8. Premature ageing in men and women can affect reproduction

THE AGEING OF FATHERS

In the 1950s Penrose of University College, London (1955), studied the ages of the parents of babies with birth defects believed to have been caused by gene mutations. He concluded:

"The influence of the father's age is shown to he of critical significance. When the effect of the father's age on incidence is appreciable, as in achondroplasia, the hypothesis of fresh gene mutation as the cause is strengthened."

Achondroplasia is the commonest form of skeletal dysplasia leading to dwarfism. An organization was formed in 1960 called, The Little People of America, open to individuals under 58 inches (147cm) in height. A team at Johns Hopkins University co-operated with the LPA to compile information of value in counselling achondroplastics and their parents (Murdock et al., 1970). In a series of 148 cases it was found that 31 achondroplastics had one or both parents affected, but in 117 cases there was no family history of achondroplasia. Father's age at the birth of the child was available in 102 cases and has been used in Figure 8.1 where an exponential curve has been fitted to the data. The risk of achondroplasia increases exponentially with the age of the father.

The prevalence of many other diseases increases with age and the increase generally follows curves that are approximately exponential over the age range used in Figure 8.1. Age-specific mortality rates from many causes also increase exponentially with age. Indeed the age-specific mortality for all diseases shown for the USA in Figure 8.2 is approximately exponential at the lower ages. Figure 8.1 relates, however, only to a small minority. Figure 8.2 relates to a minority of individuals who as a result of illness aged prematurely.

It is everyday experience that some people age more rapidly than others and this is apparent from the mortality curves from most causes of death. Figure 8.1 describes a male mutation rate that increases with age producing a condition in offspring, achondroplasia, but as it relates only to a small minority of fathers it prompts the question: Do all men have a mutation rate that increases exponentially with age or only a minority? Figure 8.2 then also prompts further questions: Are there particular diseases that accelerate ageing and also increase mutation rate thus affecting reproduction? Ageing is inevitable, but premature ageing may not be inevitable and disorders responsible for premature ageing may be prevented, or treated to delay their effects. Are the mutations in male germ cells leading to congenital disorders an inevitable and inescapable result of ageing, or can the risk of mutations be reduced in some cases by treatment of illnesses associated with premature ageing?

Achondroplasia is an example of a disorder caused by an autosomal dominant mutation with a frequency estimated to vary from 15 to 100 cases per million births in different countries. Each particular, specific autosomal dominant mutation is rare but 1,443 (+ 1,144 not fully validated) types are described in McKusick's catalogue (1988). Vogel (1984) shows graphically the increase of mutation rate for some of these diseases with patemal age. The number of such distinct dominant mutations described in the literature continues to increase. Friedman (1981) estimated that the risk of a father aged over 40 having offspring with autosomal dominant disease was between 0.3 and 0.5 per cent and that this was about one-third of all babies with diseases due to new autosomal dominant mutations.

The age of a man at the conception of his children may affect the health of his grandchildren. Thus a mutation in a father's X-chromosome carried by his daughter may not be expressed in the daughter but only in his grand-son. McKusick (1988) lists 139 (+171 not fully validated) X-linked diseases. flaemophilia A is such an X-linked disease and is shown in Figure 8.3 to increase in prevalence with the age of the sufferer 's maternal grandfather at his mother's conception (Herrmann, 1966). The rate of increase with age of these grandfathers is very similar to that for the age of the father in achondroplasia and an exponential curve fits the data well. Duchenne muscular dystrophy is another X-linked disorder. X-linked diseases can also be produced by maternal mutations but are commoner in the male. One paper from the University of Colorado says (Lubs. 1981):

"The results from this study suggest that the mutation rate in the haemophilia gene in sperm cells is two to four-fold greater in sperm than in egg cells."

Karp (1980) in the American Journal of Medical Genetics commented on the increased mutation rate with age:

"Such a high rate of gene mutation due to advanced paternal age would have considerable implications for genetic counselling. For example, should we be encouraging obstetricians to screen for 'pregnant husbands' over the age of 40 and offer them counselling, as they would do for women over 35?"

