SEARCH

SEARCH BY CITATION

Keywords:

  • intergenerational recurrence;
  • genetic mechanisms;
  • imprinting;
  • pre-eclampsia;
  • congenital defects;
  • gestational age

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. The Medical Birth Registry of Norway
  5. Gestational age
  6. Pre-eclampsia
  7. Birth defects
  8. Conclusion
  9. Acknowledgements
  10. References

Population-based data that cover reproductive health outcomes across two complete generations have recently become available in the Nordic countries. Such data enable estimation of recurrence risks from one generation to the next of different conditions such as birth defects or pre-eclampsia. Risks related to a singleton pregnancy involve the contribution of three individuals: the mother, the father and the fetus. A paternal contribution is mainly through the father’s contribution of half of the alleles of the fetus. A maternal contribution may occur in three fundamentally different ways. First, the mother provides half of the genomic alleles to the fetus, with contribution of paternal alleles completing the whole genome. Second, the mother provides the fetal environment and possible susceptibility to complications during pregnancy which she may have inherited from her mother. Finally, she provides the fetal mitochondria. Because of these different contributions, recurrence from mother to offspring is fundamentally different from recurrence from father to offspring. How recurrence risks reflect and shape the underlying contributions to overall perinatal risk is illustrated through a review of published data from Norway on gestational age, pre-eclampsia and birth defects.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. The Medical Birth Registry of Norway
  5. Gestational age
  6. Pre-eclampsia
  7. Birth defects
  8. Conclusion
  9. Acknowledgements
  10. References

After the thalidomide tragedy in the 1960s, the Scandinavian countries (Norway, Sweden and Denmark) started reproductive health registries with population coverage of births. These registries now cover almost two complete generations. Children born into cohorts in the 1970s and 1980s may be followed into adulthood and their eventual reproduction tracked. As the record of each birth typically also contains the unique person identification numbers for the child, the mother and the father, it is technically easy to perform record linkage of the record of the birth of a child in an early cohort with the birth of its offspring in more recent years. These unique identification numbers also secure a very high reliability of such linkage.

Recurrence risks across generations may then easily be generated and compared with baseline risks in the population for a series of conditions. For continuous outcome variables, correlation or regression methods may be used to estimate the degree of association across generations. For categorical outcomes, the relative risk is one measure that permits an assessment of the degree and strength of association. Although the data are available and these risks and measures of association across generations may be technically easy to estimate, they are complicated to interpret. The factors and mechanisms that contribute to recurrence from mother to offspring are fundamentally different from factors that contribute to recurrence from father to offspring. To understand and appreciate these differences, we need to start by discussing what mechanisms are involved in shaping the risk of a particular pregnancy.

A pregnancy involves three partners: the child, the mother and the father. A graphic representation of this triad, which is often used in genetics, is shown in Fig. 1. An interesting discussion of contributions of each partner, including possible genetic conflicts, to risks during pregnancy has been given by Haig.1 Reproductive health outcomes have traditionally been separated into maternal outcomes (e.g. pre-eclampsia, gestational diabetes and pelvic pain) and outcomes of the child (e.g. birthweight or birth defects). Because a pregnancy outcome may be an effect of maternal, paternal or fetal influences, it may be fruitful to look at a reproductive health outcome as a characteristic of a particular pregnancy, not of either the mother or the fetus.

image

Figure 1. Schematic representation of the different partners of a singleton pregnancy.

Download figure to PowerPoint

In short, the father's main contribution to the pregnancy is half the alleles of the child's genome. The mother contributes the other half of the child's alleles. The alleles of the father and the mother in the child may have similar effects on the pregnancy. If any of the child's genes are subject to genomic imprinting,2 alleles from the father may have an effect on the pregnancy that is different from the alleles from the mother. In addition to contributing alleles to the child's genome, the mother's own alleles may have independent effects on the pregnancy. Her capacity to carry a pregnancy and to interact with the fetus is important in the pregnancy. In addition, the mother also contributes with the child's mitochondria. The effects of these influences on pregnancy are reflected in the pattern of generational exchange of perinatal risks between two generations.

