Pathophysiology of hypercalciuria in children

Albright et al. [1] first introduced the term idiopathic hypercalciuria to describe patients with recurrent urolithiasis who had elevated urinary calcium excretion without concomitant hypercalcemia. The etiology of hypercalciuria is complex given that urinary excretion of calcium is the end result of an interplay between three organs—namely, the gastrointestinal tract, bone, and kidney—which is further orchestrated by hormones, such as 1,25-dihydroxyvitamin D₃ (1,25-(OH)₂D₃), parathyroid hormone (PTH), calcitonin, fibroblast growth factor (FGF-23), etc. Often, a primary defect in one organ induces compensatory mechanisms in the remaining two organs, such as increased absorption of calcium in the gut secondary to a primary renal loss. Hypercalciuria can be either idiopathic or secondary. In this review, idiopathic hypercalciuria and the recent developments in hereditary renal tubular disorders associated with hypercalciuria is discussed, providing an insight into the pathophysiology of hypercalciuria. The role of diet in hypercalciuria and the clinical relationship of hypercalciuria with urolithiasis and bone mineral density (BMD) is also discussed.

Hypercalciuria is the most common metabolic abnormality detected in children with stones, causing mostly the formation of calcium oxalate stones and to a lesser extent calcium phosphate stones or a mixture of the two [2‐4]. The reported incidence for urolithiasis in Icelandic children is 6.3/100,000 children <16 years of age [2]. In adults, Curhan et al. [5] reported 14–27% hypercalciuria in a cohort of control population identified from the three large studies: Nurses’ Health Study I (NHS I), NHS II, and Health Professional Follow-up Study, whereas it was 25–38% in stone formers in the same cohort. Coe et al. [6] assessed that 5% of American women and 12% of men will develop a kidney stone at some time in their life. A positive family history appears to be the single most important risk factor after controlling for known dietary factors [7]. In children with hypercalciuria, the prevalence of urolithiasis in the family is 69% [8]. Reed et al. [9] mapped a defect in three families with severe absorptive hypercalciuria to 1q23.3-q24, and they subsequently sequenced a putative gene (homologous to rat soluble adenylate cyclase gene). They identified 18 base substitutions in the putative gene, four of which increased the relative risk of absorptive hypercalciuria by 2.2- to 3.5-fold [10]. Vezzoli et al. [11] found single nucleotide polymorphism Arg990Gly in the calcium-sensing receptor (CASR) gene to account for 4.1% of total variance in calcium excretion and 12.6% of total variance in calcium excretion if independent variables of sodium excretion, body weight, serum creatinine, and enteral absorption of strontium were added to the multiple regression model. Imamura et al. [12] and Giuffre et al. [13] described three unrelated children with hypercalciuria who have 4q33-qter and 4q31.3-qter deletion, respectively, which raises the potential for a putative gene for hypercalciuria in that region. At this point, hypercalciuric trait is suspected to be polygenic and requires the interaction of genetic and/or environmental factors [14, 15]. A familial clustering of idiopathic calcium nephrolithiasis is frequently observed, most often compatible with an autosomal dominant transmission, but the quantitative genetics of urine calcium excretion has not been established. Loredo-Osti et al. [16] believe that either a mixed codominant/ polygenic model or a single-gene codominant model best determines the estimated inheritable attribute for idiopathic hypercalciuria, and thus it should be feasible to genetically map the quantitative trait locus for idiopathic hypercalciuria.

Calcium exists in three distinct pools in the body, where it is tightly regulated. The largest pool is that in the skeleton in the molar range, followed by the extracellular calcium pool in the millimolar range, and the third is in the intracellular space, which contains no more than 1 μm of calcium in an adult.