This writer concluded that the "primary educational thrust" should be aimed at the non-pregnant, prepregnant population to encourage child-bearing before age 35:

"Some of them might plan differently if they knew that there existed a risk of undiagnosable, serious genetic disease due to advanced paternal age. It seems to make sense in terms of the overall prevention of birth defects to advise men as well as women to have their children, if possible, before the age of 35 or 40. Such an approach would probably benefit not only the immediate offspring, but subsequent generations as well."

This general advice to parents is repeated in other medical journals (Jones et al., 1975; Stoll et al., 1982). This advice is not, however, appropriate for all men or the majority of men if the increase in risk with age is a consequence of the premature ageing of a minority. The advice is also not useful to parents already in their late 30s or to their medical advisers. The question that may perhaps be useful is whether the prospective father belongs to the minority with an exceptionally high mutation rate and if so whether treatment is possible. This question can only be useful for prospective fathers and it is too late when wives are already pregnant.

DISEASES WHICH CAUSE PREMATURE AGEING

Mutation rate varies more between individuals from causes unconnected with age than it does by ......increase in age during the reproductive years. An increase in the average number of chromosomal anomalies with age has been shown in somatic cells (Schneider, 1978, 1980). An increase in the average number of micronuclei or chromosomal fragments in lymphocytes with age has also been found (Fenech & Morley, 1985). However these studies show great vanations between individuals. Some men in their 20s have more micronuclei than other men in their 60s and 70s. Furthermore the number of micronuclei only about doubled between the ages of 30 and 60 in Fenech and Morley's investigations and the "increase in mutation rate with donor's age" was attributable to a minority with high rates. It has also been shown in in vitro experiments that the proliferative capacity of human cells declined with the age of the donor, but that the cells of some donors in their 60s and 70s had a greater proliferative capacity than other donors in their 20s and 30s (Martin et al., 1970). Doubt is also thrown on these and other experiments by their choice of "healthy" individuals as donors because by doing so they were excluding most of the individuals with high mutation rates and with rates increasing most rapidly with age. The evidence suggests that much if not all the increase in rates with age is attributable to individuals who are not healthy or are manifestly ill.

The ageing of germ cells might be thought of as proceeding throughout adult life at a rate which is independent of other body systems. The evidence suggests on the contrary, as discussed in earlier chapters, that the integrity of germ cells depends throughout life upon the hormonal and nutrient content of the blood and furthermore that germ cells can be seriously damaged and mutated by viruses at any age. It is therefore to be expected that the ageing of germ cells is associated with the ageing of oth2er body systems. A disease affecting a particular body system begins the premature ageing associated with chronic disease and an increased mutation rate. The increased human mutation rate associated for example with multiple sclerosis is illustrated in Figure 8.4 (Emerit & Marteau, 1971; Unglaub-Leisten et al., 1975).

The raised mutation rates associated with ulcerative colitis and Crohn's disease are illustrated in Figure 8.5 (Emerit et al., 1972, 1974). Rheumatoid arthritis and a variety of other rheumatic diseases have been reported to be associated with raised mutation rates (Sherer et al., 1981). There is a variety of uncommon inherited diseases for example Fanconi's anaemia which are associated with chromosomal instability and raised mutation rate (Schroeder-Kurth et al. 1989). These particular diseases have in common that they are associated with microscopic lesions leading to sclerosis. The repair of all tissues, and wound-healing, requires DNA and RNA synthesis and the raised mutation rate is evidence of a reduced rate of DNA synthesis and repair. Emerit et al. (1974) found that the blood serum from patients with scleroderma is mutagenic in the culture of cells from normal individuals. An aberration in the composition of blood serum, resulting from a disease which produces a raised mutation rate, may damage germ cells before conception and may continue to disturb reproduction after conception.

