PREVENTIVE
MEDICINE
4, 509-5 17 (1975)
Concordance for Mortality with Special Reference to lschaemic Heart Disease and Cerebrovascular Disease’ A Study on the Swedish
Twin Registry
UL.F DE FAIRE,~ LARS FRIBERG,AND TORBJ~RNLUNDMAN Department of Medicine, Karolinska Institutet at Serajimerlasarettet, Department of Environmental Hygiene, Karolinska Institutet, and the Swedish National Environmental and Protection Board, Stockholm, Sweden
The Swedish Twin Registry contains about 11 000 same-sexed twin pairs born between 1886 and 1925 with both members alive when the registry was formed in 1961. During the years 1962 to 1973, 2780 deaths occurred. 727 deaths were due to ischaemic heart disease (IHD), 345 due to cerebrovascular disease (CVD), and 727 due to cancer. The rate of concordance for the whole twin population revealed a significantly (p < 0.05) higher concordance rate for IHD among the male monozygotic (MZ) pairs as compared to the dizygotic (DZ) pairs (15.8% versus 8.0%). The corresponding figures for the female pairs were 11.0% (MZ) and 7.5% (DZ), respectively. With regard to death in CVD and cancer, the rates of concordance were about the same for MZ and DZ pairs in both males and females. When subgrouping was made for age groups, the difference in concordance rate for IHD in males was still more pronounced for the younger age group, born 1901- 1925, (16.1% versus 5.4%). These data may indicate the existence of a genetic determination on death in IHD, especially in males, whereas a genetic determination on death in CVD and cancer seems more uncertain.
In the etiology of ischaemic heart disease (IHD) and cerebrovascular disease (CVD), most interest has been focused on environmental influences, since these can be modified and controlled. That environmental factors play an important part, especially in the development of IHD, indeed seems to be strongly documented. It is also commonly recognized that IHD has some hereditary basis that comes into play with the environmental factors (3, 33). Several studies have demonstrated a familial aggregation which regard to IHD (16, 25, 26, 27, 28, 32), whereas the more sparse data on familial clustering of CVD (2, 22, 23, 25) are not very impressive. Prospective studies have also disclosed familial aggregation of IHD (10, 17,29). However, modest familial concentrations may be found for most common diseases for which they have been sought (11, 24). The genetic mechanisms are probably partly transmitted through underlying biological factors such as cholesterol and blood pressure. All available evidence supports the assumption that these factors are polygenetically controlled, but some 1 Presented at the Conference on Cardiovascular Disease Epidemiology, Tampa, Florida, March 1975. 2 Address correspondence to U. de Faire, Department of Medicine, Serafimerlasarettet, 112 83 Stockholm. 509 Copyright @ 1975 by Academic Press, Inc. All rights of reproduction in any form reserved.
510
DE FAIRE,
FRIBERG
AND
LUNDMAN
extremes of a factor, e.g., type II-A hyperlipoproteinemia of Fredriksson and Lees, may exhibit single gene control (14). Although several models for computing the genetic effect have been developed from family studies (12, 13, 30), studies on twins may be the most appropriate way for quantitating the relative role of these factors. Individual case reports on twin concordance for IHD and CVD are occasionally published. Although of minor importance, they give rise to a suspicion of a hereditary influence. The occurence of IHD and CVD has so far been studied in only two representative large-scale twin samples, namely the Danish and the Swedish ones. When the Danish Twin Registry was formed in 1954, it contained 8000 pairs born during 1870-19 10. By January 1, 1968, 352 fatal “coronary occlusions” had occurred (19). The rate of concordance calculated by the twin proband method (1) was found to be significantly higher for MZ pairs than DZ pairs both for males (39 versus 26%) and females (44 versus 14%). The authors concluded that environmental factors might be “strong enough to outweigh the genetic influences in men but not in women.” Since the establishment of the Swedish Twin Registry, both comprehensive questionnaire studies (4,7, 8) and studies on subsamples (9, 20, 21) have been carried out. The genetic influence on the end result, IHD-death and CVD-death, seems to be of special interest. Data on a 12-yr mortality follow-up are now available. For comparison, data on cancer concordance will also be presented. MATERIAL
The Swedish Twin Registry was set up during the years 1959-1961. At that time it contained about 10 900 pairs, which covered 95% of all Swedish samesexed twin pairs born in this country between 1886 and 1925 and in which both members were alive when the registry was formed. The compilation procedure and the demographic structure of the twin series have been described by Cederlof (5). The zygosity of the twins in the registry has been determined with the aid of similarity questions contained in a questionnaire mailed to each twin (5). Mortality has been followed since 1962; up to December 1973, 2780 deaths have occurred. METHODS
