METHODS

Information gathered from medical records, direct examination, and self-report has been collected in a computer database as part of an ongoing, long term study of the natural history of Turner syndrome begun in 1977.^2,65,66 Subjects were ascertained through self-referral, advertisement, review of hospital records, and referral for diagnosis or management from paediatricians, obstetricians, perinatologists, endocrinologists, internists, and family practitioners.

All subjects have had karyotype confirmation of their clinical diagnosis. A minority has had two or more tissues analysed. In most of these, the two cell lines were amniotic fluid cells and confirmatory postnatal peripheral lymphocyte karyotyping.

I performed a search in PubMed using: Turner syndrome, sex chromosome, 47,XXX, triple X, and triple X females as search terms. I used the bibliographies from the citations generated to find case reports published before the inception of the electronic databases, as well as for any publications overlooked by the search strategy used. Reports in English, Italian, French, and Spanish were reviewed. I discarded any case reports without karyotype confirmation. Complete information was not available for all subjects. I analysed the results using the appropriate denominator for each feature.

RESULTS

Table 1 lists the ages and reasons for the diagnosis of Turner syndrome, comparing the subjects from our clinic population and those in the published reports to our clinic population with 45,X and 45,X/46,XX chromosome constitutions.

The women in the X/XX/XXX group described in the published reports were older than the subjects in the other groups. This is most likely because a significant proportion, one quarter, of these women were diagnosed in adult life, during evaluation of recurrent pregnancy loss. Among our clinic populations, prenatal diagnosis was a more common avenue of diagnosis than for the published cases. This reflects the nature of our clinic, which is a referral centre for such cases. We are less likely to be involved in the evaluation of recurrent pregnancy loss or fetal wastage and thus less likely to ascertain adult patients with Turner syndrome through this route.

Oedema is the primary reason for the diagnosis of Turner syndrome in infancy. Its absence as a clinical feature in the females with a triple XXX cell line constitutes a major difference between these subjects and those with 45,X or 45,X/46,XX. Although oedema in the newborn period was not recognised, in utero oedema most likely had been present, based on the presence of nuchal webbing in a similar proportion of the subjects with mosaicism for a 47,XXX cell line and those with monosomy X alone.

Approximately a third of patients with Turner syndrome are diagnosed during evaluation for short stature. Among the published X/XX/XXX cases, short stature accounted for only 14% of diagnoses, despite the fact that almost two-thirds of the women were below the 3rd centile in height.

Tables 2 and 3 provide data regarding those features of Turner syndrome hypothesised most likely to be different in subjects with mosaicism for XXX: intellectual function, height, and fertility.

I did not try to establish the presence or absence of specific learning disabilities known to be associated with Turner syndrome, such as difficulties with spatial reasoning, processing, etc,^67–72 as there was inadequate information available for most of the published cases. Only frank mental retardation, as defined by a tested IQ below 70, inability to function independently in adult life, or as stated as such in the case reports, was considered. None of the females in our clinic population with a 47,XXX cell line are mentally retarded, compared with a prevalence of 8-9% in our “control” groups. However, the number of patients with mosaicism for 47,XXX is small. The proportion of published cases with mental retardation and 45,X/47,XXX (42.8%) is much higher than for the other groups. However, one was ascertained as part of a screening evaluation of residents in an institution for the mentally retarded,⁴⁵ one had a brother and a half sister with similar delays, both of whom had been extensively evaluated without a diagnosis,⁴⁴ and a third woman was karyotyped at the age of 25 during an evaluation for seizures, mental retardation, and structural brain malformations.⁶¹ One subject was described as having normal mentality but was slightly slow⁴⁷ and two others were evaluated for pregnancy loss; normal intellectual function was implied but not clearly stated.^41,53 I did not include these three women in the analysis.

Short stature was seen in the majority of published cases. Some of our 45,X/47,XXX patients have received treatment with androgens or human growth hormone. Two of the three treated were above the 3rd centile in height. Three of eight who had not been treated were above the 3rd centile. None of our 45,X/46,XX/47,XXX patients had been treated with these agents; four of six were above the 3rd centile.