Pak et al. [89] introduced a tripartite classification of absorptive, renal, and resorptive hypercalciuria for idiopathic hypercalciuria. Over the years, investigators have further modified the classification based on urine calcium excretion, serum phosphate, and serum PTH secretion during fasting and after a calcium load [90]. It is now postulated that idiopathic hypercalciuria can occur from either or in combination with an (1) increased intestinal calcium absorption mediated either by a direct increase in calcium absorption (type 1 absorptive hypercalciuria) or through excess 1,25-(OH)₂D₃-mediated calcium absorption (type II absorptive hypercalciuria); (2) decreased renal absorption of either calcium (renal hypercalciuria) or phosphorus (type III absorptive hypercalciuria); (3) enhanced bone resorption (resorptive hypercalciuria) [91‐93]. Maierhofer et al. [94] showed that administration of 1,25-(OH)₂D₃ while eating a normal-calcium diet in healthy subjects led to an increase in intestinal calcium absorption and an increase in urinary calcium excretion, concluding that the key components of idiopathic hypercalciuria are related to calcitriol. They also showed that increased calcitriol administration in humans on a calcium restricted diet leads to negative calcium balance from increased urinary calcium loss mediated through increased bone resorption [95]. Similarly, vitamin D toxicity gives rise to hypercalcemia and hypercalciuria by stimulating intestinal calcium absorption. It is important to remember that hypercalciuria usually precedes hypercalcemia as an indicator of vitamin D overdose [96]. A fair number of investigators have observed that blood calcitriol concentration is, on average, higher in patients with idiopathic hypercalciuria or inappropriately normal for the condition compared with healthy subjects. Kaplan et al. [97] found elevated calcitriol levels in one third of absorptive hypercalciuria patients and normal values in two thirds, which may be considered as inappropriately high given the presence of relative hypoparathyroidism. Zerwekh and Pak [98] observed that in both renal and absorptive hypercalciuric subjects on thiazides, urine calcium excretion was normalized, but only the renal hypercalciuric group showed a decrease in intestinal hyperabsorption, PTH, and calcitriol level, whereas no changes were observed in the absorptive hypercalciuric subjects. These results would support different inciting events in development of hypercalciuria, namely, primary calcium leak in renal hypercalciuria and abnormal 1,25-(OH)₂-D₃ metabolism in absorptive hypercalciuria.

Data from NPT2a ⁻/⁻ mice that lack the Na⁺/Pi cotransporter have provided further insight into the role of 1,25-(OH)₂-D₃ in the development of hypercalciuria. The primary defect in tubular phosphate absorption in NPT2a ⁻/⁻ mice stimulates calcitriol synthesis by the kidney, which in turn increases intestinal absorption of calcium and phosphate and inhibits PTH secretion, resulting in hypercalciuria [45]. A disruption in the 1 α-hydroxylase gene in NPT 2a⁻/⁻ mice decreases urinary calcium excretion and prevents development of nephrolithiasis, signifying the importance of increased calcitriol synthesis in the development of hypercalciuria [99]. Similarly, in subjects with hypophosphatemic tubular disorder such as X-linked hypophosphatemic rickets (PHEX) and autosomal dominant hypophosphatemic rickets (FGF-23) with decreased production of calcitriol due to the inhibitory effect of FGF-23, no hypercalciuria is observed until therapy with calcitriol is instituted [100, 101]. Studies on the genetic hypercalciuric stone-forming (GHS) rat model suggest a role for an increase in number and/or function of vitamin D receptors (VDR) in enterocyte [102, 103]. Favus et al. [104] found an elevation in tissue VDR level in patients with idiopathic hypercalciuria with normal calcitriol level.