The mortality has been established as follows. The total twin registry is matched regularly against a total death registry for Sweden at the Central Bureau of Statistics (once a month since 1971). This matching provides access to the death certificate and the name of the physician who signed it. For twins born during 1886-1900, the cause of death has been taken from the death certificates only, but for twins born during 1901-1925, hospital records, autopsy protocols, and information from general practitioners have been collected as well. The cause of death has then been judged according to these multiple sources (9, IS) and classified in line with the 1965 revision of International Statistical Classification (ISC) of Diseases, Injuries, and Causes of Death. For the present report, the causes of death have been grouped in the following way (ISC numbers in parentheses): IHD (410.00-414.99, 795.99),
FORUM:
CARDIOVASCULAR
DISEASE
511
EPIDEMIOLOGY
TABLE 1 CAUSES OF MALE DEA-~HHSACCORDING TO AGE Born 1886-1900 (No.) (%I All causes IHD CVD Cancer, total Cancer, lung Cancer, other Suicides Accidents Other
703 227 96 146 24 122 7 26 201
Born 1901-1925 (No.) (%)
100.0 32.3 13.7 20.8 3.4 17.4 1.0 3.7 28.6
717 219 48 176 40 136 46 66 162
100.0 30.5 6.7 24.6 5.6 19.0 6.4 9.2 22.6
Born 1886-1925 (%I (No.) 1420 446 144 322 64 258 53 92 363
100.0 31.4 10.1 22.7 4.5 18.2 3.7 6.5 25.6
CVD (430&O-438.99), cancer of the lung (162.01-162.19), cancer other forms (140.01-161.99, 163-239.99), suicides (E950.9-E959.9), accidents (800.00-999.99, E807.0-E949.9, E960.9-E999.9), and other causes (conditions not specified here). The concordance rate given is the so-called pair-wise concordance rate (1, 3 1). Results are also given for the expectancy of coincidence, based on the square of the prevalence rate, i.e., the probability of a pair’s being concordant due to chance alone. Comparisons are then made between the observed and expected coincidences. In the comparison of MZ and DZ concordance rates, the chi-square test was used. RESULTS
Tables 1 and 2 show the causes of death for men and women according to age groups. The distribution of the various causes seems to be fairly proportionate with that of the causes of death for the general population of Sweden. IHD is the major cause of death for men in both age groups, whereas cancer is the predominant cause of death for women, mainly due to the excess cases among those
9wEuM
lw
TWIW nEoM9v
1aeoopbr
*oRIAUN
fomw-UP
5?
Er pain’
FIG. 1. Mortality
follow-up
DEATH MS” Pam
of the Swedish Twin Registry.
512
DE FAIRE,
FRIBERG TABLE
AND
LUNDMAN
2
CAUSES OF FEMALE DEATHS ACCORDINGTO AGE Born 1886-1900 (No.1 m All causes IHD CVD
787 199 137 170 10 160 3 13 265
Cancer, total Cancer, lung Cancer, other Suicides Accidents Other
Born 1886-192.5 (No.1 (%I
Born MI-1925 (No.1 (%I
100.0 25.3 17.4 21.6 1.3 20.3 0.4 1.7 33.7
573 82 64 235 5 230 23 27 142
100.0 14.3 11.2 41.0
100.0
1360 281 201 405 15 390 26 40 407
0.9 40.1 4.0 4.7 24.8
20.7 14.8 29.8
1.1 28.7 1.9 2.9 29.9
born during 1901- 1925. Among the female twins born during 1886-1900, IHD has outweighed cancer as the cause of death. The relative distribution of causes of death seems to be about the same for MZ and DZ twins. Of the male MZ twins, 3 1.9% had died from IHD as compared to 3 1.6% of the DZ twins; the corresponding proportions for CVD and cancer deaths among the male MZ and DZ twins were 9.5 versus 10.2% and 23.2 versus 22.6%, respectively. The proportions for IHD among females were 22.5% among MZ twins and 19.8% among DZ twins; for CVD and cancer the figures were 15.2 versus 14.6% and 29.0 versus 30.0%, respectively. At the end of December 1973, 2780 deaths, 1420 male and 1360 female, had taken place (Fig. 1). At that time 259 male and 222 female pairs had been death concordant, i.e., both members in a pair had died, and 85 of the male and 79 of the female pairs were also concordant with respect to cause of death as listed in Tables 1 and 2. Death concordance rates for the whole twin population are given in Fig. 2. Male MZ twins revealed a significantly (t, < 0.05) higher concordance rate for IHD than the male DZ twins (15.8 vs 8.0%). A similar trend is noted
22 PO p-Ml5
2 5
7 s
as.
as.
Cmmrdmt
@ain
10 12 as.
5 9 n.9.
tin ltd.
Cmmordut
win
FIG. 2. Death concordance rates for twin pairs born during 1886-l 925; (left) males, (right) females.
FORUM:
CARDIOVASCULAR
DISEASE
513
EPIDEMIOLOGY
%
IN8
CVD
n M2 0 DZ
CMCER CANCER
12 13
1.8.
1
1
N.S.
5
6
Cmmdant
8
pain
9
N.S.
Il.%
4
9
2 9
as.
N.S.
3. Death concordance rates for twin pairs born during 1886-l 900; (left) males, (right) females.
FIG.
among the female twins, but the difference (11.0 versus 7.5%) was not significant. With respect to death from CVD and cancer, MZ and DZ twin pairs displayed only slight differences. By looking at both of the age groups, born 1886-1900 and 1901-1925 (Figs. 3 and 4), it is obvious that the difference regarding IHD between the male MZ and DZ twins is still more pronounced among the younger pairs, where it was significant at the 5% level (16.1 versus 5.4%). For the females, the reverse age trend was noted. With regard to CVD and cancer, the differences in concordance rates between MZ and DZ pairs were sometimes somewhat more marked (cf. cancer in females) than for the twin population as a whole, but the differences were not significant. Observed and expected coincidence rates for mortality according to age groups are shown in Tables 3-5. The actually observed number of concordant pairs in relation to the expected coincidence are also given. When the whole twin population (Table 3) is considered, it appears that the observed numbers are considerably higher than the expected ones irrespective of zygosity group or cause of death. The difference between the zygosity groups is not very impressive for CVD and cancer but is relatively pronounced for the male IHD deaths Y.
25
25
1
20
FIG.
