SEARCH

SEARCH BY CITATION

Keywords:

  • androgen theory of autism;
  • cancer;
  • infantile autism;
  • testosterone

Abstract

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. ACKNOWLEDGMENT
  7. REFERENCES

Aim:  The objective of the current study was to compare the prevalence of three testosterone-related cancer diseases in the mothers of 111 individuals diagnosed as children with infantile autism (IA) with a matched comparison group of mothers of 330 children from the general population.

Method:  All mothers were screened through the nationwide Danish National Hospital Register. We inquired about breast-, uterine-, and ovarian cancer diseases during an observation period of 27 years.

Results:  At follow up a similar proportion of case and control mothers had a diagnosis of any cancer disease: 6.3% vs 8.5%. In addition, no single cancer disease was significantly more frequent among mothers of children with IA.

Conclusion:  Our study provides no support for an (eventually androgen-mediated) increased rate of cancer in mothers of persons with IA.

INFANTILE AUTISM (IA) is a heterogeneous neurodevelopmental disorder that is characterized by deficits in socialization and communication and by the presence of repetitive and stereotyped patterns of behaviors and interests. In ICD-10,1 IA/childhood autism is classified as one of a group of related disorders collectively termed pervasive developmental disorders or autism spectrum disorders (ASD), of which it forms the main prototype.

The etiology of ASD is poorly understood. In a recent review covering 34 epidemiological surveys, 30 studies reported male : female ratios among children with ASD.2 The male : female ratio varied from 1.3 to 16.0, with a mean male : female ratio of 4:1. The constant preponderance of male individuals has yet to be explained, however understanding the basis for the male : female ratio may be an important step towards understanding the etiology of the disorder.3

In a series of papers, Baron-Cohen and colleagues have advanced a novel theory known as the extreme male brain theory of autism and the androgen theory of autism. It suggests that the behaviors seen in ASD are an exaggeration of typical sex differences and that exposure to high levels of prenatal testosterone could be involved in the sex differences in social behavior in the general population and in the male vulnerability to ASD.4 In many respects, Baron-Cohen's extreme male brain theory makes sense and has generated much new interdisciplinary research.5 Both the androgen theory and the extreme male brain theory predict that women with ASD might manifest physical masculinization, and be more vulnerable to conditions associated with elevated levels of androgens.6 In addition, there is some evidence for elevated fetal testosterone levels in parents of children with ASD: like their children, they have lower 2D : 4D (second and fourth digit) ratios (assumed to be a proxy for fetal testosterone) than expected,7 suggesting skewed hormone levels in children with ASD and their parents.

In a recent study6 The Testosterone Medical Questionnaire was completed by 54 women with ASD (50 had Asperger syndrome), 74 mothers of children with ASD, and a comparison group of 183 mothers of typically developing children to test whether women with ASD have an increased rate of testosterone-related medical conditions and behavioral traits, and to see whether mothers of children with ASD and other close relatives show similar abnormalities, as part of the broader autism phenotype. They found an increased rate of medical conditions and behavioral traits conventionally seen as associated with elevated androgen levels in women with ASD. Compared to the controls, significantly more mothers of children with ASD reported, among other diseases, a history of breast cancer and uterine cancers, and a family history of ovarian and uterine cancers. These results suggest current hormone deviances in women with ASD and their mothers and in other close relatives of individuals with ASD. In conclusion, Ingudomnukul et al.6 call upon further studies to eventually replicate these findings via clinical examination and medical records checks.

In the following we are reporting the results of a study comparing the rates of selected testosterone-related cancer diseases in the mothers of 111 individuals diagnosed as children with IA and the mothers of 330 control children from the general population, using data from the nationwide Danish National Hospital Register (DNHR) covering an observation period of 27 years.