From studies done on GHS rats, it seems that the role of bone in the development of hypercalciuria appears to be as equally important as the intestines. While on a low-calcium diet, GHS rats continue to have markedly elevated urine calcium excretion exceeding their dietary intake, suggesting a role for increased bone resorption. Krieger et al. [105] showed that bone in GHS rats is sensitive to 1,25-(OH)₂D₃-induced bone resorption, and Bushinsky et al. [106] showed that alendronate decreases urine calcium excretion in GHS rats on a low-calcium diet to a level below their dietary intake (Fig. 4). Weisinger et al. [107] found alendronate to decrease urine calcium excretion in adults with hypercalciuria, and Freundlich and Alon [abstract to be presented at 14th International Pediatric Nephrology Association (IPNA) 2007 meeting] recently reported preliminary similar outcome in osteopenic hypercalciuric children. Cytokines are known to induce bone resorption and inhibit bone formation and may play a role in the rare of “resorptive hypercalciuria”, in which a primary bone disorder is believed to be the inciting defect. Ghazali et al. [108] found cytokines such as interleukin (IL)-1β, IL-6, tumor necrosis factor (TNF)-α, and granulocyte, macrophage stimulating factor to be increased in hypercalciuric calcium-stone-forming subjects with increased bone loss. Similar findings have been reported by Pacifici et al. [109], which allude to a role for cytokines in the development of hypercalciuria.

Hypercalciuria is a complex polygenic trait, and it is possible that absorptive and renal forms of hypercalciuria may represent a continuum of a single disease [110, 111]. When adults with idiopathic hypercalciuria are placed on a low-calcium diet, there is a continuum from those who are in “positive calcium balance”, suggesting a component of direct increased intestinal calcium absorption, to those in “negative calcium balance”, suggesting other mechanisms of hypercalciuria [110]. The lack of evidence of increased bone turnover in children with hypercalciuria suggests that renal and absorptive hypercalciuria may not be distinct physiologic entities [112]. An oral calcium loading test was popular in the past to differentiate between the different forms of idiopathic hypercalciuria; it has recently come under challenge [91, 113]. In children, Aladjem et al. [114] reevaluated calcium-loading tests after an interval of 3–7 years in children who were initially diagnosed as having absorptive or renal hypercalciuria and found a different result in more than half of the children studied. However, the classification, although not practical, has allowed investigators to develop a structured approach to the understanding of hypercalciuria.

A cross-sectional study from the Third National Health and Nutrition Examination Survey (NHANES III) showed that men (data weaker for women) with kidney-stone history had lower femoral neck BMD than did men without kidney stones after adjusting for age, body mass index (BMI), ethnicity, and other potential confounders, with a concomitant higher prevalence of wrist and spine fractures [115]. In a prospective study, Asplin et al. [116] found that the severity of urine calcium excretion best predicted bone loss in idiopathic hypercalciuric stone formers. Vezzoli et al. [117] observed lower BMD in hypercalciuric compared with normocalciuric stone-forming women, even in the presence of increased intestinal calcium absorption documented by strontium absorption. Similarly, Pietschmann et al. [118] found lower spinal BMD in hypercalciuric compared with normocalciuric stone formers. In contrast, Jaeger et al. [119] and Barkin et al. [120] found no difference in BMD between normocalciuric and hypercalciuric stone formers.

When hypercalciuric stone formers were studied based on Pak’s classification, the decrease in BMD was more frequent and greater in patients with renal hypercalciuria than in those with absorptive hypercalciuria [120, 121]. Other studies observed no reduction in BMD in absorptive hypercalciuria [122, 123]. The overall trend in the literature on adults with hypercalciuria, with or without stone disease, would suggest that they have lower BMD, but the results may vary based on subgroup analysis such as normocalciuric versus hypercalciuric stone formers or between renal versus absorptive hypercalciuria.

Penido et al. [124] found lower BMD in children with idiopathic hypercalciuria, and in a subsequent study [125], they found that these findings were more marked in children with hypocitraturia in addition to the hypercalciuria. Garcia-Neto et al. [126] had also made similar observation in an earlier study but made an interesting observation of negative linear correlation between age and bone mineral content in children with idiopathic hypercalciuria, namely, the Z score for BMD was much lower in older children with hypercalciuria, which raises the issue of whether adult osteoporosis has its origin in childhood. Similarly, Freundlich et al. [127] showed that reduced BMD was present in children with hypercalciuria and concomitantly found a high incidence of both hypercalciuria and reduced BMD in their asymptomatic mothers. The data in both children and adults indicate that the risk for bone loss is present in patients with hypercalciuria, and its origin may lie in childhood; hence, one must consider monitoring bone density as a proxy for calcium balance in children.

