaths could be attributed to consanguinity in the
Pakistani group compared with 0.1 percent in the
Norwegian population. The attributable risk associated
with <13 years of maternal education accounted for
17 percent of early deaths in the Norwegian group, but
was not significant for the Pakistani group (table 5).
•u
c
a
in
0)
'c
c
lie Pa kis
5
•6>
1§
O
O) CVI
O M K) N
»-'
C\j ^
T^ T^ ^
«-c\i
t^rocMM-
Y-^
c
£oo
£
3
cc
cn
cn
CO
•-
y-
f-
p
•*
iv- c\i in
cji cji
d) ci
oi oi cvi
co co 4
2
odd
2
c
£a
a
•a
o
tilIbi h and infa
c
kistani
do
at
r
p
^
CE
O
co w
v- ^
p cn oq p
^ d d ~
I
CD
II
O
cn
o
"^
ffl
"S
CE
£
O,
E
in
CO C O
S
ST
o o
3O
?
m
Z
c
~
cq CD q
O
igian
nfiden
way,
Ii
cU
i
d d
i
Ig
o
Rate
witl
:birl
o
3
S
.5
c
CD
J
u
o a
CD
II
to c
n
.2.
. 3
CO
0)
t
O) 0)
c - c
o
u
eo
1
SB
CO
CO
O
co
i
3
T3
CD
8 ».
2"
s ?r s K ~ "7 ° V, ^
:£'
28
Am J Epidemiol
lit
o.E
£c
co
c
o
dd
ter
•i5
"J
2
«
evi in cn
11
II
CO
m
E o co
en »
^*
455
Vol. 148, No. 5, 1998
DISCUSSION
The strength of this study is that it was populationbased, included two well-defined subpopulations with
different histories and frequencies of consanguinity,
and included a large number of births from consanguineous parents. The most prominent findings were:
1) Consanguinity influenced the risk of early death
independent of maternal education.
2) Consanguinity was similarly associated with early
death in the two populations despite different frequencies and histories of consanguinity.
3) Mortality rates were the same for children with
non-consanguineous parents in both populations, even
though the Pakistani parents originate from a country
with high mortality rates and have lower socioeconomic status in Norway.
4) High frequencies of consanguinity had a large
impact, measured as attributable risk, on public health
in a population with low mortality rates.
5) In agreement with other studies (2, 7), consanguinity influenced stillbirth only slightly, while the risk of
infant death increased significantly in both populations.
These findings are of general interest because the
results for the Pakistani population in Norway may
reflect similar trends for other populations where consanguinity is favored and early mortality is declining.
456
Stoltenberg et al.
TABLE 4. Odds ratios (OR) with 95% confidence intervals (Cl) for stillbirth and infant death in the
ethnic Pakistani group relative to the ethnic Norwegian group: births in Norway, 1967-1993
Stillbirths
Logistic regression
model controlling for
Infant deaths
OR
95% Cl
OR
95% Cl
Norwegian
Pakistani
1.0
1.3
1.1-1.6
1.0
1.7
1.4-2.1
Norwegian
Pakistani
1.0
1.2
1.0-1.5
1.0
1.1
0.9-1.4
Norwegian
Pakistani
1.0
1.0
1.0-1.0
1.0
1.3
1.1-1.6
Norwegian
Pakistani
1.0
1.4
1.1-1.7
1.7
1.4-2.1
5. Consanguinity, maternal education, and
year of birth
Norwegian
Pakistani
1.0
0.9
0.7-1.1
1.0
0.9
0.7-1.1
6. Consanguinity, maternal education,
maternal age, parity, and year of birth
Norwegian
Pakistani
1.0
1.0
0.8-1.2
1.0
0.9
0.7-1.2
1. Year of birth
2. Consanguinity and year of birth
3. Maternal education and year of birth
4. Maternal age, parity, and year of birth
1.0
TABLE 5. Proportion of early deaths attributable to consanguinity and low maternal educational levels in the ethnic Norwegian
and ethnic Pakistani groups, 1985-1993
Group and
variable
Norwegian
Consanguinity^
Matemal education <13 years
or missing
Pakistani
Consanguinity^
Matemal education <13 years
or missing
No. of
births
456,056
No. of
early
deaths
Prevalence
(%)
of risk
factor
Odds
ratio*
95% C l t
Attributable
risk
95% Cl
7,302
14
0.1
1.9
1.1 to 3.2
0.1
-0.02 to 0.2
346,338
4,219
1,736
5,795
75.9
1.3
1.2 to 1.3
16.7
12.7 to 20.7
104
60
41.1
2.0
1.3 to 3.0
29.0
12.7 to 45.1
4,009
101
95.0
1.6
0.5 to 5.2
37.0
-50.4 to 100.0
455
* All odds ratios are adjusted for maternal age and parity in addition to consanguinity and maternal education,
t Cl, confidence interval.
t The category consanguinity includes first cousin and closer relations in the Norwegian group and all consanguineous relations in the
Pakistani group.
