NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Institute of Medicine (US) Committee on Understanding the Biology of Sex and Gender Differences; Wizemann TM, Pardue ML, editors. Exploring the Biological Contributions to Human Health: Does Sex Matter? Washington (DC): National Academies Press (US); 2001.

Cover of Exploring the Biological Contributions to Human Health

Exploring the Biological Contributions to Human Health: Does Sex Matter?

Show details

3Sex Begins in the Womb


Sex differences of importance to health and human disease occur throughout the life span, although the specific expression of these differences varies at different stages of life. Some differences originate in events occurring in the intrauterine environment, where developmental processes differentially organize tissues for later activation in the male or female. In the prenatal period, sex determination and differentiation occur in a series of sequential processes governed by genetic and environmental factors. During the pubertal period, behavioral and hormonal changes manifest the secondary sexual characteristics that reinforce the sexual identity of the individual through adolescence and into adulthood. Hormonal events occurring in puberty lay a framework for biological differences that persist through life and that contribute to variable onset and progression of disease in males and females. It is important to study sex differences at all stages of the life cycle, relying on animal models of disease and including sex as a variable in basic and clinical research designs.

All human individuals—whether they have an XX, an XY, or an atypical sex chromosome combination—begin development from the same starting point. During early development the gonads of the fetus remain undifferentiated; that is, all fetal genitalia are the same and are phenotypically female. After approximately 6 to 7 weeks of gestation, however, the expression of a gene on the Y chromosome induces changes that result in the development of the testes. Thus, this gene is singularly important in inducing testis development. The production of testosterone at about 9 weeks of gestation results in the development of the reproductive tract and the masculinization (the normal development of male sex characteristics) of the brain and genitalia. In contrast to the role of the fetal testis in differentiation of a male genital tract and external genitalia in utero, fetal ovarian secretions are not required for female sex differentiation. As these details point out, the basic differences between the sexes begin in the womb, and this chapter examines how sex differences develop and change across the lifetime. The committee examined both normal and abnormal routes of development that lead individuals to become males and females and the changes during childhood, reproductive adulthood, and the later stages of life.


One of the basic goals of biologists is to explain observed variability among and within species. Why does one individual become infected when exposed to a microbiological agent when another individual does not? Why does one individual experience pain more acutely than another? Sex is a prime variable to which such differences can be ascribed. No one factor is responsible for variability, but rather, a blend of genetic, hormonal, and experiential factors operating at different times during development result in the phenotype called a human being.

As suggested by the reproductive processes of some species and punctuated by recent successful efforts at cloning of some species, sexual reproduction is not necessary for species perpetuation. Debate exists on why sexual reproduction has evolved. Most biologists agree that it increases the variability upon which evolutionary selection can operate; for example, variability would allow some offspring to escape pathogens and survive to reproduce. This theory is not without its critics (Barton and Charlesworth, 1998). The contribution of genetics to sex differences has been described in Chapter 2. Here the focus is more on the endocrine and experiential bases for the development and expression of sex as a phenotype.

Different species of vertebrate animals have evolved different pathways to determine sex, but it is interesting that in all cases two sexes emerge with distinctly different roles in the social and reproductive lives of the animals (Crews, 1993; Francis, 1992). In all vertebrates the genetic basis of sex is determined by meiosis, a process by which paired chromosomes are separated, resulting in the formation of an egg or sperm, which are then joined at fertilization. Variations in the phenotypic characteristics of the different sexes are determined during development by internal chemical signals. The process can be influenced by external factors such as maternal endocrine dysfunction or endocrine disrupters, as well as fetal endocrine disorders and exogenous medications (Grumbach and Conte, 1998).

Nongenomic Sexual Differentiation and Sexual Flexibility

Nongenomic sexual differentiation has evolved in several species of fishes and reptiles. In these species, sex results from external signals. For example, temperature during embryogenesis is the cue acting on autosomal genes to result in adult males and females in several species. In many species of flounder, for instance, elevated temperatures of the water in which the larval fish develop results in a higher proportion of males (Yamamoto, 1999). Similarly, in several turtle species the incubation temperature of the eggs influences the sex ratio of the animals (Crews et al., 1989).

In some species, sex determination can be delayed until well after birth or the sex can even change after the birth of an organism. One fascinating study found that several species of fish develop sexual phenotypes as a result of the fish's social rank in a group (Baroiller et al., 1999; Warner, 1984). The blue-headed wrasse is a polygynous coral reef fish with three phenotypes that vary in size, coloration, reproductive organs, physiology, and behavior (Godwin et al., 1996; Warner and Swearer, 1991). These phenotypes are females, initial-phase males, and terminal-phase males. As a result of changes in the social role, a fish can progress rapidly through these phenotypes. Upon the disappearance of a terminal-phase male, the behavior of the largest female in the group converts to male-like behavior in minutes and the fish shows full gonadal changes in days.

The belted sandfish (Sermnus subligarius) stands out as one of the most remarkable demonstrations of vertebrate sexual flexibility. This coastal marine fish is a simultaneous hermaphrodite (Cheek et al., 2000). Its gonads produce both sperm and eggs, and each fish has the reproductive tract anatomies of both sexes simultaneously. Within minutes each individual can show three alternative mating behaviors—that is, female, courting male, or streaker male—along with the appropriate external color changes (Cheek et al., 2000). A streaker male awaits the peak moment during the courtship of male and female morphs and then streaks in to release sperm at the moment of spawning. The sperm compete with the courting male's sperm. Partners can switch between male and female roles within seconds and may take turns fertilizing each other's eggs. The frequency with which an individual plays the female or male role is, in part, a function of size. Larger fish are more likely to play the male role more often.

In contrast, mammalian sex determination is more directly under the control of a single internal event: fertilization. Under normal conditions, the direction of sexual development is initiated and determined by the presence or absence of a Y chromosome.

Intrauterine Environment

In mammals, once genetic sex has been determined and the fetus begins its development, the fetal environment, especially hormones, can result in significant modifications of the genetically based sex. The effect of prenatal hormones on later anatomy, physiology, and behavior are most clearly demonstrated in several animals showing the “intrauterine position effect” (vom Saal et al., 1999). In litter-bearing mammals such as mice, rats, gerbils, and pigs, each pup shares the uterus with several others, some of which are of a different sex. Significant differences among females occur if the fetus is located between two males or with a male on one side or with no male on either side. Testosterone is produced by fetal males and can masculinize adjacent females to various degrees. Thus, not only do individuals vary as a result of genetic variability, but they can also vary as a result of prenatal hormonal organizational effects (see additional discussion in Chapter 4). Extensive studies with the female mouse have revealed that adult anatomical structures, such as the genitalia and sexually dimorphic parts of the brain, and the rate of reproductive development vary as a result of proximity to males in the womb (Vandenbergh and Huggett, 1995).

Studies with animals suggest that hormonal transfer between fetuses can influence later anatomical, physiological, and behavioral characteristics. Some data from studies with humans, recently summarized by Miller (1998), suggest that a similar phenomenon occurs in mixed-sex twins. His review of the literature reveals a number of characteristics apparently influenced by transmission of testosterone from the male twin to the female twin. For example, (1) dental asymmetry is also a characteristic of females with male co-twins (the right jaw of the male has larger teeth) (Boklage, 1985), (2) spontaneous otoacoustic emissions are at an intermediate level in females with male co-twins (the rates of clicking sounds produced in the cochlea usually differ between males and females) (McFadden, 1993), and (3) the level of sensation seeking appears to be higher in females with male co-twins than in those without male co-twins (Resnick et al., 1993). These studies suggest that, as in rodent models, testosterone transferred to human female fetuses can have masculinizing effects on anatomical, physiological, and behavioral traits.

