Composition of urinary stones in children: clinical and metabolic determinants in a French tertiary care center

As the epidemiology of urolithiasis is constantly evolving, analyzing the composition of stones is crucial to better understand the determinants of lithogenesis. The aim of this study was to describe the composition of stones of pediatric patients in a tertiary center. Clinical and metabolic data from all pediatric patients with at least one stone that was analyzed by Fourier transformed infrared spectroscopy (FTIR) in the Hospices Civils de Lyon between 2013 and 2017 were retrospectively collected. A total of 111 patients (sex ratio 1.4:1) were included; their median ([IQR]) age was 7.5 (3.1–10.5) years. The main component of stones was calcium oxalate (weddellite for 34 (31%) stones, whewellite 23 (21%)), calcium phosphate (carbapatite 32 (29%), brushite 6 (5%), amorphous calcium phosphate 3 (3%)), struvite 5 (5%), cystine 4 (4%), uric acid 2 (2%), and ammonium acid urate 2 (2%). A total of 20 (18%) stones were pure and 24 (22%) were infectious. Carbapatite stones were the most frequent in patients < 2 years and calcium oxalate stones in patients > 2 years old. Metabolic abnormalities (most frequently hypercalciuria) were found in 50% of tested patients and in 54% of patients with infectious stones. Congenital anomalies of the kidney and/or urinary tract (CAKUT) or neurogenic bladder were present in 9/24 (38%) patients with infectious stones and 12/16 (76%) patients with bladder stones. Conclusion: This study confirms that calcium oxalate stones are the most frequent among pediatric patients, which could reflect the nutritional habits of predisposed patients. In contrast, infectious stones are less frequent and occur mostly in association with anatomic or metabolic favoring factors. What is Known: • Incidence of kidney stones is increasing among children. • Composition of kidney stones in children is constantly evolving under the influence of genetic, nutritional, environmental, and infectious factors. What is New: • Infectious stones are less frequent and occur mostly in case of associated anatomic or metabolic factors, particularly in older patients. • Stone composition and accurate metabolic analysis are guiding genetic screening in suggestive situations.


Introduction
Lithogenesis occurs when urine is supersaturated with stone-forming salts [1,2]. Multiple factors can cause this supersaturation, and some of them are dramatically impacted by the overall health status, notably by hydration and nutrition [3]. Pediatric cases of urolithiasis are less common than adult ones, but their incidence has increased over the past decades in high-income countries [4]. Changes in nutritional habits are thought to partially explain this increase [5]. Some genotypes, especially in children, are considered as additional risk factors for lithogenesis (alleles implicated in innate metabolic pathways, e.g., in the vitamin D, calcium, and phosphate metabolic pathways) [6]. Recent papers have also suggested that intestinal absorption of stone promoters is influenced by microbiome diversity [7]. In parallel, improvements in antenatal screening and management of congenital abnormality of kidney and/or urinary tract (CAKUT) have led to a decreased frequency of infectious stones [8].
Urolithiasis is responsible for recurrent pain and urinary tract infection (UTI) in children [9]. Moreover, patients with stones present a 2-fold increased risk of developing chronic kidney disease (CKD) compared to the general population [10]. This risk is even higher in monogenic stone diseases such as primary hyperoxaluria or cystinuria, which are responsible for at least 10% of urolithiasis in children. Because they require specific management and therapies, they must be diagnosed as early as possible [6].
As the epidemiology of urolithiasis is constantly evolving, analyzing the composition of stones is crucial to better understand the determinants of lithogenesis in order to adapt prevention and management strategies. Thus, the aim of the present study was to describe the composition of stones of pediatric patients in a tertiary university center and their clinical and biological presentation, to illustrate the current epidemiological trends of urolithiasis in children.