METHODS

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. ACKNOWLEDGMENT
  7. REFERENCES

Subjects

The sample was taken from the patient population attending the Departments of Child Psychiatry in the university hospitals of Copenhagen and Aarhus. The clinics provided services to the entire population of Demark. Case records of all children seen between 1960 and 1984 given a diagnosis of ‘childhood psychosis’– the ICD-88 term for IA or a ‘borderline condition’, were rediagnosed in 1985 in accordance with ICD-9 criteria.9 Based on this rediagnosis, 118 children (85 boys and 33 girls) with a diagnosis of IA were identified.10 The mothers of these children make up the case group in this study. However, seven case mothers were excluded because information was not available. Two mothers had died before the DNHR was established 1 January 1977. One mother died due to pancreas cancer in 1970, and another mother died in a car accident in 1962. Another two mothers had lived abroad during the entire observation period and one mother emigrated in 1975. The mothers of two adopted children were also unknown.

A comparison group was drawn from the Danish Central Persons Register.11 Wherever possible, each autistic child was matched with three control children by sex, time of birth, place of birth (region), and social group. The mothers of these children make up the comparison group in our study. The final study sample therefore included 111 case mothers and 330 control mothers.

At index admission in childhood, the mean age of the IA group (82 boys and 29 girls) was 5.4 years (SD 2.5). All but one medical record included information on intelligence (IQ): 48 (43%) subjects had an IQ < 50; 30 (27%) had an IQ between 50 and 69; and 32 (29%) had an IQ > 69.

All mothers were screened through the DNHR,12 using the citizen's identity number, which ensures a definitive identification. This enables the histories of the two groups to be followed with respect to in-patient admissions (day or night care) to medical hospitals since January 1977 and out–patient consultations at medical hospitals since January 1995, within the 27-year period 1977–2003. Mean age at follow up was 65.6 years (median 64.5 years; range 46.5–90.9 years). In the comparison group the respective figure is 65.6 years (median 64.9 years; range 46.3–88.8 years).

For individuals registered in the DNHR, we ascertained the place of admission, the dates of admission and discharge, together with primary and secondary discharge diagnoses. The register provides diagnoses according to ICD-88 and from 1 January 1994 according to ICD-10.1 In order to simplify data presentation all diagnoses are listed as ICD-10 diagnoses.

In earlier studies, higher levels of testosterone have been associated with the development of breast cancer,13 ovarian steroid tumors14 and the most common form of uterine cancer, endometrial cancer.15 People in the study population were classified with a history of a cancer disease if they had been admitted to a hospital or had been in out-patient care in a hospital-based clinic with one of the three following diagnoses before 31 December 2003: breast cancer (ICD-8 code: 174.xx and ICD-10 code C50.x); uterine cancer (ICD-8 codes: 180.xx, 182.xx and ICD-10 codes C53.x, C54.x and C55.x); and ovarian cancer (ICD-8 code: 183.xx and ICD-10 code C56.x). These diseases were chosen because they were all included in the study of Ingudomnukul et al.6 and are well-defined conditions. Patients with the above-mentioned diagnoses would probably always be admitted for in-patient care.

Statistical analysis

Data handling, statistical analysis and calculation of odds ratios were performed using STATISTIX for Windows.16 Comparisons between groups were based on the two-sample t-test for continuously scaled data and Fisher's Exact Test for dichotomous data. A probability level <0.05 (two-tailed test) was used to indicate significant differences between groups.

Ethics

The Danish Data Inspectorate approved the study protocol.

RESULTS

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. ACKNOWLEDGMENT
  7. REFERENCES

Maternal testosterone-related cancer diseases

The frequency of the three studied cancer diseases in mothers of children with IA and control children diagnosed any time during the 27-year observation period is given in Table 1.

Table 1.  Frequency of breast-, uterine-, and ovarian cancer in mothers of children with infantile autism (IA) and control children
 IA (n = 111)Controls (n = 330)POR95% CI
  • *

    ICD-101 diagnoses.