Spontaneous menarche is much more likely in subjects with mosaicism for an XXX cell line, with or without the presence of a normal 46,XX component (70%), than in women with 45,X without mosaicism (11%). Women with mosaicism for a 47,XXX cell line were also more likely to have spontaneous menarche than those with 45,X/46,XX (70% compared with 34%). A significant number of these women did not maintain normal menstrual cycles and secondary amenorrhoea or premature menopause were common occurrences.

Pregnancy is clearly more probable for women with the presence of either a 46,XX or a 47,XXX cell line or both, compared with those with monosomy X alone. Successful pregnancy outcome was least likely in the X/XX/XXX group, with less than 50% of pregnancies resulting in livebirths. Again, this may reflect how the subjects were ascertained, rather than a true a priori risk.

There was inadequate information provided for the published case reports to assess reliably the risk for other features known to be associated with Turner syndrome, including cardiac and renal malformations, thyroid disease, otological complications, etc. There were no obvious differences in the presence of these features between those with a 47,XXX cell line and those with monosomy X or 45,X/46,XX in our clinic population.

Establishing phenotype-karyotype correlations in Turner syndrome has been problematical. Many have been proposed and few confirmed. Overt chromosomal mosaicism for monosomy X and a second or third cell line occurs in almost 50% of patients with Turner syndrome and it is difficult to assess the relative contribution of each cell line to each organ system. All 45,X/46,XX subjects do not share the same proportion of normal:abnormal cells in all tissues and the proportion in one tissue does not necessarily predict that in others. Further, there may be selection in tissue culture for one cell line over another, so that the results in the laboratory do not represent the true in vivo situation. Moreover, occult mosaicism may mask differences conferred by the presence of cell lines with other X chromosome or Y chromosome abnormalities in ostensibly monosomic 45,X subjects. Most patients with Turner syndrome have been diagnosed based on karyotyping of only one tissue, usually peripheral blood. Thus, mosaicism for a chromosomally normal or differently aneuploid cell line may have remained undetected in the majority of cases.

Despite these limitations, the demand for prognostic counselling in Turner syndrome has increased, especially for the more uncommon karyotypes and in mosaicism, and even more so when the sex chromosome aneuploidy is detected prenatally.

DISCUSSION

In general, females with Turner syndrome, gonadal dysgenesis with sex chromosome aneuploidy, have oedema at birth or show physical evidence of in utero lymphoedema, are short, with adult height below the 3rd centile, and have streak gonads. These three features are the most consistently seen hallmarks of the condition. One-third of females with Turner syndrome are diagnosed at birth because of the presence of lymphoedema; two-thirds of those with a 45,X karyotype are so diagnosed.² Many of the infants without postnatal oedema show stigmata of prenatal oedema such as webbing of the neck and anteverted ears. Short stature is an almost invariable feature of the condition. Almost no predictions regarding final adult height can be made based on karyotype. Girls with mosaicism for a 46,XX cell line show a statistically insignificant taller mean adult height.⁷³ Those with a deletion of the long arm of the X (Xq−) are much more likely to have normal stature and to be diagnosed for primary amenorrhoea. Gonadal dysgenesis is a feature in almost all girls with Turner syndrome. Spontaneous menses occurs in approximately 10% of the entire population of girls with Turner syndrome. Almost 25% of girls who are mosaic for a normal 46,XX cell line will have spontaneous menarche.² A much smaller proportion in each group will maintain menses and even fewer will be fertile. Other physical features common to Turner syndrome, including congenital heart disease, kidney malformations, multiple naevi, etc, show no consistent relationship to karyotype.²

Triple XXX or 47,XXX occurs in 1/1000 of liveborn females.⁷⁴ Most of the information regarding these patients comes from several long term follow up studies of subjects ascertained through newborn screening programmes or from case reports.^3–8,75–8 Although there are relatively few studies, in general, it is believed that subjects with 47,XXX are more likely to have verbal processing deficits, a decrement in IQ predicted by sib achievement, and global delays. Height is usually normal and menses and fertility have not been recognised to be adversely affected. Caution must be used in making blanket predictions, as the total number of patients studied has been small, and most girls with 47,XXX are likely to go undetected or unreported.

