Changes in twinning rate: Italy 1950–1996

P. Astolfi1, L. Ulizzi2 and L.A. Zonta1,3

1 Department of Genetics and Microbiology ‘A.Buzzati-Traverso’, University of Pavia and 2 Department of Genetics and Molecular Biology ‘C.Darwin’, University La Sapienza, Rome, Italy 3 To whom correspondence should be addressed at: zonta{at}ipvgen.unipv.it


    Abstract
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
BACKGROUND: In the wide-ranging debate about the potential monitors of population fertility, twinning rate has been considered a candidate. In the developed countries, a decline in spontaneous twinning rate began around 1950 and continued until the late 1970s. The decrease in mean maternal age at delivery and the number of children per family have been considered as possible reasons for the decrease. The rise in twinning rates, which began in the 1980s, is probably due to the spread of assisted reproduction, at least for older mothers. METHODS: The temporal trend in twinning rate and the relationship with maternal age at delivery and birth order in the Italian population were analysed. RESULTS: The recent increase in twinning rate can be mainly attributed to the recourse to assisted reproduction techniques, as demonstrated by the twinning frequency among older primiparae. The twinning rate recently observed in young mothers approximates to the value of the 1950s, which is assumed to represent a ‘natural’ fertility condition. CONCLUSIONS: On the assumption that twinning is an indicator of fertility, and hence of population fitness, the present data seem to suggest that natural fertility has been slightly increasing in the Italian population.

Key words: birth order/maternal age/population fertility/secular trend/twinning rate


    Introduction
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Natural selection is thought to optimize the reproductive success of a given species by maximizing the number of surviving offspring produced during lifetime. On this assumption, a litter size is predicted to be optimal when it matches the highest number of survivors (Haukioja et al., 1989Go). In man, as well as in most species of Old World monkeys and anthropoid apes, single births may be considered the rule. It is interesting to note that the practice of monitoring pregnancies with ultrasonographic techniques has shown that about 40% of spontaneous dizygotic twin pregnancies end in single births (Landy and Keith, 1998Go). This ‘vanishing twin’ phenomenon, which has been viewed as part of the general fetal wastage in human reproduction (Boklage, 1995,Go quoted in Landy and Keith, 1998Go), possibly suggests that while multiple conceptions are not a rare event, multiple pregnancies are strongly selected against, especially during the early stages of embryonic development.

Although a relatively rare event, twin births have always been present in all human populations, with frequencies varying from 0.6 to 4.5% in different ethnic groups. Inter-population differences in twinning rate can be mainly attributed to dizygotic (DZ) twins, since the frequency of monozygotic (MZ) twins shows little variability between different populations (Bulmer, 1970Go; Vogel and Motulsky, 1986Go). Moreover, an investigation carried out on relatively isolated Finnish groups supported the hypothesis that differences in twinning rates are maintained by natural selection (Lummaa et al., 1998Go).

That genetic components can be involved in DZ twinning has been suggested in several familial studies (Bulmer, 1970Go; Lewis et al., 1996Go; Meulemans et al., 1996Go), and more recently some evidence of the association of DZ twinning to a chromosome 3 region has been found (Busjahn et al., 2000Go). Some authors have suggested that MZ twinning might also be under genetic influence (Parisi et al., 1983Go; Bortolus et al., 1999Go).

Besides showing inter-population variability, in most Western countries twinning rates have varied extensively over the past 50 years: the temporal trend, which began to decrease in the early 1950s, displayed a reversal of tendency in the late 1970s, thereby reviving the interest of epidemiologists and geneticists in twinning and its possible causes (James, 1972Go; Parisi and Caperna, 1981Go; Allen and Parisi, 1990Go; Parazzini et al., 1991Go; Imaizumi, 1997Go, 1998; Wood, 1997Go).

As several studies have shown a positive association between twinning incidence, in particular DZ, and maternal age and parity, the decrease in mean maternal age at delivery and the reduction in the number of children have been considered to be at least partly responsible for the downward trend (Bulmer, 1970Go; Eriksson and Fellman, 1973Go; James, 1975Go, 1982; Elwood, 1978Go; Bonnelykke, 1990Go; Parazzini et al., 1991Go; Campana and Roubiceck, 1996Go; Bortolus et al., 1999Go). The recent drastic increase in twinning and in multiple births has been mainly attributed to ovulation-inducing treatments and assisted reproduction techniques, which have been adopted to counter subfertility problems (Parazzini et al., 1994Go; Jewell and Yip, 1995Go; Imaizumi, 1997Go, 1998; Wood, 1997Go).