In humans, the metabolic stress of pregnancy increases the incidence of gestational diabetes in susceptible women. Transgenerational passage of diabetes may contribute to the higher incidence of impaired glucose tolerance, obesity, and hypertension in the offspring of diabetic mothers and to the prevalence of diabetes in such human communities as the Pima Indians (Cho et al., 2000; Silverman et al., 1995). This passage of a disease condition across generations by non-genome-dependent mechanisms emphasizes the importance of good maternal care and health during pregnancy. Although males will also be affected by a hyperglycemic environment during fetal life and will themselves have an increased risk of diabetes in adulthood, they do not provide the womb environment during the critical phases of fetal development of the next generation. Thus, males do not pass the tendency across generations (Cho et al., 2000; Nathanielsz, 1999; Silverman et al., 1995).

Low birth weight or small body size at birth as a result of reduced intrauterine growth are associated with increased rates of coronary heart disease and non-insulin-dependent diabetes in adult life (reviewed by Barker [2000]). The “fetal origins hypothesis” proposes that undernutrition during critical periods of fetal growth can force the fetus to adapt by altering cardiovascular, metabolic, or endocrine functions to survive. (Note that debate continues as to whether the association is truly causal [Kramer, 2000; The Lancet, 2001; Lumey, 2001].) These changes, such as redistribution of blood flow, changes in the production of fetal and placental hormones involved in growth, and metabolic changes, can permanently change the function and structure of the body. For example, offspring who were exposed in utero to maternal famine during the first trimester of development had higher total cholesterol and low-density lipid cholesterol levels and a higher ratio of low-density lipid to high-density lipid cholesterol levels, all of which are risk factors for heart disease (Roseboom et al., 2000). This altered lipid profile persisted even after adjustments for adult lifestyle factors such as smoking, socioeconomic status, or use of lipid-lowering drugs. Male offspring had higher rates of obesity at age 19 years, but maternal malnutrition during early gestation was associated with a higher prevalence of obesity in 50-year-old women (Ravelli et al., 1999).

Such permanent alterations in body structure or functions may have effects on future generations as well. Studies show that when a female fetus is undernourished and subsequently of low birth weight, the permanent physiological and metabolic changes in her body can lead to reduced fetal growth and raised blood pressure in her offspring (Barker at al., 2000; Stein and Lumey, 2000). Furthermore, in birth cohorts of males with spina bifida who had been exposed to prenatal famine, the relative risk of death was 2.5-fold greater than that in similarly affected female offspring (Brown and Susser, 1997). These traits in the offspring were not affected by the father's size at birth.


The remarkable accumulation of knowledge over the past five decades and new and continuing insights in the field of sex determination and sex differentiation represent major landmarks in biomedical science. No aspect of prenatal development is better understood. Advances in embryology, steroid biochemistry, molecular and cell biology, cytogenetics, genetics, endocrinology, immunology, transplantation biology, and the behavioral sciences have contributed to the understanding of sexual anomalies in humans and to the improved clinical management of individuals with these disorders. Major contributions to this understanding have stemmed from studies of patients with abnormalities of sex determination and differentiation and the recent advances emanating from molecular genetics. These advances, considered together, illustrate that a failure in any of the sequential stages of sexual development, whether the cause is genetic or environmental, can have a profound effect on the sex phenotype of the individual and can lead to complete sex reversal, various degrees of ambisexual development, or less overt abnormalities in sexual function that first become apparent after sexual maturity (Grumbach and Conte, 1998; Wilson, 1999).

Sex Determination

Sex determination and sex differentiation are sequential processes that involve successive establishment of chromosomal sex in the zygote at the moment of conception, determination of gonadal (primary) sex by the genetic sex, and determination of phenotypic sex by the gonads. At puberty the development of secondary sexual characteristics reinforces and provides more visible phenotypic manifestations of the sexual dimorphism. Sex determination is concerned with the regulation of the development of the primary or gonadal sex, and sex differentiation encompasses the events subsequent to gonadal organogenesis. These processes are regulated by at least 70 different genes that are located on the sex chromosomes and autosomes and that act through a variety of mechanisms including those that involve organizing factors, gonadal steroids and peptide hormones, and tissue receptors. Mammalian embryos remain sexually undifferentiated until the time of sex determination.

An important point is that early embryos of both sexes possess indifferent common primordia that have an inherent tendency to feminize unless there is active interference by masculinizing factors (Grumbach and Conte, 1998).

It has been known for more than four decades that a testis-determining locus, TDF (testis-determining factor), resides on the Y chromosome. About 10 years ago, the testis-determining gene was found to be the SRY (sex-determining region Y) gene (Ferguson-Smith and Goodfellow, 1995; Koopman, 1999; Koopman et al., 1991; O'Neill and O'Neill, 1999; Sinclair et al., 1990; Swain and Lovell-Badge, 1999), which is the primary sex determinant, as it is the inducer of differentiation of the indifferent gonad into testes and hence is the inducer of male sexual development. SRY is expressed in 46,XY gonads in Sertoli cell progenitors at the stage of sex cord formation, but unlike the mouse, in which SRY expression is brief, SRY mRNA persists in Sertoli cells at 18 weeks of gestation (Hanley et al., 2000). As discussed in Chapter 2, the human SRY gene is located on the short arm of the Y chromosome and comprises a single exon that encodes a protein of 204 amino acids including a 79-residue conserved DNA bending and DNA binding domain: the HMG (high-mobility-group) box.

The mechanisms involved in the translation of genetic sex into the development of a testis or an ovary are now understood in broad terms (Figure 3–1).

Figure Icon


Permission was not granted to electronically reproduce figure 3–1 from In: Williams Textbook of Endocrinology, 9th ed. J.D.Wilson, D.W.Foster, H.M.Kronenberg, and P.R.Larsen, eds. Philadelphia: W.B.saunders Co. This figure is available in the (more...)

It is known that a variety of autosomal and X-chromosome-linked genes, literally a cascade of genes that exert complex gene dosage balancing activities, are involved in testis determination. All major sex-determining genes have been shown to be subject to a dosage effect. In the human, the SRY protein is detected at an early age of gonadal differentiation in XY embryos, where it induces Sertoli cell differentiation. In the human adult, it is present in both Sertoli and germ cells. In embryonic and fetal life, the evidence suggests that the SRY gene product regulates gene expression in a cell-autonomous manner. The precise molecular mechanisms by which SRY triggers testis development are unknown, nor is it yet known how SRY is regulated. The genetic sex of the zygote is established by fertilization of a normal ovum by an X-chromosome- or Y-chromosome-bearing sperm.

Genes Contributing to Sex Determination

Apart from SRY, a number of autosomal and X-chromosome-linked genes have been identified and have a critical role in male or female sex determination, the testis- and ovary-determining cascades (Roberts et al., 1999) (Table 3–1). In the human, heterozygous mutations or deletion of the Wilm's tumor (WT1) gene located on chromosome 11p13 results in urogenital malformations as well as Wilm's tumors. Knockout of the WT1 gene in mice results in apoptosis of the metanephric blastema, with the resultant absence of the kidneys and gonads. Thus, WT1, a transcriptional regulator, appears to act on metanephric blastema early in urogenital development.

TABLE 3–1. Some Genes Involved in Human Sex Determination.


Some Genes Involved in Human Sex Determination.

