1 Introduction

Beckwith–Wiedemann syndrome (BWS; OMIM #130650) is an overgrowth disorder primarily characterized by macrosomia, macroglossia, and abdominal wall defects. Clinical diagnosis is based on a clinical scoring system and the variability in phenotype has led to the recent introduction of the term Beckwith–Wiedemann spectrum (BWSp) (Brioude et al. 2018). Approximately 80%–85% of clinically diagnosed individuals exhibit (epi)genetic aberrations in imprinting center regions 1 (IC1) and 2 (IC2) on chromosome 11p15, affecting the expression of monoallelically expressed (imprinted) genes. The main molecular subtypes are loss of methylation at KCNQ1OT1:TSS-DMR (IC2-LOM; approximately 50% of cases), gain of methylation at H19/IGF2:IG-DMR (IC1-GOM; 5%–10% of cases), and paternal uniparental isodisomy of chromosome 11 (patUPD11; approximately 20% of cases). A smaller number of individuals have intragenic pathogenic variants in the CDKN1C gene (Brioude et al. 2018).

Pre- and postnatal overgrowth occurs in 45%–65% of individuals with BWSp (Wang et al. 2019). Postnatal growth is in the middle to upper percentiles but is described to slow in late childhood (Shuman et al. 1993). Therefore, adult height is often a concern for parents.

In pediatric healthcare, growth charts derived from healthy children and adolescents serve as crucial tools for healthcare professionals, facilitating the evaluation of a child's growth in both preventive and clinical settings. Additionally, growth charts have been developed for various genetic syndromes, for example, Turner syndrome or Marfan syndrome, offering healthcare professionals the means to plot an individual's growth data and make meaningful comparisons with individuals sharing the same syndrome (Lauffer et al. 2023; Rongen-Westerlaken et al. 1997). These specialized (syndrome-specific) growth charts play an important role in detecting aberrant growth patterns and are instrumental in providing patients and parents with insights into anticipated future growth, including adult height prediction. Currently, there are no syndrome-specific growth charts available for BWSp up until adulthood, and to date, limited data on adult height outcomes have been reported. Brioude et al. published the largest cohort of 407 individuals with BWSp with available data on adult height (Brioude et al. 2018). Compared with the French reference population, mean adult height was +1.8 standard deviation score (SDS) and 51.4% of the individuals with BWSp reached an adult height above +2.0 SDS (Brioude et al. 2013). Mean adult height was highest in the IC1-GOM group (2.8 ± 1.4 SDS) followed by CDKN1C variants (2.0 ± 1.6 SDS), IC2-LOM (1.7 ± 1.2 SDS), and the patUPD11 group (1.0 ± 1.8 SDS).

Thus far, two studies have developed BWSp growth charts. Mussa et al. developed fetal growth charts using growth data of 89 fetuses with BWSp, showing that all four molecular BWSp subgroups had a birth weight SDS and length SDS significantly higher than the general population (Mussa et al. 2016). Choi et al. developed growth charts using growth data of 51 Korean children with BWSp from 1 year up until 12.5 years of age (Choi et al. 2024). Compared with the reference population, males with BWSp exhibited greater mean height at all ages, whereas females showed a greater mean height only before the age of 7 years.

To optimize care for individuals with BWSp, the goal of this study was to develop BWSp specific growth charts for children aged 0–18 years. Additionally, we sought to explore (epi)genotype-phenotype differences with respect to growth.

2 Materials and Methods

3 Results

Table 1 shows the characteristics of the included individuals and data of the BWSp cohort. A total of 3107 height measurements, 3414 weight measurements, and 1401 head circumference measurements were analyzed. The distribution of the country of investigation and BWSp subtypes of this cohort are given in Table 1. Growth parameters at birth and age 18 years are shown in Table 2.

3.1 Height-for-Age Charts

Investigating the impact of the country of investigation, which serves as a proxy for ethnicity, we found notable differences in height among 18-year-olds. In the Netherlands, male children with Beckwith–Wiedemann spectrum (BWSp) were significantly taller than their counterparts in the United States by 2.9 cm (standard error (SE) 0.3 cm, p < 0.001). For females, again Dutch children were significantly taller compared to those in the United States by 2.4 cm (SE 0.3 cm, p < 0.001). This prompted us to normalize the Dutch height data. We used the Dutch reference height-for-age charts (Schonbeck et al. 2013) to calculate height standard deviation score (HSDS) values for each height measurement. These HSDS values were then converted to corresponding height values on the WHO reference standards, effectively normalizing the data to ensure comparability across different populations. The normalized height values were used for growth chart modeling.

