The art and science of twins.

A joint meeting of the Section of Paediatrics and Child Health and the Forum on Maternity and the Newborn of the Royal Society of Medicine, Tuesday 20th March 2007.

Chair, morning session: Professor Lewis Spitz, President of the Section of Paediatrics & Child Health, RSM
Chair, afternoon session: Dr. Luke Zander, retired General Practitioner and teacher and founder of the Forum

Twin registries and their value.
Dr Catherine Derom (CD), Center for Human Genetics, Faculty of Medicine, Catholic University of Leuven.

Twins are generally recognized as being a valuable resource for research not only into multiple birth itself, but importantly into the causes of phenotypic variation in humans. Comparing monozygotic twins, who are genetically identical, with dizygotic twins who share half of their genes permits the experiment of nature versus nurture. This has led to the establishment of many twin registries worldwide which differ in the collection of data, in their sample sizes, and in their primary research interests. Conditions studied vary widely, and cover virtually the whole spectrum of health. Registers should be based on populations which may be local, regional or national, and unbiased in terms of disease, an objective likely to be frustrated by ascertainment from hospitals, which is more likely to include severely affected concordant twins. Volunteer based ascertainment can be very efficient when it advertises for twins, as opposed to diseased twins; by advertising for twins just because they are twins, large-scale population-based twin registries have been established in the USA, Australia and the UK. In the Scandinavian countries the unique system of population registration provides excellent opportunities to trace almost total twin populations, and then investigate these twins for a variety of phenotypes. Two entire issues of the journal Twin Research were devoted to twin registries and they give an excellent overview of those available throughout the world (Busjahn & Hur 2006).

The value of twin registries resides not only in their databases but in the possibility of maintaining contact with twins in order to identify phenotypes. Contact is made by means of newsletters, meetings, and through the internet; it is wise to establish rules to ensure funding, good planning, ethical approval for new studies, and to avoid the over-exploitation of twins, who are generally very interested in participating in research.

The identification of zygosity.

Different sex twins are dizygotic (DZ)
Monochorionic twins are monozygotic (MZ) - two thirds of MZ twins.
Twins of the same sex and dichorionic or with unknown chorionicity require genetic markers to establish zygosity; these include blood grouping and DNA fingerprinting, identical in MZ but different in DZ twins, when at least two differences in genetic markers are needed to avoid confusion with somatic mutations.

Accurate determination of zygosity and chorionicity are essential in all multiple births; it is of medical importance both pre- and postnatally. The parents and children ask for it, and it has been prescribed as their right. It enables awareness of possible inheritance of specific diseases (important should organ transplantation later be considered), malformations and twin-to-twin transfusion; each affected person must be able to identify himself whether in a social setting or at school and in the workplace, and a knowledge of zygosity is vital for this. In the scientific field comparisons of the zygosity of multiples with singletons contribute to the understanding of the roles of genetics and environment in development. A serious error to be avoided is the assumption that two chorionic sacs and separate placentas are proof of dizygosity; when similarity questionnaires are used there are 5% of false diagnoses of zygosity. These 5% are important since the lack of similarity in the appearances of twins may be due to some abnormality in one of the twins (such as cerebral palsy) and that abnormality may be the result of placentation and zygosity.

As a result of the modern methods of treatment of infertility the rate of twin birth in Belgium has almost doubled (East Flanders Prospective Twin Survey 1964 - 2005). The incidence of DZ twins in the iatrogenic group was 94%, compared with 53% where the conceptions were spontaneous. Perinatal mortality was much higher in MC twins than in DC twins; chorionicity has a much stronger effect here than zygosity, especially before birth. MZ/DC twins (15% of those conceived spontaneously) have the same risks as DZ twins. There is of course a significant increase in the numbers of low birthweight babies, whose illnesses and abnormalities are a burden on neonatal intensive care units and health services in general. There is a preponderance of girl babies among the very rare monoamniotic monochorionic twins; because of X-inactivation female tissues are a mosaic of two cell populations, and X-chromosome inactivation patterns can be very different in MZ/DC twin girls with significance for the co-incidence of X-related traits.

Twin registries are very valuable resources as tools for many forms of research, especially genetics; they rely on accurate diagnosis.


Busjahn A. and Hur Y-M. 2006. Twin registries: An ongoing success story. Twin Research and Human Genetics 9:6 705-705(1)

Derom CA, Vlietinck RF, Thiery EW, Leroy FO, Fryns JP, Derom RM. The East Flanders Prospective Twin Survey (EFPTS). Twin Res Hum Genet. 2006; 9(6):733-8.


