Archives of in Childhood 1997;77:305–309 305

Androgen insensitivity syndrome: a survey of Arch Dis Child: first published as 10.1136/adc.77.4.305 on 1 October 1997. Downloaded from diagnostic procedures and management in the UK

R M Viner, Y Teoh, D M Williams, M N Patterson, I A Hughes

Abstract single gene disorder that results in a normal Objective—A two year survey of androgen female phenotype, and typically presents in insensitivity syndrome (AIS) to assess early adult life with primary amenorrhoea, current diagnostic and management although a significant number present with strategies. inguinal hernias in infancy. The significance of Methods—Cases were ascertained by in- the relative rarity of bilateral inguinal hernias in clusion on the British Paediatric Surveil- girls is often not appreciated; it is estimated lance Unit monthly report card for 24 that 1–2% of such infants have complete AIS.4 months. The diagnosis is confirmed in both forms of Results—Fifty one of 139 notifications AIS by a male karyotype and normal testoster- were confirmed as AIS; 29 cases were one production and metabolism, in the pres- complete AIS and 22 cases partial AIS. ence of normal testicular histology, and gener- Seventy six per cent of complete AIS pre- ally the absence of müllerian duct remnants. sented with an inguinal hernia, and half Partial AIS is a clinically heterogeneous dis- the complete AIS patients had an estab- order and presents a more complicated diag- lished family history of the disorder. Pres- nostic problem.25 The degree of androgen entation in the partial AIS group was unresponsiveness is variable, so the aVected through ambiguous or undermasculinised infant presents at birth with undermasculinised genitalia; 59% of partial AIS were raised external genitalia of varying severity. Charac- as male. teristically, there is micropenis with chordee, a Conclusions—The importance of karyo- bifid scrotum, and a perineal urethral opening, typing girls with inguinal hernias is con- and gonads that may or may not be palpable firmed, and further attention should be within the bifid scrotum. The partial AIS phe- given to genetic counselling for families of notype must be diVerentiated from other con- complete AIS patients. A large number of ditions with a 46XY karyoptype, which can cases were misreported as partial AIS, give rise to the same genital abnormality such emphasising the importance of undertak- as defects in testosterone biosynthesis, 5á- reductase deficiency, mixed gonadal dysgen-

ing a comprehensive diagnostic evaluation http://adc.bmj.com/ in intersex states. A large percentage of esis, and true hermaphroditism.67 Accurate children with partial AIS were raised as early diagnosis is important and has a profound boys despite severe genital undermascu- bearing on the sex of rearing, genetic counsel- linisation, indicating the current lack of ling, and subsequent management; a trial of validated measures that predict genital androgen treatment in early infancy to assess response to androgen treatment. The the growth of penile erectile tissue may provide management of AIS is discussed and diag- important information on androgen respon-

nostic guidelines provided to improve the siveness. There is a lack of knowledge regard- on October 2, 2021 by guest. Protected copyright. ing the criteria used to decide the management Department of diagnostic yield in AIS. Paediatrics, (Arch Dis Child 1997;77:305–309) of infants with partial AIS, particularly with University of respect to sex of rearing. Analysis of the cases Cambridge, Keywords: androgen insensitivity syndrome; inguinal identified through our survey suggests that Addenbrookes hernia; intersex guidelines for clinicians on the appropriate Hospital diagnostic and management strategies for AIS Y Teoh patients, particularly in the case of suspected D M Williams A two year national survey of the androgen partial AIS, would be helpful. M N Patterson insensitivity syndrome (AIS) was undertaken I A Hughes via the British Paediatric Surveillance Unit to University College determine existing diagnostic procedures and Methods London Hospitals, management strategies. A pilot study of SURVEY PROCEDURE Great Ormond Street patients with the partial form of AIS had shown Cases were ascertained by active surveillance Hospital, and that data on gonadal histology, anatomy of the via inclusion on the surveillance unit monthly Middlesex Hospital, internal genitalia, and steroid hormone con- report card for a two year period. Previous London 1 K M Viner centrations were incomplete. AIS defines a studies using multiple ascertainment methods female or ambiguous phenotype in a 46XY have suggested that this reporting procedure Correspondence to: male with testes and normal testosterone alone provides between 60–80% case ascer- Professor I A Hughes, production and metabolism.2 As androgens are tainment rates.8 Cases were defined as an Department of Paediatrics, Addenbrookes Hospital, essential for normal male sexual development infant or child under 16 years of age with a Level 8, Box 116, Hills Road, and fertility, defects in androgen action are 46XY karyotype and with either normal female Cambridge CB2 2QQ. associated with abnormal sexual differentiation external genitalia (defined as complete AIS) or Accepted 10 June 1997 and fertility.3 Complete AIS is a rare X linked abnormal external genitalia that are consistent 306 Viner, Teoh,Williams, Patterson, Hughes

