Metabolomics of Plasma in XLH Patients with Arterial Hypertension: New Insights into the Underlying Mechanisms

X-linked hypophosphatemia (XLH) is a rare genetic disorder that increases fibroblast growth factor 23 (FGF23). XLH patients have an elevated risk of early-onset hypertension. The precise factors contributing to hypertension in XLH patients have yet to be identified. A multicenter cross-sectional study of adult patients diagnosed with XLH. Metabolomic analysis was performed using ultra-performance liquid chromatography (UPLC) coupled to a high-resolution mass spectrometer. Twenty subjects were included, of which nine (45%) had hypertension. The median age was 44 years. Out of the total, seven (35%) subjects had a family history of hypertension. No statistically significant differences were found between both groups for nephrocalcinosis or hyperparathyroidism. Those with hypertension exhibited significantly higher levels of creatinine (1.08 ± 0.31 mg/dL vs. 0.78 ± 0.19 mg/dL; p = 0.01) and LDL-C (133.33 ± 21.92 mg/dL vs. 107.27 ± 20.12 mg/dL, p = 0.01). A total of 106 metabolites were identified. Acetylcarnitine (p = 0.03), pyruvate p = (0.04), ethanolamine (p = 0.03), and butyric acid (p = 0.001) were significantly different between both groups. This study is the first to examine the metabolomics of hypertension in patients with XLH. We have identified significant changes in specific metabolites that shed new light on the potential mechanisms of hypertension in XLH patients. These findings could lead to new studies identifying associated biomarkers and developing new diagnostic approaches for XLH patients.


Introduction
X-linked hypophosphatemia (XLH) is a genetic disorder caused by mutations that inactivate the X-linked phosphate-regulating endopeptidase homolog (PHEX) gene.This condition increases fibroblast growth factor 23 (FGF23), a hormone that regulates the sodium-dependent phosphate cotransporters NPT2a and NPT2c in the renal proximal tubules.As a result, XLH patients experience hypophosphaturia, hypophosphatemia, and reduced synthesis of active vitamin D (1.25(OH) 2 vitamin D) [1].XLH affects 1 in every 20,000 individuals, classifying it as a rare disease according to European criteria [2].XLH exhibits a broad spectrum of symptoms, spanning from isolated hypophosphatemia to severe manifestations, including growth retardation, rickets, lower extremity deformities, bone and muscle pain, osteoarthritis, enthesopathy, and an elevated risk of bone fractures.The clinical manifestations vary between the pediatric and adult populations and among individuals within the same family [3].
Although controversial, some studies indicate a potential link between XLH patients and an elevated risk of early-onset hypertension, along with a greater prevalence of hypertension compared to the general population.In XLH patients, it has been suggested that hypertension may be associated with the presence of secondary or tertiary hyperparathyroidism, nephrocalcinosis, and/or impaired renal function [4,5].Furthermore, there is evidence of the influence of FGF23 on renal sodium reabsorption and its effect on vessels and cardiomyocytes, which could potentially contribute to the presence of hypertension and left ventricular hypertrophy in this population [1,6].In addition, there is evidence in observational studies that suggests that vitamin D level influences blood pressure and the risk of hypertension in patients with chronic kidney disease and normal subjects.However, these data have not been confirmed in randomized controlled trials [7].
Hypertension is a significant public health burden and one of the main cardiovascular risk factors and causes of mortality in the general population [8].As we know, hypertension is a complex condition characterized by the involvement of multiple physiological pathways and organ systems.Metabolomics is one of the "omics" sciences that quantifies the changes in multiple metabolites at once and provides a more precise and complete image of an organism's molecular structure and function.Several studies have successfully applied metabolomic techniques to characterize hypertension and suggest possible etiological pathways involving lipid, amino acid, and alcohol metabolism.As metabolites are closely linked to phenotypes, studying metabolomics can be an effective way to better understand the biology of hypertension [9].
Despite all this, it remains to be established whether the factors that could cause hypertension in this disorder are mainly due to XLH, secondary to its complications, the treatment, or other proteins and metabolites involved in phosphate homeostasis.For all this, metabolomics could be helpful to explore the associated metabolic alterations and better characterize the pathophysiology of hypertension in patients with XLH.In this study, we performed a systematic metabolomic analysis in XLH patients with and without hypertension.Our study aimed to explore the metabolomic signatures of hypertension in patients with XLH, identifying metabolites that can provide information on the pathophysiology of this disease and future lines of research.