Generational data on reproductive health outcomes reflect the different contributions of mother, father and child in a particular pattern. It is clear that environmental conditions that persist in a family over generations may contribute to recurrence of particular outcomes from one generation to the next. Parents who smoke, for instance, may more often have children who become smokers.3 This paper will, to some degree, ignore such environmental effects. Its purpose is to demonstrate how different effects of the genomes of the mother, the father and the child are reflected in generational data. Any additional effects of the environment should not be forgotten in interpretation of real data (although they are not emphasised in this review).

A graphical representation of generational data is depicted in Fig. 2. The pregnancy that led to the birth of a female infant who later became a mother is shown in the upper left corner. Similarly, the pregnancy that led to the birth of a male infant who later survived to reproduce is shown to the right. The pregnancy of the next generation is represented in the lower half of the figure. According to the figure, generational data following boys who become fathers (right branch of the figure) are fundamentally different from generational data that follow girls who become mothers (left branch of the figure).

image

Figure 2. Schematic representation of genetic effects that are involved in generational data of reproductive health outcomes.

Download figure to PowerPoint

The main genetic influence that shapes recurrence from father to offspring is alleles that worked through the father when he was a fetus and are passed on to his offspring and work in his partner's pregnancy too. Effects of alleles inherited from the father may be different from effects of those inherited from the mother. By a mechanism called genomic imprinting,2 alleles from the father have been shown in experimental studies to contribute more to placental growth.4 A father also passes on a Y-chromosome to his sons, which may contribute to specific father–son recurrence.

The genetic influences that shape recurrence from mother to offspring are even more complex. First, alleles that worked through the mother when she was a fetus are passed on to the offspring and work through her fetus on her own pregnancy too. If genes are imprinted, the effect of these alleles may be different from the alleles from the father. Second, some alleles that worked in the mother in the first generation are passed on to the girl and work through her when she becomes pregnant. Third, mitochondria are transmitted through the maternal line and may contribute to recurrence from mother to offspring. Whether an effect of mitochondria is working through the mother or the fetus would be difficult to identify by epidemiological methods.

A more formal representation of the effects in Fig. 2 involves genetic correlations.5 The purpose of this paper is, however, not formal estimation of heritability due to these different components, but rather to allow a better epidemiological interpretation of generational recurrence risks. We discuss these different components through a variety of examples from our own experience.

It is important to remember that the recurrence that is seen between generations is conditional on reproduction. Affected children who grow up and have children may be a selected subset of the children born with the condition. For some birth defects, for example, selection is so strong that it may be an important factor in understanding how risk is passed on from one generation to the next.

The Medical Birth Registry of Norway

  1. Top of page
  2. Summary
  3. Introduction
  4. The Medical Birth Registry of Norway
  5. Gestational age
  6. Pre-eclampsia
  7. Birth defects
  8. Conclusion
  9. Acknowledgements
  10. References

All examples presented here are based on data from the Medical Birth Registry of Norway. The registry started capturing all births in the country in 1967. It contains information on medical conditions of the mother before and during pregnancy, details about the delivery and characteristics of the newborn child. Any diagnoses of the child that are established within the first week of life are supposed to be recorded. Notification is undertaken by midwives, and is compulsory by law.6 Each birth is identified by a unique national identification number for the child, the mother and the father. These ID numbers make it technically easy and reliable to link the birth record of a boy in an early cohort with the birth of a child where the boy is recorded as the father many years later. Similarly it is easy to identify girls who later become mothers.

The registry contained 2.3 million births over a time period of 38 years from 1967 to 2003. The numbers of father–offspring and mother–offspring pairs that may be linked within the registry vary, depending on the specific time period used in a particular study and what criteria of selection are being used. Details are given for each example below.

