The current ‘dramatically’ high paternal ages at childbirth are not unprecedented

Abstract

There is strong individual-level evidence that late fatherhood is related to a wide range of health disorders and conditions in offspring. Over the last decades, mean paternal ages at childbirth have risen drastically. This has alarmed researchers from a wide range of fields. However, existing studies have an important shortcoming in that they lack a long-term perspective. This article is a step change in providing such a long-term perspective. We unveil that in many countries the current mean paternal ages at childbirth and proportions of fathers of advanced age at childbirth are not unprecedented. Taking the detected U-shaped trend pattern into account, we discuss individual- and population-level implications of the recent increases in paternal ages at childbirth and highlight important knowledge gaps. At the individual level, some of the biological mechanisms that are responsible for the paternal age-related health risk might, at least to some degree, be counterbalanced by various social factors. Further, how these individual-level effects are linked to population health and human cognitive development might be influenced by various factors, including technical advances and regulations in prenatal diagnostics.

Introduction

Research has cumulated substantial individual-level evidence that health outcomes of children tend to decline with increasing paternal age at childbirth. Medical studies have suggested that the offspring of older fathers are more prone to a wide range of disorders and conditions (Yip et al., 2006; Oldereid et al., 2018; Yatsenko and Turek, 2018; Du Fossé et al., 2020). Using historical as well as contemporary demographic data, Arslan et al. (2017) found that children of older fathers have higher mortality and lower reproductive success. A study by Hayward et al. (2015) suggested that this paternal age effect also affects subsequent generations. Demographic data for the Western hemisphere indicate that the mean paternal age at childbirth (hereafter MPAC) has increased significantly in recent decades (e.g. Human Fertility Collection for 19 countries). Moreover, there is evidence that the proportion of fathers who are of advanced paternal age at childbirth (>45 years, from here: APAC) has risen considerably. Based on these recent MPAC and APAC trends, scholars from various fields have raised concerns that they will have negative implications for population health and human cognitive development (Bray et al., 2006; Saha et al., 2009; Gajos and Beaver, 2017; Khandwala et al., 2017; Nybo Andersen and Urhoj, 2017). These concerns are further fueled by demographic research showing that also the maternal age at childbirth has increased significantly in recent decades (e.g. Human Fertility Database).

However, an important shortcoming of existing studies on MPAC developments is that they are usually restricted to a few high-income countries and that they fail to provide a long-term perspective (Bray et al., 2006; Nybo Andersen and Urhoj, 2017). This makes it difficult to assess whether the current levels are indeed exceptionally high. Our study is motivated by research on long-term trends in maternal ages at childbirth, which shows that the low maternal mean ages at childbirth recorded in the mid-20th century were rather unusual and that in earlier periods, maternal ages at childbirth were substantially higher. Applying such a long-term perspective to fathers is much more difficult compared to mothers because data on paternal ages at childbirth are much more sparse. This is the case for recent periods, and especially for periods prior to 1960. This study constitutes a step change in identifying and harmonizing such data based on state-of-the-art demographic methodological standards and integrating them into a single database. To our knowledge, this new LPAC database (long-term trends in paternal age at childbirth) is by far the biggest global database that has ever been created to document MPAC trends in a comparative perspective, with some data series covering more than a century. It currently contains data from 4172 country-year observations for 140 countries. We complemented this database with an R-Shiny-App for visualization, including data tables and supplement information, which is freely accessible online (https://lpac-database.shinyapps.io/Data/)

Recent ‘dramatic’ increases in a historical perspective

Our results demonstrate that the high paternal ages currently being reported are not unprecedented (see Fig. 1). In all study populations for which long-time series are available, MPAC and APAC exhibit virtually the same U-shaped pattern. This trend in MPAC came about through major changes to family structures and fertility patterns, which we can discuss here only briefly (for more details, see Hajnal, 1965; Ruggles, 2015). We focus in this discussion on Europe and North America, where long-term social trends are well documented. However, our global comparison shows that the U-shaped pattern is also observed in other parts of the world (e.g. in Chile, Japan, and Australia). Around 120 years ago, the high MPAC in large parts of western Europe was linked to (perceived and real) limits to population growth (Smith, 1980) in combination with religious norms that discouraged limits to fertility (Lesthaeghe, 1980). These factors led to the emergence of traditions of delaying marriage until higher reproductive ages (Hajnal, 1965; Towner et al., 2016) to limit the period in which couples could reproduce.