Some authors have suggested that DZ twinning incidence, in so far as it is an indicator of multiple conceptions, is related to the population fertility. However, to consider natural twinning and the ratio of dizygotic/monozygotic twins as reliable monitors of the population fertility, or even of the population reproductive health, is controversial (James, 1995Go, 1997, 1998; Tong and Short, 1998Go; Tong, 2000Go).

In the present study, the temporal trend of twinning rates in the Italian population was updated to the latest available year, and the pattern within the European context was examined. The changes in some biodemographic factors affecting population fertility were then analysed in order to investigate their possible role in the changes in twinning rate. In Italy, as well as in most developed countries, the cultural modifications in women’s reproductive strategies over the past 50 years have drastically influenced the demographic scenario, first through shrinkage in offspring number, and second by childbearing postponement (De Sandre et al., 1997Go; Ulizzi et al., 1997Go; Astolfi et al., 2000Go). Therefore, the evolution of the relationship of twinning with maternal age and birth order in three critical periods was studied, corresponding to: (i) the beginning of the decreasing pattern; (ii) the minimum reached; and (iii) the recent situation.


    Materials and methods
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
The 1950–1996 Italian data on twinning and other multiple pregnancies were drawn from yearly Italian National Statistical Institute (ISTAT) publications; rates were computed on total births, regardless of the baby’s viability. European twinning rates were drawn from a previously published report (Imaizumi, 1997Go), except for UK rates, which were obtained elsewhere (Wood, 1997Go).

The 1951–52 and 1979–81 twin data by mother’s age and birth order were drawn from ISTAT publications, and the latest available data were acquired directly from ISTAT at the authors’ request. Data for two or three consecutive years were pooled so as to obtain comparable sample sizes and to minimize yearly fluctuations. Mother’s age was recoded into five classes (<25, 25–29, 30–34, 35–39 and >=40 years), and births of third and higher orders were pooled.

Weinberg’s rule was applied to estimate the proportion of DZ twins from the proportion of opposite-sex twins and the newborn sex ratio, as is usually carried out whenever sex-pairing is the only indication of zygosity (Cavalli-Sforza and Bodmer, 1971Go).

Second-order polynomials were fitted to the European yearly twinning rates to depict better the 1975–95 patterns; the rates of increase over the last 10 years were estimated by linear regression analysis (SPSS 10.7 statistical package).

The t-test was applied on arcsin transforms to verify the differences between two twin rates (Sokal and Rohlf, 1981Go).


    Results
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Figure 1 shows the temporal trends during the second half of the 20th century in Italy of: (a) the overall twinning rate; (b) the proportion of DZ twinning; and (c) the multiple birth rate. The similarity of the curves in Figure 1a and b indicates that the overall pattern can be attributed mainly to double conceptions. Both patterns of twinning and multiple births (Figure 1a and c) show a decrease of about 30% over the first 30 years. The increase, observed from the early 1980s up to 1996, is less striking with regard to twinning than multiple births: twinning increased by 25%, while multiple births more than tripled their frequency.



View larger version (20K):
[in this window]
[in a new window]
 
Figure 1. Italian temporal trends of a) twin births per 1000 pregnancies, b) estimated DZ proportion among twin births and c) multiple (>=3) births per 1000 pregnancies.

 
As for twinning, some increase seems to have occurred also in MZ twins (Derom et al., 1987Go), with the estimated frequencies corresponding to 4.05 and 4.37%Go in 1979Go–81 and 1993–94 respectively.

Second-order polynomials were fitted to the yearly data to depict better the temporal trends in some European countries (Figure 2): the variation explained by the fitting varies from a minimum in Austria (89.3%) to a maximum in the UK (98.8%). The rate minimum value was estimated to have occurred in 1975 in Sweden, and 5 years later in Italy and Austria. The slope of the linear fitting over the past 10 years shows that the rate of increase declines monotonically from Northern (0.458 in Norway) to Southern European countries (0.175 in Italy).



View larger version (25K):
[in this window]
[in a new window]
 
Figure 2. Quadratic interpolation fitted to the yearly twin birth rates per 1000 pregnancies of European countries.