IND
CVD
CANCER
p-D.05
I.&
N.S.
7.
20
11.1.
N.S.
1.9.
4. Death concordance rates for twin pairs born during 1901-1925; (left) males, (right) females.
514
FRIBERG AND LUNDMAN
DE FAIRE,
TABLE 3 OBSERVED AND EXPECTED COINCIDENCE RATES WITH RESPECT TO DEATH IN IHD, CVD, AND CANCER IN TWIN PAIRS BORN DURING 1886-1925
IHD Coincidence Observed MZ DZ Expected MZ DZ Observed/expected MZ DZ
Males
CVD Females 10
22 20 3.93
5.6 3.3
Females
Males
Cancer Females
2 3
5 9
7 9
1.27 1.89
0.35 0.63
0.58 1.03
2.08 3.12
2.11 4.34
7.9 6.3
5.7 4.8
8.7 8.8
3.4 2.9
2.4 2.5
12
6.11
Males
5
11
TABLE 4 OBSERVED AND EXPECTED COINCIDENCE RATES WITH RESPECT TO DEATH IN IHD, AND CANCER IN TWIN PAIRS BORN DURING 1886-1900
IHD Coincidence Observed MZ DZ Expected MZ DZ Observed/expected MZ
DZ
Males
CVD Females
Females
Males
Males
CVD,
Cancer Females
12 13
1 1
4
5
9
8
6
2 8
6.18 10.20
3.33 4.99
0.74 1.73
1.40 2.50
2.98 3.37
1.85 4.04
1.9 1.3
2.4
1.4 0.6
2.9 3.2
1.7
1.1
1.8
2.0
8
1.8
TABLE 5 OBSERVED AND EXPECTED COINCIDENCE RATES WITH RESPECT TO DEATH IN IHD, CVD, AND CANCER IN TWIN PAIRS BORN DURING 1901-1925
IHD Coincidence Observed MZ DZ Expected MZ DZ Observed/expected MZ DZ
Males 10
CVD Females
7
2 3
0.98
0.14
1.82
0.21
10.2
14.5
3.8
14.5
Males 1 2
Females 1
Males
Cancer Females
1
2 3
3 3
0.05
0.07
0.57
0.10
0.13
1.32
0.91 1.81
20.0 20.0
13.4 1.5
3.5 2.3
3.3
1.7
FORUM:
CARDIOVASCULAR
DISEASE
EPIDEMIOLOGY
515
(5.6 and 3.3, respectively). Subgrouping according to age group (Tables 4 and 5) leads to ratios between the observed and expected values that are generally smaller for the older age groups, especially for IHD and CVD. With respect to the younger twins born 1901-1925, usually high ratios are seen for some of the groups. It must be remembered that the subgrouping of the twin pairs according to age groups gives a limited number of concordant pairs, especially for CVD, and that these figures must be interpreted with great caution. With regard to IHD, the observed number of concordant male deaths in relation to the expected number is seen to be more than 2.5 (10.2 and 3.8) times higher among MZ as compared to DZ pairs. DISCUSSION
The findings of the present report are in accordance with earlier results with respect to fatal coronary occlusion in Danish male twins (19). As regards females, our results diverge from the Danish follow-up, which found a more pronounced difference in concordance rate between MZ and DZ pairs for females than males (19). To some extent this divergence could probably be explained by the relatively few concordant female pairs in both of the studies. The findings of the present study are especially interesting in light of the recent results on 205 death discordant twins from the Swedish Twin Registry (9). When myocardial infarction, angina pectoris, pathologic Q-wave, and ST depressions in connection with exercise were successively included in the criteria for IHD, the prevalences of these combinations were revealed to be higher among the cotwins whose partners had died from IHD than among those whose partners had died from other causes. This applied to both the male and female groups, but the difference was more pronounced for the male pairs, especially between the MZ pairs. As could be expected, most of the biologic risk factors measured showed the same trend, but the differences were small, thus indicating that the genetic influence also may be transmitted through unknown biologic factors. Only slight differences were noted with regard to the environmental factors of smoking, extra work, and physical inactivity. These two studies can be interpreted as showing that both the development of IHD and death from it are under a relatively strong genetic influence, which is most clearly demonstrated in males under the age of 70. With regard to death from CVD, the rates of concordance were about the same for MZ and DZ twins, but the number of observed concordant pairs was far above what could be expected from chance alone in both MZ and DZ twins, especially in the younger age groups. These data point to a relatively strong, mainly environmentally governed family influence on death from CVD. The number of concordant deaths is as yet too small to warrant more definite conclusions. As regards cancer mortality, the present findings confirm the results of an earlier investigation on morbidity and mortality based on the Swedish Twin Registry (6). Only slight differences in death concordance rates were found between MZ and DZ pairs. Comparisons between expected and observed coincidence of death showed that the actual within-pair coincidence exceeds the random expec-
516
DE FAIRE,
FRIBERG
AND
LUNDMAN
tancy in both MZ and DZ twins by about 2-3 times. These data indicate a moderate family influence, which is in agreement with other twin studies (6,18). ACKNOWLEDGMENTS This investigation was supported by grants from the Swedish National Association against Heart and Chest Diseases, the Swedish Medical Research Council, and the American Medical Association, Education and Research Foundation.
