Growth patterns in individuals with CDKL5 deficiency disorder

To compare growth in individuals with cyclin‐dependent kinase‐like 5 (CDKL5) deficiency disorder with population norms and to investigate the effect of gastrostomy on growth.

Cyclin-dependent kinase-like 5 (CDKL5) deficiency disorder (CDD), once thought to be a variant of Rett syndrome, is a rare, X-linked developmental and epileptic encephalopathy 1 with an estimated prevalence of 1 in 40 000 to 60 000 live births (based on four cases in a Scottish population-based study). 2 In 2004 two reports were published, each identifying instances where mutations in the CDKL5 gene were associated with infantile spasms and severe neurodevelopmental impairment. 3,4 DD is an Xlinked dominant condition with an estimated approximate female:male ratio of 5:1. 5 CDD displays the common features of a genetically caused developmental and epileptic encephalopathy, namely refractory seizures, and significant developmental delay.The minimum diagnostic criteria, suggested in 2019, include epilepsy in the first year of life, both cognitive and motor delay, and a pathogenic variant in the CDKL5 gene.5 CDD is distinguishable from Rett syndrome by early-onset seizures and more severely impaired gross motor and global developmental delay, often associated with early hypotonia.Other comorbidities affecting the gastrointestinal and respiratory systems as well as sleep disturbances occur in both disorders in varying proportions. 6ome subtle dysmorphic features with a distinct facial, limb, and hand phenotype, not seen in Rett syndrome, may be present in CDD.7 Cortical visual impairment is also common and its presence has been shown to correlate inversely with level of developmental achievement.8 Despite the high prevalence of gastrointestinal symptoms, often associated with feeding problems, 1 the pattern of growth in CDD has not yet been comprehensively studied other than through some minimal data relating to head growth.Using data from the International Rett Syndrome Phenotype Database, InterRett, it was shown that head growth deceleration had occurred in 25 of 42 individuals with a pathogenic CDKL5 variant but only three of whom would be classified as having microcephaly.7 In a mainly European study, published in 2018, 30 out of 32 individuals with a pathogenic CDKL5 variant had a normal head circumference at birth but subsequent head growth was not reported.9 The recently published international clinical guidelines for CDD 10 recommend that, if there is a high risk of dysphagia, gastrostomy should be used to ensure long-term nutrition in CDD.This is consistent with guidance provided by the European Society for Paediatric Gastroenterology, Hepatology and Nutrition for children with neurological impairment, such that gastrostomy has increasingly been used to help with feeding difficulties in long-term neurological disability over the past 50 years.11,12 Australian population-based data have shown that all-cause and epilepsy hospitalizations in children with severe intellectual disability decreased following gastrostomy.13 However, there was no lessening of admissions for lower respiratory tract infections 13 and, compared with matched controls, survival a year after gastrostomy insertion did not improve over time.14 It has been shown both in two very small studies that medication absorption 15 and ketogenic diet stability 16 may improve following gastrostomy, and in CDD there is evidence of better caregiver emotional well-being.17 Rett syndrome is known to be associated with compromised growth, affecting weight, length, and head circumference.18 In an Australian study investigating the impact of gastrostomy in Rett syndrome, body mass index (BMI) was greater in individuals with gastrostomy placement than in those without.19 In the absence of available published data on this topic for individuals with CDD, we aimed to compare growth with population norms and to investigate the effect of gastrostomy on growth.

Data source and study population
The International CDKL5 Disorder Database was used as the data source for this longitudinal study with repeated cross-sectional design.Individuals with a pathogenic or likely pathogenic CDKL5 variant and whose families had provided any anthropometric measurement in a baseline questionnaire (administered between 2012 and 2021) and/or a follow-up questionnaire (administered between 2018 and 2019) were included.

Outcome variables
The main outcome variables were parent/carer-reported length/height, weight, and head circumference.The respondents provided the values in metric (centimetres, kilograms) or imperial (feet, inches, pounds, ounces) units; the latter were converted to the former for analytical purposes.In addition to a question asking for the current measurements at, or close to, the time of questionnaire completion, the respondents were also encouraged to report previous measures, if available, thus providing a longitudinal profile for some individuals.For all measurements, the age at measurement was determined from the date of collection.For measurements taken at home, practical advice on how to perform the measurements was provided in the questionnaire.Otherwise, measures were probably gleaned from recent assessments during clinic visits or hospital admission.
Additional outcome variables included BMI and age-and sex-standardized z-scores of the anthropometric measures (including BMI).BMI is calculated by dividing weight in kilograms by the square of height/length in meters, measured on the same day.The standardized z-score was calculated using the British 1990 growth reference fitted by the LMS method, 20 and individuals of age within the valid range (length/height-for-age, 0-23 years; weight-for-age, 0-23 years; head-circumference-for-age, 0-18 years [males] and 0-17 years [females]) were included in the z-score analysis (Tables S1 and S2).For comparison, the age-and sexmatched anthropometric values at the 50th centile of the reference population were also determined for the study individuals.