Diabetes is one of the ageing diseases that increases mutation rate in men. Fairbum et al. (1982) from the University of Oxford summarized 7 studies which showed that between 35 and 59 per cent of male diabetics were impotent. Sexual disturbances were found in men with only mild metabolic disturbance who were leading a normal life. Diabetic men have depressed hormonal responses, as illustrated in Figure 8.6 for the gonadotrophin responses to LH-RH in male diabetics (Distiller, 1975). The reduced androgen levels result in reduced spermatogenesis and in particular in a reduced rate of replication of spermatogonia. Diabetes has been listed as one of the disorders responsible for premature ageing (Goldstein, 1978). Ando et al. (1984) from the University of Calabria and Ghent reported depressed levels of the androgens testosterone and dehydroepiandrosterone in diabetics and concluded:

"Diabetes could have a similar effect to that of ageing suggesting a precocious alteration in both gonadal and adrenal secretion in male diabetics."

Both the incidence of diabetes and the mortality increase approximately exponentially with age like the incidence of autosomal dominant diseases with father's age. United States returns for all races show an incidence in 1982-5 of 5.8 per 1,000 for males under 45 rising to 52.6 per 1,000 in the age range 45 to 64. Male mortality from diabetes in the USA is shown in Figure 8.7 to increase approximately exponentially with age up to about age 50.

TABLE 8.1

COMPARATIVE PREVALENCE OF DIABETES AND OF THYROID DISORDERS IN MEN AND WOMEN, COUNTY DURHAM, 1977, 2,779 PEOPLE. .
percentages
femalemale
overt hyperthyroidism 1.9-2.70.16-0.23
overt hypothyroidism 1.4-1.9 0.10
subclinical hypothyroidism 7.52.8
overt diabetes 0.8
high fasting glucose level 1.8

Source: Tunbridge et al., 1977.

Diabetes is only one example of an endocrine disorder that affects the outcome of pregnancy. A survey was conducted of 2,779 individuals in Whickham, County Durham to determine the prevalence of thyroid disorders (Tunbridge et al., 1977). Table 8.1 compares the prevalence of thyroid disorders and diabetes. This study suggests that thyroid disease may be comparable in prevalence to diabetes. Bakke et al. (1975) at the University of Washington showed that partial removal or impairment of the thyroid in male animals before mating to a normal female produces birth defects in the offspring. Some of these defects, when compatible with survival, were found to be heritable. Thyroid disease, like diabetes, is mutagenic in male as well as female. How ever the levels of both clinical and sub-clinical disorder at which male mutation rate is seriously increased is not well enough established to estimate the comparative importance of diabetes and thyroid disorders. In cases of overt hyperthyroidism drug treatment is usual and effective. The drugs given, such as propylthiouracil, if given in excess do, however, produce hypothyroidism and therefore can be mutagenic if dose is not very well controlled.

Thyroid disorders, diabetes, multiple sclerosis, ulcerative colitis, rheumatoid arthritis, pemicious anaemia are associated with raised mutation rates. These are diseases that have an increasing prevalence with age and are responsible for at least an important part of premature ageing. The fathers who are ageing prematurely can identify themselves or be identified and can be helped. There are treatments for most of the ageing diseases that can at least reduce risks.

An increase in mutation rate with age cannot be explained by any increased exposure to radiation or smoke or pyrolysed protein or mutagenic chemicals. A simple linear accumulation of mutations with age would not produce the exponential curves associated with the risk of mutagenic disease from the ageing of fathers or mothers. The exponential curve is the most basic of all curves generated by circular processes. Diabetes mellitus is one example of a mutagenic disease that affects not only germ cells but is thought to be caused partly by somatic mutation, in this case within the immune system, causing autoimmune disease within the pancreas (Adams et al., 1984). Mutation is part of the process of premature ageing. The conclusion of the studies briefly reviewed here is that chronic diseases in men should receive attention in anticipation of a pregnancy equal to that advocated presently for a woman.

THE PREMATURE AGEING OF WOMEN CAN AFFECT CHILDBEARING

An association between diabetes and birth defects has been the subject of reports for more than a century. Lecorcbe (1885) wrote many years before the advent of insulin:

"If diabetes does not always completely prevent fertilization, it seems to cause profound impairment of the products of conception, subverts nutrition, shortens life or causes developmental defects incompatible with life."