Gestational age

  1. Top of page
  2. Summary
  3. Introduction
  4. The Medical Birth Registry of Norway
  5. Gestational age
  6. Pre-eclampsia
  7. Birth defects
  8. Conclusion
  9. Acknowledgements
  10. References

Little is known about which partner in the pregnancy triggers the delivery, whether it is the mother or the fetus. By using generational data for fathers and offspring it should be possible to establish whether there are any genetically determined effects through the fetus.

In a recent paper,7 we studied 77 452 pairs of boys and girls from the first cohorts of the registry who later became couples and parents with follow-up to 2003. We excluded preterm births, non-singleton births and deliveries by caesarean section. Fathers who came from pregnancies of long duration tended to have offspring with pregnancies of relatively long duration. Gestational length of the offspring increased on average 0.58 [95% CI 0.48, 0.67] days for each week increase in father's gestational age. This would support an effect of the fetus and fetal alleles in determining the time of delivery.

The tendency of mothers who came from pregnancies of longer duration to have pregnancies of longer duration was, however, stronger. Offspring gestational age increased 1.22 [95% CI 1.21, 1.32] days for each additional week in mother's gestational age, twice the effect of the father's gestational age. According to Fig. 2, there are several possible explanations for the difference: first, the effect of maternal alleles in the fetus may be stronger than effects of paternal alleles (which may not be very likely). Second, there may be maternal effects mediated by alleles passed on from mother to mother. The third possibility is that there could be strong mitochondrial effects (which may also be unlikely). If the generational effect from mother to offspring of 1.22 days per week is the sum of a fetal and a maternal effect, and the generational effect from father to offspring of 0.58 days per week is only the fetal effect, the fetal effect could be of similar strength as the maternal effect. In other words, the fetus and the mother may be assumed to play equally important roles in determining time of delivery.

A simple interpretation of generational data emerges if we ignore imprinting and mitochondrial effects (and environmental effects): mother–offspring recurrence is determined by the sum of fetal and maternal influences, while the father–offspring recurrence is determined only by the fetal influences.

Pre-eclampsia

  1. Top of page
  2. Summary
  3. Introduction
  4. The Medical Birth Registry of Norway
  5. Gestational age
  6. Pre-eclampsia
  7. Birth defects
  8. Conclusion
  9. Acknowledgements
  10. References

Generational data on pre-eclampsia from Norway was presented in a recent study that followed 286 945 boys who later were recorded as fathers in the birth registry and 438 597 girls who later were recorded as mothers in the registry.8

Fathers who came from a pre-eclamptic pregnancy had a 1.5-fold risk [95% CI 1.3, 1.7] of fathering a pre-eclamptic pregnancy. This finding clearly supports a role of fetal genes in triggering pre-eclampsia. In contrast, mothers who came from a pre-eclamptic pregnancy had a 2.2-fold risk [95% CI 2.0, 2.4] of pre-eclampsia in her own pregnancy. This higher mother–offspring recurrence indicates an additional role of the mother and maternal alleles in pre-eclampsia. Similar to gestational age, the mother–offspring association was about twice the father–offspring association, indicating that maternal effects may be similar in strength to fetal effects.

A note on mitochondrial effects

In another study based on the Medical Birth Registry we found no important effects of mitochondria on pre-eclampsia.9 In an extended linkage across generations we identified two types of half-sisters who were recorded as mothers in the registry (Fig. 3). The first set of half-sisters (maternal half-sisters) had the same mother but different fathers (left panel). The second set (paternal half-sisters) had the same father but different mothers (right panel). The main difference between these two sets is that maternal half-sisters (and their children) have identical mitochondria while paternal half-sisters (and their children) have unrelated mitochondria. The identical mitochondria did not add anything to the recurrence risk from one sister to the other (relative risk of 1.8 for paternal half-sisters and 1.6 for maternal half-sisters); consequently we may assume that mitochondrial effects do not add significantly to the risk.

image

Figure 3. Schematic representation of pregnancies of maternal and paternal half-sisters and the relative recurrence risk of pre-eclampsia from one sister to the other.