Data quality
Data quality may have influenced the results. There
were no indications of differential selection of fetuses
between the two populations, but differential selection
between fetuses with consanguineous and nonconsanguineous parents can not be ruled out. A linkage with Statistics Norway ensures an almost complete
registration of births and infant deaths. Registration of
stillbirths with more than 28 weeks of gestation are
considered to be similarly complete, while there is
some uncertainty about stillbirths with 16-28 weeks
of gestation. However, the proportion of early stillbirths was the same in the two populations, and independent of maternal education, maternal age, and parity. The proportion of early stillbirths increased from
the first to the last time period in both population
groups, probably reflecting more complete recording
of stillbirth.
Underreporting or misclassification of consanguinity is probable in both populations. To investigate the
validity of the data, all records were sorted in families
and reanalyzed based on the assumption that if at least
one child in a family was recorded with consanguineous parents all the siblings had consanguineous parents. As a result, the proportion of children with consanguineous parents increased from 40.4 percent to
54.5 percent among the Pakistanis, and the proportion
with first cousin parents increased from 0.1 percent to
0.2 percent among the Norwegians. The corresponding
relative and attributable risks for stillbirth and infant
death changed only slightly, and the confidence intervals were overlapping. All results were similar to
Am J Epidemiol
Vol. 148, No. 5, 1998
Stillbirth, Infant Death, and Parental Consanguinity
those presented here (data not shown). The authors
chose to use the conservative estimates of frequency
of consanguinity in the present study, although the
family approach described above is probably more
valid (18, 19).
In the analysis, all other consanguineous relations
than first cousins were combined into one category
(Appendix table). The composition of the category
"other consanguineous relations" varied between the
two populations, as well as the results. This reflects
that most of the Norwegians with "other consanguineous relations" were more distantly related than first
cousins, while for the Pakistanis a majority were classified as "non-specified relations," which probably
includes many first cousins or even closer relations,
such as double first cousins.
Information on maternal educational level was missing for a large percentage of the Pakistanis. However,
analysis without this group yielded results that were
almost identical, but with broader confidence intervals. Mothers who had missing information on education resembled those with the lowest level of education
on frequency of consanguinity (table 2) and early
death, and on the distribution of maternal age and
parity. These associations with other variables supported the use of maternal education as a measure of
socioeconomic conditions in this study. The small
proportion of mothers with higher education, and the
large proportion with missing data on education may
have hidden a relation between maternal education and
stillbirth or infant death in the Pakistani group. However, there were similar problems with other measures
of socioeconomic conditions, such as father's education, and we chose to use maternal education as a
simple measure of socioeconomic conditions because
of its well-documented influence in the Norwegian
group (31, 32). All the analyses were performed with
paternal education instead of maternal education, and
with both maternal and paternal education. The results
for the influence of education on the risk of early death
varied somewhat, while the results for the influence of
consanguinity were nearly identical independent of
choice of measure of socioeconomic conditions.
Consanguinity and early death
In this study, the effect size of consanguinity on
stillbirth and on all early deaths was the same in the
two groups, but slightly higher for infant death in the
Norwegian first-cousin group compared with either of
the Pakistani consanguineous groups. The differences
were not statistically significant. Based on findings in
a British study (33), one would expect that the degree
of inbreeding among the Pakistanis was greater than
estimated for first cousins due to generations of conAm J Epidemiol
Vol. 148, No. 5, 1998
457
sanguinity. The present data do not indicate that Pakistani first cousins are more closely related than Norwegian first cousins.
Mortality rates among children with non-consanguineous parents in Norway were low by all international and historical standards, and it has been
suggested that the influence of consanguinity on
early death should be greater than in populations
with higher mortality rates (2). In this study, relative
risks of infant death associated with consanguinity
were higher than the median relative risk of several
other studies, but well within the range observed in
these studies (2).
