Beginning with the 1968 Olympic Games and continuing through 1998, the International Olympic Committee (IOC) required on-site gender verification of female athletes via laboratory-based genetic testing. The rationale was to detect male impostors who would have an unfair competitive advantage based on superior size, strength, and speed associated with androgen-enhanced skeletal muscle mass. Additionally, the granting of eligibility certificates served to silence public innuendoes about the sexual identity of competitors. This policy filtered down to non-Olympic competitions, with no guarantee of quality control.
Concerns about the scientific appropriateness of sex chromatin testing (and later polymerase chain reaction [PCR]-based methods) for gender verification were often expressed throughout this time period (2,4). In fact, sex chromatin analysis had been discarded as a common diagnostic tool by geneticists by the 1970s (15). Several medical and professional societies endorsed resolutions or adopted policy statements in the 1980s and 1990s calling for the elimination of gender verification, including the American Medical Association (AMA), the American Academy of Pediatrics, the American College of Physicians, the Endocrine Society, the American Society of Human Genetics, and the Canadian and Australian genetic societies (4).
In 1998, the IOC’s Athlete’s Commission called for discontinuation of the laboratory-based system for mandatory gender verification of female athletes. In the summer of 1999, the IOC conditionally rescinded its 30-year requirement for on-site gender screening of all women entered in female-only events at the Olympic Games, starting with the Sydney Games in the year 2000 (7,16). Rather, intervention and evaluation of individual athletes by appropriate medical personnel could be employed if there was any question about gender identity.
NORMAL SEXUAL DIFFERENTIATION
Sexual differentiation begins with the chromosomal sex assignment, the heterogamete (XY) being male and the homogamete (XX) being female (8,18). Embryos of both sexes develop similarly for approximately the first 40 days of gestation, at which point the uncommitted gonad begins differentiating into a testis or ovary. Presence of a Y chromosome triggers testis development. Differentiation is initiated by actions of the SRY gene, a single gene on the short arm of the Y chromosome; however, several other genes are necessary for normal testicular development and regression of the female primordia (17). It is unclear whether there are analogous “ovarian-determining” genes, or whether ovarian development is a default pathway in the absence of testicular development.
The male or female urogenital tracts and external genitalia are formed from the wolffian and müllerian ducts, respectively, based on the type of gonad formed and hormones secreted from the fetal gonads. Formation of the male phenotype is based on the action of two hormones secreted by the fetal testis, anti-müllerian hormone and testosterone. After secretion, testosterone undergoes further conversion to dihydrotestosterone, which promotes development of the male urethra, prostate, penis, and scrotum through actions on typical androgen receptors. In the absence of the testes (or androgen secretion), the female phenotype emerges. Normally, phenotypic sex conforms to chromosomal sex, but there are several conditions in which this is not the case. Disorders of sexual development may be chromosomal, gonadal, or phenotypic in origin.
Disorders of chromosomal sex.
These occur when the number or structure of the X- or Y-chromosomes is abnormal as in (rare) true hermaphroditism (46,XX or 46,XY or mosaics); Klinefelter syndrome (47,XXY or 46,XY/47,XXY); XX males (46,XX); Turner syndrome (45,X or 46,XX/45,X); and mixed gonadal dysgenesis (46,XY/45,X or 46,XY) (8,18). The latter is the second most common cause of ambiguous genitalia in the newborn. Mosaicism for a Y-bearing cell line is responsible for most instances. Affected individuals usually have a testis on one side and a streak gonad on the other. The phenotype varies depending on the proportion of XY cells and their distribution. A majority of these individuals are raised as females.
Disorders of gonadal sex.
These occur when chromosomal sex is normal but differentiation of the gonads is abnormal, resulting in conditions in which gonadal sex does not correspond to chromosomal sex (8,18).
Pure gonadal dysgenesis is a disorder in which phenotypic females have gonads and genitalia characteristic of gonadal dysgenesis (i.e., bilateral streaks, infantile uterus and fallopian tubes, and sexual infantilism), but they attain normal height and have few if any somatic abnormalities. Either a 46,XX or 46,XY karyotype may be present. Individuals with XY gonadal dysgenesis have normal female internal and external genitalia that fail to mature at puberty because only thin streaks of ovarian tissues are present, which are unable to secrete sufficient estrogen to promote development of secondary sex characteristics. Both XX and XY varieties can result from single mutations that are presumed to involve genes essential for gonadal development. About 15% of 46,XY women have either a deletion or a mutation in the SRY coding sequence. Others could have mutations in SRY outside the coding sequence.