with the partial AIS phenotype, that is, esection. Also, as part of a separate study, if a micropenis/clitoromegaly, bifid scrotum, peri- partial AIS patient was undergoing surgery to Arch Dis Child: first published as 10.1136/adc.77.4.305 on 1 October 1997. Downloaded from neoscrotal hypospadias. Clinicians were asked the external genitalia, clinicians were asked to to report any child under their care who satis- provide (after parental consent) a genital skin fied the case definitions, including those newly biopsy specimen as part of the surgical diagnosed in the past month, and those procedure in order to establish fibroblast children diagnosed before the start of the study cultures. Systematic case ascertainment was and who were still 16 years of age, or less, at the confined to the surveillance unit reporting time of the study. scheme. Ethical approval was granted by the Once a case was notified, the referring clini- Cambridge local ethics committee. cian was asked to complete a questionnaire detailing further information about the clinical Results phenotype, family history, imaging studies of Table 1 details notifications during the survey internal genitalia, gonadal histology, results of period. Fifty one of the total 139 notifications endocrine investigations, and management were confirmed as cases of AIS. There was decisions on sex of rearing and hormone treat- insuYcient information to make a diagnosis ment. Determination of diagnosis was made from 37 notifications. The diagnostic criteria from information provided in consultation with for complete AIS and partial AIS were fulfilled referring clinicians. If a case of AIS was in 29 and 22 cases, respectively. confirmed as part of separate studies, referring clinicians were asked to collect blood samples COMPLETE AIS from the index case and family members for All complete AIS patients had normal female DNA extraction, the venesection being coordi- genitalia and were reared as females. Presenta- nated with other investigations requiring ven- tion in complete AIS was predominantly by the discovery of a hernia in an apparently female Table 1 Cases notified to surveillance unit (n=139) infant (76%); other cases were identified on the basis of a family history of the X linked disor- Total notifications No der, or from the determination of a karyotype Exclusions (n=63) for other reasons (table 2). Regardless of the Over age at study 21 Duplicate notifications 5 mode of presentation, inguinal hernia was Unconfirmed cases—no return of questionnaire 37 present in 28 (90%) of complete AIS patients; Cases reported and questionnaire received (n=76) nine (31%) were unilateral, 17 (59%) bilateral. Confirmed complete AIS 29 Confirmed partial AIS 22 Table 3 shows details of gonadal position and Total AIS 51 histology. One or two palpable testes were Other diagnoses: present in approximately 80% of complete AIS Abnormal or dysplastic testes 6 6 patients. Gonadal histology was consistent with Biosynthetic defect (17-hydroxydehydrogenase 1 AIS in all patients in whom histological deficiency) examination of the testes was performed (table Drash syndrome 1 Hypogonadism 1 3). Apart from the report of a ‘vestigial’ uterus Hypopituitarism 1 in one complete AIS patient, female internal http://adc.bmj.com/ Hypospadias 2 Smith-Lemli-Opitz syndrome 2 genitalia were absent in all AIS patients. WAGR syndrome 1 Unknown diagnosis 4 PARTIAL AIS Drash=Wilms’ tumour, male pseudohermaphroditism, and Presentation in partial AIS was through nephropathy. ambiguous or undermasculinised genitalia. In 59% of the partial AIS cases, the sex of rearing Table 2 Mode of presentation in CAIS (n=29) was male. Factors that appeared to aVect sex of rearing in partial AIS were the degree of labio- on October 2, 2021 by guest. Protected copyright. No (%) of cases scrotal fusion, site of urethral opening, whether Bilateral hernia 13 (45) the testes were palpable, and phallus size. All Unilateral hernia 9 (31) Family history 4 (14) infants with unfused labia (n=4) were raised as Amniocentesis 1 (3) female and a further five infants with fused Adoption karyotype 1 (3) labia were also raised as female. All those raised Unknown l (3) as male (n=13) had fused labioscrotal folds. Micropenis was reported in 54% of infants Table 3 Details of gonad position and histology; values are number (%) raised as males, although equal numbers were reported as having normal phallus size. App- Partial AIS roximately 80% of partial AIS patients had one Complete AIS Reared male Reared female or two palpable testes. Infants with palpable (n=29) (n=13) (=9) testes at diagnosis were more likely to be raised Gonad position male, and only one infant without palpable tes- Bilateral intra-abdominal 6 (21) 1 (8) 3 (33) tes was raised as a male (table 3). Bilateral inguinoscrotal 14 (48) 10 (77) 6 (67) One inguinal + one intra-abdominal 9 (31) 2 (15) 0 Unknown 0 0 0 FAMILY HISTORY Gonad histology Table 4 summarises information concerning Bilateral normal testes 12 (41) 1 (8) 7 (78) One normal testis* 8 (28) 0 0 family history. Approximately half the cases of Abnormal 0 0 0 complete AIS had an established family history Not done 0 8 (62) 0 of the disorder. There were four pairs of related Unknown 9 (31) 4 (30) 2 (22) complete AIS patients, and in four cases a his- *Unilaterally normal testis—other testis not biopsied or not identified. tory suggestive of androgen insensitivity was Androgen insensitivity syndrome 307