Diabetes in women is prominent among the diseases of premature ageing that affect reproduction (Kram & Schneider, 1978). In the USA the reported incidence of diabetes in women in 19799781 was 6.9 cases per 1,000 women under age 45 rising to 55.1 cases for women aged 45 to 64 (US National Center for Health Statistics, 1987). The steep increase in incidence with age is reflected in the exponential increase in mortality with age of women from diabetes shown in Figure 8.7.

Diabetes is an endocrine disorder and might be expected to disturb gonadotropin secretion in women as in men. Diabetes does in fact depress the hypothalamic-ovarian axis as illustrated in Figure 8.8 showing a depressed response to 100 p.g of LH-RH by 22 diabetic women (Distiller et al., 1975). All except two were on insulin therapy, but were tested before the morning dose which was delayed. The depression of gonadotrophin levels prompts many questions about the effects of diabetes on reproduction. Is the premature increase of womens' infertility with age partly a consequence of diabetes? How far does diabetes increase the risk of birth defects? Is diabetes partly responsible for the increased risk of chromosomal abnormalities including Down's syndrome with maternal age? Is diabetes mutagenic? A Leader in the Lancet said (22 March 1980):

"If preimplantation, and perhaps even preconceptual, stages of development are vulnerable to the effects of maternal diabetes then clearly the management of diabetes needs to precede conception."

This quotation poses the question as to whether diabetes does affect female germ cells.

Japanese-American teams at Niigata and Boston Universities were the first to report that diabetes is mutagenic in female mice and interferes with meiosis causing diverse chromosomal abnormalities (Endo & Ingalls, 1968). Diabetes in female mice was found to be associated with chromosomal breakage, polyploidy and aneuploidy in the cells of embryos removed 19 days after mating. The percentages of the cells with aneuploidy are shown in Figure 8.9.

The mice were made diabetic with alloxan and were mated within two weeks of treatment. Ninety-eight per cent of the control mice were fertile but only 44 out of 200, or 22 per cent of the treated mice. Twenty per cent of the diabetic mice had one or more malformed foetuses including 13 with serious malformations, for example cranioschisis, spina bifida, hernia and cleft palate. Two malformed foetuses had over 90 per cent of abnormal chromosomes. The alloxan intake was adjusted so that it did not cause infertility on the one hand and was not too low on the other hand to produce any effect (Endo, 1966). The diabetes produced casualities only at critical levels of severity and caused infertility if more severe. In a later Japanese study of diabetes in mice, also using alloxan, ova were harvested only 3bd days after mating at about the 16-cell stage. The aneuploidy, the polyploidy and chromosomal gaps and breaks were found to be present already at this stage (Yamamoto et al., 1971). The authors stress that ovulation was disturbed and most chromosomal anomalies probably had their origin before mating.

The effects of diabetes on ovulation in female rats have been studied without using diabetogenic drugs at the Institute of Physiology in Buenos Aires by Chieri et al. (1969) who removed part of the pancreas. Diabetes reduced the rate at which ova matured within the ovary before ovulation and also reduced the number of ova ovulated. The ovulation rate in these experiments was quickly restored by insulin. Furthermore insulin restored ovulation rates more study concluded:

"Insulin therapy was able to restore the number of eggs to a normal level suggesting that the ovulatory disturbance was related critically to the insulin deficit."

Ovulation in these animal experiments was unaffected by high or low blood glucose levels if the insulin levels were normal. It was the insulin levels that were found to be critical during ovulatory maturation and conception. In the culture of cells in the laboratory insulin is also a critical and essential component of the culture medium for most types of cell including human cells.

Because diabetes depresses the hypothalamic-gonadal axis in both men and women it would be expected to cause infertility. Before the discovery of insulin from 95 to 98 per cent of diabetic women were reported to be infertile (Gellis & Hsia, 1959). Insulin is highly effective in restoring fertility. It is, however, much easier to restore fertility than to ensure a satisfactory birth outcome.

Diabetic management has reduced the perinatal death rate impressively in recent years, but there have been many reports of increases in malformation rates among survivors. Ballard et al. (1984) discussed this history and listed a number of studies up until 1980 reporting malformation rates ranging up to 20 per cent with a typical figure around 10 per cent for the children of diabetic mothers. Ballard reported a major malformation rate of 16.8 per cent for the years 19709778 from the University of Cincinnati for diabetic pregnancies that had only been managed after diagnosis of pregnancy. Diabetic management needs to precede conception to prevent birth defects.