Characteristics of the Study Participants
The baseline characteristics of twenty adult patients diagnosed with XLH are outlined in Table 1.The study included 9 (45%) XLH patients with high blood pressure and 11 (55%) XLH patients with normal blood pressure.The gender distribution was equal, with 50% men and 50% women.Among the hypertensive group, six (66%) were men, compared to four (40%) in the normotensive group.There were no statistically significant differences between genders.
The median age of all individuals was 44 (33.2-54.7)years.The median weight, height, and body mass index (BMI) were 73 (50.7-85.5)kg, 158 (150-170.5)cm, and 27.5 (22-32.2) Kg/m 2 , respectively.The diagnosis of XLH was made in childhood for ten subjects (50%), with five in each group.The median age at hypertension diagnosis was 35 years.Out of the total, seven (35%) subjects had a family history of hypertension, three (33.3%) in the hypertensive group and four (40%) in the non-hypertensive group.Nephrocalcinosis was present in three (15%) subjects, two (22.2%) of whom had hypertension, and one (10%) did not.In the group with hypertension, four (44.4%) had hyperparathyroidism compared to three (30%) in the normal-blood-pressure group, resulting in a total of seven (35%) patients.Nonsteroidal anti-inflammatory drugs were taken by four (20%) subjects, of which three (30%) did not have hypertension, and only one (11.1%)did.No statistically significant differences were found between hypertensive and normotensive subjects for all these variables.Out of the subjects with hypertension, six (66.6%) were taking one antihypertensive drug, two (22.2%) were taking two drugs, and one (11.1%)was taking three drugs.Regarding antihypertensive treatment, seven (77.7%) hypertensive patients were on ACE inhibitors/ARBs, five (55.5%) on calcium channel blockers, one (11.1%) on beta-blockers, and one (11.1%) on diuretics.

Metabolomic Results
A comprehensive list of 106 identified metabolites is presented in Table 2.The table provides detailed information on the identified metabolites, including their chemical composition and properties.The Volcano graph showed that acetylcarnitine, pyruvate, ethanolamine, and butyric acid significantly differ between the blood samples of XLH patients with hypertension and those of XLH non-hypertensive patients (Figure 1).Box and whisker plots were used to represent the relative intensities of selected metabolic variables with possible clinical relevance (Figure 2).PCA score plots (Figure 3) were constructed for all metabolites and the discriminative ones, showing that the model discriminates between the two groups.From this model, the most important discriminant variables were selected according to their VIP score (>1), a jack-knife confidence interval that did not include zero, and an FC greater than 1.2.There was no correlation observed between acetylcarnitine, pyruvate, butyric acid, or ethanolamine and the levels of PTH, vitamin D, phosphorus, or FGF23 (Supplementary Materials).

Discussion
Hypertension is a chronic condition and a significant risk factor for cardiovascular diseases (CVD).It is estimated that around 20% of adults worldwide are affected by hypertension, of which 90% of cases are classified as "essential" due to an unknown underlying cause [10].This percentage is higher in XLH patients, who, despite being young, are at a higher risk of morbidity, mortality, and cardiovascular diseases.The prevalence of hypertension in our study comprises 45%, a figure similar to other previously described studies [4].However, this is controversial, as other cohorts do not find similar results, and the exact pathophysiological mechanisms involved are still unclear.
The development of hypertension is influenced by a combination of genetic and environmental factors, with age, obesity, and physical inactivity being typical risk factors,