CD: I could have wished that we had taken cord blood from the outset of this research.
Prof. Dr. Robert Derom, an obstetrician, trained the midwives with whom he worked to carry out rigorous examinations of the placentas, dissecting the membranes to establish chorionicity. Subsequently the midwives have trained each other.

The fascination of twins - – their contribution to genetic and social research.
Dr Lynn Cherkas (LC), Twin Research Unit, King's College London, St Thomas' Hospital Campus

The St. Thomas' Adult UK Twin Registry was established in 1993 by Professor Tim Spector, and the database now includes about 10,000 adult twins, with an age range of 16 to 85 years, mean 50 years. Unusually for a twin registry there is an approximately equal number of identical (MZ) and non-identical (DZ) twins. The registry was set up by a rheumatologist with an interest in the genetics of osteoporosis and osteoarthritis in females, with the result that at present most (80%) of our twins are female.

Our main objectives are to estimate heritabilities for common diseases and traits in adults and to discover the genes that influence these traits. We hope to provide a framework for epidemiological studies, to develop a statistical methodology, and to promote the wide use of the data. All 10,000 twins are sent self-completion questionnaires once or twice annually, inquiring into clinical and behavioural traits and lifestyles e.g. smoking and alcohol consumption. Four or five pairs of twins visit the unit at St. Thomas's Hospital daily when blood samples are collected and physiological measurements made; these include among others blood pressure and bone density. Several times a year a larger number of pairs attend twin days in London and around the UK, when blood is taken and fewer observations made.

In this way over 4000 phenotypes have been collected, covering a huge range of diseases, injuries, medical and surgical treatment and physiological performance features such as the perception of pitch, cognitive function and memory.

Using twins we can estimate the proportion of differences between individuals explained by genetic differences (the heritability of traits), separating genetic, family environment and unique environment components. Twins brought up together share the same family environment; if their experiences are not shared, for example by illness or separation, their environments are unique. In 1883 Sir Francis Galton's studies of twins had led him to the conclusion that for many traits nature prevails enormously over nurture.

Comparing the similarity of MZ (sharing of 100% of their genes) and DZ (sharing on average 50% of their genes) twins allows us to assess the contribution of genetic and environmental factors to clinical and behavioural traits. The Peas in the Pod questionnaire which we use asks about similarity during childhood and provides 95% accuracy. “Are your twins as like as peas in a pod, or do they look like sisters or brothers?” (Peeters et al. 1998; Martin & Martin 1975). In cases of doubt we have used 18 DNA markers and sometimes genome scans, providing almost 100% accuracy.

An example of evidence for a genetic component of a trait is cataracts : graphical representation of the similarity in measures for MZ and DZ twins significantly demonstrates the closer correlation in the former, indicating a genetic component.

Such studies have shown 90% heritability of freckles and short sight, and 50% heritability for migraine and the onset of the menopause. No evidence in our studies has been found for genetic components in irritable bowel syndrome or the appreciation of humour, although they are surprisingly present in social attitudes and infidelity.

To discover the genes which influence traits we use methods of linkage and association. Linkage is a way of assessing whether certain chromosomal regions are more likely to be shared by siblings who exhibit the same trait than would be due to chance; where that is the case we conduct more detailed searches of those regions for candidate genes. We use genetic association to test whether genotype frequencies are different between 2 groups (usually diseased subjects and healthy controls). Such studies are based on the principle that genotypes are measured "directly", i.e. by sequencing the actual genetic code.

A further aim is to provide a framework for epidemiological studies using the twin design. MZ twins share genotypes, but different phenotypes (traits) can occur. Where MZ twins are discordant for disease we look for differences in lifestyle; when they are discordant for lifestyle we look for differences in diseases (discordant twin design). In this way we have found differences in bone mineral density in twins discordant for alcohol consumption. We observed in our data that Raynaud's disease, migraine, high blood pressure, and cardiovascular disease can cluster within the same individual, leading to the hypothesis that these phenotypic associations represent an underlying vasospastic disorder. Our multivariate analyses demonstrated a common genetic factor underlying the association between these four traits. However it must be emphasized that genes are not deterministic, but provide the potential for a disease or trait, and environmental factors have an important effect on the expression of genes; for example when reviewed after ten years an individual with criminal genes living in a favourable environment will have a 1.5-fold risk of developing criminality. Another, brought up in unfavourable circumstances, will have a 10-fold risk, compared with the 3-fold risk of someone with normal genes brought up in an unfavourable environment.

Twin studies make several assumptions - one being that the shared family environment is equally similar for MZ and DZ twins. This has been widely debated in the literature, but studies have shown that parents are seen to react to the behaviour of twins rather than to their perceived zygosity. DZ twins may be dressed in identical clothing, and parents may try to individualise MZ twins by dressing them differently, so there is no reason to believe that MZ twins share a more similar environment than DZ twins.

