Our manuscript describes two children affected by Sotos syndrome who presented with CH, leading us to hypothesize a potential correlation between this syndromic condition and CH. The clinical, auxological, and biochemical characteristics of the two patients are summarized in Table 1.
Sotos syndrome is an overgrowth syndrome with an estimated incidence of 1 in 14,000 live births [30, 31]. It is characterized by three cardinal features: overgrowth (height and/or head circumference ≥ 2 SDS above the mean), usually of prenatal onset, learning disabilities and a distinctive facial appearance. The facial features include broad and prominent forehead with a dolichocephalic head shape, sparse frontotemporal hair, downslanting palpebral fissures, malar flushing, a long and narrow face, and a prominent chin. 15–89% of patients may exhibit behavioral disorders (most notably autism spectrum disorder), seizures, advanced bone age, cardiac anomalies, cranial MRI abnormalities, renal anomalies, joint hyperlaxity with or without pes planus, scoliosis, maternal preeclampsia, and neonatal complications [32].
Sotos syndrome is inherited in an autosomal dominant pattern. Diagnosis requires the identification of a heterozygous pathogenic variant in the NSD1 gene or a deletion encompassing NSD1. More than 95% of cases have a de novo pathogenic variant [31, 32]. NSD1 encodes a histone methyltransferase that acts as a transcriptional regulator, with both activating and repressive functions depending on cellular context. It is expressed in several tissues and organs such as the brain, kidney, skeletal muscle, spleen, thymus, and lung, though its complete function remains unclear [33, 34]. To date, no data are available in the literature regarding NSD1 expression in the thyroid.
The facial features of our female patient were not initially suggestive of classic Sotos syndrome, possibly due to ethnic variability, and became more recognizable later in life. While facial dysmorphism can support the clinical suspicion, they may not always be evident in early life. In contrast, macrocephaly is a consistent and useful diagnostic clue [35].
Although macrocephaly was already present at birth, the patient did not exhibit any other features of overgrowth initially. As shown in her growth chart (Fig. 1), overgrowth emerged around 5–6 months of age. We believe this delay may have been related to the recurrent infections and FPIES. Neurologically, only cervical hypotonia was recognized at birth; the developmental delay appeared during the follow-up, around 6 months of age. Considering that the levothyroxine treatment was started at 6 days of life and that during the endocrine follow-up TH levels remained well-controlled with appropriate dose adjustments, we attribute the neurodevelopmental delay to Sotos syndrome rather than to CH. In fact, many individuals with Sotos syndrome exhibit mild-to-severe learning disabilities, associated with behavioral issues and features of autism spectrum disorder [36].
Renal and cardiac anomalies, both present in this patient, can also be manifestations of Sotos syndrome [37]. Additionally, our patient’s neonatal complications align with features of the syndrome.
In contrast, our male patient showed overgrowth from birth, which persisted throughout follow-up (Fig. 2). He also presented with an early developmental delay, poor spontaneous motricity, and ligamentous laxity. Levothyroxine replacement therapy was initiated following the detection of altered thyroid function. However, due to inconsistency of follow-up, precise dose titration of levothyroxine was challenging. Nevertheless, despite occasional TSH rises, thyroid hormones levels always remained within normal ranges.
In the first patient, when genetic testing was performed we expected to find a gene variant responsible for CH in combination with cardiac and renal involvement, with PAX8 as the main candidate [18]. Overgrowth features were not evident yet, but emerged over time, prompting an extended overgrowth panel (including NSD1 and NFIX), which ultimately led to a definitive diagnosis. In contrast, in the second case, the onset of seizures, epileptic encephalopathy, ligamentous laxity and hypotonia guided the decision to perform a genetic test. In this case, a broad-spectrum NGS panel for encephalopathies was performed, leading to the identification of an NSD1 mutation. In neither case was Sotos syndrome the initial suspect, highlighting the diagnostic challenges this condition presents in the first months of life. Nevertheless, both patients exhibited some common characteristics, such as macrocephaly present at birth, neonatal apnea requiring non-invasive respiratory support, cardiac abnormalities (ASD in both cases), renal anomalies (hydronephrosis), and intellectual disability. Neonatal complications, renal and cardiac anomalies and particularly behavioral abnormalities are known features of Sotos syndrome [31, 32, 38].
The NSD1 gene variants identified in our patients, c.3439G > T p.(Glu1147*), a nonsense mutation in exon 5, and c.2954_2955delCT p.(Ser985Cysfs*25), a frameshift mutation also in exon 5, have both been previously reported by Saugier-Veber et al. as pathogenic variants associated with Sotos syndrome [38]. However, none of these variants were associated with CH in the literature.
Endocrine abnormalities are rarely reported in overgrowth syndromes, but may include hypothyroidism, as observed in a small number of patients affected by segmental overgrowth PIK3CA related disorders, Beckwith–Wiedemann syndrome and Weaver syndrome [39,40,41]. Specifically, limited data are currently available in literature regarding the potential association between CH and Sotos syndrome [42]. In 2005, Tatton-Brown et al. evaluated the genotype–phenotype association in 266 individuals with Sotos syndrome, reporting a case of CH due to thyroid hypoplasia with a p.R1984X NSD1 mutation [37]. That same year, Cecconi et al. described a group of 59 patients with congenital overgrowth and reported a patient with a hypoplastic thyroid and the same NSD1 variant [43]. In 2007, Saugier-Veber et al. described a case of Sotos syndrome with athyreosis in a cohort of 116 patients [38]. More recently, in 2021, Verma et al. reported a boy with a pathogenic novel NSD1 variant (c6076_6087del12: p.(Asn2026_Thr2029del) in exon 20) and CH. Imaging findings and treatment response suggested as a possible etiology for his permanent CH either thyroid hypoplasia or a partial iodine-trapping defect [29].
Our observations further support the relationship between CH and Sotos syndrome, and we hypothesize that this relationship is unlikely to be coincidental. In both our patients, the thyroid gland was in situ (with probable hypoplasia in one case), and both required ongoing treatment beyond the age of 3 years. Notably, 4 out of 6 patients with Sotos syndrome and CH described in literature (including ours) had thyroid hypoplasia, a pattern that requires further investigation.
The small sample size and lack of functional genetic analyses limit our ability to establish a definitive causal link between NSD1 mutations and CH. Further studies, including genome-wide association studies and functional assays investigating the role of NSD1 in thyroid development, are needed to better determine whether CH is a coincidental association or a clinic feature of Sotos syndrome.
In the meantime, we recommend early thyroid function screening in patients with Sotos syndrome. Moreover, Sotos syndrome should be suspected in children with CH, cognitive delay, and evidence of overgrowth.