Explanatory variables
Sex was categorized into female and male.For descriptive purposes, age was defined as the age (in years and months) at questionnaire completion.For the analysis of the measurements, age was defined as the time at which the measurement was taken.The age variable was grouped into six categories: younger than 1 year, 1 year, 2 to 3 years, 4 to 6 years, 7 to 12 years, and 13 years or older.Gastrostomy status was dichotomized as whether individuals were reported to have had a gastrostomy tube placed at the time of baseline or the follow-up questionnaire completion (with) or not (without).CDKL5 variant type was categorized as previously published. 21at this paper adds

Pre-and post-gastrostomy longitudinal analysis
Among individuals who had gastrostomy insertion performed between the time of follow-up and baseline questionnaire, their annual weight change within the period was presented.
Other gastrostomy-related changes, including the number of hospitalizations, and occurrence of specific comorbidities such as constipation, reflux, and air swallowing, were also reported and the results were grouped into three categories: decrease/ resolve, no change, and increase/new occurrence.

Statistical analyses
Descriptive statistics were used to summarize the characteristics of the study cohort.Categorical data were summarized using frequency and proportion.Continuous data such as anthropometric measures were summarized using mean and standard deviation.Age was found to be positively skewed (i.e., the distribution tail was more pronounced on the right-hand side) and was summarized using the median, first quartile, third quartile, and range.The same approach was used to describe the outcome variables by sex, age group, and gastrostomy status, and the current measurements collected from the baseline and follow-up questionnaires were used in the analytical sample.The age trend, which used current and previous data (when provided), was modelled using generalized linear regression with Gaussian distribution and identity link.The generalized estimating equations method was used to account for the clustering within individuals, and the robust standard errors and an exchangeable working correlation structure were used for parameter estimation.Age, the explanatory variable, was included in the model as a linear spline with knots at

Ethics and consent
The Human Research Ethics Committee, University of Western Australia, provided approval for this study.The consent form was embedded in the online questionnaire.All parents/carers provided consent to the research and to publication of the results.

R E SU LTS
The characteristics of the 353 children and adults with CDD whose families provided anthropometric measurements on baseline and follow-up questionnaires are shown in Table 1.The median age at ascertainment in the baseline and follow-up cohorts was 4 years 2 months (interquartile range [IQR] 1 year 10 months-8 years 6 months) and 9 years 1 month (IQR 5 years-13 years 10 months) respectively.Most of the individuals were female, born between 2010 and 2020, and residing in the USA.The most common variants were a missense/in-frame variant within the catalytic domain and a truncating variant between amino acids 172 and 781.Slightly more than one-quarter (n = 96, 27.2%) had had a gastrostomy insertion performed, and in the 85 for whom an age at gastrostomy insertion was available, the median age for this procedure was 2 years 9 months (n = 85, IQR 1 year 9 months-6 years 2 months).Their characteristics are shown in Tables S3 and S4.Only 6% of those aged 2 years and older with a gastrostomy were able to walk independently at baseline and only 7% at follow-up compared with 26.9% and 34% of those without a gastrostomy.Twenty individuals underwent gastrostomy insertion subsequent to their baseline questionnaire but before the follow-up questionnaire (Figure S1).Across the four anthropometric outcomes, females seemed to fare worse in the standardized weight and length/height, but not BMI and head circumference, than males.

Weight
Overall, the mean weight was 21.3 kg (standard deviation [SD] 12.6), and the mean weight-for-age z-score was −0.97 (SD 1.74) (Table 2).By age group, the standardized value was comparable between the cohort and the unaffected population before age 4 years, but the measure gradually deviated from the norm afterwards (Table 2).The mean weight of individuals in the gastrostomy subgroup was more than one standard deviation below the reference mean value (zscore −1.18 [SD 1.55]) (Table 2) and the age trend indicates most of the variation occurred during the first year of life (Figure 1).Compared with that of the 'without gastrostomy' subgroup, the standardized weight score of the gastrostomy group seemed to be lower in the early years of life (Figure 1 and Table 3).