SF-1 (steroidogenic factor-1) is an orphan nuclear receptor involved in transcriptional regulation. It is expressed in both the male and the female urogenital ridges as well as steroidogenic tissues, where it is re quired for the synthesis of, for example, testosterone and estrogen, and in Sertoli cells, where it regulates the anti-müllerian hormone gene (Parker et al., 1999). SF-1 is encoded by the mammalian homologue of the Drosophila melanogaster gene (FTZ-F1). Knockout of the Sf-1 gene in mice results in apoptosis of the genital ridge cells that give rise to the adrenals and gonads and, thus, a lack of gonadal and adrenal morphogenesis in both males and females. A heterozygous mutation in the human gene encoding SF-1 causes XY sex reversal (Achermann et al., 1999), which results in individuals with normal female external genitalia, streak-like gonads containing sparse and poorly differentiated tubules, and the failure of adrenal development. WT1 and SF-1 appear to play important roles in the differentiation of the genital ridge from the intermediate mesoderm. WT1 and SF-1 are expressed when the indifferent gonadal ridge first differentiates at 32 days postovulation in both female and male embryos (Hanley et al., 1999).

XY gonadal dysgenesis with resulting female differentiation has occurred in 46,XY individuals with intact SRY function but with duplication of Xp21, leading to a double dose of the DAX-1 (dosage-sensitive sex reversal congenital adrenal hypoplasia congenital-critical region on the X chromosome, gene 1) gene. On the other hand, a mutation or deletion of DAX-1 in XY individuals results in X-linked congenital adrenal hypoplasia and hypogonadotropic hypogonadism but not an abnormality in testis differentiation. Similarly, duplication of the DAX-1 gene on one X chromosome appears not to affect ovarian morphogenesis or function in 46,XX females. Targeted disruption of the Dax-1 gene in mice does not affect ovarian development. It has been suggested that Dax-1 is an “anti-testis” factor rather than an ovary-determining gene. SRY and Dax-1 appear to act antagonistically in gonadal dysgenesis (Parker et al., 1999; Roberts et al., 1999). Dax-1 expression is detected in the primate gonadal ridge days before the peak expression of SRY (Hanley et al., 2000).

Camptomelic dysplasia is a skeletal dysplasia associated with sex reversal because of gonadal dysgenesis in about 60 percent of affected 46,XY individuals. A gene for a camptomelic dysplasia, SOX-9, has been localized to 17q24.3–25.1. The products of SOX genes (for the SRY-related HMG-box gene), as a rule, are more than 50 percent identical to those of SRY genes at the amino acid level in the HMG-box region (Koopman, 1999). In the human, SOX-9 transcripts are present in the gonadal ridge of both male and female embryos (Hanley et al., 2000).

XY individuals with 9p- or 10q- deletions as well as patients with 1p32–36 duplications exhibit gonadal dysgenesis and male pseudohermaphrodism, which suggests that autosomal genes at these loci are important in the gonadal differentiation cascade. In this regard, two genes, DMRT-1 and DMRT-2, have been localized to the distal region of the short arm of chromosome 9 (Raymond et al., 1999). These genes are related to the sexual regulatory genes Dsx (double sex) in D. melanogaster and MAB-3 in Caenorhabditis elegans (or double sex- and MAB-3-related transcription factor). Their evolutionary conservation, deletion from sex-reversed males with the 9p- syndrome, and male-specific expression in early human gonadogenesis suggest that one or both genes have a role in human sex determination (Calvari et al., 2000).

WNT-4, a vertebrate homologue of the D. melanogaster polarity gene (“wingless”), is involved in the regulation of steroid biosynthesis in the fetal gonad (Uusitalo et al., 1999). Wnt-4 knockout female mice lack müllerian ducts and exhibit decreased levels of oocyte development and decreased rates of survival. WNT-4 is downregulated in the fetal testis, presumably by SRY. Consequently, testosterone synthesis occurs in the XY individual. It has recently been demonstrated that WNT-4 in humans is located on chromosome 1p35 and that duplication of WNT-4 upregulates DAX-1 expression and causes sex reversal in a 46,XY individual. 46,XX mice with homozygous disruption of the Wnt-4 gene manifest testosterone synthesis in the fetal ovary and masculinization of the wolffian ducts. This observation suggests that Wnt-4 expression in the fetal ovary inhibits gonadal androgen biosynthesis.

Organogenesis of the Testes

Until about the 12-millimeter stage (approximately 42 days of gestation), the embryonic gonads of males and females are indistinguishable. By 42 days, 300 to 1,300 primordial germ cells have reached the undifferentiated gonad from their extragonadal origin in the dorsal endoderm of the yolk sac. These large cells are the progenitors of oogonia and spermatogonia. In the absence of primordial germ cells, the gonadal ridges in the female remain undeveloped. Germ cells are not essential for differentiation of the testes (Grumbach and Conte, 1998).

There is a striking sexual dimorphism in the timing of gonadal differentiation under the influence of SRY and other testis-determining genes (Figures 3–2 and 3–3). Organization of the indifferent gonad is definitive by the 6th to 7th week of gestation; the testes develop more rapidly than the ovaries. The ovary does not emerge from the indifferent stage until 3 months of gestation, when the earliest sign of differentiation into ovaries appears: the beginning of meiosis, as evidenced by the maturation of oogonia into oocytes. The precursor of the Sertoli cell that arises from the coelomic epithelium expresses SRY, leading to differentiation of Sertoli cells, which marks testis differentiation (Capel, 2000). The Sertoli cell is the only cell in the testes in which SRY has a critical effect. Germ cells in the XY gonad are sequestrated inside the forming testis cords. Anti-müllerian hormone (AMH) (or müllerian-inhibiting factor or substance) is a member of the transforming growth factor beta family, one of the earliest known products of Sertoli cells. The organization of testicular cords regulates Leydig cells to the interstitial region between the primitive seminiferous tubules.

Figure Icon


Permission was not granted to electronically reproduce figure 3–2 from In: Williams Textbook of Endocrinology, 9th ed. J.D.Wilson, D.W.Foster, H.M.Kronenberg, and P.R.Larsen, eds. Philadelphia: W.B.saunders Co. This figure is available in the (more...)

Figure Icon


Permission was not granted to electronically reproduce figure 3–3 from In: Williams Textbook of Endocrinology, 9th ed. J.D.Wilson, D.W.Foster, H.M.Kronenberg, and P.R.Larsen, eds. Philadelphia: W.B.saunders Co. This figure is available in the (more...)

The early endocrine function of the fetal testis is the secretion by the Sertoli cells of anti-müllerian hormone (AMH), a homodimeric glycoprotein that functions as a paracrine secretion (Donahoe et al., 1987; Josso et al., 1993). It passes by diffusion to the paired müllerian ducts and induces their dissolution by apoptosis. The versatile Sertoli cell also secretes inhibin, nurtures the germ cells, expresses stem cell factor, synthesizes an androgen binding protein, and prevents meiosis. Leydig cells are first found at about 60 days of gestation. Leydig cells secrete testosterone, the regulator of male differentiation of the wolffian ducts, urogenital sinus, and external genitalia. After differentiation of the primitive testicular cords, they rapidly proliferate during the 3rd month and the first half of the 4th month. During this period the interstitial spaces between the seminiferous tubules are crowded with Leydig cells.

The onset of testosterone biosynthesis occurs at about the 9th week (Siiteri and Wilson, 1974). Human chorionic gonadotropin (hCG)-lutein izing hormone (LH) receptors are present in fetal Leydig cells by at least the 12th week of gestation, an observation that suggests that the initial secretion of testosterone at about 8 to 9 of weeks gestation is independent of hCG and fetal pituitary LH.

The concentration of testosterone in the plasma of the male fetus correlates with the biosynthetic activity of the fetal testis. Peak concentrations in the fetal circulation are 200 to 600 nanograms per deciliter (ng/ dl), values comparable to those in the adult male, and are reached by about the 16th week of gestation (Grumbach and Conte, 1998; Grumbach and Gluckman, 1994). Between 16 and 20 weeks of gestation the testosterone levels fall to about 100 ng/dl; after 24 weeks the plasma testosterone level is in the range found in early puberty. Clinical as well as biochemical data indicate that the hCG secreted by the syncytiotrophoblast of the placenta stimulates testosterone secretion during the critical period of male sex differentiation. The number of Leydig cells decreases after week 18 of gestation, probably by dedifferentiation.