AGE-DEPENDENT AND AGE-INDEPENDENT AETIOLOGY

There are many causes of raised mutation rate such as influenza, or a temporary nutritional deficiency, or exposure to a mutagenic chemical which are unrelated to parental age. An inherited propensity such as a translocation is also unconnected with parental age. It would therefore be expected that some part of the prevalence of birth defects would be independent of age and another part would be a function of age. Writers on ageing over many years have used models to represent morbidity and mortality which included one term to represent the age-independent component and a second term for the age-depend ent component, generally exponential. Greenwood (1'8), an actuary, suggested that the age-dependent term represented "physiological constitution" in contrast to the age-independent term reflecting enviromental causes, for example transport accidents. Lamson & Hook (1980) applied this model to Down's syndrome and Table 8.2 shows in 5 populations the proportion of cases that are apparently independent of matemal age and those which are age-dependent.

TABLE 8.2

PROPORTION OF CASES OF DOWN92S SYNDROME ATTRIBUTABLE TO CAUSES DEPENDENT AND INDEPENDENT OF MATERNAL AGE: SUMMARY OF 5 STUDIES.

Population independent of age dependent upon age
percentages
Massachusetts 36.3 63.7
New York 40.4 59.6
Australia 45.3 54.7
British Columbia 45.8 54.2

Source: Lamson & Hook, 1980.

Figure 8.10 is a diagrammatic representation separating the numbers of cases of Down's syndrome that would have been expected at the rate for age 20 to 24 and the numbers attributable to the ageing effect; 54 per cent in this figure were age-independent and 46 per cent age-dependent. Friedman (1981) superimposed the frequency curves of Down's syndrome and maternal age, and achondroplasia and paternal age and showed that they were very similar.

The prevalence of every type of birth defect can be roughly divided for any particular population into parental age-independent and age-dependent components. The age-dependent components are substantial for Down's syndrome and for most other birth defects associated with chromosomal abnormalities. Figure 8.11 shows the increased incidence of Down's syndrome notifications with maternal age, and Figure 8.12 the notifications of other chromosomal abnormalities with maternal age, both figures using data for England and Wales. The risk of most birth defects increases with maternal age but for most types, for example for congenital heart disease, the age-dependent component is much smaller than for birth defects associated with chromosomal abnormalities.

Diabetes is a prominent disease of premature ageing and would be expected to be particularly associated with the birth defects including chromosomal abnormalities with their large maternal age-dependent components of risk. Down's, Klinefelter's and Turner's syndromes are the three commonest onegeneration genetic diseases associated with chromosomal abnormalities. An association between diabetes and these three genetic diseases would be expected from the animal experiments described above which showed that critical levels of diabetes produced chromosomal abnormalities including trisomy and by the evidence that depression of the hypothalamic-gonadal axis can cause non-disjunction, and that diabetes is one of many disturbances that causes such depression (Hansmann, 1984).

There are studies from several countries showing an association of parental diabetes and the three major syndromes associated with chromosomal abnormality. Navarrete et al. (1967) reported from the First Obstetric Hospital, Mexico City, abnormal glucose tolerance tests of 9 out of 12 mothers of Down's syndrome babies. Milunsky (1969) of Tufts' University, having found a family history of diabetes in 209 out of 393 children with Down's syndrome, made a detailed study of the parents of 42 Down's syndrome children and found that 16 (38 per cent) showed abnormality in a glucose tolerance test compared with 6.8 per cent of 176 controls with normal children. Forbes and Engel (1963) of Harvard Medical School reported an association between diabetes and chromosomal anomalies in female babies. They studied 41 cases of Turner's syndrome and related disorders characterized by developmental failure of the female gonads. The patients all had defective or absent ovaries associated with the absence of one X chromosome or a defect in this chromosome, the results of a new mutation in one parent. In 20 out of 41 patients there was a family history of diabetes and in 14 the diabetes was on the father's side only. The incidence of diabetes in the families of the 41 patients was 10 times higher than in the general population of Oxford, Massachussetts.