Table 4 Details of family history; values are number (%) Table 5 Timing of gonadectomy and sex of rearing Arch Dis Child: first published as 10.1136/adc.77.4.305 on 1 October 1997. Downloaded from Complete AIS Partial AIS No (%) (n=29) (n=22) Complete AIS (timing of gonadectomy) Established complete AIS in Prepubertal 14 (48) family 12 (41) 0 Postpubertal 1 (3) Family history suggestive of AIS* 4 (14) 6 (27) Planned postpubertally 6 (21) No family history of AIS 13 (45) 15 (68) Unknown timing 8 (28) Unknown 0 1 (5) Not done 0 Partial AIS reared as male *This includes first and second degree female relatives with pri- Gonadectomy performed 0 mary amenorrhoea or lack of pubic hair, and male relatives with No gonadectomy 13 (100) undermasculinisation. Unknown 0 Partial AIS reared as female Gonadectomy performed 8 (88) present. In contrast, only one quarter of partial No gonadectomy 0 AIS cases had a family history suggestive of Unknown 1 (12) AIS. Associated congenital anomalies were absent in the complete AIS group, except for one patient with red-green colour blindness. Five complete AIS patients (17%), but no par- Associated anomalies were more common in tial AIS patients, raised as female were being the partial AIS group, with one example each given oestrogen replacement therapy at the of coarctation of the aorta, microphthalmos, time of the survey. Three partial AIS patients ectodermal , short stature with mental raised as male were receiving androgen replace- retardation, and chromosomal translocation ment therapy. Genital surgery other than gona- 10q:16, q26.2:q21. dectomy was performed in all the partial AIS females (vulvovaginoplasty), and in nearly half of partial AIS males; this included hypospadias

GONADECTOMY repair, orchidopexy, and release of penile chor- Table 5 shows details of when gonadectomy dee. was performed. This took place before puberty in half the complete AIS patients, whereas Discussion about one quarter had, or were due to have, This study provides an overview of the charac- removal of gonads after puberty. For the teristics and management of AIS in a large remainder, plans were unknown at the time of group of patients in the UK. The survey did the survey. Median age for gonadectomy in not provide incidence or prevalence figures for complete AIS was 8 years (mean 5.5 years), AIS because of many unconfirmed reported with a median prepubertal gonadectomy age of cases, the limitations of case ascertainment 1 year (mean 1.7 years). Gonadectomy was using a single reporting method, the incom- performed in all partial AIS patients raised as plete lifetime ascertainment of complete AIS females, and in none of those raised as males. due to the normality of the female phenotype, Median age for gonadectomy in partial AIS and confidentiality problems associated with

patients was 21 months (mean 24.1 months). reporting disorders of sexual development. http://adc.bmj.com/

Peripheral karyotype

Pelvic ultrasound on October 2, 2021 by guest. Protected copyright. Blood for: LHRH test: HCG stimulation Surgical samples DNA extraction Luteinising hormone, test Serum AMH FSH, testosterone

Short term: Long term: Genital skin: Gonadal biopsy 1500 U daily 1500 U twice weekly Fibroblast culture Histology for 3 days for 3 weeks Androgen binding Immuno- 5α-reductase assay histochemistry Plasma samples DNA, RNA extraction before/after HCG: 17-OHP Androstenedione DHA Testosterone DHT SHBG 24 hour urine: Androgen metabolites Figure 1 An investigation protocol for a patient with a partial AIS phenotype. A similar protocol can be used for complete AIS, although a human chorionic gonadotrophin test may not be necessary. AMH=antimüllerian hormone; DHA=dehydroepiandrosterone; DHT=dihydrotestosterone; SHBG=sex hormone binding globulin. 308 Viner, Teoh,Williams, Patterson, Hughes