REFERENCES 1. Allen, C., Harvald, B., and Shields, J. Measures of twin concordance. Acta Genet. Base1 17,475 (1967). 2. Alter, M. Genetic factors in cerebrovascular accidents. Trans. Amer. Neurol. Assoc. 92, 205 (1967). 3. Biorck, G. Environment disorders of the circulatory system. Bull. N.Y. Acad. Med. 35, 3 (1959). 4. Cederliif, R. Urban factor and prevalence of respiratory symptoms and “angina pectoris”: A study on 9,168 twin-pairs with the aid of mailed questionnaires. Arch. Environ. Health 13, 743 (1966). 5. Cederliif, R. “The Twin Method in Epidemiological Studies on Chronic Disease.” Karolinska Institutet, Stockholm, 1966. 6. Cederlof, R., Floderus, B., and Friberg, L. Cancer in MZ and DZ twins. Acta Genet. Med. Gemetlol.
19, 69 (1970).
7. Cederlof, R., Friberg, L., and Jonsson, E. Hereditary factors and “angina pectoh”: A study on 5,877 twin-pairs with the aid of mailed questionnaires. Arch. Environ. Health 14, 397 (1967). 8. CederlSf, R., Jonsson, E., and Kaij, L. Respiratory symptoms and “angina pectoris” in twins with reference to smoking habits: An epidemiological study with mailed questionnaires. Arch. Environ.
Health
13, 726 (1966).
9. de Faire, U. Ischaemic heart disease in death discordant twins: A study on 205 male and female pairs. Acta Med. Stand. Suppl. 568 (1974). 10. Deutscher, S., Ostrander, L. D., and Epstein, F. H. Familial factors in premature coronary heart disease-a preliminary report from the Tecumseh Community Health Study. Amer. J. Epidemiol. 91, 233 (1970).
11. Edwards, J. H. The genetic basis of common disease. Amer. J. Med. 34, (1963). 12. Edwards, J. H. Familial predisposition in man. Brit. Med. Bull. 25, 58 (1969). 13. Falconer, D. S. The inheritance of liability to certain diseases, estimated from incidence among relatives. Ann. Hum. Genet. London 29, 51 (1965). 14. Frederickson, D. L. Mutants hyperlipoproteinaemia and coronary artery disease. Brit. Med. J. 2, 187 (1971). 15. Friberg, L., Cederlof, R., Lorich, U., Lundman, T., and de Faire, U. Mortality in twins in relation to smoking habits and alcohol problems. Arch. Environ. Health. 27, 294 (1973). 16. Gertler, M. M., and White, P. D. “Coronary Heart Disease in Young Adults: a Multidisciplinary Study.” Harvard University Press, Cambridge, Mass., 1954. 17. Hammond, E. C., Gartinkel, L., and Seidman, H. Longevity of parents and grandparents in relation to coronary heart disease and associated variables. Circulation 43, 3 1 (197 1). 18. Harvald, B., and Hauge, M. “Hereditary Factors Elucidated by Twin Studies: Genetics and the Epidemiology of Chronic Disease.” U. S. Dept. of Health, Education and Welfare, Washington, D.C., 1965. 19. Harvald, B. and Hauge, M. Coronary occlusion in twins. Acta Genet. Med. Gemellol. 19, 248 (1970). 20. Liljefors, I. Coronary heart disease in male twins: Hereditary and environmental factors in concordant and discordant pairs. Acta Med. Stand. Suppt. 5 11 (1970). 21. Lundman, T. Smoking in relation to coronary heart disease and lung function in twins: A co-twin control study. Actu Med. Stand. Suppl. 455 (1966). 22. Marshall, J. Familial incidence of cerebrovascular disease. J. Med. Genet. f$, 84 (197 1).
FORUM: CARDIOVASCULAR
DISEASE EPIDEMIOLOGY
517
23. Marshall, J. Familial~incidence of cerebral hemorrhage. Stroke 4, 38 (1973). 24. McKusick, V. A., and Murphy, E. A. Genetic factors in the etiology of myocardial infarction, in “The Etiology of Myocardial Infarction” (T. N. James and J. W. Keys, Eds.). Little, Brown and Company, Boston, 1963. 25. Phillips, R. L., Lilienfeld, A. M., Diamond, E. L., and Kagan, A. Frequency of coronary heart disease and cerebrovascular accidents in parents and sons of coronary heart disease index cases and controls. Amer. J. Epidemiol. 100, 87 (1974). 26. Rose, G. Familial patterns in ischaemic heart disease. Bit. J. Prev. Sot. Med. 18, 75 (1964). 27. Russek, H. I., and Zohman, B. L. Relative significance of heredity, diet, and occupational stress in coronary heart disease of young adults. Amer. J. Med. Sci. 235, 266 (1958). 28. Shanoff, H. M., Little, A,, Murphy, E. A., and Rykert, H. E. Studies of male survivors of mycardial infarction due to “essential” atherosclerosis. I. Characteristics of the patients. Can. Med. Assoc. 1. 84, 519 (1961). 29. Slack, J., and Evans, K. A. The increased risk of death from ischaemic heart disease in first degree relatives of 121 men and 96 women with ischaemic heart disease. J.Med. Genet. 3, 239 (1966). 30. Smith, C. Recurrence risks for multifactorial inheritance. Amer. J. Hum. Genet. 23, 578 (197 1). 3 1. Smith, C. Correlation in liability among relatives concordance in twins. Hum. Hered. 22, 97 (1972). 32. Thomas, C. B., and Cohen, B. H. Familial occurrence of hypertension and coronary artery disease, with observations concerning obesity and diabetes. Ann. Intern. Med. 42, 90 (1955). 33. White, P. I). The genes, the heart and destiny. New Engl. J. Med. 256, 965 (1957).