Three years later it was reported from Denmark, Italy, Argentina and the USA that diabetic parents were overrepresented among the parents of male babies suffering from Klinefelter's syndrome, another one-generation genetic disease characterized by male infertility and generally an extra X chromosome resulting from non-disjunction at meiosis I or II. The results of the four studies are aggregated in Table 8.3. It is seen that almost a quarter of the Klinefelter children had a diabetic parent and that 12 out of 23 diabetic parents were fathers. Again the incidence of diabetes in both mothers and fathers of the Klinefelter children was significantly higher than would be expected for any normal population with at most 1 or 2 per cent of diabetics among people in the childbearing ages.

TABLE 8.3

DIABETES MELLITUS IN PARENTS OF PATIENTS WITH KLINEFELTER'S SYNDROME; SUMMARY OF 4 STUDIES

with diabetes
country number of patients fathers mothers percentage of patents with diabetic parent
Argentinea32 4 3 21.8
Denmarkb 312 625.8
Italyc 83 150.0
U.S.A.d 24 3 1 16.7
total 95 12 1124.2

Sources: a Wais & Salvati, 1966; b Nielsen, 1966; c Menzinger et al., 1966; d Zuppinger ., 1967.

The Leader in the Lancet (22 March 1980) referred to the "desirability of fastidious diabetic control" when planning a conception. The prevention of birth defects, including chromosomal aberrations, requires more careful control of diabetes before and around the time of conception than does the restoration of fertility or reduction of perinatal mortality. Preconception care of diabetics is a matter for a specialist hospital clinic and many such clinics are now providing such care. However primary care has the difficult initial responsibility for referring patients to specialist clinics for diagnosis and management. There is much evidence that in every country the coverage of such clinics is limited. Some women reporting to a diabetic clinic for the first time have suffered from diabetes for some years. Some women only reach a diabetic clinic too late when already pregnant. Other diabetic women are never referred to a diabetic clinic at all by the primary care services but refer themselves.

One of the first prepregnancy clinics for diabetes was established in 1976 at the Royal Infirmary in Edinburgh (Steel et al., 1982). Clear accounts of the experience of the Edinburgh clinic emphasize the care and persistence needed by both patients and clinic staff to achieve the tight control of blood compo sition needed. The results of such intensive diabetic control reported by Fuhrmann from the Institute for Diabetes, Karlsberg, are summarized in Figure 8.13 with a larger sample of women. Hospitalisation for a limited period before conception to achieve optimum control and to teach self-management was shown to be effective (Fuhrmann et al., 1983).

By combining the results of numbers of small studies from different countries as separated as Argentina, Denmark, Italy, and the USA, as in Table 8.3 for Klinefelter's syndrome, and from Mexico and the USA for Down's syndrome and Turner's syndrome it may be concluded that about 20 per cent of these disorders might be prevented by strict preconception control of diabetes. The data on which this conclusion is based is manifestly unsatisfactory. The cost of Down's syndrome alone in every country is formidable and would justify the collection of adequate data throwing light on its etiology.

The US Collaborative Perinatal Studies covered 23,000 white women of whom 1.8 per cent had "overt diabetes" and were on "insulin or analog therapy"; and 17.7 per cent of these 424 women had babies with major malformations (Chung & Myrionthopoulos, 1975). This study showed an increase by a factor of 3.5 in the risk of CNS malformations to 24 per 1,000 births. Zacharias et al. (1984) have reported 19.5 cases per 1,000 of babies with neural tube defects bom to white diabetic mothers and 18.0 per 1,000 to black diabetic mothers, increases in both cases more than 10 times that of the non-diabetic births of 1.4 and 0.8 per 1,000. Diabetes increases the risk of many other types of congenital malformation. In no country are the data that have been found good enough to estimate the percentage of total birth defects associated with diabetes of either father or mother. Assuming that 1.6 per cent of women who become pregnant have a degree of insulin deficiency enough to increase the risk of malformations by a factor of 5, then this deficiency would be responsible for 8 per cent of all malformations. The available data are such that this can be no more than a tentative hypothesis. The actual figure could be higher, but it is apparent that men and women who are not diabetic have most of the children with malformations and that diabetes cannot even be the only age-dependent disorder concerned.