Starting from these high MPAC levels and apart from some war-related short-term deviations, MPAC decreased rather steadily until the 1970s. Important factors that contributed to this trend were advances in agricultural production and industrialization processes, which freed populations from former limits to growth (Boserup, 1965; Lam, 2011). Thus, traditions that constrained population growth by encouraging delayed marriage became less relevant (Lesthaeghe, 1980). During this period, wages increased for men, while there were relatively few attractive career opportunities for women. In many countries, women also faced legal barriers to labor market participation (e.g. Goldin, 2006). These circumstances contributed to a socioeconomic setting in which the male breadwinner model was very dominant (Creighton, 1999). Men’s wages were often sufficiently high to support a family even at early adult ages, while many women tended to focus on childrearing tasks. It was under these conditions that the MPAC decreased to very low levels. This decrease was most pronounced in the 1950s and 1960s, which is also referred to as the Golden Age of Marriage (Owram, 1997; Peuckert, 2007). Indeed, in many countries, the average age at first childbirth for women was under age 25 in the 1960s and 1970s (Human Fertility Database).

After 1975, the direction of this trend reversed in all observed countries, and the MPAC increased sharply. These developments were linked to an interrelated mix of economic, social, cultural, and possibly also biological factors. The rise of the service sector and the transition to a knowledge economy created many new career opportunities for women and for well-educated women in particular. This was paralleled by shifts in women’s roles in societies that also fostered a higher labor market participation of women (Goldscheider et al., 2015). As a result, the male breadwinner model was gradually superseded by a dual-earner model in which both spouses were employed outside the family household (Blossfeld and Drobnic, 2001). Related to this trend, the share of women who received higher education increased substantially in many high-income countries (Van Bavel, 2012; Van Bavel et al., 2018; Grow and Van Bavel, 2020). This, in turn, influenced the transition to parenthood, as many individuals/couples postponed having children until they had finished their education and established themselves in the labor market (Ní Bhrolcháin and Beaujouan, 2012). The trend toward postponement was further driven by increasing economic and financial uncertainty (Vignoli et al., 2020). Especially in the post-war period, many western countries had low unemployment rates and rather high-income levels. From the late 1960s onward, globalization processes and the increase in women’s labor market participation resulted in more competition and more uncertainty about future employment trajectories. These trends also contributed to the postponement of fertility plans (Billari et al., 2006; Hofmann and Hohmeyer, 2013; Vignoli et al., 2020; Comolli et al., 2021). The postponement process has biological implications, as fecundity declines with age. In addition, it has been argued that there has been an overall decline in female and male fecundity (Ahmed et al., 2020; Jørgensen et al., 2021; Aitken, 2022).

The distinct increases in MPAC since the late 1970s have alarmed researchers from a wide range of disciplines, such as epidemiology, medical research, and psychology (Malaspina et al., 2005; Bray et al., 2006; Saha et al., 2009; Hamilton et al., 2015; Gajos and Beaver, 2017; Nybo Andersen and Urhoj, 2017). Their concerns appear reasonable given the strong evidence that fatherhood at advanced ages may be associated with poorer health and cognitive performance outcomes for the offspring, with implications for human development. However, based on the long-term perspective provided by this study, the recent increases in MPAC and APAC appear to be less dramatic, as the current high MPAC levels are not unprecedented. Indeed, in the western countries for which we have historic data, we find that the MPAC in the early 20th century was similar to what it is today. Furthermore, our results indicate that the proportions of APAC do not differ substantially between historical and contemporary populations. That we find U-shaped patterns also outside Europe and North America in countries such as Chile or Japan underlines that the phenomenon is not limited to the western world. Finally, a comparison of MPAC trends in western populations with those of many contemporary populations in sub-Saharan Africa north of the equator, where the MPAC levels often exceed current western levels by many years, further puts the increase in MPAC in a different light (Schoumaker, 2019).