MEDICINE
4, 509-5 17 (1975)
Concordance for Mortality with Special Reference to lschaemic Heart Disease and Cerebrovascular Disease’ A Study on the Swedish
Twin Registry
UL.F DE FAIRE,~ LARS FRIBERG,AND TORBJ~RNLUNDMAN Department of Medicine, Karolinska Institutet at Serajimerlasarettet, Department of Environmental Hygiene, Karolinska Institutet, and the Swedish National Environmental and Protection Board, Stockholm, Sweden
The Swedish Twin Registry contains about 11 000 same-sexed twin pairs born between 1886 and 1925 with both members alive when the registry was formed in 1961. During the years 1962 to 1973, 2780 deaths occurred. 727 deaths were due to ischaemic heart disease (IHD), 345 due to cerebrovascular disease (CVD), and 727 due to cancer. The rate of concordance for the whole twin population revealed a significantly (p < 0.05) higher concordance rate for IHD among the male monozygotic (MZ) pairs as compared to the dizygotic (DZ) pairs (15.8% versus 8.0%). The corresponding figures for the female pairs were 11.0% (MZ) and 7.5% (DZ), respectively. With regard to death in CVD and cancer, the rates of concordance were about the same for MZ and DZ pairs in both males and females. When subgrouping was made for age groups, the difference in concordance rate for IHD in males was still more pronounced for the younger age group, born 1901- 1925, (16.1% versus 5.4%). These data may indicate the existence of a genetic determination on death in IHD, especially in males, whereas a genetic determination on death in CVD and cancer seems more uncertain.
In the etiology of ischaemic heart disease (IHD) and cerebrovascular disease (CVD), most interest has been focused on environmental influences, since these can be modified and controlled. That environmental factors play an important part, especially in the development of IHD, indeed seems to be strongly documented. It is also commonly recognized that IHD has some hereditary basis that comes into play with the environmental factors (3, 33). Several studies have demonstrated a familial aggregation which regard to IHD (16, 25, 26, 27, 28, 32), whereas the more sparse data on familial clustering of CVD (2, 22, 23, 25) are not very impressive. Prospective studies have also disclosed familial aggregation of IHD (10, 17,29). However, modest familial concentrations may be found for most common diseases for which they have been sought (11, 24). The genetic mechanisms are probably partly transmitted through underlying biological factors such as cholesterol and blood pressure. All available evidence supports the assumption that these factors are polygenetically controlled, but some 1 Presented at the Conference on Cardiovascular Disease Epidemiology, Tampa, Florida, March 1975. 2 Address correspondence to U. de Faire, Department of Medicine, Serafimerlasarettet, 112 83 Stockholm. 509 Copyright @ 1975 by Academic Press, Inc. All rights of reproduction in any form reserved.
510
DE FAIRE,
FRIBERG
AND
LUNDMAN
extremes of a factor, e.g., type II-A hyperlipoproteinemia of Fredriksson and Lees, may exhibit single gene control (14). Although several models for computing the genetic effect have been developed from family studies (12, 13, 30), studies on twins may be the most appropriate way for quantitating the relative role of these factors. Individual case reports on twin concordance for IHD and CVD are occasionally published. Although of minor importance, they give rise to a suspicion of a hereditary influence. The occurence of IHD and CVD has so far been studied in only two representative large-scale twin samples, namely the Danish and the Swedish ones. When the Danish Twin Registry was formed in 1954, it contained 8000 pairs born during 1870-19 10. By January 1, 1968, 352 fatal “coronary occlusions” had occurred (19). The rate of concordance calculated by the twin proband method (1) was found to be significantly higher for MZ pairs than DZ pairs both for males (39 versus 26%) and females (44 versus 14%). The authors concluded that environmental factors might be “strong enough to outweigh the genetic influences in men but not in women.” Since the establishment of the Swedish Twin Registry, both comprehensive questionnaire studies (4,7, 8) and studies on subsamples (9, 20, 21) have been carried out. The genetic influence on the end result, IHD-death and CVD-death, seems to be of special interest. Data on a 12-yr mortality follow-up are now available. For comparison, data on cancer concordance will also be presented. MATERIAL
The Swedish Twin Registry was set up during the years 1959-1961. At that time it contained about 10 900 pairs, which covered 95% of all Swedish samesexed twin pairs born in this country between 1886 and 1925 and in which both members were alive when the registry was formed. The compilation procedure and the demographic structure of the twin series have been described by Cederlof (5). The zygosity of the twins in the registry has been determined with the aid of similarity questions contained in a questionnaire mailed to each twin (5). Mortality has been followed since 1962; up to December 1973, 2780 deaths have occurred. METHODS
The mortality has been established as follows. The total twin registry is matched regularly against a total death registry for Sweden at the Central Bureau of Statistics (once a month since 1971). This matching provides access to the death certificate and the name of the physician who signed it. For twins born during 1886-1900, the cause of death has been taken from the death certificates only, but for twins born during 1901-1925, hospital records, autopsy protocols, and information from general practitioners have been collected as well. The cause of death has then been judged according to these multiple sources (9, IS) and classified in line with the 1965 revision of International Statistical Classification (ISC) of Diseases, Injuries, and Causes of Death. For the present report, the causes of death have been grouped in the following way (ISC numbers in parentheses): IHD (410.00-414.99, 795.99),
FORUM:
CARDIOVASCULAR
DISEASE
511
EPIDEMIOLOGY
TABLE 1 CAUSES OF MALE DEA-~HHSACCORDING TO AGE Born 1886-1900 (No.) (%I All causes IHD CVD Cancer, total Cancer, lung Cancer, other Suicides Accidents Other
703 227 96 146 24 122 7 26 201
Born 1901-1925 (No.) (%)
100.0 32.3 13.7 20.8 3.4 17.4 1.0 3.7 28.6
717 219 48 176 40 136 46 66 162
100.0 30.5 6.7 24.6 5.6 19.0 6.4 9.2 22.6
Born 1886-1925 (%I (No.) 1420 446 144 322 64 258 53 92 363
100.0 31.4 10.1 22.7 4.5 18.2 3.7 6.5 25.6
CVD (430&O-438.99), cancer of the lung (162.01-162.19), cancer other forms (140.01-161.99, 163-239.99), suicides (E950.9-E959.9), accidents (800.00-999.99, E807.0-E949.9, E960.9-E999.9), and other causes (conditions not specified here). The concordance rate given is the so-called pair-wise concordance rate (1, 3 1). Results are also given for the expectancy of coincidence, based on the square of the prevalence rate, i.e., the probability of a pair’s being concordant due to chance alone. Comparisons are then made between the observed and expected coincidences. In the comparison of MZ and DZ concordance rates, the chi-square test was used. RESULTS
Tables 1 and 2 show the causes of death for men and women according to age groups. The distribution of the various causes seems to be fairly proportionate with that of the causes of death for the general population of Sweden. IHD is the major cause of death for men in both age groups, whereas cancer is the predominant cause of death for women, mainly due to the excess cases among those
9wEuM
lw
TWIW nEoM9v
1aeoopbr
*oRIAUN
fomw-UP
5?
