Impact of changes in human reproduction on the incidence of endocrine-related diseases

Abstract

The incidence rates of a wide range of diseases and conditions have increased over the last decades. There is controversy over the origin of these increases, whether they are caused by exposure to compounds thought to have an effect on the endocrine system, the “endocrine disruption theory”, or whether some other factor is responsible. In this analysis, the authors take a closer look at the role that changes in reproductive factors have played in this respect. They apply the relative risks of age at first pregnancy and parity or family size to a set of Dutch demographic data from 1955 and 2015 and calculate the percentage of disease increase explained. The decrease in parity over the last decades explains an increase of 26% in testicular cancer. The combination of decrease in parity and increase in maternal age at first pregnancy explains an increase of 34% in hypospadias prevalence. This combination of decreased parity and increased maternal age at pregnancy explains an increase of 24% in childhood obesity prevalence. The authors further point to a perhaps even more profound effect of the trend toward smaller families. This trend has led to an estimated doubling of the proportion of children born from subfertile couples. Since children born from subfertile couples are more likely to be preterm or of low birth weight, the incidence of these conditions must have increased as well. Low birth weight and preterm delivery are risk factors for a wide range of diseases and conditions. The changes in human reproduction over the last decades have had a profound impact on the incidence of a range of diseases and conditions in the next generation and thus provide a sound explanation for a substantial portion of the reported increases.

Key messages

• The incidence rates of a wide range of diseases and conditions have increased in the Western societies over the last decades.

• Many have argued that these increases are attributable to compounds thought to have effects on the human endocrine system: the endocrine disruption theory

• This analysis shows, however, that human reproductive factors such as maternal age at first pregnancy and parity explain substantial proportions of the reported increases

Keywords:

• Endocrine disruption
• risk factors
• reproductive factors
• demography
• subfertility
• testicular cancer
• hypospadias
• childhood obesity and epidemiology

Changing trends in endocrine related diseases

Over the past decades, the incidence rates of a number of diseases and conditions have increased in Western societies. For example, the reported age-adjusted incidence of testicular cancer in England has increased from 2.7 per 100,000 men in 1971 to 7.2 per 100,000 in 2016 (Statistics) (See Graph 1). In Northern European countries such as Denmark, Sweden, and Norway, the incidence has doubled between 1960 and 2000 (Jacobsen et al. 2006). Semen quality has been reported to be decreasing steadily (Carlsen et al. 1992; Levine et al. 2017), but this observation has been contested in the literature (Bonde et al. 2011). Endometriosis incidence rates have been reported to have increased as well. For example, Eisenberg et al. observed an endometriosis incidence increase from 5.0 per 10,000 in 2000 to 7.5 per 10,000 in 2015 in Israel (Eisenberg et al. 2018). Autism rates have also been reported to have increased over time (Rice et al. 2010; Gal et al. 2012).

Graph 1. Age-adjusted testicular cancer incidence rates per 100,000 population England, from 1971 to 2016. Source: Cancer registration data, England.

These diseases and conditions are among a larger group, including breast cancer, low birth weight, obesity, diabetes, cryptorchidism and others, thought to be related to hormonal factors. It should be kept in mind that these are reported trends. Diagnostic criteria, diagnostic tools, disease reporting, and registration processes have changed. Active screening programs have been set up. Public awareness about illness and the need to consult physicians in early stages of disease have also increased. It should therefore be acknowledged that a part of these reported rising trends is artificial. Nevertheless, it is unlikely that all these increasing disease trends are entirely artifacts.

The potential impact of endocrine-active xenobiotic compounds on human health continues to be a matter of controversy. The seminal WHO/UNEP report published in 2012 reviewed the available evidence for the theory that exposure to endocrine disruptors (EDs) is accountable for the rising incidences (Bergman 2012; Bergman et al. 2013). The ED theory argues that since the human genetic pool has not changed over the last five decades, the causes of these rising incidence must be sought in the environment, more specifically in the increasing human exposure to EDs. Both the toxicological evidence and the epidemiological evidence for the ED theory is heavily contested (Lamb et al. 2014, 2015) and much of the evidence in favor the ED theory lies in the rising incidence of diseases and the lack of any other potentially plausible causes per se. Although the WHO/UNEP 2012 report acknowledged that other factors, such as changes in nutrition, delayed pregnancy and viral disease are at play, they were ignored on the grounds that they are “difficult to identify”. We argue that there have been profound changes in human reproduction which have a major impact on the incidence of these diseases and conditions.