Discussion
Hypertension is a chronic condition and a significant risk factor for cardiovascular diseases (CVD).It is estimated that around 20% of adults worldwide are affected by hypertension, of which 90% of cases are classified as "essential" due to an unknown underlying cause [10].This percentage is higher in XLH patients, who, despite being young, are at a higher risk of morbidity, mortality, and cardiovascular diseases.The prevalence of hypertension in our study comprises 45%, a figure similar to other previously described studies [4].However, this is controversial, as other cohorts do not find similar results, and the exact pathophysiological mechanisms involved are still unclear.
The development of hypertension is influenced by a combination of genetic and environmental factors, with age, obesity, and physical inactivity being typical risk factors, and is strongly associated with inflammation, oxidative stress, and dyslipidemia [11].Not in vain, recent studies have shown that alterations in lipid and carbohydrate metabolism often accompany hypertension [12,13], highlighting the connection between hypertension and metabolic disorders.Thus, researchers have increasingly utilized metabolomic profiles to identify specific metabolites that could serve as biomarkers of oxidative stress and inflammation, conditions commonly present in hypertensive patients [14].With this same idea, our study, employing an untargeted metabolomic technique to identify differences in plasma metabolites between XLH patients with and without hypertension, identified a total of 106 metabolites, among which 4 showed significant clinical and statistical differences.
Firstly, we found higher levels of acetylcarnitine, a metabolite that plays a key role in lipid metabolism, in hypertensive XLH subjects compared to normotensive individuals.Endogenous carnitine biosynthesis occurs in the kidney, liver, and brain and comes from protein lysosomal degradation [15].The human body produces more than 1200 fatty acids (FAs) using L-carnitine and acylcarnitines [16].The essential role of L-carnitine and acylcarnitines is to facilitate the transportation of long-chain fatty acids (long-chain fatty acyl-CoA) to the mitochondria for β-oxidation degradation.In contrast, medium and short-chain fatty acids can enter the mitochondria freely through diffusion.This function is crucial to prevent the accumulation of long-chain fatty acids that can cause cellular damage [17].
The analysis of serum acylcarnitine levels serves as a measurable indicator of β-oxidation status, consistently tied to vascular inflammation in various research outcomes [18].The buildup of acylcarnitine intermediates in extracellular fluid indicates compromised βoxidation and disrupted mitochondrial metabolism, especially in aging, leading to inefficient fatty acid recycling.On the other hand, it has been suggested that an accumulation of intracellular lipids and acetylcarnitine can occur in hypertensive subjects if the absorption rate of fatty acids exceeds the rate of β-oxidation [19].Additionally, acylcarnitines constitute some of the most commonly implicated metabolites in cardiovascular diseases [20].Recent clinical studies have linked acylcarnitines to an elevated risk of atherosclerotic plaque formation, independent of conventional cardiovascular risk factors [21].Notably, Ramipril was found to significantly reduce the levels of propionylcarnitine, butyrylcarnitine, and isovalerylcarnitine (C3M, C4M, and C5M) in rats among 60 metabolites analyzed that were not affected by treatment with ACE inhibitors [22].Hence, their role as cardiovascular risk markers in this type of population remains to be studied.
Secondly, we observed lower levels of pyruvate in hypertensive subjects with XLH.There are limited data directly associating pyruvate levels with hypertension, but this correlation was positively made in a study involving restricted participants [19].Conversely, one study highlighted that decreased alanine and pyruvate levels in people with hypertension may indicate a dynamic balance between aminotransferase reactions in the muscle and the liver [23,24].While some authors propose that hypertensive patients have an elevated lactate-to-pyruvate (L/P) ratio compared to normotensive individuals [25], this ratio would reflect their response to oxygen and glucose supply variations, suggesting its role as a potential marker for tissue ischemia in anaerobic conditions.The higher L/P ratio might indicate the occurrence of such conditions in hypertensive tissues.Nevertheless, our study did not detect changes in the L/P ratio or a decrease in alanine levels.
Thirdly, elevated levels of ethanolamine were found in hypertensive patients with XLH as compared to the control group.This observation is in line with a study that involved hypertensive pediatric subjects, where ethanolamine proved to be effective in identifying children with hypertension [19].While only a few studies have focused on the role of this metabolite in arterial hypertension, existing evidence highlights its crucial involvement in the synthesis of phosphatidylethanolamine, glycerophospholipids linked to the formation of atherosclerotic lesions and endothelial damage [26].Furthermore, its impact on left ventricular function and the development of cardiac fibrosis has been substantiated [27].
Finally, the levels of butyric acid were also significantly higher in the hypertensive XLH population compared to normotensive individuals.Traditionally, cardiovascular research focused mainly on the involvement of long-chain fatty acids.However, increasing evidence indicates a significant impact of molecules produced by the gut microbiota, such as short-chain fatty acids (SCFA), on atherosclerosis and hypertension [28].Butyric acid (BA) is a significant SCFA that acts as a mediator between the intestinal microbiota and the circulatory system [29].The role of butyric acid in blood pressure BP was elucidated more than 50 years ago by intravenous administration of tributyrin, a BA prodrug [30].Another study indicates that an elevation in the concentration of BA in the colon induces a notable hypotensive effect regulated by signals from the afferent vagus nerve of the colon.This bacterial metabolite is believed to exert its hypotensive effect by inhibiting sympathetic activity and inducing a direct vasodilatory effect [31].In addition, BA suppresses the activation of NF-κB and reduces macrophage migration, thus assuming an anti-inflammatory role and improving atherosclerosis.In light of these findings, it is proposed that SCFAs, particularly butyric acid, could serve as a therapeutic target for atherosclerosis by modulating lipid metabolism and mitigating inflammation [32].
It is important to acknowledge certain limitations of this study.Firstly, the study's design prevents us from establishing a causal relationship.Secondly, the small sample size necessitates categorizing this study as exploratory and hypothesis-generating until further validation is performed with larger cohorts.It also makes a potentially high type II error.In fact, our study found no correlation between hypertension and well-known related factors with high BP in the general population, such as hyperparathyroidism, nephrocalcinosis, high FGF23 levels, obesity, or the use of nonsteroidal anti-inflammatory drugs, which XLH patients commonly use for pain control.Nonetheless, it did afford us the statistical power required to detect these observed metabolite differences.Thirdly, we did not include unaffected controls with and without hypertension.Fourthly, it is important to consider the effect of medications on blood pressure and metabolism in our analysis.In our study, most patients take minimal medication in addition to active vitamin D and oral phosphorus supplements, but, maybe due to limited statistical power, the study has not yielded significant results on this point.Likewise, the study was conducted under similar conditions, with all patients fasting.The impact of metabolic regulation on flora has not been explored in this manuscript.
Despite the outlined limitations, it is crucial to recognize the substantial strengths present in this study.Our XLH patient cohort is one of the largest published from Europe and the largest among adult XLH patients in Western Europe.Therefore, despite the potential insights gleaned from larger populations, every scientific contribution is significant in this rare disease.The field of omics sciences, such as proteomics, genomics, and metabolomics, is undergoing a surge in medical research, and we are confident that it will eventually provide us with extensive data to understand the pathophysiological mechanisms of diseases, particularly hypertension.Manipulating metabolic pathways and achieving personalized medicine represent the future of medicine.
In conclusion, the changes observed in the metabolite profile suggest that metabolomics can be a useful tool for better understanding the pathogenic mechanisms of hypertension in XLH.This can lead to earlier diagnosis, better treatment options, and the prevention of cardiovascular complications.Our study, which represents the first on hypertension in patients with XLH using a metabolomic approach, identified significant changes in specific metabolites.These findings can help us better understand this disease's biological and pathological processes and its regulatory mechanisms, which are still unclear.