Patients
Patients under 18 years with a urinary stone that was analyzed by Fourier-transform infrared (FTIR) spectroscopy between 1/1/2013 and 12/31/2017 in the biochemistry department of the Hospices Civils de Lyon (HCL) were retrospectively identified. Patients followed in the tertiary nephrology and/or urology pediatric units of the HCL were included. Some patients underwent multiple stone analyses; only the first spectroscopic analysis during the study period was considered.
Out of the 5782 stones analyzed by FTIR spectroscopy during the study period in the biochemistry department of the HCL, 202 (4%) belonged to patients under 18 years. Among them, 111 were followed in the tertiary nephrology and/or urology pediatric units of the HCL and were therefore included in the study; their sex ratio was 1.4:1, and their median ([IQR]) age at the time of spectroscopy evaluation was 7.5 (3.1-10.5) years. Family history of urolithiasis was known for 83 patients, and a first-degree history was reported for 19 (23%) of them. A total of 26 (23%) patients had CAKUT (viz., 6 with ureteropelvic junction obstruction, 6 with bladder exstrophy, 5 with kidney duplication, 3 with vesicoureteral reflux, 2 with complex anorectal malformation with vesicorectal fistula, 1 with bilateral ureterohydronephrosis, 1 with horseshoe kidney, 1 with posterior urethral valve, and 1 with posterior hypospadias ); among them, 6/26 (23%) had a Mitrofanoff cystostomy. A serious pre-existing medical condition at risk of urolithiasis was present for 18 (16%) patients: 9 had neurological impairment, 2 had chronic digestive inflammatory disease, 3 had metabolic disorders, 3 had hemato-oncological diseases, and 1 underwent cardiac transplantation. One (1%) patient had moderate CKD (eGFR 50 mL/min per 1.73 m 2 ) and 4 (4%) mild CKD (eGFR between 77 and 87 mL/min per 1.73m 2 ).
To be able to evaluate the changes in the composition of stones with age, a classification similar to the one proposed by Daudon et al. was used [11]: patients were categorized into five different age groups (group 1 ≥0 and <2 years, N=19; group 2 ≥2 and <5 years, N=23; group 3 ≥5 and <10 years, N=30; group 4 ≥10 and <15 years, N=28; and group 5 ≥15 years, N=11). The present retrospective study was approved by the local IRB (Comité d'Ethique des Hospices Civils de Lyon, session 6/7/2018).

Data collection
Clinical, radiological, and metabolic data were retrospectively collected from electronic medical records. Urine and blood metabolic evaluation was performed when patients were referred to a pediatric nephrologist, mostly within 6 months before or after the spectroscopic analysis. Blood tests included creatinine, electrolytes, 25-hydroxy vitamin D (25OH-D), 1-25-dihydroxy vitamin D (1-25OH-D), and parathyroid hormone (PTH) evaluation. Urinary tests were performed on 24-h urine collection if possible (10 patients) or on urinary samples. History of UTI was considered when reported as such in medical records.
The 2009 revised Schwartz formula was used to estimate glomerular filtration rate (eGFR) [12]. Normal urinary values for urinary calcium, oxalate, citrate, and uric acid were expressed as solute/creatinine ratio and depended on age, as previously referenced [13][14][15]. Hyperphosphaturia was defined when the ratio of maximum transport of phosphate to eGFR (TmP/GFR) was lower than 1.15 mmol/L [16]. Normal value for urinary cystine/creatinine ratio was < 30 μmol/mmol [17]. Patients with homozygous cystinuria had a ratio > 150 μmol/mmol. An intermediate ratio may account for heterozygous patients [18].
The composition and morphology of stones were determined using the morphoconstitutional analysis [1,19]. Briefly, after a macroscopic and a microscopic description, calculi were crushed and diluted with potassium bromure for analysis by Fourier-transformed infrared spectroscopy (FTIR). The composition was deciphered by comparison on spectral databases. Stones were described depending on their main component; for example, a "carbapatite stone" refers to a stone for which the main component is carbapatite.

Infectious stone definition
According to Daudon et al., stones were considered as infectious stones either when they contained even a small amount of struvite or in the case of calcium phosphate stones when they contained amorphous calcium phosphate and whitlockite or when the carbonatation rate of carbapatite was > 15% [19].

Statistical analysis
Quantitative variables were expressed as median (interquartile range, IQR). They were compared with nonparametric Mann-Whitney U tests. Qualitative variables were expressed as count (percentage); they were compared using χ 2 tests. A p value <0.05 was considered significant.