 
To obtain a better understanding of the role that the cultural changes in family planning might have played in twinning, the relationship with mother’s age and birth order was analysed, focusing on three specific periods. The early 1950s was chosen as the start of the decreasing trend, the early 1980s for the minimum, and the mid-1990s for the latest available data. Twin births in terms of birth order and of average mother’s age at delivery are described in Table I: maternal age followed a U-shaped pattern, and among the primiparae the 1994 value even exceeded that of the 1950s. Contemporarily, the percentage of first-born twins increased monotonically and almost doubled.


View this table:
[in this window]
[in a new window]
 
Table I. Twin births described in terms of maternal age and birth order. All information refers to the pairs for which mother’s age and birth order were available
 
The trends in twinning rates according to mother’s age for first, second, and higher birth orders are shown separately for the three periods in Figure 3. The graph for 1951–52 (Figure 3a) shows the well-known relationship of spontaneous twinning rate with mother’s age and birth order (Bulmer, 1970Go). The curves for first and second birth orders were wholly super-imposable, while that for orders >=3 was higher but almost parallel. In each curve, the value observed for mothers aged 35–39 years was about twice that found in the youngest group. In the graph for 1979–81 (Figure 3b), the effect of birth order seemed to disappear, and the relationship with mother’s age was reduced in all birth orders. Moreover, the minimum values observed in the younger mothers were lower than those in the preceding period. A complex picture emerged for 1994–96 (Figure 3c); the most striking feature was displayed by firstborns, with the curve being almost super-imposable on that of 1951–52 for mothers aged <35 years, but rising sharply thereafter. With regard to second-born twins, the overall pattern was similar to that seen in 1979–81. In newborns of third and higher orders, the pattern was almost constant in mothers aged <35 years, and decreased thereafter.



View larger version (20K):
[in this window]
[in a new window]
 
Figure 3. Twin births per 1000 pregnancies by mother’s age and birth order in the Italian population during the second half of the last century. The 1951–52 period is representative of ‘natural’ twinning condition; the 1979–81 period corresponds to the minimum twinning rate in the temporal trend; the 1994–96 period is the latest available data.

 
It seemed interesting then to compare the trends of twinning rates in the key age groups; that is, in young (<25 years) and old (>=40 years) primiparae (Table II). The increase in rate was highly significant (P < 0.001) not only in the old but also in the young group from 1979–81 to 1994–96; however, in relation to the value of the 1950s, the rate in older mothers during 1994–96 increased by >45%, while that in younger mothers was not significant.


View this table:
[in this window]
[in a new window]
 
Table II. Twinning rates, s.e. and twin births (n) in the three periods for young (<25 years) and old (>=40-year) primiparae women
 

    Discussion
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
It is well known that, besides the genetic components, environmental factors play a significant role in twinning and that intra-population variability in twinning rates is associated with several biological factors, such as maternal gonadotrophin levels and sperm count (Bulmer, 1970Go; James, 1975Go, 1997; Campana and Roubicek, 1996Go). Both genetic and non-genetic components are difficult to evaluate, and the analysis of the joint evolution of twinning rate and some biodemographic factors may represent a useful approach towards understanding their relative roles. Some authors have also suggested that twinning rate may provide information on the population fertility (James, 1997Go; Tong, 2000Go).

The variation in twinning and multiple birth incidence that occurred in Italy during the second half of the 20th century was consistent with the falling and rising patterns observed in other developed countries (Figures 1 and 2). During the same period, the attitude of Italian women toward family life underwent strong changes that affected the overall population fertility. Higher levels of education and a desire for economic and social success, as well as difficulties in finding a permanent job, played a major role in modifying the reproductive strategies, as is evidenced by delayed parenthood and the reduction in family size (De Sandre et al., 1997Go; Astolfi et al., 2000Go). In the last generation, mean maternal age at delivery increased from about 27 to over 29.5 years, while the number of children per woman of reproductive age fell to 1.20 in 1996, one of the lowest values in Europe.

In the present study, three critical periods in the temporal pattern of twinning rate were chosen, corresponding to the beginning of the decline, the minimum reached, and the recent situation, in order to study its association with the population fertility through the evolution of the relation of twinning with mother’s age and parity (Figure 3).