Study Design and Participants
This multicenter cross-sectional study was conducted on a population of adult patients diagnosed with XLH and under follow-up at the Hospital Universitari i Politècnic La Fe and other areas of the Valencian Community.The systematic identification of this cohort was performed using electronic medical records, following a methodology previously published by our group [33].The population was stratified into two groups based on the presence or absence of hypertension.Demographic, clinical, and laboratory variables, as well as treatments used, were collected.Clinical data and treatments were collected through electronic health records, which were later validated and extended through medical interviews.The study was conducted according to the guidelines of the Declaration of Helsinki and approved by the Institutional Review Board and Ethics Committee of La Fe Health Research Institute (protocol code 2021-016-1, date of approval 27 January 2021).Informed consent was obtained from all subjects involved in the study.

Measurement of Blood Pressure and Covariables
Body weight and height were measured twice, and the mean values were used to estimate BMI, calculated as weight in kilograms divided by height in square meters.Hypertension was defined as systolic blood pressure (BP) ≥ 130 mm Hg and/or diastolic BP ≥ 80 mm Hg according to the 2017 ACC/AHA guidelines [10].In all cases of hypertension, the diagnosis had already been made by a primary care physician or another specialist prior to the commencement of our study.BP measurements were obtained during medical visits, with the primary objective of validating the diagnosis and assessing its management status.It was measured following standard protocols and by trained nurses or physicians using validated devices.The participants were in a relaxed, sitting position for 5 min and avoided caffeine, cigarettes, alcohol, and physical activity for 12 h before the clinic visit.At the medical examination, three BP measurements were recorded at 5 min intervals, and the mean BP was calculated based on the last two values.Patients diagnosed with secondary arterial hypertension were excluded.Nephrocalcinosis was defined as diffused calcium deposition within the kidney, identifiable through ultrasonography, whereas renal dysfunction was determined according to KDIGO criteria, eGFR < 60 mL/min/1.73m 2 , and/or the presence of structural damage [34].Hyperparathyroidism (HPT) was diagnosed in cases exhibiting consistently elevated intact parathormone (PTH) levels (>65 pg/mL) for six months or more.

