This retrospective study was performed at the Department of Clinical Genetics and the Department of Paediatrics and Adolescent Medicine at the Hong Kong Children's Hospital. Subjects with 45,X/46,XY mosaicism identified on karyotype studies performed from January 1990 to December 2023 were included. Data on date of birth, age, karyotype, phenotype, sex of rearing, growth and pubertal assessment, surgical history of gonadectomy and gonadal tumor, as well as systemic comorbidities were retrieved from the electronic medical record and written or printed medical record files. External genitalia scores (EGS) were assigned to all subjects with available data on their initial genital appearance [8]. Height data were converted to age and sex-specific z-score according to the HK 2020 growth study [9]. Reaching near-adult height was defined as having less than 1cm of growth in the preceding year. Subjects were excluded if postnatal karyotype analysis was not performed or if they were lost to follow-up after initial diagnosis. Males and females in this study were defined by the sex of rearing.

Prenatal testing, including amniocentesis, chorionic villi sampling, and cordocentesis for karyotype, was included. Both prenatal samples and postnatal peripheral blood samples were karyotyped with G-bands at 550 banding resolution and a minimum of 30 cells counted. Fluorescence in situ hybridization and microarray results were also included when available.

All statistical analyses were carried out using IBM SPSS Statistics ver. 27.0 (IBM Co., USA). Descriptive statistics were reported in absolute numbers and percentages. All variables were assessed for normality before analysis and expressed as mean±standard deviation (SD) for parametric and median (interquartile range [IQR]) for nonparametric distributions. Differences in means were tested using Student t-test for parametric data and Mann-Whitney U-test for nonparametric data. Fisher exact test was employed to detect differences in categorical variables. Pearson correlation coefficient was used to determine the correlation between the proportion of cells with monosomy X and clinical phenotype. In addition, multiple regression analysis was conducted to identify potential confounding factors associated with EGS. Statistical significance was defined by a 2-sided P-value <0.05 for all analyses.

This study was approved by the institutional review board of Hospital Authority Central (PAED-2024-318).

Sixty-eight patients were identified to have 45,X/46,XY mosaicism. Among them, 35 (51.5%) were raised as females and 33 (48.5%) were raised as males. The demographics of these patients are detailed in Table 1. The reasons of referral for genetic testing included short stature (26.5%), abnormal genitalia (26.5%) and delayed puberty or amenorrhea (19.1%). There was no significant difference in the age of diagnosis between those raised as females and those raised as males (12.8±9.5 vs. 11.2±15.8, P=0.617). Prenatal diagnosis accounted for 22.1% of cases. The median age of last evaluation was 23.4 (IQR, 12.8–37.3) years and the duration of follow-up was 13.2 (IQR, 4.2–23.1) years.

Abnormal external genitalia were observed in 30.9% of patients (EGS=0.5–11.5), whereas 44.1% had normal female genitalia (EGS=0) and 25% had normal male genitalia (EGS=12). Patients reared as females had EGS scores ranging from 0 to 3.5, whereas patients reared as males had EGS scores ranging from 4.5 to 12. The external genitalia and gonadal outcomes are shown in Fig. 1. The proportion of cells with monosomy X in the peripheral blood karyotype did not show a significant correlation with male genital phenotype of lower EGS (P=0.070).

Cytogenetic analyses were categorized into 3 distinct groups: 45,X/46,XY mosaicism (70.6%), 45,X/46,XY mosaicism with 3 cell lines (5.9%) and structural Y abnormalities (23.5%) (Fig. 2A). Structural Y abnormalities were further subdivided into isodicentric Yq (62.5%), isodicentric Yp (25%), isochromosome Yp (6.3%) and marker Y (6.3%). A partial karyotype and ideogram of isodicentric Yq11.23 and isodicentric Yp11.2 are shown in Fig. 2B and C, respectively. For isodicentric Yq, there are 2 copies of SRY (sex-determining region Y), whereas for isodicentric Yp the SRY region is deleted.