Putting the U-shaped trend in perspective: implications and knowledge gaps

Given the evidence presented above, should we be less concerned about the contemporary upward trends in MPAC and APAC? The answer to this question is complex and might depend on whether the question relates to the individual or population level. Since biological and social factors can interact in multiple ways, the observation that today’s MPAC and APAC levels are comparable to those at the beginning of the 20th century is not automatically a reason to be less concerned. As mentioned above, medical and epidemiological research has clearly shown that the offspring of older fathers are more prone to a wide range of disorders and conditions (Yatsenko and Turek, 2018). To a certain degree, genetic mechanisms such as the paternal age-related accumulation of de novo mutations in sperm cells appear to be responsible (Kong et al., 2012). However, it is neither clear to what extent changes in MPAC affect health and human cognitive development at the population level nor to what extent biological disadvantages related to paternal age may be counterbalanced (or enhanced) by factors on the individual level. One reason for this gap is limited exchange between researchers or disciplines focused on individual-level mechanisms and those investigating population-level phenomena and implications.

For the individual level, it should be noted that standards of living and medical care have generally improved over time. Historical contexts are thus not fully transferable to contemporary contexts. Overall, we still have a limited understanding of how biological disadvantages related to paternal age may be counterbalanced by social resource accumulation. In many studies on the effects of parental age, the role of social resources in offspring health is ignored. However, the few studies that have examined these aspects have identified strong effects. For instance, in a study on Finnish cohorts born 1936–1950, the negative effects of advanced paternal age are explained by early paternal loss, which is associated with a significant loss of social resources (Myrskylä et al., 2014). Children may benefit biologically from having a younger father, because of, for example the genetic quality of his sperm is higher (Kong et al., 2012). However, children of older fathers might benefit from their father’s experience, including his better financial and social circumstances. The benefits of social resource accumulation may vary over time and place. They also depend on family size, as children’s development in larger families is more likely negatively affected by resource dilution and higher risks of cross infections. Notably, family size has declined in many countries in recent decades (Downey, 2001).

Further, it is important to note that the epidemiological transition (Omran, 1998) has not only increased life expectancy but has also significantly affected the speed of aging and of biological processes in the human body. For instance, the social consequences of having a 50-year-old father may have changed substantially over time because attritional processes in the body have slowed down due to better standards of living. Questions regarding long-term changes in the relationship between biological, social, and chronological age remain underexplored. Most social science studies use chronological age as the operational variable to account for age effects, even though biological (Butler et al., 2004) and social age often appear to be better but more complicated predictors for many outcomes (Brown et al., 2017). The use of different concepts of age can therefore lead to substantial variation in outcomes of studies on long-term developments. Additional research on the intersection between (molecular) epidemiology and sociology is needed to understand the effects of biological and social aging on offspring outcomes.

At the population level, it should be noted that there is still a significant knowledge gap regarding how individual-level effects are linked to population-level outcomes (Billari, 2015). Evolution seems to act on older fathers, with the offspring of older fathers exhibiting higher mortality in historical populations and having fewer children in contemporary populations (Arslan et al., 2017). However, the extent to which this selection affects genetic health at the population level, especially in today’s societies, is largely unknown. To improve our understanding of these mechanisms, demographic studies on the population-level effects of higher MPAC on offspring (health) outcomes in historical and in contemporary populations with high MPAC, such as those in sub-Saharan African countries north of the equator, are needed.