Changes in human reproduction

Over the recent decades, profound changes have occurred in the demographic constellation of the populations in the Western world. After the baby boom in the fifties and sixties, the introduction of birth control, family planning and increased prosperity, couples have started to delay pregnancy and to prefer smaller families. Over the past five decades, fertility rates have dropped substantially, leading to a significant increase in smaller families and thus in the proportion of firstborn children. For example, between 1970 and 1990, the proportion of firstborn children from mothers aged 30 to 39 years more than doubled in the United States (Tough et al. 2002). Families with two children or less have become the norm in Western countries. Perhaps the best way to illustrate these profound changes is to look at some demographic data routinely collected by national bureaus of statistics. Since we were most familiar with those from the Netherlands, we have chosen to use these (See Graph 2).(http://statline.cbs.nl/StatWeb/publication/?PA=71090ned). In 1950, the proportion of firstborn babies was 27%. In 2015, this proportion had risen to 46%. Even more strikingly, in 1950, 31% of all children that were born, were fourthborn or beyond. In 2015, this figure had dropped to 5%. The ratio of firstborn children to 3⁺ children has changed from about 1 to 1 in 1950 to nearly 10 to 1 in 2015. Similarly, maternal age at first pregnancy has also changed profoundly. Where about 6% of all children born in 1950 were first pregnancies from mothers of 30 years of age or over, this figure had risen to 23% in 2015.

Graph 2. Proportional differences in parity, in the Netherlands, in 1950 and 2015, data from Statistics Netherlands.

Impact of changes in human reproduction on the incidence of testicular cancer, hypospadias, and childhood obesity

If these reproductive factors are risk factors for certain diseases, then it is only logical that their incidences have increased as well. The endocrine disruption topic is extremely complex, since it involves the combination of a wide range of diseases and conditions, and potentially a wide range of risk factors may be at play. Where to begin a literature search for risk factors on the 18 diseases and conditions potentially associated with a wide range of risk factors as mentioned in the WHO/UNEP report on endocrine disruption? Structure in the literature overview was created by first focusing on testicular cancer. Testicular cancer was selected because it features prominently in the WHO/UNEP report and because of its large increase in the reported incidence over the last five decades. As second health outcome, hypospadias was selected, since reproductive health outcomes also play an important role in the WHO/UNEP report as being related to endocrine active compounds. Thirdly, childhood obesity was selected because of its continued attention in the literature and the ED controversy. The selection was made a priori to the literature searches.

The PubMed literature database was searched for publications reporting associations between testicular cancer, hypospadias, childhood obesity, and reproductive factors, including maternal age, parity and delayed pregnancy in relation to the health impact on the siblings. This was achieved by means of simple combinations of terms such as “testicular cancer” AND “risk factor*” AND “review”. Additional searches were performed using terms describing human reproductive factors such as “maternal age”, “parity”, and “delayed pregnancy” AND “sibling”.

Next, we applied the relative risks reported in the literature to the demographic data from 1950 and 2015 (or 2016 if 2015 was unavailable). Based on these relative risks and the demographic data, we calculated how much the incidence of the diseases and conditions would rise as a result of these changes in risk factors. The calculations are basically a weighted sum of the product between the proportion of each stratum and the relative risk for the disease in that stratum, according to the formula: $$\mathrm{I}\mathrm{E}=\mathrm{ }\frac{\sum _{i=1}^n({\mathrm{R}\mathrm{R}}_i\mathrm{ }\times \mathrm{ }{p}_{1i})-\mathrm{ }\sum _{i=1}^n({\mathrm{R}\mathrm{R}}_i\mathrm{ }\times \mathrm{ }{p}_{0i})}{\sum _{i=1}^n({\mathrm{R}\mathrm{R}}_i\mathrm{ }\times \mathrm{ }{p}_{0i})}$$ where IE is the proportional increase in incidence explained by these factors, n is the number of categories for the risk factors included in the summary, RR_i is the category-specific relative risk, p_1i is the proportion of subjects in category i in the final exposure distribution, and p_0i is the corresponding proportion in the initial exposure distribution.

By applying the category-specific relative risks to these demographic data of 1950 and 2015, one can estimate the proportion of the change in disease incidence that is explained by that factor. This was done for the demographic factors birth order, maternal age, and parity. Unfortunately, the routinely collected statistics data do not include Body Mass Index (BMI) or paternal age, so these factors could not be taken into account.