Both male and female individuals with BWSp exhibited increased linear growth across infancy, childhood, and adolescence when compared to the reference population, as illustrated in Figure 1A,B. The linear growth charts plateaued for males at 16 years of age and for females at 15 years of age, aligning with the patterns observed in the reference populations. In Figure 2, this enhanced growth is further visualized by comparing height SDS (HSDS) to the WHO reference over time. In both males and females, HSDS values remained consistently above the WHO reference.

We further analyzed GAMLSS height-for-age models (including data of children with BWSp in the Netherlands) by adding the BWSp genotype as a variable to examine the effect on mean height per genotype at age 18 years. For this analysis, IC2-LOM was set as the reference level. The analysis indicated that while there were differences in height associated with specific genotypes, these were not large (Table 3). Specifically for males, the IC1-GOM subgroup was associated with a slight increase in height compared to the IC2-LOM subgroup, suggesting that individuals with this genotype may be slightly taller by 18 years of age. Conversely, the patUPD11 subgroup was associated with a reduction in height, and the CDKN1C subgroup showed a slight decrease in height, but this difference was not statistically significant. In the analysis focused on females, we observed evident differences in height at 18 years of age, though these were also modest. Notably, individuals of the CDKN1C subgroup were found to be slightly shorter, the IC1-GOM group was further reduced in height, and the patUPD11 subgroup showed the most considerable decrease in height compared to the IC2-LOM subgroup. Given the limited amount of data for each subgroup at 18 years of age, these observations did not allow for reliable estimations of adult height for each subgroup.

TABLE 3. Results of the BWSp genotype analysis.

	Males			Females
	Difference	Standard error	p	Difference	Standard error	p
Height
patUPD11 compared to IC2-LOM	−1.8 cm	0.2 cm	< 0.001	−2.4 cm	0.2 cm	< 0.001
IC1-GOM compared to IC2-LOM	1.9 cm	0.4 cm	< 0.001	−1.8 cm	0.3 cm	< 0.001
CDKN1C variant compared to IC2-LOM	−0.1 cm	0.6 cm	0.8	−1.3 cm	0.6 cm	0.04
Weight
patUPD11 compared to IC2-LOM	0.02 kg	0.1 kg	0.9	−0.2 kg	0.1 kg	0.01
IC1-GOM compared to IC2-LOM	0.8 kg	0.2 kg	< 0.001	0.03 kg	0.1 kg	0.8
CDKN1C variant compared to IC2-LOM	0.4 kg	0.2 kg	0.1	−0.6 kg	0.2 kg	< 0.01
Head circumference
patUPD11 compared to IC2-LOM	−0.02 cm	0.1 cm	0.8	−0.6 cm	0.1 cm	< 0.001
IC1-GOM compared to IC2-LOM	0.1 cm	0.2 cm	0.7	−0.6 cm	0.2 cm	< 0.001
CDKN1C variant compared to IC2-LOM	−0.6 cm	0.01 cm	0.02	0.1 cm	0.3 cm	0.7

3.2 Weight-For-Age and BMI-For-Age Charts

Looking at the impact of the country of investigation, we did not observe significant weight differences between the children in the United States, the Netherlands, and Italy for both males and females. Therefore, all weight data were analyzed together.

Weight gain in both males and females with BWSp from birth is depicted in Figure 3A,B. The weight-for-age charts exhibited a consistent upward trajectory compared to the reference population, indicating sustained enhancement with age. This suggests that the increased weight in individuals with BWSp is not solely attributed to a temporary surge but rather reflects a persistent pattern of increased weight gain compared with the general population. The mean weight SDS for females with BWSp consistently aligned with +1 SDS at all ages. In males, the mean weight SDS consistently registered at +1.2 SDS across all ages. Similarly, in both males and females with BWSp, BMI demonstrates an early and sustained increase, as illustrated in Figure 4A,B. Notably, the SDS for both sexes consistently coincided with +1 SDS.

Investigating the impact of genotype (Table 3), with IC2-LOM as the reference level, for males we found a significant difference with IC1-GOM, where children with IC1-GOM were marginally heavier at age 18. IC2-LOM was not statistically different from the CDKN1C and the patUPD11 subgroups. For females, the CDKN1C and patUPD11 subgroups were marginally less heavy.