Maternal education and early death
In agreement with previous results from a subset
of the present data (31, 32) we found that low
maternal education significantly increased the risk
of early death in the Norwegian group. Early mortality rates are high in Pakistan (34, 35). However,
the association between maternal education and
early death in the Pakistani group in Norway was
uncertain, although the relative risk estimates indicated a similar trend as in the Norwegian group for
infant death. Possible reasons for the differential
importance of maternal education are that 1) maternal education is not an appropriate measure of social
class among Pakistanis in Norway, 2) only a small
percentage of the Pakistanis had higher education,
3) the large proportion of missing information conceals a relation between maternal education and
early death in the Pakistani group, or 4) the influence of socioeconomic conditions in the Pakistani
population is really lower than in the Norwegian
population. Probable mediators between low maternal education and early mortality include nutritional
deficiencies, infections, and cigarette smoking. It is
not known whether these risk factors vary for the
two populations, but clinical experience indicates
that cigarette smoking and alcohol consumption is
extremely rare among Pakistani women in Norway,
and significant nutritional differences have been
shown in studies on vitamin D and iron (36, 37).
In addition, considerable theoretical problems arise
when comparing the effects of education across ethnic
groups and over time, and especially when maternal
education is the only indicator of social class.
In conclusion, the present study indicates that the
most important aspect of the relation between social
class and early mortality among Pakistanis in Norway
is the greater frequency of consanguinity among those
with lower education. Analyses based on better measures of social class will probably not change these
results dramatically, but can add useful information on
458
Stoltenberg et al.
the relations between consanguinity, social class, and
early mortality.
ACKNOWLEDGMENTS
This study was supported by a grant from the University
of Oslo and the Norwegian Ministry of Health and Social
Affairs.
The authors thank Dr. Jennifer Harris and Jak Jervell for
their valuable comments and cooperation.
REFERENCES
1. Vogel F, Motulsky AG. Human genetics. 3rd ed. Berlin:
Springer-Verlag, 1997.
2. Khoury MJ, Cohen BH, Chase GA, et al. An epidemiologic
approach to the evaluation of the effect of inbreeding on
prereproductive mortality. Am J Epidemiol 1987;125:251-62.
3. Khlat M, Khoury M. Inbreeding and diseases: demographic,
genetic, and epidemiologic perspectives. Epidemiol Rev 1991;
13:28-41.
4. Bittles AH, Mason WM, Greene J, et al. Reproductive behavior and health in consanguineous marriages. Science 1991;
252:789-94.
5. Ahmad WIU. Reflections on the consanguinity and birth outcome debate. J Public Health Med 1994;16:423-8.
6. Bittles AH. The role and significance of consanguinity as a
demographic variable. Popul Devel Rev 1994;20:561-84.
7. Bittles AH, Grant JC, Shami SA. Consanguinity as a determinant of reproductive behaviour and mortality in Pakistan.
Int J Epidemiol 1993,22:463-67.
8. Yaqoob M, Gustavson KH, Jalil F, et al. Early child health in
Lahore, Pakistan. II. Inbreeding. Acta Paediatr Suppl 390
1993;82:17-26.
9. El-Hazmi MAF, Al-Swailem AR, Warsy AS, et al. Consanguinity among the Saudi Arabian population. J Med Genet
1995;32:623-6.
10. Al-Awadi SA, Moussa MA, Naguib KK, et al. Consanguinity
among the Kuwaiti population. Clin Genet 1985;27:483-6.
11. Khlat M. Consanguineous marriages in Beirut: time, trends,
spatial distribution. Soc Biol 1989;35:324-30.
12. Al-Ghazali LI, Dawodo AH, Sabarinathan K, et al. The profile
of major congenital abnormalities in the United Arab Emirates
(UAE) population. J Med Genet 1995;32:7-13.
13. Hamamy HA, Al-Hakkak ZS. Consanguinity and reproductive
health in Iraq. Hum Hered 1989;39:271-5.
14. El-Shafei A, Rao PSS, Samdhu AK. Congenital malformations and consanguinity. Aust NZ J Obstet Gynaecol 1986;26:
168-72.
15. Hafez M, El-Tahan H, Awadalla M, et al. Consanguineous
matings in the Egyptian population. J Med Genet 1983;20:
58-60.
16. Khoury SA, Massad D. Consanguineous marriage in Jordan.
Am J Med Genet 1992;43:769-75.
17. Basaran N, Sayly BS, Basaran A, et al. Consanguineous marriages in the Turkish population. Clin Genet 1988;34:339-41.
18. Darr A, Modell B. The frequency of consanguineous marriage
among British Pakistanis. J Med Genet 1988;25:186-90.
19. Bundey S, Alam H, Kaur A, et al. Why do UK-bom Pakistani
babies have high perinatal and neonatal mortality rates? Paediatr Perinat Epidemiol 1991;5:101-14.