Length/height
Like weight, the mean length/height of the individuals with CDD (111.4 cm [SD 27.7]) was lower than that of the reference population, as reflected in the mean length/ height-for-age z-score of −0.65 (SD 1.94) (Table 2).The progressive shift away from the norm after 4 years of age was similarly apparent (Table 2).The mean z-score of the gastrostomy subgroup was −1.06 (SD 1.97) (Table 2), and its age trajectory was generally similar to that of the subgroup without gastrostomy except, perhaps, in adulthood (Figure 2 and Table 3).

BMI
The mean BMI of the study cohort was 16.4 kg/m 2 (SD 3.4), and the difference from the norm was less than that of length/height or weight alone (mean BMI-for-age zscore −0.81 [SD 1.91]) (Table 2).The deviations were below the average of the reference cohort for most age groups and were within one standard deviation except in the 13 years d Includes 20 individuals who had gastrostomy performed between the time of follow-up and baseline questionnaire (see Figure 1).
or older subgroup (Table 2).In the gastrostomy subgroup, the mean z-score was −0.95 (SD 1.75) (Table 2), and the age trend indicated a probable improvement in the score in this subgroup compared with those without gastrostomy in the age range between 2 and 3 years (Figure 3 and Table 3).

Head circumference
The mean value of head circumference was 48.5 cm (SD 3.6), which was more than two standard deviations below the average of the unaffected population (mean head circumferencefor-age z-score −2.12 [SD 1.73]) (Table 2).The standardized score was around −1 during infancy and then declined with increasing age and eventually stabilized during adolescence (Table 2).Individuals with gastrostomy had a mean z-score of −2.58 (SD 1.58) (Table 2).The age trajectories of the with and without gastrostomy subgroups were different in the early years but tracked with each other after 7 years of age (Figure 4 and Table 3).Additional plots, including boxplots of the standardized values and trend plots of the raw values superimposed on the British 1990 growth reference, for the four anthropometric measurements by subgroups (e.g.age group, sex) are shown in Figures S2 to S9.

Pre-and post-gastrostomy longitudinal analysis
The effects of gastrostomy on the 20 children and adults who had the procedure performed between questionnaires are shown in Table 4.The median age at gastrostomy was 3 years 6 months (IQR 2 years 6 months-9 years) and the mean duration between questionnaires was 3 years 8 months (SD 1.8) (Table 4).There was a small gain in weight over time (mean value 1.9 kg per year [SD 1.0]), but the change was slightly below par (mean weight-for-age z-score −0.22 per year [SD 0.53]).Apart from the number of hospital admissions, which showed a decrease in nearly half of the sample (7 out of 15), F I G U R E 1 Scatterplot and trend of weight-for-age z-score over age in 339 children and adults with CDKL5 deficiency disorder, overall and by gastrostomy status.Note: linear spline with knots at 1, 2, 4, 7, and 13 years.Abbreviation: CDKL5, cyclin-dependent kinase-like 5. although there was also an increase for an equivalent number, there was no apparent improvement or deterioration in most individuals concerning constipation, reflux, air swallowing, and use of rescue medications (Table 4).