Diet can have a significant impact on calcium handling by the renal tubules. Urinary calcium excretion is significantly affected by sodium, protein, potassium, phosphorous, and calcium in the diet. Ninety percent of calcium absorption occurs as a paracellular event in the proximal tubule and TALH, which places calcium absorption at the mercy of sodium absorption. An increase in calcium delivery to the distal nephron for transcellular absorption can overwhelm the distal nephron, leading to obligatory hypercalciuria. An increase in either oral or intravenous sodium chloride inhibits net renal tubular calcium absorption and is used with beneficial effect in the treatment of hypercalcemia to increase urinary calcium excretion. The average consumption of salt in industrialized countries is 10 g (or 170 mmol Na)/day per person as determined by urinary 24-h excretion in the INTERSALT study [128]. Nordin et al. [129] have shown that approximately 1 mmol (or 40 mg) calcium is excreted for every 100 mmol (or 2.3 g) of sodium. There is a reproducible linear positive correlation between urinary sodium (a surrogate for dietary intake) and calcium excretion in stone formers as well as in normal individuals [130]. Thus, a diet high in sodium can lead to hypercalciuria [131].

An increase in dietary protein intake increases net acid excretion because of the release of protons from oxidation of sulfur in the amino acids methionine, cysteine, and cystine [132]. Conversely, dietary potassium found mostly in the form of potassium salts of metabolizable organic anions in vegetable and fruits, reflects the dietary intake of actual bicarbonate or potential bicarbonate, which reduce net acid excretion [133]. Urine calcium excretion increases as net acid excretion increases; hence, it rises progressively as the protein intake increases. The increment in urinary calcium excretion is ∼0.04 mmol (∼1.6 mg) calcium per gram of protein. The increase in calcium excretion with dietary protein is more marked in calcium-stone formers than in healthy subjects [91, 134]. Similarly in healthy subjects, an increase in dietary calcium increases urine calcium excretion by 6–7% of the dietary intake increment, whereas the change in calcium-stone formers is almost twice for the same increase in calcium intake [133]. A severe dietary phosphate deprivation induces increased calcium excretion, probably by activating the vitamin D endocrine system and thereby enhancing intestinal calcium absorption when calcium is available in the diet, or bone resorption when dietary calcium is low [94, 95, 135].

In the management of hypercalciuric stone formers, close attention must be paid to dietary intake and corrections made for dietary errors. Dietary hypercalciuria linked to excessive intakes of sodium, protein, or calcium or to deficiency in potassium or phosphate intake is diagnosed when calcium excretion is high while the patient is on his/her usual diet and normalizes during optimal dietary conditions. Thus, it appears that the more society deviates from the traditional balanced diet with optimal intake of protein, salt, fruits, and vegetables, replacing them with sodium-rich fast foods and artificial drinks, accompanied by a decrease in intake of potassium-rich fruits and vegetables, and increase in protein intake, the higher is the risk for “physiological hypercalciuria” [136].

Hypercalciuria in children can present as nonglomerular hematuria (gross or microscopic), noninfectious dysuria, urinary frequency and dysuria, abdominal and back pain, or with urolithiasis [137, 138]. It can be intermittent or persistent, a transient phenomenon or associated with a family history of urolithiasis. Once hypercalciuria is detected in a child, a secondary etiology should be considered, as successful correction of hypercalciuria in such cases depends on eradication of the primary cause. An evaluation for the rare monogenic disorders characterized by hypercalciuria should be considered in the presence of positive family history, failure to thrive, growth retardation, rickets, acid-base disturbances, renal dysfunction, proteinuria, electrolyte imbalance, dysmorphic features, or poor response to therapy.