Clinical presentation
There was a significant difference in the main symptoms leading to urolithiasis diagnosis between age groups (p=0.003): the proportion of UTI was the highest (58%) in the first age group (group 1) and decreased with age; the proportion of renal colic and pain symptoms increased with age and was the highest (73%) in the oldest age group (group 5). The proportion of fortuitous diagnosis reached about 20% in each age group (Fig. 1). Initial or intercurrent UTIs were present in 54 (49%) patients.
The composition of stones seemed to be different between age groups (Fig 2a) but not between genders (Fig. 2b-c): the proportion of carbapatite stones decreased with age from 8/19 (42%) in group 1 to 1/11 (9%) in group 5; the proportion of calcium oxalate stones increased with age from 7/19 (37%) in group 1 to 9/11 (82%) in group 5. More precisely, the proportion of weddellite stones was similar in all groups (21 to 39%) whereas whewellite stones were absent in the youngest groups of age, were found in group 3, and were predominant in groups 4 and 5. There was a significant difference in the composition of stones between age groups relative to the 3 most frequent components (whewellite, weddellite, and carbapatite, p=0.004; Table 1).

Metabolic evaluation
Adapted metabolic urinary evaluations were performed for 70 (63%) patients. Out of the 70 patients tested, 35/70 (50%) had at least one metabolic abnormality, most frequently hypercalciuria (Table 2). A total of 6/70 (9%) patients displayed Fig. 1 Clinical presentation according to age groups. Data are presented as % of total number of patients of each age group. Proportions were compared using chi-square test: p = 0.03 Fig. 2 Main components of urolithiasis according to age groups (all patients (a), girls (b), boys (c)) , 70 patients with complete or targeted metabolic evaluation). Data are presented as % of total number in each age group. Statistical analyses were not performed for a-c because of a low number of patients in each group. W1, whewellite; W2, weddellite; CA, carbapatite; BRUSH, brushite; STR, struvite; ACP, amorphous calcium phosphate; CYS, cystine; UA, uric acid; AAU, ammonium acid urate; G1, group 1 elevated urinary cystine/creatinine ratio, and among them, 3 patients (all with cystine stones) had a significant elevation of this ratio between 150 and 370 μmol/mmol. The three remaining patients had a moderately elevated ratio (from 34 to 57 μmol/mmol) and stones of another composition.
A metabolic evaluation was performed for 20 of the 32 patients with predominant carbapatite stones (considered infectious stones or not depending on carbonatation rate and associated components) and revealed a metabolic abnormality for 13/20 (65%) of them.
A total of 17 patients underwent a genetic screening driven by clinical and biological findings, and pathogenic mutations were identified in 7/17 (41%) of them. The four patients with cystine stones had homozygous cystinuria (3/4) or hypotonia/ cystinuria syndrome (1/4). The 3 patients with moderately elevated cystine/creatinine ratio were not evaluated for heterozygous cystinuria. One patient had primary hyperoxaluria, and two harbored heterozygous mutations in the SLC34A3 gene encoding the sodium/phosphate co-transporter Npt2c (for both of these patients genetic screenings were driven by increased 1-25OH-D).