DISCUS SION
We have shown that weight, height, BMI, and head circumference in those with CDD were all below the British general population norms for the equivalent age.This was especially striking for head circumference.The disparity was particularly pronounced for all anthropometric measurements after 4 years of age except for BMI.This is important because BMI partly takes into account the relation between weight and height of relevance if these children were for some reason constitutionally small.Interestingly, growth in individuals with gastrostomy placement was slightly more compromised than in those without but potentially less so for BMI where some possible beneficial effect of gastrostomy was seen in children aged 2 to 3 years.As expected, those with a gastrostomy were less likely to be walking and, consequently, may have reduced muscle bulk which could also have contributed to their lower weight.In a small subset of 20 individuals whose gastrostomy was placed in the period between baseline and follow-up questionnaire, there was only a small gain in weight, no reported improvement in other gastrointestinal symptoms, and mixed changes in the overall number of hospitalizations over that period.We believe this is the first study to investigate growth parameters, including head circumference in CDD, using a relatively large sample of 353 individuals.The age trend analysis could be affected by the lack of longitudinal data, given that only a proportion of families could complete a follow-up questionnaire or provide previous measurements.Nevertheless, the cross-sectional data were supplemented with historical measurements reported by the families in a separate but related question.As a result, the average number of measures per individual was around two.We are unsure of the exact proportion of families who took the measurements themselves rather than using those obtained at clinic or hospital visits.However, there is no reason to believe that family measurements were systematically biased.
Nearly one-third of our group had gastrostomy, a similar proportion to the 30% of children with Rett syndrome with gastrostomy in the Australian Rett Syndrome Database 19 where poor growth is recognized as a feature of the disorder. 18CDD is characterized by complex multimorbidity and a combination of symptoms including difficulties with feeding and refractory epilepsy (necessitating the administration of multiple antiseizure medications) 1 which, in additional to poor growth, could have F I G U R E 3 Scatterplot and trend of BMI-for-age z-score over age in 288 children and adults with CDKL5 deficiency disorder, overall and by gastrostomy status.Note: linear spline with knots at 1, 2, 4, 7, and 13 years.Abbreviations: BMI, body mass index; CDKL5, cyclin-dependent kinase-like 5. indicated the need for gastrostomy.A gastrostomy may also be provided to assist in the delivery of the ketogenic diet for seizure management in CDD. 16Although the age at gastrostomy had not been provided for all individuals, which would have been optimal, there was a small subset whose families completed a follow-up questionnaire where gastrostomy placement had occurred between the two questionnaires, ideally providing a valid method of assessing its effects.However, the analysis was limited by the small sample size, and we are uncertain of its representativeness.We elected to use Cole's British 1990 growth reference centiles for weight, height, BMI, and head circumference as our normative data because the US Centers for Disease Control and Prevention growth charts 22 do not provide information for head circumference beyond the age of 3 years, or for height and BMI below the age of 2 years.The choice was also partly related to our interest in understanding growth in children younger than 2 years, a period for which more intensive clinical follow-up can be provided.We recognize that there is a need for disorderspecific growth curves similar to those of Rett syndrome to enable effective screening and monitoring, and further research is warranted.To our knowledge, there have been no other studies evaluating growth in CDD specifically or the impact of gastrostomy.However, in Rett syndrome, growth charts have been  generated from US data also using Cole's British 1990 growth reference centiles. 18Australian population-based data for Rett syndrome have also shown that those with suboptimal weight who had a gastrostomy placed had a higher BMI but not an improved weight or height than those with no gastrostomy. 19This was an important reason for also including BMI in this current study given we had used weight for age rather than weight for height in our analysis and this was somewhat validated by our findings.Previous qualitative research has extolled positive effects of gastrostomy placement in Rett syndrome. 23In another qualitative study (n = 21), parents of children with intellectual disability have described improved child health and community participation, along with easier care routines for themselves enabling their better community participation. 24In CDD gastrostomy has been shown to have varying effects on caregiver health and was associated with better emotional but not physical well-being. 17The long-term impact of gastrostomy placement on both individuals with CDD and their families still remains unclear.
Head growth deceleration following a normal head circumference at birth was one of the original necessary criteria for Rett syndrome. 25In our 2012 study where we compared the characteristics of 86 individuals with pathogenic CDKL5 variants with 920 females with MECP2 mutations, we found similar proportions with evidence of head growth deceleration (63.3% vs 59.5%) although only 3 out of 42 with CDD would be classified as having microcephaly. 7Our current findings show some consistency although we have a higher proportion with microcephaly with z-scores somewhat lower than in a recent comparison of CDD (n = 53) with three other developmental encephalopathies. 26There the head circumference z-score was lowest in FOXG1 followed by Rett syndrome, then CDD, and closest to normal in MECP2 duplication disorder.Otherwise, the only research we could find relating to head growth in other developmental epileptic encephalopathies was a study of 139 patients with Dravet syndrome which also found that there was head growth deceleration after the age of 24 months. 27It is difficult to know whether the pattern we are seeing with head growth is a reflection of the overall growth pattern of height and weight or whether it is acting independently.However, early evidence from neuroimaging in a small number of patients with CDD suggests a monotonic downward trend of relative grey matter volume in parallel with the disease course, 28 evidence supported by an even more recent study. 29Moreover an animal study published in 2002 suggested that antiseizure medication-induced apoptotic neurodegeneration could be a possible mechanism to explain continuing reduction in brain mass associated with early postnatal exposure to antiepileptic therapy. 30Both scenarios are possible and would be consistent with evidence of a true growth abnormality.
a Between the time of follow-up and baseline questionnaires.
b Between the 1-year period before and after gastrostomy.
Our study shows that in CDD there is a decline in weight and height for age after the age of 4 years, with the decline most pronounced in those aged 13 years and over.As expected, there was much less of a decline from the norm with BMI.The dissociation between simple and composite measures indicates that height growth declined faster than weight, and the pattern was similarly observed in Rett syndrome. 18Relative to the norm, the anthropometric measures were generally lower in the gastrostomy group compared with the non-gastrostomy group, and a probable benefit of gastrostomy in terms of narrowing the gap in standardized weight and BMI was only noted in the early years and was not further sustained.However, there may also be other longer-term effects (benefits and/or side effects) of gastrostomy not identifiable by our study design.We acknowledge the potential difficulty today in further assessing the impact of gastrostomy on nutrition in a randomized control trial.The pattern with head circumference differed from the other parameters with an initial decline which stabilized in adolescence.However, with a median value of more than two standard deviations below the norm this is indicative of microcephaly and could be considered a new recognized feature of this disorder, especially in adults.