In this regard, also technical advances in prenatal diagnostics and medical care affecting population health might be relevant (Wilkinson, 2015). Medically indicated abortions because of fetal impairment and disability are common in most high-income countries. Because they are considered high-risk, later-life pregnancies often receive increased medical attention, which improves the chances of detecting parental age-related fetal impairments. Health-selective abortion practices might also reduce the number of individuals in a population whose health is affected by advanced parental age. For instance, in Europe, the estimated reduction in the prevalence of live births of children with Down syndrome due to medically indicated abortion has been, on average, 54%, ranging from 0% in Malta to 83% in Spain (de Graaf et al., 2021). On the other hand, due to improvements in medical treatments, health-impaired infants (e.g. preterm births) have access to treatments that allow them to survive the neonatal period, in part, with no or only very limited negative consequences for their future health trajectories. To better understand how these potentially countervailing factors play out at the population level, more longitudinal studies that also take the maternal and the paternal age at childbirth into account are needed.

And finally, there might be interaction effects that can affect both, the individual as well as the population levels. A particularly important question is how maternal and paternal age at childbirth interact in shaping offspring health. Since it is methodologically extremely difficult to disentangle the effects of maternal and paternal age in monogamous societies (Khandwala et al., 2018; Phillips et al., 2019; Thompson, 2019), researchers have mainly investigated the impact of paternal and maternal ages on offspring health separately, while studies focusing on interactions between paternal and maternal ages at birth are rare (de La Rochebrochard and Thonneau, 2002). The risk of adverse offspring outcomes, such as birth complications and fetal impairment, increases exponentially for mothers after age 35 (de La Rochebrochard and Thonneau, 2002; Schmidt et al., 2012). In addition, there is evidence that advanced paternal ages have negative effects not only on offspring but also on mothers (Khandwala et al., 2018). Therefore, it is unclear whether the negative effects of parental age on the health of offspring are additive, or whether they reinforce each other if parental ages continue to increase. Currently, the population-level impacts appear to be modest due to the relatively low prevalence of advanced paternal age (Khandwala et al., 2018). However, while fathers are, on average, significantly older than mothers, there is significant variation in spousal age gaps across populations (Schoumaker, 2019; Dudel and Klüsener, 2021). More research is needed to determine to what degree trends in maternal ages at childbirth influence the impact of MAPC and APAC on offspring health.

Conclusion and outlook

This article provides for the first time a comprehensive overview over long-term trends in MPAC and APAC. This allows us to demonstrate that in many countries, the current MPAC and APAC levels are not unprecedented. Our finding puts many prominent studies that have looked at recent strong MPAC and APAC increases (Malaspina et al., 2005; Bray et al., 2006; Saha et al., 2009; Gajos and Beaver, 2017; Khandwala et al., 2017; Nybo Andersen and Urhoj, 2017) in a different perspective. Moreover, we argue that a true assessment of the implications of the current increases in MPAC and APAC on individual risks and how these relate to population-level health and human cognitive development is currently limited due to large knowledge gaps. Future research should aim to close these important gaps, which requires interdisciplinary research efforts that can take guidance from this article.

Data availability

All data are available online: https://lpac-database.shinyapps.io/Data/.

Acknowledgements

We would like to thank: Sigrid Gellers-Barkmann for corresponding with the official statistical offices to obtain data on male fertility; Josep Sottile Perez for helping to digitize the United Nations Demographic Yearbooks; and Ruben Arslan, Alyson van Raalte, and Nikola Sander for their helpful comments.

Authors’ roles

Conceptualization: K.P.W.; methodology: K.P.W. and S.K.; investigation: K.P.W. and S.K.; visualization: K.P.W.; project administration: K.P.W.; and writing of original draft: K.P.W. and S.K.

Funding

The authors declare that no specific funding was used.

Conflict of interest

The authors declare that they have no competing interests.

References