Hypercalciuria, like blood pressure, is defined in children as urinary calcium excretion of >4 mg/kg per day on a “statistical” basis, whereas in adults, hypercalciuria is defined as >250 mg/day in women and >300 mg/day in men as an “outcome” value observed in most calcium-stone formers. The statistical cutoffs of 24-h urine >4 mg/kg per day or urine calcium/creatinine ratio >0.21 and their clinical implications have been recently addressed in detail by Butani and Kalia [139]. They raise many questions about planning a strategy for therapy in children. We believe that only symptomatic hypercalciuric children should be treated with pharmacological agents, whereas nonpharmacological intervention (vide infra) can be used more liberally.

When idiopathic hypercalciuria is confirmed in symptomatic children, we recommend as the first step to assess whether dietary manipulation can normalize calcium excretion. We recommend a Dietary Reference Intake for protein and calcium that is not excessive in salt (2.0–2.4 g) per day and supplemented with at least the recommended daily allowance of five to six servings of fruits and vegetables (3.0–3.5 g potassium) per day. Compliance with these dietary recommendations can be assessed by measuring urine Na⁺/K ratio, which should be <2.5. The dietary implications of salt, protein, and fruits and vegetables in hypercalciuria are well known, but we are cognizant of the fact that children may not fully comply with such dietary manipulations nor with the traditional recommendation of high fluid intake [4]. If in 4–6 weeks hypercalciuria persists, treatment with potassium citrate at 1–1.5 mEq of potassium per kilogram per day is recommended. If the child fails to tolerate potassium citrate or hypercalciuria fails to correct, a thiazide diuretic can be added [4, 136]. Chlorothiazide 15–25 mg/kg per day or hydrochlorothiazide 1.5–2.5 mg/kg per day can be used. In the past, it was proposed that thiazide-induced hypocalciuria occurred from volume contraction, which through increased proximal sodium absorption would increase the passive calcium absorption. Costanzo et al. [140] showed that acute administration of chlorothiazide in the tubular lumen stimulated transcellular calcium transport in the DCT. The earlier observation made by Costanzo et al. was recently confirmed by Jang et al. [141], who showed that thiazides increased the expression of TRPV5 and calbindin-D(28K) and decreased expression of sodium-chloride cotransporter in the DCT, leading to increased calcium absorption in the DCT. Children on long-term thiazide diuretics will need to be monitored for dyselectrolytemia, hyperlipidemia, and hyperglycemia. One can consider adding amiloride, as it further increases the hypocalciuric effect and decreases potassium and magnesium loss. Contrary to past practice, dietary restriction of calcium is not recommended in children with hypercalciuria, as it puts the growing child at risk for negative calcium balance and poor bone mineralization. It may also increase urinary excretion of oxalate from increased gastrointestinal absorption of oxalate resulting from decreased luminal calcium present to bind with oxalate. For risk of possible negative calcium balance, drugs such as sodium cellulose phosphate, a nonabsorbable ion-exchange resin used for complexing intestinal calcium, are not used in children. Phosphate salts can be used in children with hypercalciuria due to tubular phosphate leak. In children with hypercalciuria secondary to renal tubular acidosis, potassium citrate is the drug of choice for treatment of hypercalciuria. At times, this may need to be supplemented by sodium bicarbonate and calcium-sparing diuretics.

In summary, a better understanding of the rare inherited renal tubular disorders associated with hypercalciuria has improved our understanding of calcium handling by the kidney and development of hypercalciuria. On the other hand, our understanding of the more common idiopathic hypercalciuria, probably inherited as a polygenic trait and affected by the environment, remains dismal, and even more so in children. Many questions remain open: Is dietary manipulation adequate for all children, or should a subset of children be offered anticalciuric therapy given that dietary manipulation will not suffice or is not needed in them? Should all children be treated with anticalciuric therapy? Once an anticalciuric therapy is initiated, then for how long should it be continued? The data on BMD is suggestive of poor bone health in hypercalciuria; therefore, should all children have a dual-energy X-ray absorptiometry (DXA) scan with all its known pitfalls in children? Should DXA findings be considered in planning therapy for hypercalciuria in children? There are many such questions with respect to hypercalcuria in children that need to be addressed. We encourage the pediatric nephrology community to further address these issues in the hope of developing evidence-based care.