Er pain’
FIG. 1. Mortality
follow-up
DEATH MS” Pam
of the Swedish Twin Registry.
512
DE FAIRE,
FRIBERG TABLE
AND
LUNDMAN
2
CAUSES OF FEMALE DEATHS ACCORDINGTO AGE Born 1886-1900 (No.1 m All causes IHD CVD
787 199 137 170 10 160 3 13 265
Cancer, total Cancer, lung Cancer, other Suicides Accidents Other
Born 1886-192.5 (No.1 (%I
Born MI-1925 (No.1 (%I
100.0 25.3 17.4 21.6 1.3 20.3 0.4 1.7 33.7
573 82 64 235 5 230 23 27 142
100.0 14.3 11.2 41.0
100.0
1360 281 201 405 15 390 26 40 407
0.9 40.1 4.0 4.7 24.8
20.7 14.8 29.8
1.1 28.7 1.9 2.9 29.9
born during 1901- 1925. Among the female twins born during 1886-1900, IHD has outweighed cancer as the cause of death. The relative distribution of causes of death seems to be about the same for MZ and DZ twins. Of the male MZ twins, 3 1.9% had died from IHD as compared to 3 1.6% of the DZ twins; the corresponding proportions for CVD and cancer deaths among the male MZ and DZ twins were 9.5 versus 10.2% and 23.2 versus 22.6%, respectively. The proportions for IHD among females were 22.5% among MZ twins and 19.8% among DZ twins; for CVD and cancer the figures were 15.2 versus 14.6% and 29.0 versus 30.0%, respectively. At the end of December 1973, 2780 deaths, 1420 male and 1360 female, had taken place (Fig. 1). At that time 259 male and 222 female pairs had been death concordant, i.e., both members in a pair had died, and 85 of the male and 79 of the female pairs were also concordant with respect to cause of death as listed in Tables 1 and 2. Death concordance rates for the whole twin population are given in Fig. 2. Male MZ twins revealed a significantly (t, < 0.05) higher concordance rate for IHD than the male DZ twins (15.8 vs 8.0%). A similar trend is noted
22 PO p-Ml5
2 5
7 s
as.
as.
Cmmrdmt
@ain
10 12 as.
5 9 n.9.
tin ltd.
Cmmordut
win
FIG. 2. Death concordance rates for twin pairs born during 1886-l 925; (left) males, (right) females.
FORUM:
CARDIOVASCULAR
DISEASE
513
EPIDEMIOLOGY
%
IN8
CVD
n M2 0 DZ
CMCER CANCER
12 13
1.8.
1
1
N.S.
5
6
Cmmdant
8
pain
9
N.S.
Il.%
4
9
2 9
as.
N.S.
3. Death concordance rates for twin pairs born during 1886-l 900; (left) males, (right) females.
FIG.
among the female twins, but the difference (11.0 versus 7.5%) was not significant. With respect to death from CVD and cancer, MZ and DZ twin pairs displayed only slight differences. By looking at both of the age groups, born 1886-1900 and 1901-1925 (Figs. 3 and 4), it is obvious that the difference regarding IHD between the male MZ and DZ twins is still more pronounced among the younger pairs, where it was significant at the 5% level (16.1 versus 5.4%). For the females, the reverse age trend was noted. With regard to CVD and cancer, the differences in concordance rates between MZ and DZ pairs were sometimes somewhat more marked (cf. cancer in females) than for the twin population as a whole, but the differences were not significant. Observed and expected coincidence rates for mortality according to age groups are shown in Tables 3-5. The actually observed number of concordant pairs in relation to the expected coincidence are also given. When the whole twin population (Table 3) is considered, it appears that the observed numbers are considerably higher than the expected ones irrespective of zygosity group or cause of death. The difference between the zygosity groups is not very impressive for CVD and cancer but is relatively pronounced for the male IHD deaths Y.
25
25
1
20
FIG.