AC K NO W L E D GE M E N T S
We express our thanks and appreciation to all the families who have participated in the International CDKL5 Disorder Database.

F U N DI NG I N FOR M AT ION
The International Foundation for CDKL5 Research for support of the International CDKL5 Disorder Database and from the Orphan Disease Centre, University of Pennsylvania (CDKL5-19-D-101-3) for analytical support.

C ON F L IC T OF I N T E R E S T S TAT E M E N T
Both HL and JD have consulted for Avexis, Anavex, GW, Newron, and Neurogene on unrelated subject matter.They have funding from the National Institutes of Health related to this subject matter.They have consulted for Ovid Therapeutics and Marinus on related subject matter.JD has consulted for Acadia, Ultragenyx, and Taysha for related subject matter.SA has received funding from GW Pharmaceuticals, Norvartis, PTC Therapeutics, Boston Scientific, Nutricia, UCB, BioMarin, LivaNova, Medtronic, Desitin, Ipsen, Orion, CDKL5 UK, TSA, and the National Institute for Health Research.

SU PP ORT I NG I N FOR M AT ION
The following additional material may be found online:

F I G U R E 2
Scatterplot and trend of length/height-for-age z-score over age in 334 children and adults with CDKL5 deficiency disorder, overall and by gastrostomy status.Note: linear spline with knots at 1, 2, 4, 7, and 13 years.Abbreviation: CDKL5, cyclin-dependent kinase-like 5.

F I G U R E 4
Scatterplot and trend of head circumference-for-age z-score over age in 235 children and adults with CDKL5 deficiency disorder, overall and by gastrostomy status.Note: linear spline with knots at 1, 2, 4, 7, and 13 years.Abbreviation: CDKL5, cyclin-dependent kinase-like 5.

Figure S1 :
Gastrostomy status flow chart Figure S2: Boxplot of weight-for-age z-score in 338 children and adults with CDD, by age group and gastrostomy status Figure S3: Scatterplot and trend of raw weight value over age, overall and by gender (including the British 1990 growth reference) Figure S4: Boxplot of length/height-for-age z-score in 333 children and adults with CDD, by age group and gastrostomy status Figure S5: Scatterplot and trend of raw length/height value over age, overall and by gender (including the British 1990 growth reference) Figure S6: Boxplot of BMI-for-age z-score in 278 children and adults with CDD, by age group and gastrostomy status Figure S7: Scatterplot and trend of raw BMI value over age, overall and by gender (including the British 1990 growth reference) Figure S8: Boxplot of head circumference-for-age z-score in 232 children and adults with CDD, by age group and gastrostomy status Figure S9: Scatterplot and trend of raw head circumference value over age, overall and by gender (including the British 1990 growth reference) Characteristics of the 353 children and adults with CDD.
T A B L E 1a Nine families completed the initial questionnaire but did not provide any anthropometric data.b c Includes families from Ireland, Hungary, India, Italy, New Zealand, Spain, Belarus, Belgium, Bulgaria, Luxembourg, Mexico, Norway, Poland, Saudi Arabia, and Turkey.

n 1 = 349, n 2 = 338 a ) Length/height, cm (n 1 = 343, n 2 = 333 a ) Raw value
Distribution of weight, length/height, BMI, and head circumference in 353 children and adults with CDD, by sex, age group, and gastrostomy status.
T A B L E 2
Mean (SD), n Mean (SD), n Mean (SD), n Mean (SD), n Mean (SD), n Mean (SD), n a Within the British 1990 growth reference age range.b

Table S1 :
Characteristics of the 353 children and adults with CDKL5 deficiency disorder, by inclusion/exclusiona status (weight and length/height)

Table S2 :
Characteristics of the 353 children and adults with CDKL5 deficiency disorder, by inclusion/exclusion status (BMI and head circumference)

Table S3 :
Characteristics of the 96 CDKL5 deficiency disorder children and adults with a gastrostomy insertion