Disorders of phenotypic sex.
These disorders (pseudohermaphroditism) occur in 46,XX or 46,XY individuals with appropriate gonadal sex but in whom development of the urogenital tract is inappropriate for the chromosomal and/or gonadal sex (8,18). Female pseudohermaphroditism occurs in 46,XX women with bilateral ovaries but with variable virilization of the urogenital tract because of androgen excess during fetal life. This may be associated with congenital adrenal hyperplasia (most commonly 21-hydroxylase deficiency), developmental disorders of müllerian ducts, or other nonadrenal enzymatic deficiencies (see Fig. 1).
Male pseudohermaphroditism is caused by defective virilization of the 46,XY embryo and can result from defects in androgen synthesis, androgen receptors, or müllerian duct regression. Impaired testosterone synthesis accounts for approximately 20% of cases (see Fig. 1) (18). Defects in androgen action associated either with steroid 5-α-reductase deficiency or receptor disorders are most common. Individuals with androgen insensitivity syndrome are genetically male because they possess both an X and Y chromosome, but their tissues cannot respond to androgens and they develop phenotypically as (sterile) women. The syndrome is caused by mutations of the androgen receptor gene carried on the X chromosome. These XY females are taller than average women. The uterus is absent and the vagina is only one-third normal length. Male internal genital ducts remain undeveloped. Normal-sized testes are usually found in the pelvis or at the inguinal ring. Most commonly, androgen insensitivity is complete, but partial sensitivity may be present in about 10% of patients. These individuals exhibit normal patterns of pubic hair and minor virilization of the external genitalia, physiologic changes that offer no competitive advantage.
GENDER VERIFICATION OF WOMEN ATHLETES
The practice of mandatory gender verification of women athletes arose in part because three world champion athletes who competed as women in the 1930s and 1940s (and a World Cup champion skier from the 1960s) subsequently underwent sex reassignment surgery to become males. One individual (Herman Ratjen) who competed for Germany in the women’s high jump in the 1936 Berlin Games (as Dora) later admitted he was a male impostor forced into this role by Nazi Germany. A world-class “female” runner at 400 and 800 meters in the 1960s also was later discovered to be male (12). Additionally, an autopsy report noted a mosaic sex chromosome pattern and ambiguous sex organs in the 1932 Olympic women’s 100-meter sprint champion.
In the 1950s and early 1960s, questions regarding the femininity of highly successful but “masculine” female track and field athletes from the Soviet Union and Eastern Bloc persisted. Coupled with the increasing popularity of women’s sports and striking improvement in athletic achievements by women, efforts were made to ensure that women competing at international events were in fact women and that “athletes were competing on an equal basis considering their physical status” (9). Both performance-enhancing drugs and sex impostors were to be prevented by on-site testing.
Gender verification was accomplished before international competitions in 1966 and 1967 by physical inspection and/or direct gynecologic examination, a practice that was soon replaced by laboratory-based genetic tests. The IOC officially mandated gender verification for female athletes preceding competition in IOC-sanctioned events beginning in 1968 and continuing through 1998 with laboratory-based genetic tests. From 1968 until 1992, buccal smears were analyzed for the presence of sex chromatin (i.e., inactive X or Barr bodies and fluorescent Y-body chromatin material). The Barr body is formed after inactivation of one of the two X chromosomes in female cells. Many other international and national competitions also adopted this model (9).
Concerns about the appropriateness of sex chromatin testing for gender verification were voiced continuously in the 1970s and 1980s but had little impact on IOC governance (2). These tests detected athletes who were feminine but who happened to have an XY chromosomal pattern, including male pseudohermaphrodites with complete or partial androgen insensitivity and patients with variants of XY gonadal dysgenesis. Sex chromatin tests screened out phenotypic women who had genetic differences that afforded no unusual physical advantage for sports, while potentially missing XX men and women with medical conditions such as congenital adrenal hyperplasia that could offer competitive advantages. At least 13 women were excluded from athletic competition between 1972 and 1990 using sex chromatin testing; many others with abnormal sex chromatin tests “retired” or opted to forgo further assessment to avoid public scrutiny (5,11).