Blood Sample Collection and Preparation
All patients included in the study were scheduled for a single medical visit at the research unit within the Nephrology Service of Hospital Universitari i Politècnic La Fe.A comprehensive evaluation was carried out during this visit, including blood tests, urine sampling, and kidney ultrasounds.All patients observed a minimum 8 h fasting period before undergoing these assessments.Blood samples were collected using ethylenediaminetetraacetic acid (EDTA) tubes and processed within 30 min to prevent platelet activation, protein degradation, and the decomposition of thermolabile compounds.Intact FGF23 levels were determined using an enzyme-linked immunosorbent assay (ELISA) test kit from Kainos Laboratories, Tokyo, Japan.Total cholesterol, high-density lipoprotein cholesterol (HDL-C), low-density lipoprotein cholesterol (LDL-C), and triglycerides were assayed using enzymatic procedures.The CKD-EPI formula was used to calculate the estimated glomerular filtration rate (eGFR), while the remaining biochemical variables were examined following the established protocols and best-practice guidelines outlined by Universitari I Politècnic La Fe Hospital.
For the metabolomic study, serum samples were promptly centrifuged at room temperature within 15 min of extraction, using the Eppendorf Centrifuge 5702 model at 5000 rpm for 10 min.The resulting serum was then divided into 250 µL aliquots, each labeled with an anonymized numerical code, before being frozen at −80 • C using a New Brunswick Scientific Ultra-Low Temperature Freezer V 410 Premium, ensuring optimal conditions for further analysis.
Upon thawing, 20 µL of serum was combined with 180 µL of cold methanol, an organic solvent that dissociates the protein-bound metabolites.The samples were then shaken at −20 • C for 30 min, inducing protein precipitation, before being centrifuged at 16,000 rpm for 20 min at 4 • C to separate the cellular fraction from the plasma.Then, 90 µL of the upper cleaned was collected from each sample and transferred to an HPLC vial and combined with 10 µL of a standard solution containing internal standards (a mixture of 20 µM caffeine-d9, leucine enkephalin, reserpine, and phenylalanine-d5).
Quality control samples (QCs) were prepared by combining 10 µL from each extract.Blank samples, prepared to replace the extract with ultrapure water, were used to identify artifacts from reagents, the tube, and other materials.Finally, samples, QCs, and blank samples were injected randomly into the chromatographic system.To monitor the stabilities of the instrumental system and the instrumental drift, QC samples were injected into every 7th sample in each sequence, and the blank samples were performed at the end of the sequence.

Untargeted Metabolomics Based on UPLC-Q-ToF Mass Spectrometry
The metabolomic analysis was performed using Ultra-Performance Liquid Chromatography (UPLC) equipment coupled to a high-resolution mass spectrometer (MS) with Orbitrap UPLC-QExactive Plus detector (UPLC-TOF/MS-Orbitrap, QExactive Plus MS) available at the Analytical Unit of the Instituto de Investigación Sanitaria La Fe (IISLaFe).All reagents and chemicals were purchased from Sigma Aldrich (St. Louis, MO, USA) and Fisher Scientific (CAN, Waltham, MA, USA).
Chromatographic separation was performed using an HILIC UPLC X bridge BEH amide column (150 mm × 2.1 mm, particle size 2.5 uM, Waters, Wexford, Ireland).The column and auto counter were set to 25 • C and 4 • C, respectively, with an injection volume of 5 µL.The total running time of the chromatogram was 25 min at a flow rate of 105 µL/min.Metabolites were eluted with gradients of mobile phase A consisting of H 2 O and 10 mM ammonium acetate and mobile phase B in acetonitrile (ACN).During the first 2 min, the eluent composition was 90% mobile phase B and 10% mobile phase A; at minute 2, 85% B; from minute 3 to minute 8, 75% B; from minute 8 to minute 10, 70% B; from minute 10 to minute 13, 50% B; from minute 13 to minute 16, 25% B; from minute 16 to minute 22, 0% B; from minute 22 to minute 25, again 90% B reaching the starting conditions.A mass detector (Orbitrap) was used in Full scan in both positive and negative ESI modes with two different events, the first scan event between 70 and 700 Da and a second event in the mass range of 700-1700 Da.
Different measures were taken to ensure the analysis's quality and reproducibility and mitigate intra-batch variability.The reagent blank was analyzed at the beginning of the sequence to identify sampling artifacts and exclude contaminating signals.Subsequently, a random injection sequence was performed that included the analysis of 5 quality control (QC) samples to condition the column and equipment (these data were not used for data analysis).Sample analyses and quality control were conducted at a 5:1 ratio to monitor and control variability resulting from contamination or changes in the column.Any signal detected with an intensity similar to the quality control and blank samples was eliminated.