Diabetes is a slowly progressive autoimmune disease in which the cells of the immune system, notably the T-cells, slowly destroy the islet cells of the pancreas that produce insulin (Powers & Eisenbarth, 1985). Diabetes is only one autoimmune disease and others include hypothyroidism, Graves' disease, multiple sclerosis, pernicious anaemia, arthritis. The prevalence of these disorders overlaps with diabetes and some 10 to 12 per cent of women diabetics have been reported to develop thyroid disease and 2 or 3 per cent to develop pernicious anaemia (Beral et al., 1984). Another report shows 23 per cent of 144 diabetic women as developing arthritis of autoimmune origin (Cruickshanks et al., 1984). The autoimmune origin of diabetes prompts the question as to whether there is any evidence of an association between the birth of children with chromosomal abnormalities and other autoimmune diseases.

Hypothyroidism is among the autoimmune diseases overlapping with diabetes. Case histories of an association of Down's syndrome and hypothyroidism go back to Stolzner (1919). Myers (1938) in Canada found 9 times the prevalence of thyroid disease among the mothers of Down's syndrome children compared with controls. Benda (1949), writing about Down's syndrome, said:

"Thyroid anomalies are so frequently seen that they are an important link in the chain of events. The common denominator is a threshold of sterility."

Thyroid (T3) deficiency like insulin deficiency slows down ovulatory maturation causing infertility, and there is a grey area between infertility and normal ovulation in which meiosis I can be slowed down producing chromosomal abnormalities. Ek (1959) examined 41 mothers of Down's children and found that the thyroid gland was pathologically enlarged in 12, but none had any severe physical disease; the frequency of goitre in the same part of Sweden was 1.1 per cent. The mothers of the 41 Down's children had an average protein bound iodine (PBI) significantly higher than normal. Fialkow et al. (1971) reported 17 per cent of 177 mothers of children with Down's syndrome as having clinical thyroid disease compared with 6 per cent of controls, who were mothers of children with a variety of disorders mainly mental retardation but who were not known to cany chromosomal abnormalities. Thirty-four per cent of the mothers of Down's syndrome children had abnormal levels of thyroid antibodies compared with 15 per cent of controls. Nielsen (1972b) listed 5 different studies with a total of 348 mothers of Down's syndrome children of whom 30.7 per cent had increased thyroid antibodies compared with 13.8 per cent of controls. There is an association between Down's syndrome and maternal thyroid deficiency. Thyroid disease is mutagenic in animals as already noted (Bakke et al., 1975). Partial thyroidectomy in female rats before mating produces a range of congenital malformations (Langman & Faassen, 1955). Thyroid hormone (T3) is essential like insulin in cell culture. The Whickham survey concludes (Tunbridge et al., 1977):

"There is clearly a large reservoir of autoimmune thyroid disease in the community and clinically obvious hyperthyroidism and hypothyroidism are only the extreme ends of the spectrum."

The percentage of women with raised thyrotrophin (TSH) levels increased with age from 4.0 per cent under age 25 to 17.4 per cent over 75, the increase being steepest between 35 and 55, and 12.3 per cent of women had palpable and visible goitres.

Both thyroid disease and diabetes point to autoimmune disease in general as one important cause of chromosomal abnormalities. It cannot however be assumed that the hormonal deficiencies associated with these diseases act directly on germ cells. They depress the hypothalamic-gonadal axis and this alone can disturb meiosis and produce a range of chromosomal abnormalities. There are many papers that discuss the origins of autoimmune disease from which it may be concluded that the pan of the genome concemed with the immune system is particularly susceptible to damage by mutagenic influences. As most if not all autoimmune diseases increase mutation rate they are self generating but could also be initiated by viruses or other mutagens. There are reported cases of viral diseases, for example congenital rubella, causing diabetes and viruses can induce diabetes experimentally in mice. Viruses are, however, only one cause of mutation.

Conclusions