Phenotypic characterization of X-linked hypophosphatemia in pediatric Spanish population

X-linked hypophosphatemic rickets (XLH) (OMIM 307800) (ORPHA 89936) is the most common hereditary rickets [1‐5] with an estimated prevalence of 1:20,000 [6, 7]. It follows an X-linked dominant transmission [8]. The disease is caused by a defective function of PHEX gene [1, 9‐13], leading to elevated circulating concentrations of fibroblast growth factor 23 (FGF23) [14], relatively low levels of 1,25 dihydroxyvitamin d [1,25(OH)₂D], hyperphosphaturia secondary to decreased proximal tubular reabsorption of phosphate and hypophosphatemia [8, 10, 15]. Classical, conventional treatment of XLH is based on the administration of phosphate supplements and 1-alpha hydroxylated derivates of vitamin D [16]. The wider availability of genetic studies and the recent development of an anti-FGF23 antibody, burosumab, as novel and promising therapy [10, 17] have resulted in a growing current interest for XLH.

We here report the clinical manifestations at diagnosis and follow-up of a large series of Spanish patients included in the online database RenalTube [18]. This study is justified at least by the following reasons: (1) XLH is a rare disease and few publications provide data on large series of patients; (2) XLH has a broad phenotypic variability and additional information is required to better characterize the clinical spectrum of the disease and to explain why the number of cases diagnosed usually does not correspond to the estimated prevalence of the disease; (3) It is important to share data of patients with genetically confirmed XLH in order to facilitate the finding of a potential phenotype—genotype correlation and to have current data that can be compared for the assessment of the new therapies.

Forty-eight patients included in the RenalTube database with the diagnosis of XLH confirmed by defect-of-function variants found in the PHEX gene were analyzed. All variants had been identified as pathogenic. Sixteen patients (33%) had variants with strong evidence of pathogenicity (nonsense, frameshift, deletions) while the other 32 (67%) harbored variants with very strong evidence of pathogenicity (SNPs). Demographic and genetic data from patients are shown in Table 1. Patients were from 39 families and were being followed in pediatric nephrology units of 17 Spanish hospitals (Fig. 1). Fifteen patients were males and 33 females. Median age at diagnosis was 2.0 (IQR 2.6) years and the age ranged from 3 months to 8 years 2 months.

Presenting manifestations are shown in Table 2 for each patient. Bone deformities and radiological signs of active rickets were the most frequent findings leading to diagnosis. Ten patients were diagnosed because of family screening. Age at diagnosis of these patients was no different from that of the rest of the series as it ranged from 0.5 to 8 years with a median age of 1.04 years.

Twenty-two out of40 patients (55%) in whom the height was registered presented growth retardation (height ≤ 2 SDS). Patients’ height (X ± SEM) was − 1.89 ± 0.19 SDS (n = 40) (Fig. 2). In 87% (35/40) the height was below the 50th percentile. Weight was − 0.88 ± 0.14 SDS (n = 41) and body mass index 0.2 ± 0.15 SDS (n = 40).

Biochemical findings at diagnosis are shown in Table 3. Mean values (± SEM) of available data were serum phosphate 2.7 ± 0.1 mg/dl; − 2.81 ± 0.11 SDS, (n = 41), alkaline phosphatase (892 ± 84 mU/ml) (n = 39), 1,25(OH)₂D62 ± 7 pg/ml (n = 34), parathyroid hormone (PTH)70 ± 7 pg/ml (n = 33), and tubular phosphate reabsorption (TPR) 69 ± 4% (n = 26).

No differences were found between males and females for clinical manifestations, growth impairment or biochemical data at diagnosis. Likewise, no genotype–phenotype correlation was found. Actually, even patients within the same family presented different severity of clinical and biochemical manifestations.

Growth and biochemical variables of 26 patients after a median follow uptime of 7.42 years (IQR = 11.26) are shown in Table 4, Figs. 3 and 4. Anthropometric data were − 1.94 ± 0.16 SDS for height (n = 24), − 0.82 ± 0.10 SDS for weight (n = 22) and 0.14 ± 0.19 SDS for BMI (n = 22). Comparison of data from patients with information both at diagnosis and last follow-up showed mean variations of 0.13 ± 0.23 SDS for height (p > 0.05) (n = 20), 0.35 ± 0.14 SDS for weight (p = 0.02) (n = 20) and 0.13 ± 0.20 SDS for BMI (p > 0.05) (n = 20).

Mean SDS for serum phosphate was − 2.72 ± 0.20 (n = 25). Alkaline phosphatases, 1,25 (OH)₂D and PTH levels were 525 ± 82 mU/ml (n = 14), 53 ± 7 pg/ml (n = 15) and 68 ± 8 pg/ml (n = 26) respectively. Mean tubular phosphate reabsorption was 65 ± 3% (n = 22). Comparison between diagnosis and last follow-up data revealed a variation of 0.20 ± 0.28 SDS for serum phosphate (p > 0.05) (n = 20), 5 ± 7 pg/ml for 1,25 (OH)₂D (p > 0.05) (n = 10), 0 ± 11 pg/ml for PTH (p > 0.05) (n = 16) and − 7 ± 8 for tubular phosphate reabsorption (p > 0.05) (n = 11).