Opposing efforts were limited in impact, in part by a lack of information on the frequency of positive results, diagnoses, and follow-up. The situation began changing with the celebrated case of the Spanish national champion hurdler Maria Patino, who was disqualified from competing at the 1985 World University Games. Ms. Patino had male pseudohermaphroditism with complete androgen insensitivity. She was the first woman to publicly protest her disqualification and was eventually reinstated (1). In 1990, the International Amateur Athletics Federation (IAAF) called for an end to required genetic screening of female athletes and in 1992 adopted an approach designed to prevent only male impostors from competing (6,10). The IAAF recommended that the “medical delegate” shall have the ultimate authority in all medical matters, including the authority to arrange for the determination of the gender of the competitor if that approach is judged necessary (10).
The IOC failed to heed this example and beginning in 1992 replaced sex chromatin testing with DNA analysis for Y chromosome material using PCR amplification of chromosomal DNA extracted from nucleated cells (13,15). At the 1992 Summer Olympics in Barcelona, 2406 female athletes were screened for DNA located on the Y chromosome (DYZ1 repeat). Positive samples were reanalyzed for the presence of the SRY gene. Eleven were positive for DYZ1; of these, five were positive for SRY (14).
The 1996 Summer Games in Atlanta used a comprehensive process for screening, confirmation of testing, and counseling of affected individuals (3,4). Eight of 3387 female athletes had positive test results for SRY, 7 had partial or complete androgen insensitivity, and the other had undergone gonadectomy and was presumed to have 5-α-reductase deficiency. Two of the seven individuals with androgen insensitivity had not undergone gonadectomy. All of these female athletes were allowed to compete. Overall, the prevalence of male pseudohermaphroditism has been estimated to be 27 in 11,373 or 1 in 421 through five Olympic Games preceding Sydney, compared with an estimated incidence of 1:20,000 to 1:40,000 in the general population (4).
The shift to PCR-based techniques replaced one diagnostic genetic test with another but did not alleviate the problems. Positive results still stigmatized women with such conditions as androgen insensitivity, XY mosaicism, and 5-α-reductase deficiency. Both sex chromatin and SRY tests identify individuals with genetic anomalies that yield no competitive advantage (7,16,18). Therefore, finally in 1999, the IOC conditionally rescinded its 30-yr requirement for on-site gender screening of all women entered in female-only events at the Olympic Games, starting with Sydney in 2000. Rather, intervention and evaluation of individual athletes by appropriate medical personnel could be employed if there was any question about gender identity. This change has not been made permanent.
Gender verification has long been criticized by geneticists, endocrinologists, and others in the medical community (2). The combination of invalid screening tests, failure to understand the problems of intersex, the discriminatory singling out of women based only on laboratory results, and the stigmatization and emotional trauma experienced by individuals screened positive prompted organized objection among medical professionals toward gender verification in sports.
Genuine sex-impostors have not been uncovered by laboratory-based genetic testing; however, gender verification procedures have resulted in substantial harm to a number of women athletes born with relatively rare genetic abnormalities affecting gonadal development or the expression of secondary sexual characteristics (2). The application of this practice is discriminatory and based on a false assumption of unfair advantage. One major problem was unfairly excluding women who had a birth defect involving gonads and external genitalia (i.e., male pseudohermaphroditism) but who had partial or complete androgen insensitivity. Additionally, these tests fail to exclude all potential impostors (e.g., 46,XX males). Individuals with sex-related genetic abnormalities raised as females have no unfair physical advantage and should not be excluded or stigmatized, including those with 5-α-steroid-reductase deficiency, partial or complete androgen insensitivity, and chromosomal mosaicism.
No men posing as women have been detected at the Olympics or other international events at which X chromatin analysis or SRY testing has been performed; therefore, gender screening based on finding Y chromosomal material should be abandoned. Of interest, however, is the apparent high frequency of male pseudohermaphroditism in elite class female athletes. The presence of Y chromosomal material is associated with increased height in these athletes. Mean heights are 175 cm or about 69 inches, which is 10 cm taller than normal female controls and only 2 cm less than normal male controls. Although some assume that this anomaly may provide potential physical advantages, none of the characteristics, including increased height, are outside of the normal traits exhibited by genetically typical (46, XX) females.
Finally, the current practice of urine testing to exclude doping requires that voiding be observed by an official to verify that a sample from a given athlete has actually come from his or her urethra. Such a practice, performed according to uniform standards, would seem to obviate the possibility of male impostors successfully competing and winning. The policy advocated by the IAAF and conditionally adopted by the IOC protects rights and privacy for athletes while safeguarding fairness of competition and should become the permanent approach. Consequently, as recommended by the Council on Scientific Affairs, the following was adopted as AMA policy in December 2001: The AMA declares its opposition to the use of laboratory testing for genetic sex as a basis for verification of gender in athletic competition and urges the IOC to make permanent the 1999 conditional ban on this practice for all future competitions.