Histological data of the gonads were available for 49 patients, comprising 43 gonadectomies and 6 gonadal biopsies. The overall gonadal tumor rate within the cohort was 22.4%. Details of the 11 patients with gonadal tumors are described in Table 2. All except one female had undergone gonadectomy and the rate of gonadal tumor in females was 17.6% (6 of 34). In males presenting with abnormal external genitalia, the gonadal tumor rate was 38.5% (5 of 13), whereas no tumors were detected in males with normal external genitalia. The mean EGS was significantly lower in males with tumors compared to those without tumors (6.6 vs. 10.4, P=0.002) (Supplementary Table 1). Among the males with retained gonads, surveillance ultrasonography (USG) of the testes was performed for 17 patients (70.8%), with no abnormal histology found in those with suspicious USG findings.

Twenty-one males and 31 females reached postpubertal age at the time of the study. For the 22 females who had gonadectomy performed after pubertal age, 4 had spontaneous thelarche up to Tanner stage II–III (18.2%) but none of them had spontaneous menarche. All males with preserved gonads, including 2 with EGS of 5.5 and 7 only, had spontaneous puberty. Three of them (with EGS ranging from 11.5 to 12.0) ultimately required testosterone replacement in their 30s and 40s for declining testosterone levels. Median of follicle-stimulating hormone levels of males with preserved gonads was 13.6 (IQR, 7.2–23.9) IU/L. Fertility data was limited, with one male demonstrating azoospermia on semen analysis.

The overall final adult height z-score in the cohort was -2.6±1.2. Female final adult height z-score was -2.2±1.6 while that for males was -3.2±1.2. Near-adult height was available for 42 patients, averaging 147.7±5.6 cm for females (n=26) and 151.0±8.3 cm for males (n=16). Seventeen subjects (including 6 males and 11 females) received recombinant GH therapy commencing at a mean age of 10.9±4.0 years. They were treated for 4.1±2.7 years and all had reached their final adult height. GH use and height outcomes are summarized in Table 3. There was no significant difference in the final height of males who were treated with GH compared to those who were not treated (152.3±8.6 cm vs. 150.5±8.5 cm, P=0.660). The 2 individuals on GH treatment with retained gonads did not develop gonadal tumors upon serial biopsy or USG testes. Females who used GH showed better increment in height z-score, though the final adult height was also not statistically significant compared to the untreated females (149.3±3.7 cm vs. 146.5±6.7 cm, P=0.210).

Regular thyroid function tests were performed in 72.1% of subjects (n=49) with abnormalities detected in 14.3%. Renal ultrasound was performed in 64.7% of subjects (n=44), out of which 9.1% showed various renal anomalies. Echocardiogram was done at diagnosis in less than half of the cases (n=30), with abnormalities detected in 20% of the scans. The majority of the subjects did not have a formal hearing test (n=21) or evaluation of orthopedic manifestations (n=27). Cardiometabolic health screening, including evaluation for hypertension, diabetes and dyslipidemia, was performed in 45.6% (n=31) to 54.4% (n=37) of the cohort, and these metabolic complications were respectively found in 12.9%, 19.4% and 21.6% of the subjects (Table 1).

Among the cohort, 10.3% had learning or psychological comorbidities including intellectual disability, autism spectrum disorder, attention deficit hyperactivity disorder and dyslexia. Two subjects developed gender identity disorder (GID). Their EGS were 7 and 12 respectively. They were reared as males and later expressed desire to transition.

Fifteen patients were diagnosed with 45,X/46,XY mosaicism prenatally (Supplementary Table 2). Their median age at last follow-up was 6.3 (IQR, 3.4–15.9) years. Eleven (73.3%) had normal male genitalia and had no gonadal tumors detected upon surveillance. The 4 females had gonadectomy performed at a median age of 1.6 (IQR, 1.0–2.8) years and none of them had gonadal tumor.

Previous surgical cohorts reported a higher gonadal tumor risk of 50%–55% [10], predominantly in males with undervirilization or hypospadias requiring surgery. Our current study, which included males with a broader spectrum of phenotypes, revealed a lower tumor risk of 38.5% despite a longer follow-up interval with more adults. This is likely due to the inclusion of more individuals with milder undervirilization (25% of males who had gonadal biopsy did not require any surgical intervention for genital atypia) compared to the existing literature. In fact, Cools et al. [11] also reported a tumor risk as low as 13% in those in mild undervirilization (external masculinization score EMS≥7). Among the 11 patients with gonadal tumors, all had either female phenotype or were males with abnormal genitalia (EGS 6–7.5; P=0.002) (Supplementary Table 1). In contrast, no malignancy was observed in those with normal male phenotype. This is consistent with findings in previous literature whereby gonadal tumor risk is highest in individuals with less masculinized genitalia and abnormally located gonads due to immature and poorly differentiated gonadal tissue, while those with normal male phenotype do not typically develop malignancy [11-13]. Hence regular sonographic assessment and gonadal biopsy at the end of puberty are advised, especially for those with lower EGS. On the other hand, the role of gonadal biopsy in patients with normal male phenotype remains uncertain and warrants further studies [14,15]. In this regard, the limited sensitivity of ultrasound in detecting premalignant lesions should be recognized, and that contrast magnetic resonance imaging might be necessary in suspicious cases especially with nonpalpable testes [16].