Testicular cancer

Boys born from nulliparous mothers (firstborn boys) are at an increased risk for testicular cancer, compared to boys from multiparous mothers, irrespective of the gender of the siblings. This association is found in several studies (Swerdlow et al. 1987; Prener et al. 1992; Moller and Skakkebaek 1996; Sabroe and Olsen 1998; Wanderas et al. 1998; Westergaard et al. 1998; Bevier et al. 2011), but not all (Moller and Skakkebaek 1997), and it seems to be the case in particular in the earlier birth cohorts. Other frequently reported risk factors are physical length (Dieckmann et al. 2008), age at puberty (later age decreases testicular cancer risk and average age at puberty has decreased) (Weir et al. 1998), low birth weight (Brown et al. 1986), and maternal age (Moller and Skakkebaek 1997). In a case-control study on 259 cases of testicular cancer and 238 controls in the United Kingdom, Swerdlow et al. observed a statistically significant association between birth order and testicular cancer risk. Taking the firstborn son as reference, the son born as second child had a relative risk of 0.92, as third born child of 0.75 and of fourth- or higher born child of 0.3 (Swerdlow et al. 1987). These relative risks were applied to the demographic data from 1950 and 2015 (See Table 1). By calculating the weighted sum of the relative risks in 1950 and 2015, it is shown that an increase of 26% in testicular cancer is attributable to the temporal change in birth order.

Table 1. Impact of changes in birth order distribution on testicular cancer incidence in 1950 and 2015 in the Netherlands.

Birth order	Relative risk (Prener et al. 1992)	p children born in 1950	Product 1950	p children born in 2015	Product 2015
1st	1.0 (Reference)	.272	0.272	.455	0.455
2nd	0.9	.246	0.2214	.368	0.3312
3rd	0.7	.171	0.1197	.127	0.0889
≥4	0.3	.312	0.0936	.049	0.0147
∑		1	0.7067	1	0.8898
% increase explained		(0.8898–0.7067)/0.7067 = 0.259 = 26%

Changes in birth order (or family size) between 1950 and 2015 explain a 26% increase of testicular cancer in boys born in these years.

Hypospadias

Hypospadias is one of the most common birth defects, with a prevalence of around 0.3% to 0.5% (Carmichael et al. 2007). There is a substantial body of evidence linking hypospadias to risk factors such as parity and maternal age. The study by Carmichael et al. quantified the relative risks for hypospadias for the combination of maternal age and parity (Carmichael et al. 2007). Since national statistics for the Netherlands were available for that particular combination, we applied the relative risks to these data from 1950 and 2016 (See Table 2).

Table 2. Impact of changes in maternal age and parity on the prevalence of hypospadias in the Netherlands in 1950 and 2016.

Parity maternal age	Relative risk (Carmichael et al.2007)	p children born in 1950	Product	p children born in 2016	Product
1+ < 30	1 (Reference)	.256	0.256	.137	0.137
1 + 30–34	2.4	.222	0.5328	.232	0.5328
1 + 35+	2.5	.251	0.6275	.177	0.4425
0^a < 30	2.8	.212	0.5936	.223	0.6244
0 30–34	5.5	.042	0.231	.160	0.88
0 35+	8.2	.017	0.1394	.069	0.5658
∑		1	2.3803	1	3.1825
% change explained		(3.1825–2.3803)/2.3803 = 0.337 = 34%

^aChildren born from nulliparous mothers.

The combination of the risk factors Parity and Maternal age explains a 34% increase in hypospadias prevalence from 1950 to 2016. Carmichael et al. further investigated the additional risk of a maternal BMI equal or over 26, which in total is 1.3 and also provides the Relative Risk for the strata specific for BMI. It is evident that the prevalence of mothers with a BMI of 26 or over has increased from 1950 to 2016. However, there are no national statistics available for the combination of parity, maternal age, and BMI and thus their impact cannot be taken into account in our analysis.

Childhood obesity

Thirdly, there is ample evidence that parity and maternal age play a role in childhood obesity, again with the highest risk in firstborn children and from older mothers. In a prospective cohort study, children were followed from birth up to the age of 6 (Gaillard et al. 2014). The authors reported associations between birth order and childhood obesity at age 6. Again, the stratum-specific relative risks were combined with the demographic data from 1950 and 2015 to calculate the percentage of increase in prevalence explained (See Table 3).

Table 3. Impact of parity on the prevalence of childhood obesity at age 6 between children born in 1950 and 2015 in the Netherlands.