Eight out of 24 patients with renal ultrasounds at last follow-up presented nephrocalcinosis (Table 5).

This study provides a current description of the phenotypic characteristics of a large cohort of Caucasian pediatric patients with XLH genetically confirmed. The sample is a broad representation of the Spanish children with XLH, coming from several hospitals scattered through the country and provides data at diagnosis and after a median follow-up of 7.42 years. The study confirms that growth retardation, bone deformities and active lesions of rickets are the main presenting manifestations of the disease, within a wide spectrum of symptoms. No significant differences were found between males and females as for the severity of the disease. It is of interest that a broad spectrum of PHEX gene variants, all of them already described as pathogenic, was found and no mutation was specifically prevalent in Spanish population. All variants were classified as pathogenic according to the American College of Medical Genetics and Genomics consensus [19]. There was a high phenotypical variability even among family members harboring the same mutations, suggesting that other genes and environmental factors may affect the severity of XLH, as reported by other authors [20, 21].

In addition, the study shows that conventional treatment with phosphate supplements and vitamin D metabolites does not lead to persistent correction of hypophosphatemia or reduction of renal wasting of phosphate and does not modify the circulating levels of calcitriol. Unfortunately, this study does not provide information on circulating FGF23 levels, given its retrospective design. At the last follow-up visit, 11 out of 25 patients had serum PTH values mildly elevated. In XLH, development of hyperparathyroidism is thought to be related with the pharmacological administration of phosphate [22]. Eight patients developed nephrocalcinosis during the follow-up period, a finding linked to the administration of phosphate and vitamin D that usually does not result in subsequent clinical complications [16].

Though conventional treatment has been described to heal active signs of rickets and may improve bone deformities [23], this study confirms that it does not lead to catch-up growth. Mean height Z score of the group of patients remained low, − 1.89 at diagnosis versus − 1.94 at the last visit, although Fig. 3 indicates that the individual patients’ response varied from marked improvement to worsening of growth impairment. Two patients, VII.3 and XV.1, transiently received growth hormone treatment and their heights improved + 2.19and + 0.66 SDS, respectively. It is of note that 4 out of 16 patients had BMI greater than + 1.00 SDS at the last follow-up visit. This percentage of 25% corresponds to the normal distribution of reference population and it indicates that tendency to overweight and obesity was not found in the group of XLH patients here reported, unlike other series that have recently drawn attention to these complications likely related with the sedentary life and restricted mobility of these patients [11]. In this regard, a slight but significant increase in weight was found during the follow-up period in our series.

Our study presents methodological limitations inherent to the retrospective analysis and to the fact that patients’ information was extracted from a database in which some data were missing and cannot be recovered. It is also of note the lack of information on the degree of adherence to medication of each patient as well as the different monitoring protocols among the participating centers. However, it is an observational clinical study describing a large cohort of Spanish pediatric patients with genetically confirmed XLH and it provides current and interesting information on the clinical and biochemical features of the disease, at diagnosis and follow-up after conventional treatment. Our findings could be used as reference for further studies using burosumab treatment.

The RenalTube database including 48 patients, 15 males and 33 females, with the diagnosis of XLH confirmed by defect-of-function mutations found in the PHEX gene was retrospectively reviewed to obtain demographic information and clinical and biochemical manifestations at diagnosis and at the last annual follow-up. Genetic information was confirmed and formatted according to Genome Reference Consortium Human Build 38 patch release 13 (GRCh38.p13) [24]. Variants were analyzed in silico and classified according to recommendations from the consensus of the American College of Medical Genetics and Genomics and the Association for Molecular Pathology [19] as pathogenic, likely pathogenic, benign or likely benign. Results for the age are presented as median and interquartile range (IQR). Other variables are presented for the group as mean (X) ± standard error of the mean (SEM). Z score (SDS) of anthropometric values was calculated using Spanish age and sex-matched reference values [25]. Patients with height ≤ 2 SDS were considered to have longitudinal growth retardation according to World Health Organization standards [26]. Reference values for biochemical parameters were obtained from the laboratory of the Hospital Universitario Central de Asturias (HUCA) [27].

All patients received treatment with phosphate supplementation (dose range of phosphorus element: 27–90 mg/kg/day at last follow-up) and vitamin D metabolites (dose range: 0.25–1.5 µg/day at last follow-up), according to the criteria and indications given by their physicians (Table 5). Two patients (VII.3 and XV.1) received growth hormone treatment. None of them received burosumab treatment.

Information in RenalTube database was downloaded and formatted to an Excel database. All fields but reasons for consultation and genetic information were multichoice or numeric format.

Chi squared test was used to analyze differences between sex for binary (Yes/No) fields (growth retardation, bone deformities). F-test was used to assess variance equality between sex for anthropometric and biochemical values. T-test was used to analyze differences between sex for anthropometric and biochemical values. Paired T-test for means was used to analyze differences between diagnosis and last follow-up for anthropometric and biochemical values. T-test for unequal variances was used to compare age at diagnosis for patients with and without family history of the disease.

Phenotype—genotype correlation was assessed by isolating the most severe phenotypes (lowest serum concentrations, most severe growth retardation, highest levels of alkaline phosphatases) and comparing genetic mutations in these patients looking for big deletions, SNPs with entirely different amino acids or nonsense mutations.