Twin studies also assume that twins are representative of the general population, so that results are generalisable to the population at large. A large study comparing twins and singletons for a wide range of measures (Andrew et al. 2001) showed that the only difference was the lower birthweight of twins. However, this is not reflected in an increased incidence of cardiovascular disease in later life as is the case with singletons (Morley and Dwyer 2005). Interestingly, the incidence of suicide is lower in twins. Parents need to be reassured that the increased risk of primary ovarian failure in twins is very small (1% to 5%). There is no difference in the prevalence of disease between mono- and dichorionic adult twins, and mortality rates in twins and singleton populations are similar.

Twins represent a remarkable natural experiment, showing us that nearly all the functions of our minds and bodies are driven to some extent by the genes of our ancestors. Our uniqueness as individuals results from the variability of the 30,000 genes within all of us and in the different ways in which they react with our environment.

The special relationship of twins to their contribution to social research.

The knowledge of their zygosity is very important to twins; a pair of MZ twins were assured by their mother that they were not identical because there were 2 placentas at birth, and experienced discomfort about their identity. They were thrilled when they later learned that their zygosity had been wrongly assigned - MZ twins can have two placentas; “the discovery was the missing piece of the jigsaw”.

Recently we have conducted both qualitative and quantitative studies to investigate the relationship between twins, looking at the dynamics between them and exploring how identical and non-identical twins perceive the physical and emotional bonds between them. We found that many MZ twins think of themselves as individuals and are usually aware that their genetic identity is modified by their environment; however some identical twins find it harder to establish their own identity. Twins also feel that it is others who regard and treat them as one individual. About half of those twins responding to a questionnaire were aware of the dominance of one of them, and this is usually ascribed to the birth order with first born twins usually being larger and more dominant; for this reason some parents pretend that both were born simultaneously.

Twins do tend to feel special and pleased to have close and caring friends in their co-twins; a small proportion believe in a telepathic connection with their twin. However, not all twins have such a close relationship; and whilst some twins are almost inseparable, others are barely on speaking terms. Twins dislike being addressed as “twin” rather than by name, their individuality being ignored, and they resent being compared with their twin, particularly at school, and sharing the birthday and presents. One pair reported an attempt (successfully repelled) during wartime to fob them off with a single ration of food.

Having experienced genetic sameness, the views of identical twins towards human reproductive cloning were of particular interest to one of our collaborators, Dr. Barbara Prainsack, and a summary of the results of her study are presented here (Prainsack et al 2007). MZ twinning is central to the debate on human reproductive cloning, although identical twins are more alike than clones, sharing a simultaneous conception, pre- and postnatal environments and a common birth. A large sample of MZ and DZ twins were presented with an extensive questionnaire on this topic, and of particular interest were the responses to the statement “Human cloning should be allowed for medical purposes e.g. to save the life of an older sibling”. MZ twins and those without children were significantly more likely to be in favour of cloning in this situation, perhaps because MZ twins are aware that the being genetically identical does not make you the same individual . (Prainsack et al 2006)


Andrew T, Hart DJ, Snieder H, de Lange M, Spector TD, MacGregor AJ. Are twins and singletons comparable? A study of disease-related and lifestyle characteristics in adult women. Twin Research 2001 Dec;4(6):464-77

Prainsack, B., and Spector, T.D. (2006) Twins: A Cloning Experience. Social Science & Medicine 63:10, 2739-2752

Prainsack B, Cherkas LF, Spector TD. Attitudes towards human reproductive cloning, assisted reproduction and gene selection: a survey of 4600 British twins. Hum Reprod. 2007 Aug;22 (8):2302-8

Peeters H, Gestel SV, Vlietinck R, Derom C, Derom R. Validation of a telephone zygosity questionnaire in twins of known zygosity. Behav Genet. 1998; 28:3 159-63. Martin NG, Martin PG. The inheritance of scholastic abilities in a sample of twins. Ascertainment of the sample and diagnosis of zygosity. Annals of Human Genetics 1975; 39: 213-218


A delegate quoted Mark Twain: “I was a twin. One of us died; I have never been sure which one”. LC: When one of twins has died the surviving twin often has a sense of loss.

The importance of twin studies in research.
Caroline Edmonds (CE), Research Fellow, MRC Childhood Nutrition Research Centre, Institute of Child Health

In his “Inquiries into Human Faculty and its Development” of 1875 Galton addressed the issue of Nature versus Nurture; when we study whether MZ twins are more alike than DZ twins raised in the same family environment we are looking to confirm the inheritance of traits. Classical twin studies examine the contribution of genes and the environment to a given trait.  An alternative use for them is to view MZ twins as a natural experiment, the study of which can control for many confounding factors.  My example is the effect of low birth weight (indicating suboptimal intrauterine growth) on cognitive function in childhood (Richards et al. 2001 & 2002).