Birth order	Relative risk (Gaillard et al. 2014)	p children born in 1950	Product 1950	p children born in 2015	Product 2015
1st	1.0 (Reference)	.272	0.272	.455	0.455
2nd	0.8	.246	0.1968	.368	0.2944
3d	0.6	.171	0.1026	.127	0.0762
≥4	0.35	.312	0.1092	.049	0.01715
∑		1	0.6806	1	0.84275
% increase explained		(0.84275–0.6806)/0.6806 = 0.2382 = 24%

The change in the demographic factor Birth order between 1950 and 2015 explains an increase of 24% in childhood obesity.

Reproductive factors and the 18 diseases and conditions in the WHO/UNEP report

It seems that a substantial part of the increase in incidence of testicular cancer, hypospadias, and childhood obesity can be explained by reproductive factors alone. But what about the other diseases and conditions mentioned in the WHO/UNEP report? The report lists 18 diseases and conditions that may be associated with endocrine disruption from exposure to environmental chemicals (Bergman 2012). We explored the PubMed database for journal articles reporting associations between reproductive factors and these 18 diseases and conditions. The findings of these explorative searches are summarized in Table 4.

Table 4. Reproductive factors reported in PubMed to be associated with any of the 18 diseases and conditions mentioned in the WHO/UNEP report on endocrine disruptors.

Disease or condition	Reproductive factor
Low semen quality	Obesity (Sharpe 2010)
Cryptorchidism	Higher risk in firstborn boys (Biggs et al. 2002)
Hypospadias	Increased risk with higher maternal age, lower risk with parity (Gaillard et al. 2014)
Preterm birth	Preterm birth rates are higher in mothers with advanced age and in primiparous mothers (Fuchs et al. 2018)
Low birth weight	Low birth weight more frequent in primiparous mothers (Suzuki et al. 2016)
Thyroid disruption	Higher rate of thyroid dysgenesis in children from mothers with advanced maternal age (Kirmizibekmez et al. 2012)
Thyroid cancer	Highest risk in firstborn children (Bevier et al. 2011)
Breast cancer	Nulliparity and older age at first pregnancy are well established risk factors (Pathak et al. 2000)
Endometrial cancer	Highest risk in firstborn girls (Bevier et al. 2011)
Ovarian cancer	Parity reduces the risk for ovarian cancer (Braem et al. 2010)
Prostate cancer	Higher risk associated with older paternal age (Zhang et al. 1999)
Testicular cancer	Highest risk in firstborn boys (Swerdlow et al. 1987)
Early onset breast development	Obesity (Kaplowitz et al. 2001)
Obesity	Elevated childhood obesity in firstborn children and from mothers with advanced maternal age (Gaillard et al. 2014)
Type II diabetes	Firstborn children have greater reduced insulin sensitivity (Ayyavoo et al. 2013).
ADHD	Higher risk with older paternal age (D’Onofrio et al. 2014), but inconsistently found
Autism	Advanced paternal age (Reichenberg et al. 2006) lowest risk in boys with birth order ≥4 (Wang et al. 2017)
Anogenital distance	No relevant human reproductive factors found in PubMed

For 15 out of the 18 diseases and conditions listed in the WHO/UNEP report, associations with reproductive factors were reported in the literature. Although we did not perform a full systematic search, the overview shows that reproductive factors such as parity and maternal age, and sometimes paternal age, play are role in the incidence of these diseases and conditions. Many of the listed conditions have interrelations, cryptorchidism, for example, is a risk factor for testicular cancer, and low birth weight is a risk factor for autism (Schendel and Bhasin 2008; Gardener et al. 2011) and cryptorchidism.

Impact of smaller family size on children born from subfertile couples

Today, families with two children or less are the norm in Western societies and family size is much more a matter of the parents’ desire than a biological fertility issue. Decades ago, a subfertile couple would be likely to have no more than two children, while a fertile couple would have four or more. Today, fertile couples have a similarly small number of children as subfertile couples. Thus, where in the past subfertile couples would contribute relatively few children to the next generation compared to fertile couples, they now contribute equally. This development has substantial consequences for disease rates in the next generation, since a couple’s subfertility is a risk factor for disease incidences in the next generation. There is substantial evidence that births from subfertile couples are at an increased risk for preterm birth, low birth weight, and perinatal death (Luke 2017).

