Linear Growth in Children With Cystic Fibrosis in the Netherlands Born Between 1997–2004; Results of a Multicenter Cohort Analysis

ABSTRACT

Objectives

Short stature has been associated with reduced life expectancy in people with Cystic Fibrosis (pwCF). We aimed to evaluate linear growth and final height in a Dutch cohort of children with CF, diagnosed in early childhood and now aged ≥ 18 years and identify risk factors for impaired linear growth.

Methods

A multicenter longitudinal retrospective cohort study was performed in pwCF born between 1997 and 2004, before implementation of newborn screening (NBS). Anthropometric measurements and CF‐related risk factors for poor growth (pulmonary infections, malnutrition, CF‐Related Diabetes [CFRD], CF‐related liver disease [CFLD]) were obtained annually from ages 0.5 to 10 years and biannually from ages 10 to 18. Measurements were converted to Height‐For‐Age‐For‐Target‐Height (HFA‐TH) Z‐scores. Differences in HFA‐TH Z‐scores between pwCF and healthy standards, and risk factors associated with linear growth were analyzed.

Results

A total of 128 pwCF (60 males) were included. Most patients did not receive modulator‐therapy during pubertal growth. In boys, mean HFA‐TH Z‐scores at age 18 years (HFA‐TH₁₈) were lower in comparison to healthy standards (−0.66 [0.96], p < 0.001). In girls at age 18, a normal mean HFA‐TH z‐score was found (−0.18 [0.78]). Development of CFRD and a greater change in BMI Z‐scores between 0.5 and 6 years of age (ΔBMI_0.5–6) were associated with lower HFA‐TH₁₈ Z‐scores in boys. In both sexes, pulmonary function and BMI Z‐scores were positively associated with linear growth.

Conclusion

Boys with CF may have impaired final height, especially those with CFRD or a ΔBMI_0.5–6. Glucose metabolism and nutritional status should be monitored closely in pwCF, as these factors may contribute to impaired linear growth.

Abbreviations

BMI

body mass index

CF

cystic fibrosis

CFLD

cystic fibrosis‐related liver disease

CFRD

cystic fibrosis‐related diabetes

EMM

estimated marginal means

FEV1%pp

percentage of predicted forced expiratory volume in 1 s

HFA

height‐for‐age

HFA‐TH

height‐for‐age for target‐height

LMM

linear mixed models

pwCF

people with cystic fibrosis

TH

target height

Z‐scores

standard deviation scores

1. Introduction

Nutrition is an important pillar of cystic fibrosis (CF) treatment. Malnutrition in people with CF (pwCF) may be due to malabsorption, increased energy expenditure, or illness [1]. Several studies have shown improved outcomes and survival in well‐nourished pwCF in comparison to those with malnutrition [2, 3, 4, 5].

Not only body weight, but also body height is important for pwCF. Height not only serves social purposes, but impaired height has also been associated with smaller lung volumes [6] and reduced life expectancy in pwCF [7]. Poor growth velocity and short stature may be related to pulmonary infections, malnutrition, CF‐related diabetes (CFRD), or CF‐related liver disease (CFLD) [8, 9].

In a previous single center study, we described final height in boys with CF to be impaired (mean height‐for‐age for target height [HFA‐TH] Z‐score of −0.57) with a possible association between early body mass index (BMI)‐increase (between 1 and 6 years of age) and final height [10]. In an additional study, we explored the association of early adrenal activation to early BMI‐increase, and hypothesized that this might explain the limited pubertal growth spurt and final height [11].

Here, we aim to evaluate growth and BMI trajectories (from age 0.5 to 18 years) and final height (after achieving the age of 18) in a larger, national cohort of boys and girls with CF. Secondary, we aimed to assess risk factors for deprived final height, including early BMI‐increase and CF‐related diseases.

2. Methods

2.1. Study Population and Design

A multicenter longitudinal retrospective cohort study was conducted in pwCF born between 1997 and 2004. All children had received medical care in the Wilhelmina Children's Hospital, Utrecht, ErasmusMC‐Sophia Children's Hospital, Rotterdam, or MosaKids Children's Hospital, Maastricht in The Netherlands. CF diagnosis was based on clinical presentation and confirmed by sweat‐chloride tests and genetic testing.