The presentation of three quarters of com- plete AIS patients through inguinal hernias, Arch Dis Child: first published as 10.1136/adc.77.4.305 on 1 October 1997. Downloaded from nearly half of which were unilateral, emphasises Key messages the importance of considering AIS in any x A karyotype is essential infants with female infants with hernias. Estimates of the suspected hernias incidence of AIS in such infants have ranged x Further attention to genetic counselling from 1–12%,49suggesting that any girl with an is needed for complete AIS patients to inguinal hernia should have a karyotype improve case ascertainment in subse- performed. quent generations Nearly half the patients had a proved family x Management of suspected partial AIS history of complete AIS, yet this was the mode should involve a specific sequence of of presentation in only 14% of cases, suggest- clinical, radiological, biochemical, and ing that more attention be given to genetic molecular investigations (fig 1) counselling for the families of complete AIS patients. There were a number of examples when an aVected sibling had previously Establishing a diagnosis of complete AIS is undergone a herniorrhaphy but the underlying straightforward with few other conditions pro- diagnosis had not been recognised. Karyotyp- ducing a similar phenotype. However, Leydig’s ing female infants with first or second degree cell hypoplasia can be mistaken for complete relatives with complete AIS should be consid- AIS and may also present with ambiguous ered by clinicians. Information about the man- genitalia or only isolated micropenis. Loss of agement of complete AIS is mainly limited to function mutations in the luteinising hormone cases presenting before puberty as the survey receptor gene have recently been reported to was restricted to age 16 years and below. The cause this syndrome.17 Thus, in the absence of results indicate a trend towards prepubertal an increased basal testosterone, a human gonadectomy in complete AIS, the majority of chorionic gonadotrophin stimulation test is cases having surgery in infancy (median age 12 indicated to verify normal androgen produc- months). However, oestrogen replacement tion in complete AIS. therapy had not been started in late childhood The diagnosis and management of partial or early adolescence in several cases, raising AIS is surrounded by inaccuracy and confu- concerns about an increased risk of osteoporo- sion. Many cases are interpreted as resulting sis in later life.10 The risk of malignant transfor- from dysgenetic testes despite the lack of infor- mation of the gonads in adult life is well mation on gonadal histology. Since the partial documented,11 but medical opinion remains AIS phenotype may also be the expression of divided about whether the testes should remain several other disorders, it is important to inves- in place until after feminisation has occurred at tigate cases as thoroughly as possible. Pelvic puberty.2 ultrasound assessment, to identify the presence Cases of partial AIS presented as expected or absence of müllerian structures, is particu- because of abnormal genitalia at birth. Condi- larly important as it can be done quickly, in tions misreported initially as partial AIS contrast with some biochemical tests where the http://adc.bmj.com/ included testicular dysgenesis and examples of results may be delayed. Figure 1 illustrates a abnormal genitalia forming a component of suggested flowchart for the investigation of known syndromes such as Denys-Drash,12 partial AIS phenotype. Smith-Lemli-Opitz,13 and Wilms’ tumour, ani- Some form of androgen insensitivity can ridia, genital anomalies, and mental retardation reasonably be assumed when an XY male has (WAGR).14 Other cases were associated with normal testes as based directly on histological specific defects in androgen production. These examination and indirectly on the absence of findings emphasise that the diVerentiation of female internal genitalia (normal antimüllerian on October 2, 2021 by guest. Protected copyright. partial AIS from other causes of undermascu- hormone and receptor),18 and evidence of linisation requires the demonstration of normal adequate Leydig cell function based on the testicular histology, adequate testosterone pro- androgen response to human chorionic gona- duction after human chorionic gonadotrophin dotrophin stimulation. In these circumstances, stimulation, and normal androgen metabolism. it is reasonable to assume that the partial AIS Sex assignment in partial AIS was, not surpris- phenotype may be associated with a dysfunc- ingly, aVected by the size of the phallus and the tional androgen receptor.23 Even so, in many presence of palpable testes. However, a surpris- isolated cases of the partial AIS phenotype, ingly high percentage of partial AIS cases were androgen binding studies and mutational raised as male even though there was severe analysis of the androgen receptor gene are undermasculinisation of the external genitalia. normal.19 The sex hormone binding globulin response to Further androgen related genetic mecha- androgens has been advocated as a simple in nisms involved in male sex diVerentiation and vivo marker of a functional defect in the andro- development remain to be defined. Detailed gen receptor in partial AIS patients.15 Such a studies of intersex conditions can play a funda- marker if validated in relation to functional mental part in identifying genes involved in sex androgen receptor assays and clinical outcome, determination and diVerentiation.20 The clini- would be useful to guide sex of rearing cal investigation of intersex patients such as decisions. The phenotypic expression of partial those with AIS, the development of a national AIS can be so variable within aVected families database of AIS patients, and continuing that the severity of undermasculinisation may research into the molecular genetics of these be unpredictable.16 disorders is necessary for progress in the man- Androgen insensitivity syndrome 309