Fetal pituitary gonadotropins are essential for the continued growth and function of the fetal testis after the early period of sex differentiation. Fetal pituitary LH seems necessary in concert with hCG for the normal growth of the differentiated penis and scrotum during the latter half of gestation and for descent of the testes. Fetal Leydig cells differ from adult Leydig cells in their morphologies, their regulatory mechanisms, and their lack of desensitization to high doses of hCG and LH. Figure 3–4 correlates the pattern of testosterone, hCG, and fetal pituitary LH and follicle-stimulating hormone (FSH) concentrations during gestation with the histological changes in the fetal testis.

Figure Icon


Permission was not granted to electronically reproduce figure 3–4 from In: Williams Textbook of Endocrinology, 9th ed. J.D.Wilson, D.W.Foster, H.M.Kronenberg, and P.R.Larsen, eds. Philadelphia: W.B.saunders Co. This figure is available in the (more...)

In sum, organogenesis of the testis involves successive differentiation of the Sertoli cell and the seminiferous tubules with envelopment of the extragonadally derived germ cells by Sertoli cells, development of the tunica albicans, appearance of Leydig cells, and differentiation of the mesonephric tubules into ductule efferentes, which connect the seminiferous tubules and network with the epididymis to provide the pathway for sperm transport at the ejaculatory duct system (Grumbach and Conte, 1998).

Organogenesis of the Ovaries

In the absence of testis-determining genes, the gonadal primordium has an inherent tendency to develop as an ovary, provided that germ cells are present and survive. The indifferent stage persists in the female fetus weeks after testis organogenesis begins. There is, however, continued proliferation of the coelomic epithelium and primordial germ cells, which gradually enlarge and become oogonia. Steroid biosynthesis by the fetal ovary is meager in early and midgestation and appears to arise from hilar interstitial cells in the ovarian primordium at about the 12th week of gestation. Both female and male human fetuses are bathed in estrogens of placental origin. The fetal ovary does not contribute significantly to circulating estrogens, which in the fetus are almost exclusively of placental origin, nor does it secrete AMH. The ovary has no documented role in differentiation of the female genital tract (Grumbach and Auchus, 1999).

At about the 11th to 12th week of gestation, long after differentiation of the testis in the male fetus, germ cells in the ovary begin to enter the meiotic prophase, which characterizes the transition of oogonia to oocytes and marks the onset of ovarian differentiation. The Wnt-4 gene, at least in the mouse, acts as a suppressor of the differentiation of steroidogenic cells in the fetal ovary.

Differentiation of the Genital Ducts

At the 7th week of intrauterine life, the fetus is equipped with both male and female genital ducts derived from the mesonephros. Müllerian ducts serve as the analog of the uterus and fallopian tubes, whereas the mesonephric or wolffian ducts have the potential to differentiate further into the epididymis, vas deferens (or ejaculatory ducts), and seminal vesicles. During the 3rd fetal month, either the müllerian or the wolffian ducts complete their development and involution occurs simultaneously in the opposite structure.

More than 50 years ago Alfred Jost, the French developmental endocrinologist, demonstrated that secretions from the fetal testis played a decisive role in determining the direction of genital duct development. In the presence of functional testes, the müllerian structures involute and undergo programmed cell death and the wolffian ducts complete their development; in the absence of testes, the wolffian ducts do not develop and müllerian structures differentiate. The regression of the müllerian ducts and the stabilization and differentiation of wolffian ducts are mediated by different secretions by the fetal testis: a glycoprotein, AMH, secreted by the fetal Sertoli cells and a sex steroid, testosterone, synthesized by fetal Leydig cells.

Female development is not contingent on the presence of an ovary because development of the uterus and tubes occurs if no gonad is present. However, the müllerian ducts fail to differentiate in the absence of the mesonephric ducts, which serve as the analog for both the male urogenital tract and the metanephros (primordial kidney). The influence of the fetal testis on duct development is exerted locally and unilaterally; if one testis is removed in an early stage of development, the oviduct develops normally on that side but müllerian duct regression occurs on the side of the intact testis. It is through the secretion of AMH by the fetal Sertoli cells that apoptosis of the müllerian ducts is induced, which leads to their degeneration.

Although müllerian duct involution is not androgen dependent, the differentiation of primitive wolffian ducts into the epididymis, vas deferens, and seminal vesicles requires testosterone and the androgen receptor. Differentiation of the wolffian ducts to form the epididymis, vas deferens, and seminal vesicles is testosterone dependent (the wolffian ducts lack the enzyme 5α-reductase type 2, which converts testosterone to a more potent androgen, dihydrotestosterone) (Wilson et al., 1981). Experimental data and studies with humans with steroid 5α-reductase type 2 deficiency provide additional evidence that testosterone, not dihydrotestosterone, mediates the differentiation of the wolffian ducts. This is in striking contrast to the dihydrotestosterone-dependent differentiation of the urogenital sinus and genital tubercles, which express steroid 5α-reductase 2 even before the testis has developed the capacity to synthesize testosterone. Thus, testosterone leads to the development of the internal genitalia and dihydrotestosterone leads to the development of the external genitalia (see Figures 3–1, 3–2, and 3–3).

In patients with ambiguous genitalia, male genital ducts are well differentiated only in those who have testes. Females with congenital adrenal hyperplasia do not display wolffian duct differentiation, even though their external genitalia may be highly virilized in utero. It is the critical role of the testes in male duct development to provide high local concentrations of testosterone. Male duct development is therefore deficient, even though testes may be present, in patients with severe defects in steroid biosynthesis and in XY patients whose tissues are unresponsive to testosterone (Grumbach and Conte, 1998).

Differentiation of External Genital and Urogenital Sinus

At the 8th fetal week the external genitalia of both sexes are identical and have the capacity to differentiate in either direction. They consist of the urogenital slit bounded by periurethral folds and more laterally by labioscrotal swellings. The urogenital slit is surrounded by genital tubercles consisting of corpora cavernosa and glans. The mucosa-lined urethral folds may remain separate, in which case they are called labia minora, or they may fuse to form a corpus spongiosum enclosing a phallic urethra. The fleshy labioscrotal swellings may remain separate to form labia major a, or they may fuse in the midline to form the scrotum and the ventral epidermal covering of the penis. The distinction between the clitoris and penis is based primarily on size and whether or not the labia minora fuse to form a corpus spongiosum.

By the 50-mm crown-rump stage, male and female fetuses can be distinguished by inspection of the external genitalia; in the male, the urethral folds have fused completely in the midline to form the cavernous urethra and corpus spongiosa by the 12th to 14th weeks of gestation. Penile length in the male increases linearly at about 0.7 millimeter/week from the 10th week to normal term. A 12-fold increase occurs from 0.3 centimeter (cm) at 12 weeks to 3.5 cm at term, a rate of growth about 3.5 times that of the clitoris.

The urogenital sinus separates from a common cloaca in early fetal life. It is thought that the müllerian duct contributes to the upper part of the vagina. There is disagreement about the relative contribution of the müllerian ducts and the urogenital sinus to the vagina, but the contact and interaction of the fused müllerian ducts with the urogenital sinus are essential for normal development of the vagina. In female development, proliferation of the vesicovaginal septum pushes the vaginal orifice posteriorally so that it acquires a separate external opening; thus, no urogenital sinus as such is preserved. In male development, the vaginal pouch is usually obliterated when the müllerian ducts are reabsorbed, although a vestigial blind vaginal pouch known as the prostatic utricle can sometimes be demonstrated. The prostate gland and the urethral glands of Cowper in the male are outgrowths of the urogenital sinus, in which male differentiation is mediated by dihydrotestosterone and requires the presence of androgen receptors (Grumbach and Conte, 1998).