Only patients with severe CF genotype (two class I, II, or III mutations) were included. Patients were excluded if final height (defined as the highest height achieved after a < 2 mm increase in height in the last 6 months) was not yet achieved at age 18 or if the biological height of the parents was unknown. Other exclusion criteria were the presence of non‐CF‐related comorbidities that might cause growth retardation, such as congenital heart‐disease.

Height and weight measurements were obtained annually during regular clinical visits from the moment of CF diagnosis until the age of 10 and biannually from the age of 10 until age 18 at the outpatient clinic.

2.2. Definitions

2.2.1. Anthropometric Measurements

Until the age of 2 years, length was measured with a measuring table in the lying position. After age 2 years, height was measured using a stadiometer to an accuracy of 0.1 cm. Body weight was measured using digital scales to the nearest 0.1 kg.

Standard deviation score values (Z‐scores) for height‐for‐age (HFA) and BMI were determined according to the latest Dutch nationwide growth study (2010) [12]. To account for the genetic growth potential, the height of biological parents were used to calculate the target height (TH), and TH Z‐scores of the patients were determined with formulas, based on the nationwide Dutch Growth study [13]. Finally, the Z‐scores for HFA‐accounted‐for‐target‐height (HFA‐TH) were calculated by subtracting the TH Z‐scores from the HFA Z‐scores [14].

The HFA‐TH at age 18 (HFA‐TH₁₈) was used as the final height. Change in BMI Z‐scores during early childhood was defined as a change in BMI Z‐score between 6 months and 6 years (ΔBMI_0.5–6). Rapid BMI‐gain was defined as an increase in BMI Z‐score of > 1, between 6 months and 6 years of age.

2.2.2. Clinical Variables

Pulmonary function was assessed by spirometry using the percentage of predicted forced expiratory volume in 1 s (FEV1%pp). The highest measured FEV1%pp in the period 3 months before or after the annual anthropometric measurements was used for analysis.

Pancreatic insufficiency was defined as the use of pancreatic enzyme replacement therapy.

Prolonged glucocorticoid use was defined as use of glucocorticoids for a period of four consecutive weeks, oral or IV, in a period of 1 year.

CFRD was Defined Present if (1) oral glucose tolerance test after 120 min exceeded 11.1 mmol/L [15] or (2) exceeded 7.8 mmol/L and was afterwards confirmed by continuous glucose monitoring during a 72 h period with glucose levels exceeding 7.8 mmol/L > 10% or (3) as use of insulin was administered.

CFLD was defined using the European diagnostic criteria from DeBray: abnormal physical exam (hepatomegaly or splenomegaly), abnormal liver serologies, and/or abnormal abdominal ultrasound (two out of three) [16]. Children treated with ursodeoxycholic acid (Ursochol) due to persistently elevated transaminase levels were also included in the CFLD group, as this was considered to reflect some degree of liver disease activity.

These clinical variables were obtained longitudinally, and their presence or absence (yes/no) was defined at each time point using a 3‐month window before or after the anthropometric measurements.

In girls, age at menarche was obtained from the medical records, if available. For boys, no pubertal measurements could be retrieved from the medical records.

Underweight, overweight, and obesity were classified according to the international cut‐offs by Cole et al. [17, 18].

2.2.3. Statistics

Descriptive statistics were used to summarize the baseline characteristics of our study population. To analyze differences in continuous variables between sexes, the independent sample's t‐test or Mann−Whitney U test was used, depending on normality. Sex differences for categorical variables were tested with the chi‐squared‐test. The one sample t‐test was used to analyze the difference between HFA‐TH during childhood and adolescence between CF children and the healthy Dutch standards. Distribution of normality was assessed graphically and by the Shapiro−Wilk non‐parametric test.

Associations between HFA‐TH₁₈ and relevant clinical parameters, including ΔBMI_0.5–6 were tested by univariate regression analyses.

Linear mixed models (LMM) were performed for HFA‐TH Z‐scores to determine relevant predictors. Age was used as a fixed covariate in all models, with variables such as FEV1%pp, BMI Z‐scores, presence of CFRD and CFLD, and prolonged glucocorticoid use as fixed covariates separately. An interaction term of “age*variable” was added to account for the effect of age on the variable. The use of the interaction term in the model was based on model validation by comparing the likelihood ratios. If statistically significant, the independent variables were combined into a new model. Depending on the outcome, the effect of the variables on the HFA‐TH₁₈ Z‐score and the estimated marginal means (EMM) for HFA‐TH₁₈ were analyzed.