IND
CVD
CANCER
p-D.05
I.&
N.S.
7.
20
11.1.
N.S.
1.9.
4. Death concordance rates for twin pairs born during 1901-1925; (left) males, (right) females.
514
FRIBERG AND LUNDMAN
DE FAIRE,
TABLE 3 OBSERVED AND EXPECTED COINCIDENCE RATES WITH RESPECT TO DEATH IN IHD, CVD, AND CANCER IN TWIN PAIRS BORN DURING 1886-1925
IHD Coincidence Observed MZ DZ Expected MZ DZ Observed/expected MZ DZ
Males
CVD Females 10
22 20 3.93
5.6 3.3
Females
Males
Cancer Females
2 3
5 9
7 9
1.27 1.89
0.35 0.63
0.58 1.03
2.08 3.12
2.11 4.34
7.9 6.3
5.7 4.8
8.7 8.8
3.4 2.9
2.4 2.5
12
6.11
Males
5
11
TABLE 4 OBSERVED AND EXPECTED COINCIDENCE RATES WITH RESPECT TO DEATH IN IHD, AND CANCER IN TWIN PAIRS BORN DURING 1886-1900
IHD Coincidence Observed MZ DZ Expected MZ DZ Observed/expected MZ
DZ
Males
CVD Females
Females
Males
Males
CVD,
Cancer Females
12 13
1 1
4
5
9
8
6
2 8
6.18 10.20
3.33 4.99
0.74 1.73
1.40 2.50
2.98 3.37
1.85 4.04
1.9 1.3
2.4
1.4 0.6
2.9 3.2
1.7
1.1
1.8
2.0
8
1.8
TABLE 5 OBSERVED AND EXPECTED COINCIDENCE RATES WITH RESPECT TO DEATH IN IHD, CVD, AND CANCER IN TWIN PAIRS BORN DURING 1901-1925
IHD Coincidence Observed MZ DZ Expected MZ DZ Observed/expected MZ DZ
Males 10
CVD Females
7
2 3
0.98
0.14
1.82
0.21
10.2
14.5
3.8
14.5
Males 1 2
Females 1
Males
Cancer Females
1
2 3
3 3
0.05
0.07
0.57
0.10
0.13
1.32
0.91 1.81
20.0 20.0
13.4 1.5
3.5 2.3
3.3
1.7
FORUM:
CARDIOVASCULAR
DISEASE
EPIDEMIOLOGY
515
(5.6 and 3.3, respectively). Subgrouping according to age group (Tables 4 and 5) leads to ratios between the observed and expected values that are generally smaller for the older age groups, especially for IHD and CVD. With respect to the younger twins born 1901-1925, usually high ratios are seen for some of the groups. It must be remembered that the subgrouping of the twin pairs according to age groups gives a limited number of concordant pairs, especially for CVD, and that these figures must be interpreted with great caution. With regard to IHD, the observed number of concordant male deaths in relation to the expected number is seen to be more than 2.5 (10.2 and 3.8) times higher among MZ as compared to DZ pairs. DISCUSSION
The findings of the present report are in accordance with earlier results with respect to fatal coronary occlusion in Danish male twins (19). As regards females, our results diverge from the Danish follow-up, which found a more pronounced difference in concordance rate between MZ and DZ pairs for females than males (19). To some extent this divergence could probably be explained by the relatively few concordant female pairs in both of the studies. The findings of the present study are especially interesting in light of the recent results on 205 death discordant twins from the Swedish Twin Registry (9). When myocardial infarction, angina pectoris, pathologic Q-wave, and ST depressions in connection with exercise were successively included in the criteria for IHD, the prevalences of these combinations were revealed to be higher among the cotwins whose partners had died from IHD than among those whose partners had died from other causes. This applied to both the male and female groups, but the difference was more pronounced for the male pairs, especially between the MZ pairs. As could be expected, most of the biologic risk factors measured showed the same trend, but the differences were small, thus indicating that the genetic influence also may be transmitted through unknown biologic factors. Only slight differences were noted with regard to the environmental factors of smoking, extra work, and physical inactivity. These two studies can be interpreted as showing that both the development of IHD and death from it are under a relatively strong genetic influence, which is most clearly demonstrated in males under the age of 70. With regard to death from CVD, the rates of concordance were about the same for MZ and DZ twins, but the number of observed concordant pairs was far above what could be expected from chance alone in both MZ and DZ twins, especially in the younger age groups. These data point to a relatively strong, mainly environmentally governed family influence on death from CVD. The number of concordant deaths is as yet too small to warrant more definite conclusions. As regards cancer mortality, the present findings confirm the results of an earlier investigation on morbidity and mortality based on the Swedish Twin Registry (6). Only slight differences in death concordance rates were found between MZ and DZ pairs. Comparisons between expected and observed coincidence of death showed that the actual within-pair coincidence exceeds the random expec-
516
DE FAIRE,
FRIBERG
AND
LUNDMAN
tancy in both MZ and DZ twins by about 2-3 times. These data indicate a moderate family influence, which is in agreement with other twin studies (6,18). ACKNOWLEDGMENTS This investigation was supported by grants from the Swedish National Association against Heart and Chest Diseases, the Swedish Medical Research Council, and the American Medical Association, Education and Research Foundation.