The induction of differentiation of the external genitalia and urogenital sinus in males is affected by dihydrotestosterone, the 5α-reductase-reduced metabolite of testosterone (Wilson, 1999). Testosterone is a prohormone, and it is delivered through the bloodstream to these target tissues, which are rich in the enzyme 5α-reductase type 2 and which can readily convert testosterone to dihydrotestosterone even before the fetal testes acquire the capacity to secrete testosterone. Dihydrotestosterone binds to the androgen receptor and initiates the events that lead to androgen action.

As in the case of genital ducts, there is an inherent tendency for the external genitalia and urogenital sinus to feminize in the absence of fetal gonadal secretions. Complete differentiation of the external genitalia and urogenital sinus in males occurs only if the androgen stimulus is received during the critical period of development. Dihydrotestosterone stimulates growth of the urogenital tubercle and induces fusion of the urethral folds and labial fold swelling during this critical period; it also induces differentiation of the prostate and inhibits growth of the vesicle vaginal septum, thereby preventing the development of the vagina (Griffin et al., 1995; Grumbach and Conte, 1998). Androgen stimulation however, can cause clitoral hypertrophy at any time during the fetal life or after birth in the female.

Table 3–2 provides some examples of variations in sexual differentiation.

TABLE 3–2. Selected Examples of Variations in Sexual Differentiation.


Selected Examples of Variations in Sexual Differentiation.


Puberty is the transitional period between the juvenile state and adulthood during which the adolescent growth spurt occurs, secondary sexual characteristics appear (resulting in the striking sexual dimorphism of mature individuals), fertility is achieved, and profound psychological changes take place. Puberty tends to be regarded as a set of physical changes arising from reactivation of the hypothalamic-pituitary-gonadotropin-gonadal apparatus (the feedback system integrating nervous and hormonal signals in the hypothalamus). These changes can be timed and measured. On the other hand, adolescence is a more general and gradual coming of age that transpires during most of the second decade of life. Physiological and hormonal processes are involved in many aspects of this growth and development, with the onset of puberty a benchmark of the passage from childhood to adolescence.

Puberty is not a de novo event but rather is a phase in the continuum of development of the hypothalamic-pituitary-gonadal function from fetal life through puberty to the attainment of full sexual maturation and fertility (Grumbach and Styne, 1998). Endocrine events recognized as adolescent puberty actually begin early in fetal life. The hypothalamic-pituitary-gonadotropin-gonadal system differentiates in function during fetal life and early infancy, is suppressed to a low level of activity during childhood (the juvenile pause), and is reactivated at puberty (Grumbach and Kaplan, 1990; Grumbach and Styne, 1998). As mentioned earlier, a significant sex difference in fetal pituitary gonadotropin levels and the high circulating testosterone levels in the male fetus through the 24th week of gestation are the most prominent features of the hypothalamicpituitary-gonadotropin-gonadal system. There is no evidence that the concentrations of estradiol or other estrogens in serum differ in male and female fetuses.

Within a few minutes after birth, the concentration of LH in serum increases abruptly (about 10-fold) in the peripheral blood of the male newborn but not in that of the female newborn. This short-lived surge in LH release is followed by an increase in the serum testosterone level during the first 3 hours that persists for 12 hours or more. In the female neonate, LH levels do not increase, and FSH levels in both males and females are low in the first few days of neonatal life. After the fall in circulating placental steroid levels, especially estrogens, during the first few days after birth, serum FSH and LH levels increase and exhibit a pulsatile pattern with wide perturbations for several months. The FSH pulse amplitude is greater in female infants, and the FSH response to hypothalamic luteinizing hormone-releasing hormone (LHRH) or gonadotropin-releasing hormone is higher in females than males throughout childhood; LH pulses are higher in males. A sex difference in plasma FSA and LH values is also present in anorchid boys and agonadal girls less than three years old.

The high gonadotropin concentrations in infancy are associated with a transient second wave of differentiation of fetal-type Leydig cells and increased serum testosterone levels in male infants for the first 6 months or so and with elevated estradiol levels intermittently in the first 1 to 2 years of life in females. The mean FSH level is higher in females than males during the first few years of life. By approximately 6 to 8 months of age in the male and 2 to 3 years of age in the female, plasma gonadotropin levels decrease to low values until the onset of puberty. Thus, the restraint of the hypothalamic LHRH pulse generator and the suppression of pulsatile LHRH secretion (and thus FSH and LH release) attain the prepubertal level of quiescence in late infancy or early childhood and earlier in boys than in girls (for reviews see Grumbach and Styne [1998] and Grumbach and Gluckman [1994]).

The juvenile pause (that interval between early childhood and the peripuberty period when the LHRH pulse generator is at a low level of activity and circulating pituitary gonadotropin levels are low) is not associated with complete suppression of pituitary gonadotropin-gonadal function. Some studies have used highly sensitive immunoassays to show that both prepubertal boys and prepubertal girls have a pulsatile pattern of serum LH and FSH concentrations, with higher concentrations during the night than during the day (see Mitamura et al., 1999, 2000). The pulses are of very low amplitude compared with the increase in the pulse amplitude that occurs with the approach of puberty. There is apparently no change (or only a modest one) in pulse frequency with the onset of puberty (Mitamura et al., 1999, 2000).

A striking sex difference has been detected in prepubertal children by a highly sensitive immunoassay for estradiol in serum. Prepubertal girls have a mean estradiol concentration of 0.6 picograms per milliliter (pg/ ml), whereas the mean value in prepubertal boys is 0.08 pg/ml (Klein et al., 1994). During prepuberty in both sexes, serum testosterone concentrations are detectable, but at a very low level. The higher concentration of estradiol in prepubertal girls is associated with about a 20 percent advancement in bone age and may be a factor in the earlier onset of puberty in girls. For example, a bone age of about 11 years in girls is the equivalent of a bone age of 13 years in boys.

In addition, striking sex differences exist in the gonadally synthesized glycoprotein hormone inhibins throughout development in boys and girls (Andersson et al., 1997; Sehested et al., 2000). Inhibin B concentrations are strikingly elevated in males for the first 2 years of life and show a striking increase from childhood levels to adult levels at the onset of puberty, whereas levels of inhibin B are low or undetectable in prepubertal girls, followed by a sharp increase through midpuberty and then a decline.

Data on the normal variations in pubertal development in the United States are becoming more plentiful but are still incomplete. In recent years striking ethnic differences in the time of onset of puberty have been detected for girls but not for boys (Biro et al., 1995; Herman-Giddens et al., 1997). In girls, two distinct phenomena occur in the development of secondary sex characteristics. The development of breasts is under the control of estrogen secreted by the ovaries; the growth of pubic and axillary hair is under the influence of androgen secreted by the adrenal cortex and the ovary. Most recent data suggest that the mean age of onset of breast development in Caucasian girls is 10.6 years, with limits between 7 and 13 years, and that in African-American girls is 9.5 years, with limits between 6 and 13 years. The onset of breast development in African-American girls is about 1 year earlier than that in Caucasian girls, even though the average age of menarche in a large cross-sectional study was different by only 0.7 year (12.2 years for African-American girls and 12.9 years for Caucasian girls) (Herman-Giddens et al., 1997).

A careful review of U.S. studies of the onset of puberty in girls in the past three decades suggests that any change in the mean age of onset of breast development in Caucasian American girls is probably no more than 1 year earlier, if in fact a decrease in the mean age has actually occurred (Grumbach and Styne, 1998; Herman-Giddens et al., 1997). The age of menarche, a well-recognized landmark of pubertal development in girls, has not changed over the past four decades (Eveleth and Tanner, 1990). In African-American girls the mean age of onset of breast development apparently is 1 year earlier; while ethnic differences in fat mass maybe a factor (Kumanyika, 1998), the nature of the discordance is uncertain. In girls (as will be discussed below) the onset of puberty, in retrospect, is marked by an increase in the growth rate even before breast development.