For statistical significance, the threshold was considered as p‐value < 0.05. Statistical analyses were conducted using the Statistical Package for the Social Sciences Computer Software (SPSS Inc. Version 27.0.0.0; IBM, Chicago, IL).

2.3. Ethical Regulations

Informed consent was obtained locally at each participating center for the use of regular care data for scientific research purposes. The Act on Medical Research Involving Human Subjects did not apply, and a waiver was provided.

3. Results

3.1. Patient Characteristics

In total, 161 patients with CF, born between 1997 and 2004, were eligible for inclusionat the three inclusion sites. Fourteen patients were excluded because of having a mild or unclassified CF genotype, another seven patients had not reached final height at age 18, six patients because growth measurements during childhood and adolescence could not be retrieved from the medical records, four had relevant non‐CF‐related comorbidities that caused growth retardation, such as coeliac disease or growth‐hormone deficiency and one patient had died before reaching the age of 18. In total, 128 pwCF (46.9%, male) could be included in our study. See Table 1 for patient's characteristics.

Table 1.

Patient's characteristics of the study population.

	Boys (n = 60)	Girls (n = 68)	p value
Age at diagnosis, yrs (median [IQR])	0.30 (0.10–0.67) (n = 54)	0.30 (0.10–2.07) (n = 64)	0.42c
Birth weight Z‐score (mean, [SD])	0.013 (0.72) (n = 21)	−0.25 (0.92) (n = 25)	0.29e
Genotype			0.037d
– Homozyguous F508del	50 (83.3)	45 (66.2)
– Heterozyguous F508del	10 (16.7)	19 (27.9)
– Other	0 (0)	4 (5.9)
Mode of diagnosisa			0.76d
– Failure to thrive	27 (45.0)	25 (36.8)
– Respiratory symptoms	16 (26.7)	19 (27.9)
– Meconium ileus	9 (15.0)	12 (17.6)
– Siblings with CF	2 (3.3)	4 (5.9)
– Gastro‐intestinal symptoms	1 (1.7)	3 (4.4)
Pancreatic insufficiency	60 (100)	66 (97.1)	0.18d
Prolonged glucocorticoid useb	16 (26.7)	24 (35.3)	0.29d
CFRD diagnosis < 18 yrs	22 (36.7)	30 (44.1)	0.39d
CFLD diagnosis < 18 yrs	29 (58.3)	29 (42.6)	0.46
FEV1%pp at 18 yrs (mean, (SD))	82.0 (18.5) (n = 57)	83.5 (15.1) (n = 67)	0.62e
CFTR‐modulator use < 18 yrs	32 (53.3)	31 (45.6)	0.38d
Mean age at menarche (mean, (SD])	—	13.3

Note: Values are counts and percentages (in brackets), unless otherwise stated.

Abbreviations: CFLD, cystic fibrosis‐related liver disease; CFRD, cystic fibrosis‐related diabetes; FEV1%pp = percentage of predicted forced expiratory volume in 1 s; Yrs = years.

^a Mode of diagnosis was missing in 10 patients (5 male, 5 female).

^b Defined as use of glucocorticoids for a period of four consecutive weeks oral or IV in a period of 1 year.

^c The Mann−Whitney U test used to differentiate in median numerical values between boys and girls.

^d The Chi‐square test used to differentiate in categorical variables between boys and girls.

^eThe independent Sample's t‐test used to differentiate in mean numerical values between boys and girls.

3.2. Linear Growth and BMI‐Trajectories

3.2.1. HFA‐TH Z‐Scores Compared Between Boys and Girls

HFA‐TH Z‐scores at diagnosis were low in boys and girls (Table 2). Mean HFA‐TH Z‐scores improved after diagnosis. At ages 1, 14, and 18, HFA‐TH Z‐scores in boys were lower than in girls. Height Z‐scores of the parents were included in the table to provide context for the HFA‐TH Z‐scores of the patients.

Table 2.

HFA‐TH Z‐scores in boys and girls at diagnosis, 1 year, mid‐childhood, mid‐adolescence, and at age 18, and height Z‐scores of parents and at age 18.