A confounding factor here is early birth; these babies average lower IQ scores (Spirillo et al. 1997; Sung et al. 1993), raising the question “Is lower IQ a consequence of growth retardation or prematurity?” Also lower socio-economic status is associated with both low birthweight (Teberg et al. 1996) and low IQ scores (Drillien 1970; Hawdon et al., Robertson et al. 1990); again, is it growth? Of course low birth weight might be associated with a heritable family trait, and the heritability of IQ is approximately 50%, another possible confounding variable. Finally males are more affected than females by low birth weight (Stevenson et al. 2000), possibly affecting IQ scores.

Research can attempt to control for confounding factors in various ways. Participants should have the same sex and family backgrounds and gestational ages (group matching). Covariants can be introduced into statistical analyses. Such measures may be insufficient, and their shortcomings avoided using the MZ twin study design, since these twins have the same gestational age, are exposed to the same family influences, are of the same sex, and have identical genetic potentials for growth and cognition.

Studying MZ twins discordant for birthweight we used the birthweight and IQ of the heavier twin as benchmarks for the birthweight and IQ that the lighter twin might have achieved. 71 MZ twin pairs were studied; 41 of them were male, mean gestation was 36.5 weeks and mean age at test eleven years and six months. We excluded any twin pairs in which there was a neuropsychiatric diagnosis such as autism; most were born at term, a few before 34 weeks. Our assessment used the Wechsler Intelligence Scales for Children (WISC-III) , in which IQ scores provide a measure of general cognitive ability. Some of the tasks assess verbal IQ, some performance IQ.

We calculated difference scores for birthweight and IQ within twin pairs to answer the question “Does failure to reach birthweight potential impact on ability to reach potential IQ”. We found that right across the birthweight spectrum a 1kg reduction in birthweight resulted in a 12.8 point decrease in verbal IQ (almost one standard deviation), a result unaffected by mean actual birthweight, gestation, sex or birth order. The weight differences did not affect performance IQ, as has been observed in other research. Omitting the seven most preterm pairs did not alter the findings.


Richards, M., Hardy, R., Kuh, D., & Wadsworth, M.E.J. (2001). Birth weight and cognitive function in the British 1946 birth cohort: longitudinal population based study. British Medical Journal, 322, 199-203.

Richards, M., Hardy, R., Kuh, D., & Wadsworth, M.E.J. (2002). Birthweight, postnatal growth and cognitive development in a national UK birth cohort. International Journal of Epidemiology, 31, 342-348.

Teberg, A.J., Settlage, R., Hodgman, J.E., King, Y., & Aguilar, T. (1996). Maternal factors associated with delivery of infants with birthweight less than 2000 grams in a low socioeconomic population. Journal of Perinatology, 9, 291-295.

Stevenson, D.K., Verter, J., Fanaroff, A.A., Oh, Q., Ehrenkranz, R.A., Shankaran, S., Donovan, E.F., Wright, L.L., Lemons, J.A., Tyson, J.E., Korones, S.B., Bauer, C.R., Stoll, B.J., & Papile, L-A. (2000). Sex differences in outcomes of very low birthweight infants: the newborn male disadvantage. Archives of Diseases in Childhood; Fetal & Neonatal Edition, 83, F182-F185.

Robertson CMT, Etches PC & Kyle JM (1990): Eight-year school performance and growth of preterm, small for gestational age infants: A comparative study with subjects matched for birthweight or for gestational age. Journal of Pediatrics, 116, 19-26.

Spirillo, A., Capuzzo, E., Piazzi, G., Baltaro, F., Stronati, M., & Ornetto, A. (1997). Significance of low birthweight for gestational age among very preterm infants. British Journal of Obstetrics and Gynaecology, 104, 668-73.

Sung, I.K., Vohr, B., & Oh, W. (1993). Growth and neurodevelopmental outcome of very low birth weight infants with intrauterine growth retardation: comparison with control subjects matched by birth weight and gestational age. Journal of Pediatrics, 123, 618-24.

Drillien, C.M. (1970). The small for date infant: Etiology and prognosis. Pediatric Clinics of North America 17, 9-24.

Hawdon, J.M., Hey, E., Kolvin, I., & Fundudis, T. (1990). Born too small - is outcome still affected? Developmental Medicine and Child Neurololgy, 32, 943-953.


CE: We will need to address twin-to-twin transfusion in future research.