Discussion
The present study confirmed the trend observed over the last decades reporting an increase in the proportion of calcium oxalate stones and a decrease in the proportion of infectious stones [11,20]. Less than a fifth of the stones analyzed here  were pure, suggesting that stone formation was multifactorial in most cases. There was a high proportion of bladder stones in this cohort. Calcium oxalate stones have become more frequent in every age group compared to the historical French cohort published by Daudon et al. and presenting data on the 1977-2007 period [11]. This is particularly true below 2 years of age with less than 10% oxalate stones reported previously compared to more than 35% here.
Calcium oxalate was predominant in 52% of stones in our cohort, which is consistent with recent reports worldwide: respectively, 48%, 49%, 54%, and at least 57% of stones mainly composed of oxalate stones were reported in Egyptian, Pakistanis, British, and American cohorts [21][22][23][24]. In the American cohort, the proportion was even higher as they reported 57% of pure calcium oxalate stones and 30% of mixed calcium oxalate/calcium phosphate stones. The second most frequent component varies depending on the countries: AAU accounts for approximately 30% of stones in lowmiddle-income countries (i.e., in the Egyptian and Pakistanis cohorts) while carbapatite and/or struvite stones were around 30% in high-income countries.
This rise of calcium oxalate over time could be explained by a strong modification of the global sociocultural environment of children worldwide. Weddellite (calcium oxalate dihydrate) and whewellite (calcium oxalate monohydrate) stones are found in conditions of hypercalciuria or hyperoxaluria, respectively, or both. These conditions can be influenced by dietary factors and intestinal absorption and are more likely to be encountered in patients with predisposing innate urinary conditions [20]. Calcium excretion is increased by high intakes of sodium, diets rich in animal proteins [25,26], and high fructose intake [27]. Changes over the last decades towards such diets might therefore be responsible for an increased incidence of calcium oxalate stones in children close to the one found in adults. It is particularly true in high-income countries as in our cohort and could explain why calcium oxalate stones are even higher in American pediatric cohorts [24,28].
Although influenced by food intake, oxalate urinary excretion is particularly affected by intestinal oxalate absorption that is increased with digestive inflammation [29]. Additionally, the gut microbiome is less diverse in pediatric calcium oxalate stone formers, with particularly low levels of oxalate-degrading bacteria [7]. Alterations of the microbiome diversity could explain the association found between the frequent use of antibiotics and stone disease and is another explanation for the rise of calcium oxalate stones [30].
While calcium oxalate stones are increasing, AAU stones disappear as health conditions improve [31]. As expected, AAU stones were rarely identified in the cohort herein (2%). The formation of AAU occurs in conditions of undernutrition: it is a major component of endemic bladder stones that represents up to 50% of the stones below 2 years in low-income countries as reported in a Pakistanis cohort [32].
Uric acid stones were rare and found herein only in two patients who were receiving multiple treatments. Compared to calcium oxalate, there is no increase of uric acid stones towards a proportion similar to the one encountered in adult population, despite more favorable diets and more frequent obesity in children. Children have a slightly higher urine pH than adults that does not favor uric acid crystallization [33]. This could explain why the well-established association between obesity and urolithiasis in adults remains discussed for children [5].
The influence of inherited factors on pediatric urolithiasis varies according to the composition of stones. Cystine stones occur, for instance, only in the case of congenital cystinuria, Fig. 3 Bladder stones compared to other localizations (UUT and not localized): history of CAKUT (a), history of UTI (b), and main components (c). W1, whewellite; W2, weddellite; wCA, carbapatite; BRUSH, brushite; STR, struvite; ACP, amorphous calcium phosphate; CYS, cystine; UA, uric acid; AAU, ammonium acid urate an autosomal recessive condition accounting for 6-8% of urolithiasis in children [34]. We report here consistently 4 patients (4%) with cystine stones. The influence of genetic predispositions for the formation of calcium oxalate or calcium phosphate stones is more variable. Some rare monogenic disorders, such as primary hyperoxaluria [35], or Dent's disease [36], induce a high lithogenic susceptibility. Recent data have revealed that a causative monogenic mutation could be detected in up to 20% of idiopathic urolithiasis and/or nephrocalcinosis [37]. However, innate risks for stone formation are frequently driven by a polygenic susceptibility or a heterozygosity in known genes. Heterozygous mutations in SLC34A3 (responsible for hereditary hypophosphatemic rickets when homozygous) identified for two patients illustrate this mildly increased risk of stone formation induced by genetic variations [38,39]. It is commonly accepted that genetic analyses are particularly indicated in case of an early start, familial history, recurrent, and/or bilateral stones or in case of extrarenal symptoms. However, as illustrated in our cohort, stones spectrophotometry and metabolic analysis are always crucial to guide genetic screenings.