	Boys (n = 60)	Girls (n = 68)	p valuea
HFA‐TH Z‐score at diagnosis	−1.41 (1.27) (n = 35)	−1.13 (1.31) (n = 41)	0.35d
HFA‐TH Z‐score 1 yrs (median, IQR)b	−0.65 (−2.11 to −0.20) (n = 44)	−0.53 (−0.90 to 0.27) (n = 49)	0.019a, c
HFA‐TH Z‐score 10 yrs	−0.50 (0.67) (n = 53)	−0.36 (0.96) (n = 61)	0.37d
HFA‐TH Z‐score 14 yrs	−0.47 (0.90) (n = 58)	−0.11 (0.80) (n = 64)	0.019a, d
HFA‐TH Z‐score 18 yrs	−0.66 (0.96)	−0.18 (0.78)	0.002a, d
HFA Z‐score 18 yrs	−0.79 (1.02)	−0.28 (1.16)	0.008d
Height Z‐score father	−0.05 (0.98)	−0.16 (1.01)	0.54d
Height Z‐score mother	−0.26 (0.88)	−0.07 (1.10)	0.27d

Note: Values are means and SDs, unless otherwise stated.

Abbreviations: IQR, interquartile range; HFA‐TH, height‐for‐age for target‐height; Yrs, years.

^aIndicates significance.

^b The median was used as the HFA‐TH Z‐scores at 1 year were non‐normally distributed.

^c The Mann−Whitney U test was used to differentiate the median HFA‐TH Z‐scores between boys and girls.

^d The independent sample's t‐test was used to differentiate the mean HFA‐TH Z‐scores between boys and girls.

3.2.2. Linear Growth in Comparison With the Dutch Average (Figure 1)

Figure 1.

Growth trajectories (longitudinal mean Z‐scores for height‐for‐age‐for‐target‐height [HFA‐TH]) in boys and girls with CF. [Color figure can be viewed at wileyonlinelibrary.com]

In boys, the mean and median HFA‐TH Z‐scores (at ages 1, 15, and 15.5) were statistically significantly lower than the national Dutch average (p < 0.001 for all ages). At age 0.5 years, the mean (SD) HFA‐TH Z‐score was impaired (−1.35 [1.19]), but subsequently increased, with fluctuations, until the age of 11, with a mean (±) HFA‐TH Z‐score of −0.38 (0.66). From this age onward, however, the HFA‐TH Z‐scores gradually decreased again, with minor fluctuations between 12.5 and 14 years. The HFA‐TH₁₈ Z‐score, was −0.66 (0.96), which was significantly lower than the national Dutch average (p < 0.001).

In girls, similar as in boys, the mean (SD) HFA‐TH Z‐score was impaired at age 0.5 years (−0.88 [1.05]). HFA‐TH Z‐scores increased until the age of 2 years, then decreased until the age of 5, to increase again until 9 years of age. From the age of 11, mean HFA‐TH Z‐scores increased until age 14.5 (−0.07 [0.75]). From that age onwards, mean HFA‐TH Z‐scores decreased slightly until a mean HFA‐TH₁₈ Z‐score of −0.18 (0.78). Mean HFA‐TH Z‐scores at ages 1, 5–12, and at age 13 years were statistically significantly lower than the national Dutch Average (all p < 0.05).

3.3. BMI Changes in Time

In the first year of life, BMI Z‐scores were found to be significantly impaired both for boys and girls (Figure 2). Both sexes showed a steep increase to normal BMI Z‐scores in young childhood (Figure 2). Despite this steep increase, all BMI Z‐scores in childhood and adolescence, except for BMI Z‐scores at ages 2, 3, and 10.5 years in boys, and at ages 2, 3, 4, and 6 in girls, respectively, were significantly lower than the national average. After the age of 3 in boys and 2 in girls, the BMI Z‐score fluctuated to eventually a mean (SD) BMI Z‐score at age 18 years of −0.46 (0.85) in boys and −0.35 (0.35) in girls, respectively. At the last moment of follow‐up, 3.3% (n = 2) of the boys and 1.5% (n = 1) of girls were underweighted, 1.7% (n = 1) of boys and 1.5% (n = 1) of girls were overweighted, and none of the children were obese.

Figure 2.