The pattern of twinning rate by mother’s age and parity observed in the early 1950s is well known and is assumed to be typical of spontaneous twinning in ‘natural’ fertility conditions (Figure 3a). A strong decrease in population fertility during the second period is therefore suggested by the abatement of birth order effects, by the low twinning rates in young mothers (aged <25 years), and by the reduction of maternal age influence (Figure 3b). The use of subfertility treatments in recent years is most likely the main cause of the increase in the population twinning rate, and the hypothesis is supported by the strikingly high frequency found among older (>=35 years) (Figure 3c) and even >=40-year primiparae (Table II). The same explanation is also provided by the sharp increase in triple and higher deliveries in Italy, which is generally found in the countries where the use of assisted reproduction has been spreading (Martin and Park, 1999Go). Moreover, in spite of the recourse to medical treatments, in Italy the overall rate in 1996 had not reached the twinning value of the 1950s, supporting the notion that spontaneous twinning had not yet recovered from the dip of the late 1970s (Figure 1). It is interesting to note that the increase in twinning rate started earlier and proceeded with higher velocity in northern Europe (Figure 2).

It is significant however that in the young (women aged <25 years) Italian primiparae, who are extremely unlikely to recourse to fertility treatments, the twinning rate has recently reached the ‘natural’ value observed at the start of the 1950s (Table II). On the assumption that only twinning rates in young mothers should be considered as a monitor of the population’s natural fertility and even reproductive health (James, 1998Go), the present results would suggest that some increase in natural fertility has been taking place in the Italian population.


    Acknowledgements
 
These studies were financed by Ministero dell’Università e della Ricerca Scientifica and by the University of Pavia and Rome.


    References
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Allen, G. and Parisi, P. (1990) Trends in monozygotic and dizygotic twinning rates by maternal age and parity. Further analysis of Italian data, 1949–1985, and rediscussion of US data, 1964–1985. Acta Genet. Med. Gemellol., 39, 317–328.[Medline]

Astolfi, P., Ulizzi, L. and Zonta, L.A. (2000) Natural selection and reproductive behavior. Hum. Biol., 72, 349–357.[ISI][Medline]

Boklage, C.E. (1995) The frequency and survival probability of natural twin conceptions. In Keith, L.G., Papiernik, E., Keith, D.M. et al. (eds), Multiple Pregnancy: Epidemiology, Gestation and Perinatal Outcome. Parthenon, New York, pp. 41–50.

Bonnelykke, B. (1990) Maternal age and parity as predictors of human twinning. Acta Genet. Med. Gemellol., 39, 329–334.[Medline]

Bortolus, R., Parazzini, F., Chatenoud, L., Benzi, G., Bianchi, M.M. and Marini, A. (1999) The epidemiology of multiple births. Hum. Reprod. Update, 5, 179–187.[Abstract/Free Full Text]

Bulmer, M.G. (1970) The Biology of Twinning in Man. Clarendon Press, Oxford.

Busjahn, A., Knoblauch, H., Faulhaber, H.D., Aydin, A., Uhlmann, R., Tuomilehto, J., Kaprio, J., Jedrusik, P., Januszewicz, H. and Strelau, J. (2000) A region to chromosome 3 is linked to dizygotic twinning. Nature Genet., 26, 398–399.[CrossRef][ISI][Medline]

Campana, M.A. and Roubiceck, M.M. (1996) Maternal and neonatal variables in twins: an epidemiological approach. Acta Genet. Med. Gemellol., 45, 461–469[Medline]

Cavalli-Sforza, L.L. and Bodmer, W.F. (1971) The Genetics of Human Populations. Freeman and Co., San Francisco.

Derom, C., Vlietinck, R., Derom, R., Van den Berghe, H. and Thiery, M. (1987) Increased monozygotic twinning rate after ovulation induction. Lancet, 1, 1236–1238.[CrossRef][ISI][Medline]

De Sandre, P., Ongaro, F., Rettaroli, R. and Salvini, S. (1997) Matrimonio e figli: fra rinvio e rinuncia. Ed Mulino Studi e Ricerche, Roma.