Data Processing and Acquisition
The data obtained during the analysis were converted into mzXML format using the Mass Converter Proteowizard program.Then, they were processed in EI-MAVEN software (Version Version 0.4.1) for alignment, noise filtering, the integration of chromatographic peaks, generation of a peak table, and identification of metabolites using an in-house library of polar compounds previously developed and used at the Analytical Unit [35].Structures, m/z spectra, and formulas were obtained from the Human Metabolome Database (HMDB, version 4.0).
The configuration parameters of EI-MAVEN for data processing were the following: ionization in automatic mode type ESI, precision q1 and q3 for MRN transitions of 0.5 amu with total filter line.Maximum differences in retention time between spikes in a group were 15 s.The smoothness threshold was 2 λ, and the asymmetry was 0.08 ρ.A weighted average of the m/z values observed in the chromatographic peak was calculated to annotate the chromatographic peaks.Cluster selection criteria for use in the alignment were at least 2 peaks in each cluster with a total limit of 1000 clusters and a clustering window of 20 scans.For peak picking, the minimum intensity was 1000 with a maximum signal-to-noise ratio of 3 and a minimum peak width of 5 scans.The peaks are assigned a clustering score.
We have applied the NetID algorithm to the dataset, creating a peak table containing the exact mass (m/z) and retention time (min).Subsequently, matches were sought with those metabolites in the database.The predefined tolerance was ±10 ppm.Peaks were considered metabolites or artifacts based on the match with the local database formulas.All adducts, fragments, or isotopes of metabolites were considered artifacts.
When comparing with blank samples, background peaks were removed if their intensity was greater than 0.5 times that of the sample.Compounds showing a coefficient of variation (CV) in the analytical response of the quality controls exceeding 30% were excluded.

Statistical Analysis
Categorical variables were presented as n (%), while continuous variables were presented as median (P25-P75).The distribution of the variables was tested with the Shapiro-Wilk test.Differences between hypertensive and normotensive groups were analyzed using independent samples t-test for normally distributed continuous data, the Mann-Whitney U test for skewed data, and the χ 2 test for categorical variables.
The relative abundances of metabolic members of XLH individuals with hypertension were compared with non-hypertensive XLH controls.A univariate statistical analysis was conducted between the case and control groups utilizing a t-test and visualized through a Volcano Plot.This analysis was executed using an in-house script within the R platform (Version 4.3.3).The methodology integrated a Fold Change (FC) approach alongside the significance derived from a paired Student's t-test for normally distributed variables.For skewed data, the analysis employed the Wilcoxon signed-rank test after confirming normality via the Shapiro-Wilk test.

Figure 1 .
Figure 1.Volcano plot illustrates significant differences in serum metabolites between XLH patients with and without hypertension.

Figure 1 .
Figure 1.Volcano plot illustrates significant differences in serum metabolites between XLH patients with and without hypertension.Int.J. Mol.Sci.2024, 25, x FOR PEER REVIEW 7 of 14

Figure 2 .
Figure 2. Box and whisker plot of significantly altered metabolites in the hypertensive group compared with the normotensive group.

Figure 2 .
Figure 2. Box and whisker plot of significantly altered metabolites in the hypertensive group compared with the normotensive group.

Figure 2 .
Figure 2. Box and whisker plot of significantly altered metabolites in the hypertensive group compared with the normotensive group.

Figure 3 .
Figure 3. PCA score plot with metabolic variations between hypertensive XLH patients and normotensive XLH patients when all metabolites and only discriminative metabolites are identified.

Figure 3 .
Figure 3. PCA score plot with metabolic variations between hypertensive XLH patients and normotensive XLH patients when all metabolites and only discriminative metabolites are identified.

Table 1 .
Demographic and clinical characteristics of the study population.

Table 2 .
List of identified metabolites.