BMI‐trajectories (Z‐scores) in boys and girls with CF. [Color figure can be viewed at wileyonlinelibrary.com]

3.4. Risk Factors Associated With Decreased Linear Growth and Impaired Final Height

In boys, a significant association was found between a greater ΔBMI_0.5–6 and HFA‐TH₁₈ Z‐scores (β = −0.24, p = 0.038). In addition, in boys with rapid BMI gain between 0.5 and 6 years (n = 15), the mean HFA‐TH₁₈ Z‐score was lower than in boys without such rapid BMI gain (n = 19) (−0.96 vs. −0.25, p = 0.02). Girls with rapid BMI gain between 0.5 and 6 years (n = 17) also had lower HFA‐TH₁₈ Z‐score than girls without rapid BMI gain, however this difference was not significant (p = 0.21) (Table 3). The growth trajectories of boys and girls with and without rapid BMI gain are shown in the supplemental data.

Table 3.

Differences in mean HFA‐TH18 Z‐scores in children with and without rapid BMI‐gain.

	Rapid BMI gaina 0.5–6
	Yes		No		p valueb
Boys
HFA‐TH18 Z‐score	−0.96 (1.07) (n = 15)		−0.25 (0.64) (n = 19)		0.02
BMI6 m Z‐score	−2.24 (0.85) (n = 15)		−0.67 (0.87) (n = 19)		< 0.001
ΔBMI0.5−6, median (IQR)	1.58 (0.27)		0.23 (0.93)		< 0.001
Girls
HFA‐TH18 Z‐score	−0.32 (0.79) (n = 20)		0.05 (0.60) (n = 14)		0.15
BMI6 m Z‐score	−1.94 (1.14) (n = 20)		0.51 (0.92) (n = 14)		< 0.001
ΔBMI0.5−6, median (IQR)	1.63 (0.63)		−0.14 (1.42)		< 0.001

Note: Values are means and SDs, unless otherwise stated.

Abbreviations: BMI, body mass index; HFA‐TH, height‐for‐age for target height; yrs, years.

^aRapid BMI gain was defined as a BMI Z‐score increase ≥ 1.

^b The independent sample's t‐test was used to differentiate the mean HFA‐TH Z‐scores between the children with and without rapid BMI gain.

In both sexes, BMI Z‐scores at the age of 6 months (BMI_6m Z‐scores) were significantly lower in children with rapid BMI gain_0.5–6 than without rapid BMI gain_0.5–6. In children with and without rapid BMI gain, no associations were found between BMI_6m Z‐scores and HFA‐TH₁₈ Z‐scores, There were no differences in ΔBMI_0.5–6 Z‐scores between boys and girls with rapid BMI‐gain or without rapid BMI‐gain (results not shown).

3.4.1. Factors Affecting Growth in Boys

In LMM analysis, the presence of CFRD was a negative predictor for the HFA‐TH₁₈ Z‐scores in boys. The EMM (SE, [95% CI]) of HFA‐TH₁₈ Z‐scores was lower in boys with CFRD (−0.81 [0.20, (−1.20 to −0.42)]) vs. boys without CFRD (−0.53 [0.19, (−0.91 to −0.15)]), corrected for the mean FEV1%pp and BMI Z‐scores at age 18 (p < 0.001). In addition, FEV1%pp and BMI Z‐scores were positively associated to HFA‐TH Z‐scores (FEV1%pp: β = 0.005, p < 0.001, BMI Z‐score: β = 0.061, p = 0.012).

The presence of CFLD was a positive predictor for the HFA‐TH Z‐scores in boys. However, this effect was redundant after the age of 11, when the HFA‐TH Z‐scores did not differ significantly between boys with or without CFLD. When adjusting for the presence of CFRD, the effect of the presence of CFLD became insignificant.

Boys with CFRD had a greater mean ΔBMI_0.5–6 than boys without CFRD (1.39 [1.7] vs. 0.66 [1.1]), however this difference was not‐significant (p = 0.12). In logistic regression analysis, the presence of rapid BMI‐gain did not have effect on the presence of CFRD (odds ratio 1.89, 95% CI 0.47–7.70, p = 0.37).

No associations were found between prolonged corticoid use and the HFA‐TH Z‐scores.

3.4.2. Factors Affecting Growth in Girls

The presence of CFLD was a positive predictor for the HFA‐TH Z‐scores (β = 0.50, p = 0.003) until the age of 15, but this effect decreased with older age (β age*CFLD = −0.026, p = 0.03). After this age, HFA‐TH Z‐scores did not differ significantly between girls with and without CFLD. The FEV1%pp and BMI Z‐scores were positively associated with the HFA‐TH Z‐scores (FEV1%pp: β = 0.005, p < 0.001, BMI Z‐score: β = 0.119, p < 0.001).