20. Balarajan R, Soni Raleigh V, Botting B. Sudden infant
death syndrome and postneonatal mortality in immigrants
in England and Wales. BMJ 1989;298:716-20.
21. 0rstavik KH, Lindemann R, Steen J. Infant death and consanguineous marriage. (Letter). BMJ 1994;308:659.
22. Fraser FC, Biddle CJ. Estimating the risks for offspring of first
cousin matings. An approach. Am J Hum Genet 1976;28:
522-6.
23. Khlat M. Social correlates of consanguineous marriages in
Beirut: a population based study. Hum Biol 1988;60:541-8.
24. Coleman DA. A note on the frequency of consanguineous
marriages in Reading, England 1972/1973. Hum Hered 1980;
30:278-85.
25. Guz K, Dedeoglu N, Luleci G. The frequency and medical
effects of consanguineous marriages in Antalya, Turkey.
Hereditas 1989; 111:79-83.
26. Stoltenberg C, Magnus P, Lie RT, et al. Birth defects and
parental consanguinity in Norway. Am J Epidemiol 1997;145:
439-48.
27. Medical Birth Registry of Norway. Annual report 1993.
Bergen: University of Bergen, Norway 1995.
28. Altaian DG. Practical statistics for medical research. London:
Chapman and Hall, 1991.
29. Clayton D, Hills M. Statistical models in epidemiology.
Oxford: Oxford University Press, 1993.
30. Levin ML, Bertell R. Re: "Simple estimation of population
attributable risk from case-control studies." (Letter). Am J
Epidemiol 1978;108:78-9.
31. Arntzen A, Moum T, Magnus P, et al. The association between maternal education and postneonatal mortality. Trends
in Norway, 1968-1991. Int J Epidemiol 1996;25:578-84.
32. Arntzen A, Moum T, Magnus P, et al. Is the higher postneonatal mortality in lower social status groups due to SIDS? Acta
Paediatr 1995 ;84:188-92.
33. Bundey S, Alam H. A five-year prospective study of the health
of children in different ethnic groups, with particular reference
to the effect of inbreeding. Eur J Hum Genet 1993;l:206-19.
34. Fikree FF, Gray RH. Demographic survey of the level and
determinants of perinatal mortality in Karachi, Pakistan.
Paediatr Perinat Epidemiol 1996; 10:86-96.
35. Khan SR, Jalil F, Zaman S, et al. Early child health in Lahore,
Pakistan. X. Mortality. Acta Paediatr Suppl 390 1993;82:
109-17.
36. Henriksen C, Brunvand C, Stoltenberg C, et al. Diet and
vitamin D status among pregnant Pakistani women in Oslo.
Eur J Clin Nutr 1995;49:211-18.
37. Brunvand L, Henriksen C, Larsson A, et al. Iron deficiency
among pregnant Pakistanis in Norway and the content of
phytic acid in their diet. Acta Obstet Gynecol Scand 1995;74:
520-5.
Am J Epidemiol
Vol. 148, No. 5, 1998
APPENDIX TABLE. Parental consanguinity, numbers and rates of stillbirth and infant death, and estimated coefficient of inbreeding (F) in the ethnic Norwegian and
ethnic Pakistani groups: births in Norway, 1967-1993
Ethnic Pakistanis
Ethnic Norwegians
5;
_k
.&.
o
en
_,.
CO
CO
00
All
Parental
consanguinity
No.
Parent, offspring, or sister-brother
2
Half siblings, uncle-niece or auntnephew, double first cousin
91
First cousin
1,890
First cousin once removed
268
Second cousin
5,239
Third cousin
4,773
Not consanguineous
1,415,421
Other nonspecified relations
3,371
Stillbirths
%
Infant deaths
Per
1,000
births
No.
Per
1,000
livebirths
No.
Stillbirths
All
Infant deaths
Per
No.
%
No.
1,000
births
Per
No.
1,000
livebirths
F
Estimated
coefficient
of
0.0
0
0
0
0
0
0.0
0
0
0
0
1/4
0.0
0.1
0.0
0.4
0.3
98.9
0.2
1
33
11
17
30
9
11
12
15
1
59
3
80
50
13,144
30
11
32
12
15
11
9
9
6
2,248
5
308
24
4,339
344
0.1
30.9
0.1
4.2
0.3
59.7
4.7
1
35
0
5
3
54
2
167
16
0
16
125
12
9
0
41
0
6
1
37
7
0
19
0
1/8
8
49
52
16,379
51
20
48
9
20
1/16
1/32
1/64
1/256
0
?
C/3
g;
3.'
I
O
CD
0)
Q.
01
O
o
co
B>
CO
c
to