In the subset of subjects with mosaicism for a monosomy X cell line who have been diagnosed with Turner syndrome, approximately 3-4% are mosaic for a 47,XXX cell line.² In our clinical experience, the recognition of this chromosome constitution prenatally can be particularly distressing for prospective parents and physicians. The most common reason for detection of 45,X/46,XX/47,XXX or 45,X/47,XXX prenatally is screening for Down syndrome. Prospective parents are usually concerned about the possibility of mental retardation and an increased likelihood of significant intellectual handicap associated with a 47,XXX cell line may be a pivotal factor in decisions regarding termination or maintenance of pregnancy. Thus, it seemed particularly worthwhile to try to establish phenotype-karyotype correlations based on postnatal survey for this subset of people.

On review of medical publications and our own clinic population, there are some major differences between those females with a 47,XXX cell line and those with 45,X only or 45,X/46,XX. The former are unlikely to present with perinatal oedema and more likely to be diagnosed in childhood for short stature, or later for primary or secondary amenorrhoea or during evaluation for pregnancy loss.

Despite the association of mental retardation with 47,XXX alone, there appears to be no significant difference in the risk for retardation in 45,X/47,XXX and 45,X/46,XX/47,XXX compared with monosomy X. None of our 16 patients is retarded. Although a much higher proportion (7/39, 18%) of published cases with 45,X/47,XXX or 45,X/46,XX/47,XXX had mental retardation, one had tubercular meningitis at the age of 11 months,³⁰ a second had neurofibromatosis,¹² another had Duchenne muscular dystrophy,³⁸ and a fourth woman had structural brain abnormalities including lissencephaly, cerebral atrophy, and pachygyria of the frontal lobe and cerebellar atrophy.⁶¹ None of these is associated with 47,XXX specifically. A fifth subject was ascertained through screening of inmates in a mental institution.⁴⁵ Bias of ascertainment must always be considered a contaminant of published data.

Short stature is not invariable in this subset of women. There are no data to suggest that treatment with growth promoting agents in this subset is more or less effective than for other karyotypes associated with Turner syndrome. I believe it is reasonable and fair to counsel parents that approximately 60% of girls with 45,X/46,XX/47,XXX or 45,X/47,XXX will have short stature.

Gonadal dysgenesis and infertility are the primary features of Turner syndrome that are a major concern for parents and, in my opinion and experience, play a significant role in the weighing of consideration for termination or continuance of a pregnancy when the prenatal diagnosis of an X chromosome abnormality is made. It is clear that the presence of an XXX cell line carries with it a greater likelihood of residual ovarian function. However, it must be understood that while the development of secondary sexual characteristics may be normal, fertility may be impaired in the majority of these women.

Prenatal diagnosis of mosaicism for a sex chromosome abnormality always carries with it uncertainty of prognosis based on the limitations of the testing itself. The detection of mosaicism in chorionic villus cells or amniotic fluid may reflect placental rather than true embryonic/fetal mosaicism and the proportion of chromosomally abnormal cells in vitro may not represent the true proportion in the embryo/fetus, nor the specific tissue distribution of the mosaicism. It has been shown for both 45,X/46,XY mosaicism^81,82 and 45,X/46,XX mosaicism⁸³ that there is a much better prognosis and greater likelihood of normality when the diagnosis is made prenatally rather than postnatally. At this time, we do not have a sufficient number of patients with mosaicism for 47,XXX that have been ascertained prenatally to come to any useful conclusions. Our three prenatally diagnosed patients, while growing within normal parameters and intellectually unimpaired, are only 1½12, 5, and 7 years old.

In summary, I recommend that prenatal and postnatal predictive counselling for subjects with 45,X/46,XX/47,XXX and 45,X/47,XXX Turner syndrome be essentially the same as for Turner syndrome in general with regard to likelihood of mental retardation, short stature, and structural abnormalities. For any particular infant, it is impossible to predict specific outcome. It is fair to suggest that residual ovarian function is possible and to caution that premature ovarian failure is common. Those women who do become pregnant, with or without medication, are at higher risk for chromosomally abnormal offspring (7/44 livebirths, two with trisomy 21, the rest with sex chromosome aneuploidy^{13,20,27,32,42}) and should be counselled appropriately.