Download figure to PowerPoint

Data on children who are paternal half-siblings from the Norwegian registry9 also confirm the interpretation of the generational data8 that the fetal contribution to risk of pre-eclampsia is of the same magnitude as the maternal. There was a 1.8-fold [95% CI 1.2, 2.6] recurrence risk between pregnancies of unrelated women whose pregnancies were fathered by the same father. This recurrence risk is probably affected by fetal alleles shared by the two offspring who are half-siblings and may be directly compared with the effect of maternal alleles reflected in the half-sister data shown in Fig. 3. This may be taken as a confirmation that a fetal effect is of the same order of magnitude as a maternal effect on pre-eclampsia.

Birth defects

  1. Top of page
  2. Summary
  3. Introduction
  4. The Medical Birth Registry of Norway
  5. Gestational age
  6. Pre-eclampsia
  7. Birth defects
  8. Conclusion
  9. Acknowledgements
  10. References

In two separate studies from Norway we followed 459 433 girls with birth defects and 486 207 boys with birth defects to look at their respective tendency to survive till adulthood and have offspring.10,11 In this number we included stillbirths. We also estimated the tendency of birth defects to recur across generations. Altogether 8192 or 1.8% of the girls in the first generation had birth defects. The number of boys with birth defects was higher (12 292 or 2.5%).

The tendency to survive and reproduce was reduced for both girls and boys with birth defects and varied substantially by severity between 24 categories of birth defects.10,11 Eighty per cent of girls with birth defects survived within the follow-up (up to 1992), while 98% of girls without birth defects survived. A similar but smaller difference was seen for boys (84% vs. 97%). The higher survival of boys with birth defects may indicate that boys are affected by more minor conditions. Among all girls, we identified that 27% had reproduced by 1997. For boys, the corresponding proportion was 17% by 1998. These numbers reflect the incomplete follow-up in our second generation data.

The proportion of children with birth defects was 3.8% for mothers who themselves were born with a birth defect, compared with 2.4% for children of mothers without birth defects (relative risk 1.6, [95% CI 1.2, 2.0]). For males with birth defects, the proportion of offspring with birth defects was 5.1%, and for males without birth defects, the proportion in offspring was 2.1%. The lower overall proportion was probably due to missing fathers' information on some stillbirths. Thus, relative risk of recurrence was 2.4 [95% CI 1.7, 2.8] for fathers. This overall recurrence from father to offspring was significantly higher than from mother to offspring (= 0.03) when adjustment is made for a slightly different composition of defects for offspring of fathers and mothers with birth defects.

We should probably be cautious in interpreting the overall difference in recurrence of birth defects from father to offspring (relative risk 2.4) and mother to offspring (relative risk 1.6). It is, however, clear that the difference leaves little room for maternal effects (Fig. 2), which should contribute a higher mother-to-offspring recurrence. The only effect in our framework that could explain a higher father-to-offspring recurrence is imprinting. The data set does not have sufficient power yet for a more specific analysis of imprinting effects for specific categories of birth defects. The two studies from Norway describe a recurrence pattern that was very specific for different categories of defects.10,11

A note on selection and attributable risks

Reduced survival and reproduction for persons born with birth defects have important effects on intergenerational exchange of birth defects. In principle, the number of children with birth defects would be higher if affected individuals had the same probability of reproducing as unaffected parents.

The contribution of parents with birth defects to the occurrence of them in the next generation may be captured by an attributable risk measure. Only 0.5% of birth defects in the next generation was attributable to mothers who themselves had birth defects. For fathers with birth defects, the attributable risk was higher (1.6%) both because more boys were born with birth defects in the previous generation, and because fathers tend to pass on more birth defects to the next generation.

The attributable risk is determined by the relative risk of recurrence and the prevalence of exposed parents (parents with birth defects). The prevalence of exposed parents may be expressed as the original prevalence of birth defects in the previous generation, times the proportion becoming parents. This makes the attributable risk a more direct measure of intergenerational exchange than the relative risk of recurrence that does not incorporate the amount of selection.