The beginning of pubertal onset in boys is marked by an increase in the size of the testes, which occurs in both white and African-American boys at a mean age of about 11 years (Biro et al., 1995), with the normal limits being between 9 and 13.5 years. It is well established that the changes in the levels of sex steroid and gonadotropin secretion may precede or anticipate for some years the onset of physical changes of puberty. The actual dimorphic physical changes of puberty are primarily the consequence of testosterone secretion by the Leydig cells in boys and of estrogen secretion by the granulosa cells in girls (Grumbach and Styne, 1998).


Leptin, a hormone produced by adipose tissue, appears to have an important permissive action in the progression into puberty and the maintenance of normal secondary sex characteristics through its effect on hypothalamic-pituitary-gonadotropin-gonadal function (Clement et al., 1998; Farooqi et al., 1999; Strobel et al., 1998). The leptin concentration in serum correlates with body mass index or percent body fat and even more highly with the absolute amount of adipose tissue. There is a striking sexual dimorphism in the circulating concentration of leptin at birth, at which time females have higher levels than males, and again in late puberty and adulthood. A sexual dimorphism in circulating leptin concentrations has not been detected during childhood, however (Horlick et al., 2000), with the rise in leptin concentrations occurring 1 year earlier in girls than in boys. The levels in boys peaked at Tanner stage 2 and decreased by Tanner stage 5. In contrast, in girls, leptin levels increased in breast stage 2 and peaked at breast stage 5 (Blum et al., 1997; Clayton et al., 1997; Horlick et al., 2000). The decreased leptin levels in late puberty in boys have been attributed to the action of testosterone

Pubertal Growth Spurt

One of the most striking sex differences in puberty is the earlier age of onset of the pubertal growth spurt and the earlier attainment of peak height velocity in girls, in contrast to the later onset of the increased rate of growth and greater peak height velocity in boys. Prepubertal height and growth velocities are similar in boys and girls. Boys reach peak height velocity approximately 2 years later than girls and are taller at the beginning of the pubertal growth spurt.

In contrast to girls, in whom the increase in height velocity is probably the earliest sign of pubertal maturation, in boys, peak height velocity does not occur until genital stage 3 or 4 of puberty (Boxes 3–1 and 3–2). The mean height difference of 12.5 cm between adult men and women results partly from the greater pregrowth spurt growth in boys, partly from the height difference at the age of takeoff (boys being taller at their later age of the beginning of the pubertal growth spurt), and partly from the greater gain in height of boys during the pubertal growth spurt (Figure 3–5).

Box Icon

BOX 3–1

Sex Differences in the Relationship of Onset of Pubertal Growth Spurt to Sexual Maturation in Girls and Boys.

Box Icon

BOX 3–2

Sex Differences in the Timing of the Onset of Estrogen Synthesis in Girls and Boys.

Figure Icon


Permission was not granted to electronically reproduce figure 3–5 from In: Williams Textbook of Endocrinology, 9th ed. J.D.Wilson, D.W.Foster, H.M.Kronenberg, and P.R.Larsen, eds. Philadelphia: W.B.saunders Co. This figure is available in the (more...)

The hormonal control of the pubertal growth spurt is complex. Growth hormone, insulin-like growth factor 1, and triiodothyronine are the principal regulators of prepubertal growth and regulate about 50 percent of the growth during the pubertal period; superimposed on this growth is the linear growth induced by estradiol in both boys and girls. Although the role of estradiol in the pubertal growth spurt in girls has been appreciated for more than 20 years, only now do new observations indicate that estradiol is the major sex steroid responsible for the pubertal growth spurt in boys as well as girls (reviewed in Grumbach [2000] and Grumbach and Auchus [1999]).

In boys, the estradiol is derived mainly from the extragonadal conversion of testosterone to estradiol in a wide variety of tissues, but there is also a small testicular contribution (Siiteri and MacDonald, 1973). Furthermore, estradiol, but not testosterone, appears to be the critical mediator of skeletal maturation and epiphyseal fusion and the major sex steroid in bone mineral accrual in boys as well as girls (Grumbach, 2000; and Grumbach and Auchus, 1999). This conceptual sea change has emanated from studies of men, women, and children with mutations in the gene encoding aromatase (Bilezikian et al., 1998; Grumbach and Auchus, 1999; Morishima et al., 1995) and from studies of one man with a null mutation in the gene encoding the estrogen receptor a (Smith et al., 1994).

There is a very striking and poorly understood difference in the prevalence of so-called idiopathic true or central precocious puberty in boys and girls. The idiopathic form is about 10 times more common in girls than in boys. In contrast to the striking sex difference in idiopathic true precocious puberty, constitutional delay in growth in adolescents (idiopathic delayed puberty) is more common in boys than in girls.

Adrenarche Versus Gonadarche

In both boys and girls, beginning before age 8 (skeletal age, 6 to 8 years), an increase in the levels of secretion of adrenal androgens and androgen precursors, called “adrenarche,” occurs. It is marked biochemically by progressive increases in plasma dehydroepiandrosterone and dehydroepiandrosterone sulfate (DHEAS) concentrations. The mechanism of activation of adrenal androgen secretion or adrenarche is independent of the mechanisms that regulate the onset of sex steroid secretion by the gonads, which is called “gonadarche” (Grumbach and Styne, 1998).

Premature adrenarche, which is more common in girls than in boys, is characterized by the precocious appearance of pubic hair or axillary hair, less commonly an apocrine odor, and comedones and acne without other signs of puberty or virilization (Grumbach and Styne, 1998). Adrenarche is premature when it occurs in Caucasian girls before age 7 or African-American girls before age 5. In boys the diagnosis is limited to those who develop pubic hair or axillary hair before the age of 9. In contrast to boys, in whom premature adrenarche is usually a benign, self-limited normal variant of puberty, girls with premature adrenarche are at increased risk (about 10-fold) for the development of insulin resistance and ovarian hyperandrogenism, in particular, polycystic ovary syndrome (PCOS) (Dunaif et al., 1992; Ibáñez et al., 1996; Morales et al., 1996; Oppenheimer et al., 1995). PCOS affects about 5 to 10 percent of women of reproductive age and is the most common endocrine disorder in women. A proportion of girls with exaggerated adrenarche (Likitmaskul et al., 1995), which is marked by higher levels of circulating DHEAS, may exhibit insulin resistance and hyperinsulinism or dyslipidemia, and beginning in late adolescence they are at increased risk for the development of ovarian hyperandrogenism and anovulation (Ibáñez et al., 1998, 1999a). Affected girls, however, usually begin gonadarche within the normal range of time.

There is a correlation between the occurrence of exaggerated adrenarche in prepubertal girls and a higher risk for ovarian hyperandrogenism at puberty. The androgen excess is a consequence of PCOS and is associated with an increased risk of metabolic complications, including type II diabetes mellitus, hypertension, dyslipidemia, and possibly, cardiovascular disease. There is a tendency for familial aggregation of women with PCOS, and evidence suggests that this heterogeneous disorder represents a complex, multifactorial trait.

A promising area of research is the increasing body of evidence that supports an association of premature adrenarche, insulin resistance, and dyslipidemia in girls with intrauterine growth retardation (Francois and de Zegher, 1997; Ibáñez et al, 1999b). Recent studies (Ibáñez et al., 2000) suggest that girls with prenatal growth restriction have at birth a smaller complement of primordial follicles than infants of appropriate weight for gestational age and have at adolescence a small uterus and ovaries. After menarche these girls tend to show ovarian hyporesponsiveness to FSH.