Members and staff of the Council on Scientific Affairs at the time this report was prepared: Roy D. Altman, M.D., Miami, FL (Chair); Scott D. Deitchman, M.D., M.P.H., Duluth, GA (Chair Elect); Myron Genel, M.D., New Haven, CN (Past Chair); Rebecca Gee, M.P.H., New York, NY; J. Chris Hawk III, M.D., Charleston, SC; Mohamed K. Khan, M.D., Ph.D., Ann Arbor, MI; Mary Anne McCaffree, M.D., Oklahoma City, OK; Carolyn B Robinowitz, M.D., Washington, DC; John F. Schneider, M.D., Ph.D., Chicago, IL; Melvyn L. Sterling, M.D., Orange, CA; Patricia L. Turner, M.D., Silver Springs, M.D.; Michael A. Williams, M.D., Baltimore, MD; Gary L. Woods, M.D., Concord, NH. Staff: Barry D. Dickinson, Ph.D. (Secretary); James M. Lyznicki, M.S., M.P.H. (Assistant Secretary); Marsha Meyer (Editor), Chicago, IL.
This report was presented as Report 3 of the Council on Scientific Affairs at the 2001 AMA Interim Meeting. The recommendation was adopted and the remainder of the report was filed.
Address for correspondence: Barry D. Dickinson, Ph.D., Secretary to the Council on Scientific Affairs, American Medical Association, 515 North State Street, Chicago, IL 60610; E-mail: firstname.lastname@example.org.
1. Carlson, A. When is a woman not a woman? Women’s Sport Fitness
March: 24–29, 1991.
2. de la Chapelle, A. The use and misuse of sex chromatin screening for “gender verification” of female athletes. JAMA 256: 1920–1923, 1986.
3. Elsas, L. J., R. P. Hayes, and K. Muralidharan. Gender verification at the Centennial Olympic Games. J. Med. Assoc. Ga. 86: 50–54, 1997.
4. Elsas, L. J., A. Ljunqvist, M. A. Ferguson-Smith, et al. Gender verification of female athletes. Genet. Med. 2: 249–254, 2000.
5. Ferguson-Smith, M. A., E. A. and Ferris. Gender verification in sport: the need for change? Br. J. Sports Med.
6. Ferris, E. A. Gender verification testing in sport. Br. Med. Bull. 48: 683–697, 1992.
8. Grumbach, M. M., and F. A. Conte. Disorders of sexual differentiation. In:Williams Textbook of Endocrinology,
9th Ed. J. D. Wilson (Ed.). Philadelphia: Saunders, 1998, pp. 1303–1425.
9. Hay, E. Sex determination in putative female athletes. JAMA 4: 39–41, 1972.
10. Ljungqvist, A, and J. Simpson. Medical examination for health of all athletes replacing the need for gender verification in international sport. JAMA 277: 850–852, 1992.
11. Puffer, J. C. Gender verification: a concept whose time has come and passed? Br. J. Sports Med. 30: 278, 1996.
12. Ryan, A. J. Sex and the singles player. Physician Sportsmed. 4: 39–41, 1976.
13. Serrat, A, and A. Garcia de Herreros. Determination of genetic sex by PCR amplification of Y-chromosome-specific sequences. Lancet 341: 1593, 1993.
14. Serrat, A., and A. Garcia de Herreros. Gender verification in sports by PCR amplification of SRY and DYZ1 Y chromosome specific sequence: presence of DYZ1 repeat in female athletes. Br. J. Sports Med. 30: 310–312, 1996.
15. Simpson, J. L., A. Ljungqvist, A. de la Chapelle, et al. Gender verification in competitive sports. Sports Med. 16: 305–315, 1993.
16. Simpson, J. L., A. Ljungqvist, M. A. Ferguson-Smith, et al. Gender verification in the Olympics. JAMA 284: 1568–1569, 2000.
17. Swain, A., and R. Lovell-Badge. Mammalian sex determination: a molecular drama. Genes Dev. 13: 755–767, 1999.
18. Wilson, J. D., J. E. and Griffin. Disorders of sexual differentiation. In:Harrison’s Principles of Internal Medicine
, 15th Ed., E. Braunwald, S. L. Hauser, A. S. Fauci, et al. (Eds). New York: McGraw-Hill, 2001, pp. 2172–2184.