Besides dietary and inherited metabolic factors, urinary infections also play an important role. Struvite stones are always caused by infection [1]. Their proportion was lower compared to historical cohorts, particularly among males younger than 2 years: in our cohort, only 1/19 stones contained struvite, whereas Daudon et al. have reported 38% of struvite-containing stones in the same age group [11]. This decrease in infectious stones partially explains why the historical male predominance in pediatric urolithiasis is currently less clear. Over the last 50 years, the trend towards an equilibration of the sex ratio in high-income countries has been illustrated by three successive British cohorts, reporting a decrease from 2.9 to 1.3 of the sex ratio between 1966 and 2015 [23,40,41]. The sex ratio found in the present study was close to the one reported in the most recent study by Robinson et al. [23], but it remains higher among patients with infectious stones. This higher sex ratio in patients with infectious stones could be partly explain by CAKUT being more frequent in males [42].
Unlike struvite stones, carbapatite stones-about one third of the stones identified herein-are not always caused by UTIs [1]: they form in alkaline urines induced by either urea-splitting bacterial infection or defect in urine acidification. When patients were screened for metabolic abnormalities, 65% of carbapatite stones were associated with metabolic abnormalities, sometimes leading to a specific diagnosis and management such as thiazide treatment in case of hypercalciuria resistant to dietary measures. Metabolic abnormalities rates were similar whether carbapatite was considered infectious or not based on Daudon criteria. It seems therefore essential to systematically screen patients for metabolic abnormalities even in case of UTI or infectious stones in order to optimize therapeutic strategies, especially for older patients without associated CAKUT and for patients with carbapatite stones for which the carbonatation rate is < 15%.
In addition, the proportion of infectious stones was lower, and the median age of patients with infectious stones was increased in this cohort compared to historical cohorts [11]. It likely reflects an improvement in the management of UTIs and CAKUT in young children over the past decades. Moreover, the high prevalence of complex CAKUT in our tertiary center is an explanation for infectious stones occurring in older patients.
Another consequence of frequent complex CAKUT was the elevated proportion of bladder stones (14%) compared to recent studies from high-income countries (mostly < 10%) [23,43,44]. The composition of these stones differs from endemic bladder stones reported in low-income countries and occurs in older patients. It outlines the need for particular attention on the risk of bladder stones during the follow-up of this specific population. Bladder stones in conditions of complex CAKUT are not always infectious, and patients must be screened for metabolic factors [45].
The present study has strengths: it is one of the largest series on pediatric urolithiasis in high-income countries of the last decade, and data collection is the most recent. However, it also suffers from limitations, mainly because of its retrospective design. The definition of UTI was limited as a history of UTI was considered when reported as such in medical records with compatible symptoms and urinary analysis. Detailed bacteriological examination was not always available. Since some patients were not addressed to nephrologists, a large proportion of patients (37%) did not undergo metabolic evaluation, despite current guidelines recommending a systematic complete metabolic evaluation in case of pediatric urolithiasis [46]. It indicated that a closer collaboration between urologists and nephrologists is worth for a better evaluation in all children with urolithiasis. As the study was conducted in a tertiary center, complex CAKUT (bladder exstrophy) and associated medical conditions were more frequent.
In conclusion, this retrospective cohort provided a recent picture of renal stone disease among children followed in a pediatric university hospital in France. The composition of stones has evolved over the last decades: infectious stones are less frequent and occur mostly in association with anatomic or metabolic factors. In contrast, calcium oxalate stones are more frequent, which could be explained by changes in nutritional habits. The high proportion of bladder stones was associated with complex CAKUT. As factors are often combined, a systematic metabolic evaluation of pediatric stone formers remains nevertheless crucial, even when the infectious etiology seems clear, and particularly in the case of carbapatite stones. Genetic screening should be considered in case of suggestive situations. Authors' contributions CR, CM, JB and ABT contributed to the conception of the work. CM made spectrophotometry analyses. LD made biological analyses and extracted the biological data. CR collected clinical/ biological data and wrote the manuscript with support of JB and ABT. JB made statisical analyses. All authors revised it critically and approved the version to be published.

Data Availability
The datasets analyzed during the current study are available from the corresponding author.

Declarations
Ethics approval This retrospective and strictly observational study was approved by the local IRB (Comité d'Ethique des Hospices Civils de Lyon, session 6/7/2018).