No associations were found between the presence of CFRD, prolonged corticoid use, or menarche age and the HFA‐TH Z‐scores.

4. Discussion

Our results confirm earlier findings by Hak et al. [10] that Dutch boys with CF have impaired final height. In our current study, we have found an association between a greater ΔBMI_0.5–6 and decreased HFA‐TH₁₈ Z‐scores in boys. Furthermore, boys with rapid BMI gain showed lower mean HFA‐TH₁₈ Z‐scores and BMI_6m Z‐scores than boys without rapid BMI gain. Our results provide new data on growth trajectories of pwCF in relation to BMI and may help to prevent decreased final height in boys, which may not only be relevant for pulmonary function and life expectancy, but also may have psychosocial effects.

Ensuring optimal longitudinal growth may be of special importance for pwCF, as it is a significant marker of their overall health status. Previous research has shown that lower HFA during the toddler years is associated with lower pulmonary function later in life and that short stature is a prognostic factor for higher mortality [7, 19]. Lastly, short stature may have negative psychosocial effects, especially in men. Therefore, monitoring and maintaining adequate growth are important aspects of good medical care.

In girls, in contrast to the boys, the mean final height did not differ from the Dutch national average, and no differences in mean final height with or without rapid BMI gain were found.

PwCF are known to have higher insulin resistance than the general population [20]. BMI‐increase in early childhood is associated with insulin resistance at the age of 8 years [21]. Our results show that a greater ΔBMI_0.5–6 negatively influences final adult height in boys with CF. We hypothesize that this may be caused by higher insulin levels, caused by the previously described insulin resistance, in combination with early BMI‐increase, which may in turn activate the adrenal glands at an early age and lead to precocious production of androgens [21]. Androgens, in turn, are converted to estrogens in adipose tissue and accelerate the maturation of growth plates in long bones, potentially resulting in premature epiphyseal closure leading to suboptimal linear growth. Our earlier findings, that higher androgen levels (DHEAS) are found in boys with CF between 8 and 9 years of age with increased bone age [11], confirm this hypothesis. Also, boys with rapid BMI increase show earlier completion of growth, whereas boys without rapid BMI increase tend to continue growing for a longer period (see Supporting Information S1: Figure 1), which also supports this hypothesis.

Presence of CFRD was negatively associated with linear growth in boys, resulting in lower mean HFA‐TH₁₈ Z‐scores than in boys without CFRD. These results are consistent with earlier research showing lower height velocity and final height in patients with CFRD than without CFRD [22, 23, 24]. As the prevalence of CFRD is similar in boys as in girls, it may be that boys are longer exposed to the effects of increased insulin and androgen levels compared to girls [11].

Presence of CFLD was positively associated with linear growth in both sexes, however no effect was seen on the HFA‐TH₁₈ Z‐scores, indicating that the effect of CFLD on linear growth primarily exhibits during earlier stages of growth. This unexpected finding might be due to the small number of children with CFLD at a younger age, as the difference in HFA‐TH Z‐scores at older ages became smaller or insignificant.

FEV1%pp and BMI Z‐scores were positively associated with linear growth in both boys and girls, emphasizing that, as may be expected, overall health is important for pwCF.

Previous studies have been conflicting regarding longitudinal growth in pwCF. Nonetheless, our results broadly align with some of the previous research. Fairly similar to our findings, two studies demonstrated lower final height in both sexes [25, 26]. Both studies showed lower peak height velocity compared to references, especially in boys, resulting in a more pronounced lower final height in boys in comparison to girls. While we did not analyze peak height velocity, our results imply the absence of a pubertal height spurt in boys by analyzing the mean HFA‐TH Z‐scores, while the mean HFA‐TH Z‐scores increased in girls during adolescence (Figure 1). This timing corresponds well with the mean age at menarche in our cohort (13.3 years), suggesting that the observed growth patterns align with normal pubertal development. Other research also reported lower final height in pwCF combined for boys and girls [27, 28]. Only one of these studies took genetic potential into account and reported a mean final height Z‐score of −0.5 [28]. On the contrary, normal adult height in comparison to references was also found in pwCF [29, 30]. Notably, the patients reached their final height at a later age than the references.