Conclusion

  1. Top of page
  2. Summary
  3. Introduction
  4. The Medical Birth Registry of Norway
  5. Gestational age
  6. Pre-eclampsia
  7. Birth defects
  8. Conclusion
  9. Acknowledgements
  10. References

The increasing availability of intergenerational data opens new opportunities in perinatal epidemiology. The new opportunities to separate maternal from fetal effects are among the most interesting. Interpretation of recurrence risks across generations may, however, be challenging and will require care in the design, analysis and interpretation of findings. Simpler and more descriptive epidemiological approaches as reviewed here should be supplemented with methods of more formal statistical estimation of genetic (and environmental) components of disease risk. Such estimation has been employed in a study of pre-eclampsia,12 and should be employed for other outcomes as well. Precise estimation of the relevant components of such models, however, may require data on other family relations in addition to intergenerational data.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. The Medical Birth Registry of Norway
  5. Gestational age
  6. Pre-eclampsia
  7. Birth defects
  8. Conclusion
  9. Acknowledgements
  10. References

Thanks to my colleagues, Dr Rolv Skjærven and Dr Allen Wilcox, for numerous discussions during the work with papers reviewed here.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. The Medical Birth Registry of Norway
  5. Gestational age
  6. Pre-eclampsia
  7. Birth defects
  8. Conclusion
  9. Acknowledgements
  10. References
  • 1
    Haig D. Genetic conflicts in human pregnancy. Quarterly Review of Biology 1993; 68:495532.
  • 2
    Barlow DP. Gametic imprinting in mammals. Science 1995; 270:16101613.
  • 3
    Taylor JE, Conard MW, Koetting O'Byrne K, Haddock CK, Poston WS. Saturation of tobacco smoking models and risk of alcohol and tobacco use among adolescents. Journal of Adolescent Health 2004; 35:190196.
  • 4
    Coan PM, Burton GJ, Ferguson-Smith AC. Imprinted genes in the placenta. Placenta 2005; 26 (Suppl. A):S10S20.
  • 5
    Pawitan Y, Reilly M, Nilsson E, Cnattingius S, Lichtenstein P. Estimation of genetic and environmental factors for binary traits using family data. Statistics in Medicine 2004; 23:449465.
  • 6
    Irgens LM. The Medical Birth Registry of Norway. Epidemiological research and surveillance throughout 30 years. Acta Obstetricia et Gynecologica Scandinavica 2000; 79:435439.
  • 7
    Lie RT, Wilcox AJ, Skjærven R. Maternal and paternal influences on length of pregnancy. Obstetrics and Gynecology 2006; 107:880886.
  • 8
    Skjærven R, Vatten LJ, Wilcox AJ, Ronning T, Irgens LM, Lie RT. Recurrence of preeclampsia across generations: exploring fetal and maternal genetic components in a population based cohort. British Medical Journal 2005; 331:877879.
  • 9
    Lie RT, Rasmussen S, Brunborg H, Gjessing HK, Lie-Nielsen E, Irgens LM. Fetal and maternal contributions to risk of preeclampsia: population based study. British Medical Journal 1998; 316:13431347.
  • 10
    Skjærven R, Wilcox AJ, Lie RT. A population-based study of survival and childbearing among female subjects with birth defects and the risk of recurrence in their children. New England Journal of Medicine 1999; 340:10571062.
  • 11
    Lie RT, Wilcox AJ, Skjærven R. Survival and reproduction among males with birth defects and the risk of recurrence in their children. JAMA 2001; 285:755760.
  • 12
    Cnattingius S, Reilly M, Pawitan Y, Lichtenstein P. Maternal and fetal genetic factors account for most of familial aggregation of preeclampsia: a population-based Swedish cohort study. American Journal of Medical Genetics A 2004; 130:365371.