REFERENCES 1. Allen, C., Harvald, B., and Shields, J. Measures of twin concordance. Acta Genet. Base1 17,475 (1967). 2. Alter, M. Genetic factors in cerebrovascular accidents. Trans. Amer. Neurol. Assoc. 92, 205 (1967). 3. Biorck, G. Environment disorders of the circulatory system. Bull. N.Y. Acad. Med. 35, 3 (1959). 4. Cederliif, R. Urban factor and prevalence of respiratory symptoms and “angina pectoris”: A study on 9,168 twin-pairs with the aid of mailed questionnaires. Arch. Environ. Health 13, 743 (1966). 5. Cederliif, R. “The Twin Method in Epidemiological Studies on Chronic Disease.” Karolinska Institutet, Stockholm, 1966. 6. Cederlof, R., Floderus, B., and Friberg, L. Cancer in MZ and DZ twins. Acta Genet. Med. Gemetlol.
19, 69 (1970).
7. Cederlof, R., Friberg, L., and Jonsson, E. Hereditary factors and “angina pectoh”: A study on 5,877 twin-pairs with the aid of mailed questionnaires. Arch. Environ. Health 14, 397 (1967). 8. CederlSf, R., Jonsson, E., and Kaij, L. Respiratory symptoms and “angina pectoris” in twins with reference to smoking habits: An epidemiological study with mailed questionnaires. Arch. Environ.
Health
13, 726 (1966).
9. de Faire, U. Ischaemic heart disease in death discordant twins: A study on 205 male and female pairs. Acta Med. Stand. Suppl. 568 (1974). 10. Deutscher, S., Ostrander, L. D., and Epstein, F. H. Familial factors in premature coronary heart disease-a preliminary report from the Tecumseh Community Health Study. Amer. J. Epidemiol. 91, 233 (1970).
11. Edwards, J. H. The genetic basis of common disease. Amer. J. Med. 34, (1963). 12. Edwards, J. H. Familial predisposition in man. Brit. Med. Bull. 25, 58 (1969). 13. Falconer, D. S. The inheritance of liability to certain diseases, estimated from incidence among relatives. Ann. Hum. Genet. London 29, 51 (1965). 14. Frederickson, D. L. Mutants hyperlipoproteinaemia and coronary artery disease. Brit. Med. J. 2, 187 (1971). 15. Friberg, L., Cederlof, R., Lorich, U., Lundman, T., and de Faire, U. Mortality in twins in relation to smoking habits and alcohol problems. Arch. Environ. Health. 27, 294 (1973). 16. Gertler, M. M., and White, P. D. “Coronary Heart Disease in Young Adults: a Multidisciplinary Study.” Harvard University Press, Cambridge, Mass., 1954. 17. Hammond, E. C., Gartinkel, L., and Seidman, H. Longevity of parents and grandparents in relation to coronary heart disease and associated variables. Circulation 43, 3 1 (197 1). 18. Harvald, B., and Hauge, M. “Hereditary Factors Elucidated by Twin Studies: Genetics and the Epidemiology of Chronic Disease.” U. S. Dept. of Health, Education and Welfare, Washington, D.C., 1965. 19. Harvald, B. and Hauge, M. Coronary occlusion in twins. Acta Genet. Med. Gemellol. 19, 248 (1970). 20. Liljefors, I. Coronary heart disease in male twins: Hereditary and environmental factors in concordant and discordant pairs. Acta Med. Stand. Suppt. 5 11 (1970). 21. Lundman, T. Smoking in relation to coronary heart disease and lung function in twins: A co-twin control study. Actu Med. Stand. Suppl. 455 (1966). 22. Marshall, J. Familial incidence of cerebrovascular disease. J. Med. Genet. f$, 84 (197 1).
FORUM: CARDIOVASCULAR
DISEASE EPIDEMIOLOGY
517
23. Marshall, J. Familial~incidence of cerebral hemorrhage. Stroke 4, 38 (1973). 24. McKusick, V. A., and Murphy, E. A. Genetic factors in the etiology of myocardial infarction, in “The Etiology of Myocardial Infarction” (T. N. James and J. W. Keys, Eds.). Little, Brown and Company, Boston, 1963. 25. Phillips, R. L., Lilienfeld, A. M., Diamond, E. L., and Kagan, A. Frequency of coronary heart disease and cerebrovascular accidents in parents and sons of coronary heart disease index cases and controls. Amer. J. Epidemiol. 100, 87 (1974). 26. Rose, G. Familial patterns in ischaemic heart disease. Bit. J. Prev. Sot. Med. 18, 75 (1964). 27. Russek, H. I., and Zohman, B. L. Relative significance of heredity, diet, and occupational stress in coronary heart disease of young adults. Amer. J. Med. Sci. 235, 266 (1958). 28. Shanoff, H. M., Little, A,, Murphy, E. A., and Rykert, H. E. Studies of male survivors of mycardial infarction due to “essential” atherosclerosis. I. Characteristics of the patients. Can. Med. Assoc. 1. 84, 519 (1961). 29. Slack, J., and Evans, K. A. The increased risk of death from ischaemic heart disease in first degree relatives of 121 men and 96 women with ischaemic heart disease. J.Med. Genet. 3, 239 (1966). 30. Smith, C. Recurrence risks for multifactorial inheritance. Amer. J. Hum. Genet. 23, 578 (197 1). 3 1. Smith, C. Correlation in liability among relatives concordance in twins. Hum. Hered. 22, 97 (1972). 32. Thomas, C. B., and Cohen, B. H. Familial occurrence of hypertension and coronary artery disease, with observations concerning obesity and diabetes. Ann. Intern. Med. 42, 90 (1955). 33. White, P. I). The genes, the heart and destiny. New Engl. J. Med. 256, 965 (1957).