agement of infants and children with intersex 9 Pergament E, Heimler A, Shah P. Testicular feminization

and inguinal hernia. Lancet 1973;ii:740-1. Arch Dis Child: first published as 10.1136/adc.77.4.305 on 1 October 1997. Downloaded from conditions. 10 VanGelderen H. Skeletal maturation in the XY female syn- drome. Clin Genet 1986;30:199-201. This survey was made possible as a result of the Surveillance 11 Verp M, Simpson J. Abnormal sexual diVerentiation and Unit of the Royal College of Paediatrics and Child Health (for- neoplasia. Cancer Genet Cytogenet 1987;25:191-218. merly the British Paediatric Association). We are grateful to unit 12 Mueller R. The Denys-Drash syndrome. J Med Genet 1994; staV for advice and support in the organisation of the study. We 31:471-7. thank all paediatricians throughout the UK for their time and 13 McGaughran J, Donnai D, Clayton P. Diagnosis of trouble in notifying intersex conditions to the unit and provid- Smith-Lemli-Opitz syndrome. N Engl J Med 1994;330: ing further information for analysis. The study was supported by 107-13. the Wellcome Trust. The work of Mrs Norma Coggins as infor- 14 Clarkson PA, Davies HR, Williams DM, et al. Mutational mation coordinator is gratefully appreciated. screening of the Wilms’ tumour gene, WT1, in males with genital abnormalities. J Med Genet 1993;30:767-72. 1 Williams DM, Evans BAJ, Hughes IA. A clinical and 15 Sinnecker GHG, Hiort O, Nitsche EM, et al. Functional biochemical analysis of 68 patients with the partial andro- assessment and clinical classification of androgen insensi- gen insensitivity syndrome (PAIS). Horm Res 1990;3(suppl tivity in patients with mutations of the androgen receptor 3):54. gene. Eur J Pediatr 1997;156:7-14. 2 Patterson MN, McPhaul MJ, Hughes IA. Androgen insensi- 16 Evans BAJ, Hughes IA, Bevan CL, MN Patterson, JW Gre- tivity syndrome. Ballière’s Clin Endocrinol Metab 1994;8: gory. Extreme phenotypic diversity in two siblings with the 379-404. partial form of the androgen insensitivity syndrome. Arch 3GriYn JE. Androgen resistance—the clinical and molecular Dis Child 1997;176:529-31. spectrum. N Engl J Med 1992;326:611-8. 17 Kremer H, Kraaij R, Toledo S, Post M, Fridman J, Hayash- 4 Jagiello G, Atwell J. Prevalence of testicular feminization. ida C. Male pseudohermaphroditism due to a homozygous Lancet 1962;i:329. missense mutation of the luteinizing hormone receptor 5 Quigley C, DeBellis A, Marschke K, et al. Androgen recep- gene. Nature Genet 1995;9:160-4. tor defects: historical, clinical and molecular perspectives. 18 Rey NJR, Lordereau-Richard I, Cate R. Anti-müllerian hor- Endocr Rev 1995;6:271-321. mone and testicular development clinical and hormonal 6 Conte FA, Grumbach MM. Pathophysiology, genetics, nosology and diagnosis of male pseudohermaphroditism. aspects. In: Dufau M, Fabbri A, eds. Cell and molecular biol- In: Hughes IA, ed. Sex diVerentiation: clinical and biological ogy of the testis. New York: Raven Press, 1994:7-15. aspects. Frontiers in . Rome: Serono Symposia, 19 McPhaul MJ, Marcelli M, Zoppi S, GriYn JE, Wilson JD. 1996;20:153-72. The spectrum of mutations in the androgen receptor gene 7 Berkovtiz G. Abnormalities of gonadal determination and that causes androgen resistance. J Clin Endocrinol Metab diVerentiation. Semin Perinatol 1992;16:289-98. 1993;76:17-23. 8 British Paediatric Surveillance Unit. Eighth annual report, 20 Shafer AJ, Goodfellow PN. Sex determination in humans. 1993. London: BPSU, 1994. Bioessays 1996;18:955-63. http://adc.bmj.com/ on October 2, 2021 by guest. Protected copyright.