The majority of previous CF cohorts concerning linear growth in pwCF reported on pwCF born in the early nineties. Our cohort exists of pwCF born in the late nineties and later. Despite improvement in medical care, the final adult height in boys is still impaired. The fact that we took the genetic potential of these children into account and adjusted the HFA Z‐scores for the TH makes this study unique and valuable.

This study also has some limitations. Firstly, this study has a relative small cohort size as we could not include all CF‐centers in the Netherlands and not all available patients were eligible for inclusion. Secondly, pubertal stages of our patients could not be systematically obtained from the patients' records to analyze the timing of puberty. By analyzing the timing of puberty, we could have tested our hypothesis on the difference in final height between boys and girls. Thirdly, the retrospective collection of data may be considered another limitation because anthropometric measurements were not available at all time points for each patient, resulting in missing data, especially in very young children. In addition, due to the retrospective observational design of the study, causality could not be established. Furthermore, the use of inhaled glucocorticoids could not be accurately collected from the patient records, making the data on their use unreliable. A further consideration is that newborn screening (NBS) for CF was introduced in the Netherlands in 2011, and all individuals in this cohort were born before its implementation. In more recent times, being diagnosed through NBS could have an impact on the growth, especially during early childhood. Finally, it should be acknowledged that CFTR modulators became available during the follow‐up period of this cohort. While a subset of participants received treatment, initiation generally occurred around or after the onset of puberty. Therefore, potential effects on growth could not be meaningfully assessed, and no specific analyses were performed. It is also known that CFTR modulators have positive effects on weight, so therefore it would be useful to also monitor growth patterns and the effects on BMI‐increase during early childhood in cohorts who receive modulator therapy from a very young age. Despite our data reflecting patients born before NBS and before the availability of CFTR‐modulators from an early age on, there are still patients who are not eligible for or do not have access to modulator‐therapy. These data could be important for this subset of patients, and further research on growth and its relation with early rapid BMI‐increase should include these patients as well.

While we underline the potential impact of rapid BMI‐increase during early childhood on linear growth, it should be acknowledged that an adequate nutritional status remains crucial for other CF‐related outcomes, such as pulmonary function and associated morbidity. Future studies in larger cohorts are needed to further examine the associations between linear growth, BMI‐increase in early childhood, and early adrenarche to support our hypothesis before drawing any definitive conclusions.

In conclusion, in our cohort with patients born before NBS and the pre‐modulator era during young childhood, boys with CF show impaired final height, preceded by suboptimal linear growth during adolescence. A greater ΔBMI_0.5–6 is related to lower HFA‐TH₁₈ Z‐scores, and the presence of CFRD is negatively related to linear growth, resulting in lower final adult height in boys. In both sexes, BMI Z‐scores and FEV1%pp are positively related to linear growth, emphasizing the importance of overall good health and nutritional status during early childhood. We recommend careful monitoring of glucose metabolism for the presence of CFRD and nutritional status, especially during early childhood, to maintain a balance between adequate nutrition and avoiding a too rapid BMI increase, as both may influence linear growth. With recent therapeutical advancements, such as NBS and CFTR‐modulator therapies, we expect to see further improvement in the overall health status of pwCF. It remains uncertain if such a rapid BMI increase is still relevant in the context of current care practices. However, not all patients can benefit from these therapeutical advancements. Therefore, future studies may focus on longitudinal follow‐up of pwCF diagnosed through NBS or treated with CFTR‐modulators from early childhood on, but also on patients not eligible for modulator‐therapy, with regard to linear growth in relation to BMI, and other disease or non‐disease related factors.

Author Contributions

Gizem Tamer: conceptualization, investigation, writing – original draft, methodology, writing – review and editing, validation, visualization, project administration, formal analysis, data curation. Hanneke van Santen: conceptualization, writing – original draft, methodology, writing – review and editing, resources, supervision. Michiel Bannier: resources, project administration, writing – original draft, writing – review, and editing. Hettie Janssens: resources, writing – original draft. Badies Manai: project administration, resources, writing – original draft. Rosanne Swolfs: writing – original draft, investigation. Kors van der Ent: writing – review and editing, conceptualization, writing – original draft, supervision, resources, methodology. Bert Arets: conceptualization, writing – original draft, methodology, supervision, resources, writing – review and editing. Hetty van der Kamp: conceptualization, writing – original draft, writing – review and editing, methodology, supervision, resources.

Funding

The authors received no specific funding for this work.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supplemental figure manuscript growth in children with CF.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.