Let us now perform some simple estimations. Luke et al. estimated that 10 to 15% of the couples are subfertile (Luke 2017). Assuming 10% subfertile couples, 2 children per subfertile couple, and 4 children per fertile couple in 1950s, we can estimate that in a hypothetical group of 100 couples there would be 10 × 2 = 20 births from a subfertile couple and 90 × 4 = 360 births from fertile couples. Thus, in the 1950s 20/380 = 5.3% of the next generation was born from subfertile couples. Today, a family with 2 children is the norm, regardless of a couple’s fertility. Therefore, in similar group of 100 couples in 2015, subfertile couples would still contribute 10 × 2 = 20 births compared to 90 × 2 = 180 births from fertile couples, resulting in 20/200 = 10% of the next generation born from subfertile parents. The trend to have smaller families has resulted in almost a doubling of the percentage of children born from subfertile couples. The doubling of the proportion of children born from subfertile couples results in an increase in the proportion of children with low birth weight and preterm birth, with all its associated health implications in later life (Virtanen and Toppari 2008; Rogvi et al. 2015; Huang et al. 2018).

Discussion

The trends of smaller families and delayed pregnancies, so common in Western societies, have resulted in profound shifts in demographic factors. The ratio of firstborn children in 1950 to 2015 has changed with a factor 8, and the percentage of firstborn children from mothers over age 35 has increased dramatically. Clearly, these shifts have their effect on disease rates in the next generation. In addition to these obvious consequences, the fact that family size today is determined by the desire of the parents, rather than their actual fertility, has a profound impact on the proportion of children born from subfertile couples. The three case studies presented here clearly indicate that a substantial proportion of the rising trends in a number of diseases and conditions is attributable to these changes in reproductive behavior.

What can be the underlying biological mechanism for these phenomena? Firstly, it is widely acknowledged that the first or nulliparous pregnancy generally has some aspects that may impact fetal development. Maternal nulliparity is a risk factor for suboptimal hemodynamic adaptations during pregnancy, which may adversely affect fetal nutrient supply (Gaillard et al. 2014). In the study by Gaillard et al., the average birth weight of babies born from nulliparous mothers was 3325 grams, compared to averages of over 3500 grams for multiparous births. Secondly, hormone levels change as people age. A study on more than 100,000 blood samples of women at age 20 had a median serum testosterone concentration of 1.4 nmol/l, which decreased to a median of 0.9 nmol/l around age 50 (Handelsman et al. 2016). A fetus in a young mother is, on average, exposed to different hormone levels than a fetus in an older mother. Thirdly, pregnancy itself, and thus parity, has an impact on maternal hormone levels. For example, a recent study (albeit on small numbers) reported that the progesterone and testosterone levels were lower in multiparous women than in primiparous women throughout gestation (Schock et al. 2016).

As stated earlier, there are many factors at play in the etiology of ED diseases, one group of these being the demographic factors. Could one, estimating the percentage of increase attributable to these factors, exclude other factors? We believe that this would be opportunistic, simply because the incidence data collected in 1950 cannot be directly compared to data from 2015. There have been too many changes to allow for a direct comparison. Diagnostic procedures have changed, classifications have been adapted, the quality and completeness have improved, and public awareness and consultation of doctors have increased as well. It is therefore impossible to assess what percentages of the temporal increases remain unexplained by the changes in demographic or other factors, but it must be clear that a change from 27% of firstborn children in 1950 to 46% in 2015, has a major impact on disease rates in the next generation, given that firstborn children, on average, have lower birth weights, higher preterm birth rates, and higher rates of being small for gestational age (Gaillard et al. 2014).

Reproductive factors, such as family size and older maternal age at first pregnancy, have fundamentally changed during the last decades. We demonstrate in our analysis that these factors are responsible for significant changes in the occurrence of ED-related diseases. Other demographic, reproductive and metabolic factors that we were not able to study, such as BMI, may play a similar role in the observed trends. For example, in the case of autism, there is substantial evidence that older paternal age increases the risk of autism in the children. In a systematic review of the literature, Wang et al. concluded that paternal age over 34 years increased the risk of autism with 32% (Wang et al. 2017). We recommend that these factors be taken into account in the ongoing controversy on endocrine disruption.

Declaration of interest

The authors declare to have no conflict of interest. No funding was received for this project. This critical review was conducted during the normal course of the authors’ employment using institutional funding. No outside funds were used to prepare the review. The review is the professional work product of the authors and the views expressed are not necessarily the views of their employers. None of the authors have appeared during the last five years in any regulatory or legal proceedings related to the contents of this paper.

Acknowledgments

The authors gratefully acknowledge the reviewers for their comments on the manuscript. Their comments and suggestions were helpful to us and contributed to the clarity and accuracy of our publication. But most of all we are grateful for the inspiring words of one of the reviewers who remarked that our piece of work gave some serious food for thought on the endocrine disruption controversy.