Any testosterone-related cancer disease (%)*7 (6.3)28 (8.5)0.550.730.31–1.71
Breast cancer (C50.x)4 (3.6)18 (5.5)0.470.650.21–1.96
Uterine cancer (C53.x–C55.x)2 (1.8)4 (1.2)1.001.500.27–8.28
Ovarian cancer (C56.x)1 (0.9)6 (1.8)0.690.490.06–4.12

In total, seven case mothers and 28 control mothers were known in the DNHR with one of the three studied cancer diseases (7/111 [6.3%] vs 28/330 [8.5%], P = 0.55, OR = 0.73, 95%CI 0.31–1.71). In addition, no single cancer disease was significantly associated with IA.

Gender and testosterone-related cancer diseases

Among the seven case mothers diagnosed with a cancer disease, six mothers had given birth to a male child with IA and one to a female child with IA (6/82 [7.3%] vs 1/29 [3.4%]). That is, mothers with a male child had a higher frequency of cancer than mothers with a female child. However, the difference is not statistically significant (P = 0.67, OR = 2.21, 95% CI 0.26–19.19). In the comparison group the respective figures are (21/243 [8.4%] vs 7/87 [8.1%], P = 1.00, OR = 1.08, 95% CI 0.44–2.64). Thus the gender of the child was not significantly associated with having a cancer diagnosis in either group.

Age at diagnosis of breast cancer

The mean ages at initial breast cancer diagnosis in the DNHR for the IA and comparison group were 55.2 years (median 51.6 years, range 41.3–76.2 years) and 56.9 years (median 56.0 years, range 34.8–77.9 years), respectively. The difference is not statistically significant (two-sample t-test with equal variance: t = −0.25, d.f. = 20, P = 0.80).

If we dichotomized maternal age at initial breast cancer diagnosis under and above 55 years of age (stipulated menopause), then 50% in both groups were diagnosed before menopause and 50% after menopause. If menopause is stipulated at 50 years of age the figures remain unchanged in the IA case group, while only five mothers in the comparison group had received a breast cancer diagnosis at 50 years of age, and 13 mothers at an older age.

DISCUSSION

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. ACKNOWLEDGMENT
  7. REFERENCES

In a recent pioneer study, Ingudomnukul et al.6 found an increased rate of a number of medical conditions and behavioral traits conventionally seen as associated with elevated androgen levels in women with ASD. In addition, mothers of children with ASD and other close relatives had some of the same abnormalities. These findings are consistent with the androgen theory of autism, suggesting that the studied conditions and ASD share a common risk factor, presumably elevated testosterone levels. In contrast to this study we were unable to find significantly increased rates of three studied testosterone-related cancer diseases in mothers of children with IA. In addition, the proportion of case mothers diagnosed with one of the three cancer diseases (breast cancer [3.6%], uterine cancer [1.8%] and ovarian cancer [0.9%]) are all lower than in the study by Ingudomnukul et al.6

Concerning breast cancer, it has been suggested13 that a high testosterone level is protective against breast cancer before menopause (offsetting the proliferative effects of estrogen), but that post menopause testosterone may increase risk through aromatization to estrogen. If true, this would suggest that only postmenopausal cases of breast cancer would be linked to high testosterone (and thus relevant to the androgen hypothesis of ASD). Premenopausal cases might actually be less frequent in mothers of children with ASD. Our results do not support this hypothesis.

Methodological considerations

Our results are based on hospital register data, covering an entire national population, and not assessed face-to-face. It is a major advantage of the current study that bias in recalling life events was avoided as all data were collected routinely, independently of the study. Furthermore, using register data gives a complete follow up on all cases and controls without the attrition that may limit ‘personal’ follow-up studies. In the questionnaire study by Ingudomnukul et al.6 the response rate was 50.3% for ASD case mothers. The authors called attention to the possibility that individuals who were affected by the studied conditions were probably more likely to fill out the questionnaire than those who were unaffected, resulting in an overestimation of the studied diseases. A limitation of our study is that the reliability of the medical data is unknown. However, a validation study dealing with a number of diagnoses in the DHNR found validity ranging from 75% to 90%.17

In conclusion, studies of testosterone-related cancer diseases in mothers of children with ASD have so far shown inconsistent results. Additional studies with larger sample sizes of different diagnostic subgroups, comparison groups matched for age and observation period, and systematically collected data on maternal disease status will be required to determine whether there is a true association between a maternal history of testosterone-related cancer diseases and ASD in their offspring.