Elwood, J.M. (1978) Maternal and environmental factors affecting twin births in Canadian cities. Br. J. Obstet. Gynecol., 85, 351–358.[ISI][Medline]

Eriksson, A.W. and Fellman, J. (1973) Differences in the twinning trends between Finns and Swedes. Am. J. Hum. Genet., 25, 141–151.[ISI][Medline]

Haukioja, E., Lemmetyinen, R. and Pikkola, M. (1989) Why are twins so rare in Homo sapiens? Am. Naturalist, 133, 572–577.[CrossRef][ISI]

Imaizumi, Y. (1997) Trends of twinning rates in ten countries, 1972–1996. Acta Genet. Med. Gemellol., 46, 209–218.[Medline]

Imaizumi, Y. (1998) A comparative study of twinning and triplet rates in 17 countries, 1972–1996. Acta Genet. Med. Gemellol., 47, 101–114.[Medline]

James, W.H. (1972) Secular changes in dizygotic twinning rates. J. Biosoc. Sci., 4, 427–434.[ISI][Medline]

James, W.H. (1975) Sex ratio in twin births. Ann. Hum. Biol., 4, 365–378.

James, W.H. (1982) Second survey of secular trends in twinning rates. J. Biosoc. Sci., 18, 497–504.

James, W.H. (1995) Are "natural" twinning rates continuing to decline? Hum. Reprod., 10, 3042–3044.[Abstract]

James, W.H. (1997) Secular trends in monitors of reproductive hazard. Hum. Reprod., 12, 417–421.[CrossRef][ISI][Medline]

James, W.H. (1998) The use of twinning rates as a reproductive health monitor in England and Wales 1988–1996. Hum. Reprod., 13, 3577–3578.[ISI][Medline]

Jewell, S.E. and Yip, R. (1995) Increasing trends in plural births in the United States. Obstet. Gynecol., 85, 229–232.[Abstract/Free Full Text]

Landy, H.J. and Keith, J.G. (1998) The vanishing twin: a review. Hum. Reprod. Update, 4, 177–183.[Abstract/Free Full Text]

Lewis, C.M., Healy, S.C. and Martin, N.G. (1996) Genetic contribution to DZ twinning. Am. J. Med. Genet., 61, 237–246.[CrossRef][ISI][Medline]

Lummaa, V., Haukioja, E., Lemmetynen, R. and Pikkola, M. (1998) Natural selection on human twinning. Nature, 394, 533–534.[CrossRef][ISI][Medline]

Martin, J.A. and Park, M.M. (1999) Trends in twin and triplet births: 1980–97. Natl Vital. Stat. Rep., 47, 1–16.[Medline]

Meulemans, W.J., Lewis, C.M., Boosma, D.I., Derom, C.A., Van den Berghe, H., Orlebeke, J.F., Vlietinck, R.F. and Derom, R.M. (1996) Genetic modeling of dizygotic twinning in pedigrees of spontaneous dizygotic twins. Am. J. Med. Genet., 61, 258–263.[CrossRef][ISI][Medline]

Parazzini, F., Tozzi, L., Mezzanotte, G., Bocciolone, L., La Vecchia, C., Fedele, L. and Benzi, G. (1991) Trends in multiple births in Italy: 1955–1983. Br. J. Obstet. Gynaecol., 98, 535–539.[ISI][Medline]

Parazzini, F., Restelli, S., Moroni, S. and Crosignani, P.G. (1994) Trend in legitimate and illegitimate multiple births. Hum. Reprod., 9, 1784–1785.[ISI][Medline]

Parisi, P. and Caperna, G. (1981) The changing incidence of twinning: one century of Italian statistics. In Twin Research 3: Twin Biology and Multiple Pregnancy. Alan R.Liss, Inc., New York, pp. 35–48.

Parisi, P., Gatti, M., Prinzi, G. and Caperna, G. (1983) Familial incidence of twinning. Nature, 304, 626–628.[ISI][Medline]

Sokal, R.R. and Rohlf, F.J. (1981) Biometry. Freeman and Co., San Francisco.

Tong, S. (2000) Dizygotic to monozygotic twinning ratio at The Royal Women’s Hospital, Melbourne 1947–1997, compared with Australian national twinning incidence. Twin Res., 3, 12–16.[CrossRef][Medline]

Tong, S. and Short, R.V. (1998) Dizygotic twinning as a measure of human fertility. Hum. Reprod., 13, 95–98.[Abstract]

Ulizzi, L., Astolfi, P. and Zonta, L.A. (1997) Natural selection in industrialized countries: a study of three generations of Italian newborns. Ann. Hum. Genet., 62, 47–53.[ISI]

Vogel, F. and Motulsky, A.G. (1986) Human Genetics, Problems and Approaches. Springer, Berlin.

Wood, R. (1997) Trends in multiple births, 1938–1995. Population Trends, 87, 29–35.[Medline]

Submitted on August 9, 2002; accepted on September 25, 2002