In sum, the evidence suggests a link between intrauterine growth retardation and the increased risk of exaggerated adrenarche followed by PCOS, including hyperandrogenism, insulin resistance, dyslipidemia (with or without obesity), and cardiovascular disease (Barker, 1995, 1997; Cresswell et al., 1997). As first advanced by Barker (1995) from observational studies, the association of impaired or disproportionate fetal growth, related to fetal undernutrition, with premature adrenarche and PCOS is another example of disorders in adolescence and adulthood that may be programmed in fetal life. Many issues, however, remain unresolved (Jaquet et al., 1999).

Sex Differences in Behavior

The hormonal and physical changes at puberty described above have implications for sex differences in behavior in early adolescence. Some behavioral changes probably result from the direct effects of gonadal hormones acting directly on the brain. For example, in early adolescence, increasing testosterone levels in boys have been associated with increasing aggression and social dominance, and changes in estrogen levels in girls have been associated with mood changes (Brooks-Gunn et al., 1994; Buchanan et al., 1992; Finkelstein et al., 1997; Olweus et al., 1980; Schaal et al., 1996; Susman et al., 1987). Some behaviors appear to relate to the absolute level of the hormone, others appear to relate to the ratio of hormone levels (for example, the testosterone level/estradiol level ratio), and others appear to relate to hormonal variation. (These associations are not always noted and probably depend on the reliability of hormone level measurements, intersubject variation, and the specific behaviors measured.) It is important to note that hormone levels themselves can be changed by behavior. For example, winning an athletic event has been shown to increase testosterone levels in males (Booth et al., 1989).

As discussed later, the rise in estrogen levels at puberty may contribute to females' superior phonological skills and may allow females with dyslexia to compensate for their reading deficiencies. There is also some suggestion that other cognitive changes in adolescence are related to hormonally induced maturation of the frontal lobe (Spear, 2000). Given the sex difference in the timing of gonadarche, there may well be sex differences in the developmental timing of behaviors subserved by the frontal lobe (including planning and judgment), although probably not in the ultimate levels of those behaviors at maturity.

The hormonal and physical changes that occur during puberty also contribute in indirect ways to differences between adolescent boys and adolescent girls. For example, the development of secondary sex characteristics in girls creates social signals that result in different responses from peers, parents, and teachers. There is a substantial literature showing that girls who mature earlier than their peers are at greater risk than girls who mature on time or later than their peers for problems during the pubertal transition and continuing into adulthood. These problems include substance use, depression, and eating disorders (Caspi and Moffitt, 1991; Graber et al., 1997). Some of these effects are mediated by the fact that girls who mature early are more likely than others to associate with older adolescents and to be treated as if they are older (including increased responsibilities from parents and increased expectations of parents). Boys who mature earlier than their male peers do not have a similarly increased risk of problems compared with the risk for boys who mature on time or later, in part because the absolute age of boys who mature early is, on average, 2 years later than that of girls who mature early and because their physical maturation gives them status among adolescents, who value athleticism and physical skill in boys.

Adolescence is associated with changing social roles, and there is good reason to believe that gender socialization intensifies at that time of life (Crouter et al., 1995; Hill and Lynch, 1983; Stein, 1976). These changes in social roles will have wide-ranging effects, including, for example, variations in interpretations of harmful stimuli and responses to injury, as discussed below.


During the long period of about 40 years of fertile adulthood, an individual's occupation(s), social roles, and lifestyle change episodically and develop slowly as experiences accumulate. Although societal norms are rapidly changing, in general it remains the case that women still predominate as caregivers and organizers with wide-ranging obligations and duties spanning the family, workplace, and leisure realms, whereas men still predominate in more focused aggressive and physically demanding activities with a relatively narrower range of social obligations. Accompanying these developing and highly individual psychosocial characteristics are the more consistent (but not constant) sex differences in anatomy, organ function (physiology), and endocrine function. Thus, on average, women, relative to men, have a higher percentage of body fat, smaller muscle mass, lower blood pressure, higher levels of estrogens and progestins, and lower levels of androgen. The challenge in understanding the significance of this vast array of sex differences for health and health care lies not so much in assessing the influence of each of these factors in isolation but, rather, in deciphering how the factors interact throughout the course of adulthood to affect each individual at any moment.

In addition, women, but not men, undergo fluctuations associated with the reproductive condition (such as the ovarian cycle and pregnancy) that influence numerous bodily functions (e.g., gastrointestinal transit time, urinary creatinine clearance, liver enzyme function, and thermoregulation), including brain function. (The effects of pregnancy, lactation, and parity are obviously important to the health of women later in their lives but are not addressed specifically in this report.)

Effects of Menopause on Women's Health

After the fertile years in women there is a 5- to 10-year period of menopause-related alterations in hormone patterns, terminating in the sharp decline in female hormone levels. As follicle depletion occurs in the ovaries, the rate of ovarian hormone production slows. The tissues most affected by reduced estrogen levels are the ovaries, uterus, vagina, breast, and urinary tract. However, other tissues such as the hypothalamus, skin, cardiovascular tissue, and bone are also substantially affected. A major challenge to the prevention of disease in older women lies in exploring the effects of both short-term and long-term reductions in ovarian hormone levels on the development of symptoms and disease.

The lack of ovarian estrogens appears to contribute significantly to the onset of several postmenopausal diseases, such as osteoporosis and cardiovascular disease, two leading causes of morbidity and mortality in older women. Much of the evidence to support the finding of a cardioprotective effect for estrogen has come from observational studies of women on estrogen replacement therapy, which has shown that estrogen users experience half as many cardiovascular events as nonusers, but numerous questions remain (also discussed in Chapter 5). An adverse influence of hormone therapy on cardiovascular risk in women with coronary heart disease has been shown during the initial year of use; however, few data are available on the effects of long-term hormone therapy (Grodstein et al., 2000). The protection conferred by estrogen has been shown to be mediated by mechanisms acting at different levels, including a beneficial effect of estrogen on plasma lipid concentrations (Lamon-Fava, 2000).

In addition, research has identified estrogen receptors in bone (reviewed in Grumbach and Auchus, 1999; Khosla et al., 1999; Prestwood et al., 2000). Declines in estrogen production correlate with rapid bone loss, which predisposes a woman to osteoporosis. Although age-related bone loss is a universal phenomenon shared by men and women, the effect of osteoporosis on women is much more profound and pervasive. Several reports have shown that combining high-calcium supplements with a regimen of hormone therapy increases the efficacy of estrogen in bone conservation. Hormone replacement by estrogen therapy or the newly developed therapy with selective estrogen receptor modulators may prevent the development of osteoporosis and its related fractures (Kamel et al., 2001).

In men, more inconsistent and complex changes in hormone metabolism, called “andropause” by some, occur over a longer period of time, on average, between the ages of 48 and 70 (Morales et al., 2000; Vermeulen, 2000). Currently, much more is known about the consequences of menopause than about those of andropause. Androgen deficiency has been shown to be associated with osteoporosis. Although testosterone replacement therapy in hypogonadal men decreases bone resorption and increases bone mass, placebo-controlled trials are needed to better define the effectiveness and risks of such therapy in older men. The effect of testosterone is, at least in part, related to its conversion in bone (Bilezikian et al., 1998; Grumbach, 2000; Khosla et al., 1998; Smith et al., 1994) to estradiol.