ACKNOWLEDGMENT

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. ACKNOWLEDGMENT
  7. REFERENCES

The study was supported by a project grant from Beatrice Haskell's Fund.

REFERENCES

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. ACKNOWLEDGMENT
  7. REFERENCES
  • 1
    World Health Organization. The ICD-10 Classification of Mental and Behavioural Disorders. World Health Organization, Geneva, 1992.
  • 2
    Fombonne E. Epidemiology of autistic disorder and other pervasive developmental disorders. J. Clin. Psychiatry 2005; 66 (Suppl. 10): 38.
  • 3
    Rutter M, Caspi A, Moffit TE. Using sex differences in psychopathology to study causal mechanisms: Unifying issues and research strategies. J. Child Psychol. Psychiatry 2003; 44: 10921115.
  • 4
    Baron-Cohen S. The extreme mail brain theory of autism. Trends Cogn. Sci. 2002; 6: 248254.
  • 5
    Knickmeyer RC, Baron-Cohen S. Foetal testosterone and sex differences in typical social development and in autism. J. Child Neurol. 2006; 21: 825845.
  • 6
    Ingudomnukul E, Baron-Cohen S, Wheelwright S, Knickmeyer RC. Elevated rates of testosterone-related disorders in women with autism spectrum conditions. Horm. Behav. 2007; 51: 597604.
  • 7
    Manning JT, Baron-Cohen S, Wheelwright S, Sanders G. The 2nd to 4th digit ratio and autism. Dev. Med. Child Neurol. 2001; 43: 160164.
  • 8
    World Health Organization. International Classification of Diseases, 8th edn. (Danish version). Sundhedsstyrelsen, Copenhagen, 1971.
  • 9
    World Health Organization. International Classification of Diseases, 9th edn. World Health Organization, Geneva, 1978.
  • 10
    Mouridsen SE, Rich B, Isager T. Parental age in infantile autism, autistic-like conditions, and borderline childhood psychosis. J. Autism Dev. Disord. 1993; 23: 387396.
  • 11
    Pedersen CB, Gøtzsche H, Møller JØ, Mortensen PB. The Danish civil registration system. A cohort of eight million persons. Dan. Med. Bull. 2006; 53: 441449.
  • 12
    Andersen TF, Madsen M, Jørgensen J, Mellemkjoer L, Olsen JH. The Danish National Hospital Register. A valuable source of data for modern health sciences. Dan. Med. Bull. 1999; 46: 263268.
  • 13
    Somboonporn W, Davis SR. Postmenopausal testosterone therapy and breast cancer risk. Maturitas 2004; 49: 267275.
  • 14
    Reedy MB, Richards WE, Ueland F et al. Ovarian steroid cell tumors, not otherwise specified: A case report and literature review. Gynecol. Oncol. 1999; 75: 293297.
  • 15
    Kaaks R, Lukonova A, Kurzer MS. Obesity, endogenous hormones, and endometrial cancer risk: A synthetic review. Cancer Epidemiol. Biomarkers Prev. 2002; 11: 15311543.
  • 16
    Analytical Software. Statistix. Analytical Software, Tallahassee, FL, 2003.
  • 17
    Mosbech J, Jørgensen J, Madsen M, Rostgaard K, Thornberg K, Poulsen TD. The national patient register: Evaluation of data quality. Ugeskr. Laeg. 1995; 157: 37413745.