Spontaneous puberty was observed in all of the males in our study, including those with severe undervirilization. In addition, we observed no relationship between neonatal EGS and later need for hormonal replacement. In fact, 4 men with EGS 6.0–7.5 had adequate testicular function even in their 30s, demonstrating that the degree of androgenization in the fetal period could not predict hormonal outcomes in the long run. Hence, regular surveillance to detect declining function in retained gonads is necessary regardless of the initial phenotype in this population [5].

The risk of gonadal tumor in 45,X/46,XY females was historically reported to be as high as 100% [17-19], prompting conventional recommendations of upfront gonadectomy upon diagnosis in females [20]. However, recent studies, including our own, have shown that the risk may be lower and that most precursor lesions exhibit slow progression to malignancy [21,22]. On the other hand, spontaneous thelarche and menarche [23], as well as successful spontaneous pregnancy and live birth, have been documented in females with retained gonads [24,25]. In our study, the rate of spontaneous thelarche was 18.4%, with one girl reaching Tanner stage III breast. Consequently, current guidelines emphasize a shift towards shared decision-making on timing of gonadectomy, weighing tumor risk against potential gonadal function and fertility in these females [22]. In this regard, our findings also highlight the diverse presentation of gonadal tumor in this group of patients raised as females, with the youngest one diagnosed with intratubular germ cell neoplasia at 1.3 years, while an adult female undergoing gonadectomy at 38 years had no malignancy. Early gonadectomy may impose physical and emotional consequences [26], yet surveillance remains challenging due to unreliable imaging and low sensitivity of existing tumor markers including alphafetoprotein and human chorionic gonadotropin. Emerging biomarkers such as microRNAs and epigenetic markers have shown promise for detecting gonadal tumor in DSDs [27,28]. Future studies to validate their clinical application is necessary, and until then, a cautious approach should be adopted in those who decided to defer gonadectomy [29].

Consistent with previous literature, growth impairment and compromised final height were observed in our study [5,6,30]. The primary etiology of growth impairment in 45,X/46,XY individuals, as in Turner syndrome, is SHOX gene haploinsufficiency in the 45,X cell line [6,7]. Given the well-documented efficacy of GH therapy in Turner syndrome and SHOX haploinsufficiency, its use in this population appears theoretically justified [31,32]. However, disappointingly, GH treatment in males with 45,X/46,XY did not yield a significant improvement in final height in our study. The response to GH treatment in males with 45,X/46,XY has been examined in only a limited number of studies, most with small sample sizes and variable dosing regimens. While short-term growth improvement has been reported in some, data at adult height are similarly discouraging [5,6,30]. Both the group from Martinerie et al. [5] and Bertelloni et al. [33] found no significant difference in adult height in those treated with GH compared to the untreated group, and the adult height of treated subjects remained suboptimal (>2SD below mean). In our cohort, only one-third of the male patients treated with GH demonstrated an improvement in height z-score. While it can be argued that they were started on GH at a relatively late age and treated for a rather short period of time, the overall efficacy of GH, at least in the short term, appears limited. While no cases of gonadal malignancy were observed in GH-treated patients with retained gonads, the theoretical risk of testicular cancer in this population, combined with the modest therapeutic benefits, necessitates a careful discussion with families weighing the potential risk and benefit of GH treatment. In this regard, further prospective randomized trials examining the effects of GH in this population are warranted.