3.3 Head Circumference-for-Age Charts

Investigating the impact of country of investigation, we did not observe significant head circumference differences between the children in the United States, the Netherlands, and Italy for both males and females. Therefore, all head circumference data were analyzed together.

Growth charts for head circumference in individuals with BWSp could only be constructed up until the age of 4 years due to a lack of sufficient data beyond this point. The mean head circumference in females consistently aligned with +1 SDS until 2 years of age, after which it gradually declined to +0.6 SDS. In males, the average head circumference remained at +1 SDS until 4 years of age as illustrated in Figure 5A,B.

In males, the CDKN1C subgroup had significantly lower head circumference than IC2-LOM at age 4 years (Table 3). In females, the IC1-GOM and patUPD11 subgroups had significantly lower head circumference.

4 Discussion

In this study, we retrospectively collected anthropometric and genetic data of 574 individuals with BWSp aged 0–18. We constructed BWSp growth charts including height-for-age, weight-for-age, BMI-for-age, and head circumference-for-age growth charts.

Our cohort includes individuals with BWSsp from the United States, Italy, and the Netherlands. The WHO growth charts are based on two studies: (1) a longitudinal study on healthy, term-born infants from six countries (Brazil, Ghana, India, Norway, Oman, United States) from 0 to 5 years under ideal conditions, and (2) a US population growth study from 5 to 19 years (de Onis et al. 2007). Therefore, the WHO reference data are suitable for comparison with US and Italian individuals because mean heights in these populations are close to WHO reference mean heights (Cacciari et al. 2006; de Onis et al. 2007). However, the WHO growth charts are not well suited as reference data for individuals from, for example, the Netherlands, where children are considerably taller (Schonbeck et al. 2013). For that reason, we adjusted Dutch BWSp data after finding a statistically significant difference in height of Dutch BWSp individuals compared to US/Italian BWSp individuals. The resulting growth charts are suitable for use in countries where growth patterns align closely with WHO standards. However, in countries where growth deviates from these standards, the HSDS-for-age charts (Figure 2) can be used as an alternative tool to evaluate linear growth since we speculate that genetic disorders affect height similarly across nations. In our study we could only investigate this for US-Italy data versus Dutch data. A detailed comparison with previous Korean BWSp data was not possible due to missing HSDS values for individuals with BWSp compared to the reference population (Choi et al. 2024). However, visual inspection of the Korean BWSp growth charts suggests that individuals with BWSp in their cohort grow approximately around +1 SDS. While this is slightly taller than the average in our population, it is broadly consistent with our hypothesis that genetic disorders like BWSp influence height similarly across populations. Future studies should further investigate this hypothesis by incorporating data from countries with shorter average heights, such as those in South-East Asia.

Reviewing the growth patterns of BWSp children in the current report, all parameters were increased compared to the WHO growth charts. Contrary to the notion that postnatal growth appears to slow around the age of 7 to 8 years, we observed enhanced growth across all ages (Figure 2) (Shuman et al. 1993). Our data confirm previous observations that individuals with BWSp had a significantly higher length, height, and weight SDS compared to the general population, both at prenatal and postnatal ages (Choi et al. 2024; Mussa et al. 2016). It should be noted that, as illustrated in Figure 2, growth rate appears to slow somewhat during the pubertal ages (compared to the WHO growth charts), whereafter it recuperates. We believe the GAMLSS modeling approach might have somewhat masked the pubertal growth spurt due to the smoothing methodology and relatively sparse data during adolescence (Figures S1–S40). Future studies with denser pubertal data or inclusion of pubertal staging may provide deeper insights into the growth trajectory during this period.