The Aging Human

The evolving effects of interactions between an individual's personal physiology and unique experiences make it difficult to assess how simply being female or male affects that individual's health as life progresses from birth through fertile adulthood into old age. Although the field of gerontology is growing rapidly, research from this perspective is meager. Most studies simply address the question of how elderly individuals differ from younger individuals, with little attention paid to how the differences might develop over time. Nevertheless, it is evident that the patterns of sex differences that exist during the long period of fertile adulthood change during old age in clinically relevant ways. For example, community prevalence estimates for chronic widespread pain and fibromyalgia show a general increase with age until about age 65, followed by a decrease, with the prevalence in women always being higher (LeResche, 1999). On the other hand, the prevalence of pain in the knee or finger joints shows a continual increase across the life span for both sexes, with no sex differences until age 50, after which the prevalence becomes higher in women (LeResche, 1999). A second example—in this case, one relevant to diagnosis—is that the symptom presentation of patients with confirmed acute myocardial infarction varies by sex, but, importantly, the pattern changes with age (Goldberg et al., 2000). Younger patients (less than age 55) were significantly more likely than older patients to complain of sweating and arm pain.

A third example—in this case, relevant to treatment—involves recent data showing sex- and age-related differences in the optimal effects of antihypertensive and antiplatelet therapies for the prevention of cardiovascular disease (Kjeldsen et al., 2000). For example, compared with treatment with a placebo, daily acetylsalicylic acid (ASA) treatment resulted in a significant reduction in the rate of occurrence of composite major cardiovascular events in younger patients (younger than age 65). Some reduction in the rate of occurrence of major cardiovascular events was also seen in ASA-treated older patients, but the reduction was not statistically significant.

From 1900 through about 1940, Americans who lived to age 65 had a life expectancy of another 11 or 12 years, regardless of sex. Since the 1940s differences in life expectancies between males and females after age 65 have emerged, and these differences favor females. Similarly, in 1900 individuals who reached the age of 85 had, on average, another 4 years of life, with very little difference between the sexes. Differences in survival began to appear in the 1960s, and these again favored females. Much of this difference can be attributed to differences in rates of death from cardiovascular disease. A breakdown by sex and age (65 to 74 years, 75 to 84 years, and 85 years and older) reveals that in each age group men have higher death rates from both heart disease and cancer than women (National Center for Health Statistics, 1999). Death rates from stroke, another leading cause of death, are more balanced between males and females.

Currently, life expectancy at birth is greater for females than males by ~6 years, but once old age has been attained, it becomes 2 to 3 years (Table 3–3). The actual life expectancy differs among ethnic groups and is, for example, notably shorter among African American than Caucasian Americans, but the consistency in the observation of an advantage for females across ethnic groups is striking (Figure 3–6). This consistent observation of greater life expectancy at birth for females has grown over time, from about 2 years in 1900 to 6 years in 1998 (Figure 3–7), with some fluctuations during the interim.

TABLE 3–3. Life Expectancy at Birth, Age 65, and Age 75 Years, United States, All Races, 1998.


Life Expectancy at Birth, Age 65, and Age 75 Years, United States, All Races, 1998.

FIGURE 3–6. Life expectancy at birth for males and females in several U.


Life expectancy at birth for males and females in several U.S. ethnic groups (data are from 1989 to 1994). Source: National Center for Health Statistics (1996) and National Institutes of Health, Office of Research on Women's Health (1998).

FIGURE 3–7. Life expectancy at birth for males and females, selected years between 1900 and 1998, United States, all races.


Life expectancy at birth for males and females, selected years between 1900 and 1998, United States, all races. Source: National Center for Health Statistics (2000a).

Although the mechanisms that underlie both the general increase in longevity and the increasing advantage for females are poorly understood, some components of the male longevity disadvantage can be identified. For example, rates of death from the major causes (both intentional and unintentional injuries and illnesses) are usually higher for males than for females at each stage of life (Leveille et al., 2000), although there are exceptions in which the rates are similar for the two sexes or are higher for females.

Stress and its hormonal consequences are complex factors that may contribute to longevity. A recent provocative suggestion is that the behavioral response to stress may differ between males and females (Taylor et al., 2000). According to this hypothesis males display the classic “fight-or-flight” response that in females is modulated to become a “tend-and-defend” response. On the basis of a meta-analysis (integrating the data from a number of independent studies), the female's response is apparently mediated by oxytocin, a hormone known to reduce stress and increase social affiliation in rodents (Carter et al., 1995; Witt et al., 1990). This proposed difference in response may have implications for the sex differences in stress-related disorders in human populations and may contribute to the longer life span of females.

The female longevity advantage, however, is not without cost. Although females live longer, those who do live longer experience more disabling health problems than males. Thus, a recent study of 10,263 older adults in three communities in the United States showed that the proportion of disabled women increased from 22 percent at age 70 to 81 percent at age 90, whereas the figures for men were 15 and 57 percent, respectively (Leveille et al., 2000). Similarly, another study showed that although the current life expectancy at age 32 is 39.45 years for men and 44.83 years for women, it becomes 31.8 years for men and 33.1 years for women when life expectancy is adjusted for “quality of life.” In other words, a 5.38-year advantage for women is reduced to 1.3 years (Kaplan and Erickson, 2000).

Clearly, studies that address the mechanisms that underlie the development of these differences as individuals age could yield important information of benefit to both sexes.



Sex differences occur throughout the life span, although their specific expression varies at different at stages of life.

  • Intrauterine environment: Some sex differences originate in events that begin in the womb, where developmental processes differentially organize tissues for later activation..
  • Early development: During the prepubertal period there are behavioral as well as subtle hormonal sex differences.
  • Puberty: During sexual maturation, hormones activate organ systems differently between males and females; these include brain anatomy and functions that were previously organized by hormones and modified by the environment.
  • Adulthood: Throughout the life span, including midlife and old age, the brain, as well as many other organs, retain plasticity and continue to be modified by gene expression, hormones, and environmental factors. These factors act in an integrative manner but may be expressed differently in males and females.

To continue to advance human health and health care, research on sex differences in health and illness across the life span is essential. Such research can be aided by information obtained from the observation and study of other species. In this regard, the committee makes the following recommendations.


RECOMMENDATION 2: Study sex differences from womb to tomb.

The committee recommends that researchers and those who fund research focus on the following areas:

  • inclusion of sex as a variable in basic research designs,
  • expansion of studies to reveal the mechanisms of intrauterine effects, and
  • encouragement of studies at different stages of the life span to determine how sex differences influence health, illness, and longevity.

Sex is an important marker of individual variability. Some of this sex-related variability results from events that occur in the intrauterine environment but that do not materialize until later in life. Current research varies in its level of attention to these matters.

The committee acknowledges that inclusion of people, animals, or cells and tissues of both sexes in all studies is not feasible or appropriate. Rather, the committee is urging researchers to regard sex, that is, being male or female, as an important basic human variable that should be considered when designing, analyzing, and reporting findings from studies in all areas and at all levels of biomedical research. Statistical methods can be used to evaluate the effect of sex without necessarily doubling the sample size of every study. In addition, it is particularly important that researchers revisit and revise approaches to studying whole-animal physiology in light of what has been learned in the past decade about major sex differences.

RECOMMENDATION 3: Mine cross-species information.

  • Researchers should choose models that mirror human sex differences and that are appropriate for the human conditions being addressed. Given the interspecies variation, the mechanisms of sex differences in nonhuman primates may be the best mimics for some mechanisms of sex differences in humans. Continued development of appropriate animal models, including those involving nonhuman primates, should be encouraged and supported under existing regulations and guidelines (see the Guide for the Care and Use of Laboratory Animals [National Research Council, 1996]).
  • Researchers should be alert to unexpected phenotypic sex differences resulting from the production of genetically modified animals.

Sex differences and their relevance to human health can be examined through the use of cross-species comparisons. The use of appropriate animal models can reveal underlying mechanisms of normal and pathogenic processes.

Copyright 2001 by the National Academy of Sciences. All rights reserved.
Bookshelf ID: NBK222286


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...