Besides short stature and risk of gonadal tumor, 45,X/46,XY individuals present with diverse health issues including cardiovascular abnormalities, metabolic and thyroid disorders, as well as renal anomalies. In fact, it is recommended that surveillance guidelines for Turner syndrome should also be applied to individuals with 45,X/46,XY mosaicism regardless of the sex of rearing [34]. Our study reveals that adherence to health complication screening remains suboptimal. A key contributing factor might be related to the fragmented nature of their medical care by multiple specialties throughout their life, who may each address isolated aspects of their condition without coordinated oversight. Results from our study revealed gaps in our current service model and underscored the need for an integrated care pathway to ensure timely surveillance for various health risks related to 45,X/46,XY mosaicism. Apart from physical health, recognizing the increased risk of psychological comorbidities is essential. Notably, within our study, 2 individuals reared as males had GID and received gender-affirming therapy. One of them was found to have 45,X/46,XY mosaicism during workup for GID. GID has been reported more often in individuals with 45,X/46,XY mosaicism raised as female, with a few case reports describing transition to male gender in adolescent years [35,36]. While there were no long-term data in those assigned males, careful attention to distress related to gender identity is imperative as part of integral care.

Since the discovery of cell-free fetal DNA in maternal plasma, NIPT has become commercially available in Hong Kong as a self-financed option since 2011, allowing earlier diagnosis of chromosomal disorders [37]. As illustrated in our study, this in turn identifies a subgroup of children with 45,X/46,XY mosaicism with relatively normal genital phenotypes and better height outcomes. However, there may be ascertainment bias as a portion of patients in the postnatal group presented with short stature. For those raised as females, earlier gonadectomy can be considered, whereas for males, surveillance can be started at an earlier age, possibly leading to better health outcomes [14,34].

Structural Y abnormalities included different breaks and rearrangements of the Y chromosome, with no known gender-specific break site regions identified thus far [38]. Our study also showed no correlation between different structural Y abnormalities with genital phenotype (50% had female phenotype with genotype carrying 2 SRY copies) (Supplementary Table 3). While the Y chromosome and SRY region have long been associated with the development of gonadal tumor, more detailed sequence mapping has now defined a gonadoblastoma locus on the Y chromosome (GBY) near the centromere with several loci extending from the proximal short arm at Yp11.1 (TSPY) to the proximal long arm at Yq11.2 (DDX3Y) as the proto-oncogenes in dysgenetic gonads [39,40]. This may explain why gonadal tumors were detected in the group of isodicentric Y patients with breakpoints in either long or short arms. Detailed molecular mapping and immunohistochemical staining of the GBY proteins were not routinely performed in our cohort. Future studies including these tests will be informative on determining if the presence or absence of the GBY locus correlates with the development of gonadal tumors.

The significance of this study lies in the large sample size with an extended duration of follow-up with such a rare clinical condition. In addition, previous literature mainly focused on postnatally diagnosed cases which may introduce ascertainment bias. In contrast, we included a subgroup of individuals diagnosed prenatally, thereby providing a more representative reflection of the heterogeneous patient population encountered in clinical practice. The inclusion of chromosomal analysis, clinical and histological outcomes also allowed a thorough understanding of how they relate to each other. Importantly, over 70% of the cohort has attained adulthood, enabling assessment of long-term outcomes such as final height and gonadal function.

There are nonetheless several limitations worth noting. The retrospective design may lead to missing data for some variables. Besides, the sample size in the prenatally diagnosed group is relatively small. Data on pregnancy termination rates for fetuses with 45,X/46,XY mosaicism and significant structural abnormalities are lacking, hence conclusions may be skewed as liveborn infants with normal structural scan tend to have milder phenotypes. In addition, this study did not assess the presence of germ cells in the histological specimens, and only one male patient had semen analysis performed. This information may provide valuable insights into the fertility potential of these individuals. Regarding assessment of the GH effect, anthropometric measurement data were incomplete especially for individuals not treated with GH, therefore adjusted longitudinal models were not feasible. Lastly, inconsistencies in screening for comorbidities within the cohort constrain the ability to draw definitive conclusions about the prevalence of various complications associated with 45,X/46,XY mosaicism.

In conclusion, this study highlights the heterogeneity of phenotypic manifestations and long-term outcomes of a large cohort of patients with 45,X/46,XY mosaicism diagnosed over 3 decades. Although a comparatively lower risk of gonadal tumors was observed, our findings emphasize the necessity for vigorous surveillance in males with abnormal genitalia and females with retained gonads. Our data also provides insights into the growth trajectory in this population and advocates the need for better adherence to complication screening, ideally within a multidisciplinary care framework. Further longitudinal studies of prenatally diagnosed cases are warranted to elucidate long-term clinical outcomes.