Adult height is often a major concern for parents of individuals with BWSp. Our study revealed that the mean adult height for males was highest in the IC1-GOM group, followed by IC2-LOM, CDKN1C, and the lowest in the patUPD11 group. For females, the highest mean height at 18 years of age was seen in the IC2-LOM group, followed by the CDKN1C, IC1-GOM, and patUPD11 group. The interpretation of the differences in growth between the four subgroups is difficult and may be due to the differently regulated timing of gene expression of the two domains during embryogenesis (Mussa et al. 2016). Interestingly, the sub-analysis of the impact of genotype on height revealed a distinct difference between males and females, highlighting how genetic influences on height may vary between sexes (Table 3). It is important to note that we could not accurately calculate adult heights for these subgroups. However, the observed differences in height among BWSp genotypes were modest, amounting to less than 3 cm. Mean (modeled) height at 18 years in our study was 180.6 cm for males (+0.6 SDS compared to the WHO reference population) and 166.3 cm for females (+0.5 SDS compared to the WHO reference population) (Table 2). Contrasting our findings, an earlier study showed the highest mean adult height in the IC1-GOM group (2.8 ± 1.4 SDS), followed by those with the CDKN1C variant (2.0 ± 1.6 SDS), IC2-LOM (1.7 ± 1.2 SDS), and the patUPD11 group (1.0 ± 1.8 SDS) (Brioude et al. 2013). The discrepancies in adult heights could potentially be attributed to continued growth in certain subtypes after 18 years of age, which our model did not account for. Another contributing factor may include that our analysis was based on modeled mean height at 18 years and did not capture adult heights, whereas the Brioude et al. study included the adult height of 35 individuals, offering a more direct measurement. Importantly, in the Brioude et al. study, the French reference charts of Sempé et al. were used for comparison and to calculate the HSDS (Brioude et al. 2013; Sempé et al. 1973). It was later postulated that these reference charts may underestimate the height of the current French population (Scherdel et al. 2015). This raises the possibility that the findings of Brioude et al. may have overestimated the average HSDS of individuals with BWSp. In that case, the French data would be more in line with ours.

The presented growth charts will facilitate personalized medical care of children with BWSp and allow providers to determine whether an individual's growth pattern is consistent with that of other individuals with the same syndrome. These charts in combination with bone age assessment will also allow for a more accurate prediction of an individual's future growth and thereby facilitate decision-making concerning epiphysiodesis in children with asymmetric lateralized overgrowth of the legs (Ruzbarsky et al. 2017).

This study presents several limitations. The heterogeneous nature of the BWSp cohort, encompassing various genetic subtypes and countries of investigation, is a potential source of variability. We believe that the analyses including the country of investigation as a variable should be interpreted with caution, since our dataset did not include detailed information on ethnicity. However, we believe that the country of investigation serves as a valid proxy for ethnicity. Second, the retrospective design of the study limited our control over data quality and the ability to collect additional relevant information, such as pubertal stages, bone age assessments, and parental heights. Third, a potential bias might be present due to the utilization of measurements from children attending specialized care clinics, which may not accurately reflect the entire BWSp population. Finally, due to the fact that growth in syndromes varies among different populations (e.g., Dutch individuals with Down syndrome are up to 10 cm taller than their peers in other countries) (Van Gameren-Oosterom et al. 2012), Dutch data needed to be normalized. Still, by combining Italian, US, and normalized Dutch data, a pooled syndrome-specific growth chart could be constructed (akin to the WHO growth standards) that may be useful in most parts of the world.

In conclusion, this study presents the first BWSp-specific growth charts. The findings offer valuable insights into growth patterns and potential genotype–phenotype correlations in individuals with BWSp. Despite the limitations, these growth charts hold promise for optimizing care and for providing a valuable tool for healthcare professionals managing the unique growth trajectories of individuals with BWSp.

Author Contributions

Saskia M. Maas: writing – original draft, data curation, project administration, conceptualization. Peter Lauffer: writing – original draft, data curation, methodology, formal analysis, project administration. Guido Cocchi: writing – review and editing, data curation. Madison DeMarchis: writing – review and editing, data curation. Andrew M. George: writing – review and editing, data curation. Alessandro Mussa: writing – review and editing, data curation. Francesco Pellegrino: writing – review and editing, data curation. Alexander M. J. Spaans: writing – review and editing, methodology, formal analysis, software. Emma C. van den Brink: writing – review and editing, data curation. Jan M. Wit: writing – review and editing, methodology. Leonie A. Menke: writing – review and editing. Jennifer M. Kalish: writing – review and editing, data curation.

Ethics Statement

Children's Hospital of Philadelphia data were collected under exempt IRB protocol IRB 19-016459. Informed consent was not necessary as all data were anonymized prior to analysis and collected by the physicians of the included individuals with Beckwith–Wiedemann spectrum. This study conforms to the Declaration of Helsinki.

Conflicts of Interest

The authors declare no conflicts of interest.