Introduction

Nephrocalcinosis (NC), defined as renal calcification, was first described by Hufnagle et al. in 1982 in premature neonates who received long-term furosemide therapy [1].

NC in preterm neonates occurs as a result of imbalance between stone-promoting and stone-inhibiting factors. The aetiology of NC is multifactorial and comprehensive and will be discussed in detail.

As nephrogenesis is not completed until 34–36 weeks of gestation, the development of the kidneys, both anatomical and functional, is not completed at birth [2]. After birth, rapid changes in functional development occur. Injury to the kidneys in this period can affect renal function later in life [3]. There is growing evidence of an association of low birth weight with low nephron numbers and subsequent risk for adult cardiovascular disease and renal insufficiency [4]. However, full consensus has not been established, as recent studies do not fully support the hypothesis that low birth weight contributes to impaired kidney function, at least not until the age of 20–26 years [5, 6]. Nevertheless, development of NC in prematurely born children may carry an additional risk of compromising renal function later in life.

We will discuss prevalence, radiological and histological diagnosis, aetiology, natural course and long-term effects, in addition to prevention and treatment of NC in preterm neonates.

Prevalence of nephrocalcinosis

NC is diagnosed in 7–64% of preterm neonates with gestational age < 32 weeks or birth weight < 1,500 g [7–16]. The wide range in prevalence of NC is a consequence of different study populations and ultrasound equipment and criteria, in addition to a moderate interobserver variation [12]. Recent studies have noted a slightly lower prevalence of 7–41% than in the original reports [12–16].

Diagnosis of nephrocalcinosis

Radiological evaluation

NC can be detected by conventional radiography, ultrasonography (Fig. 1) or computer tomography (CT). Ultrasonography is more sensitive than conventional radiography [17]. In rabbits with NC, ultrasound was more sensitive than CT (96% vs 64%), but CT was more specific than ultrasound (96% vs 85%) [18]. As CT involves a high radiation dose, it is unsuitable to detect NC in preterm neonates. Ultrasonography is a reliable method for screening and grading of young children with risk of NC, with good intra- and interobserver agreement (kappa coefficient, respectively, 0.80 and 0.76) [19]. Reproducibility of ultrasound in detecting NC in preterm neonates has a very good intra-observer agreement (kappa 0.84), but a moderate interobserver agreement (kappa 0.46) [12]. NC was located exclusively in the medulla in more than 95% of kidneys [12].

Fig. 1
figure 1

a Renal ultrasound of a preterm neonate with moderate nephrocalcinosis, with small white flecks in the tip of the pyramids. b Ultrasound of kidney of a preterm neonate with severe nephrocalcinosis. White dots almost entirely fill the pyramids

Full size image

However, increased medullary echogenicity in the preterm neonate is not exclusively found in NC, but also in other conditions, such as renal candidiasis, cytomegalovirus infection, acute renal failure, polycystic kidney disease, renal vein thrombosis, or as a transient spontaneously resolving phenomenon of unclear aetiology in the first postnatal week [20–26].

Histological evaluation

NC is defined as mineral precipitates located in the renal parenchyma. Knowledge of the pathological aspects of the ultrasound findings in preterm neonates is scarce. Histological examination of a few kidneys of preterm neonates with medullary echogenicity showed calcifications, located either within the tubules or in the interstitium, consisting of calcium oxalate or calcium phosphate crystals [1, 10, 27–29]. Renal histology of 44 infants and neonates who had died after intensive care treatment between 1972 and 1992 was compared with that of 64 infants and neonates who had died without intensive care treatment [29]. Intratubular calcium oxalate deposits were found in eight of the intensive care-treated patients, whereas two of these cases in addition showed intratubular calcium phosphate deposits, which, in one patient, extended into the interstitium. The patients without intensive care treatment, on the other hand, demonstrated minimal calcium phosphate microliths in two cases, but no oxalate crystals. This suggests that calcium oxalate crystals play an important role in NC in intensive care-treated patients.

Aetiology of nephrocalcinosis in preterm neonates

NC in preterm neonates has a multifactorial aetiology, consisting of low gestational age and birth weight [1, 7–9, 13, 14, 30], often in combination with severe respiratory disease [8, 13, 14, 30, 31], and it occurs as a result of an imbalance between stone-promoting and stone-inhibiting factors (Table 1). There is a clear correlation between prevalence of NC and low gestational age [7–9, 13, 14, 30]. Premature kidneys have relatively well-developed deep nephrons, with a long loop of Henle and probably low urine velocity. As a result, conditions are favourable for the formation of crystals, which can stick to the surface and grow and aggregate in the tubules. Under these circumstances stone-promoting factors such as hypercalciuria and hyperoxaluria, as described below, in combination with reduced stone-inhibiting factors like low urinary citrate excretion, can lead to NC.

Table 1 Aetiology of nephrocalcinosis in preterm neonates
Full size table

Of course, other well-known causes of NC, such as primary hyperoxaluria, distal tubular acidosis, glucose–galactose malabsorption, Bartter’s syndrome, Williams’ syndrome, and hypophosphatasia, can also occur in preterm neonates [32–35]. Here, we will focus on NC of prematurity.

Factors promoting nephrocalcinosis

Factors promoting nephrocalcinosis are shown in Table 1.

Hypercalciuria

Mean urinary calcium/creatinine ratio in preterm neonates varies from 2.3–2.7 mmol/mmol, which is much higher than reference values for term neonates and older children (< 0.6–0.9 mmol/mmol) [30, 36]. On the other hand, considerable variation is reported, with much lower (median 0.31 mmol/mmol) as well as much higher (mean 4 mg/mg = 11.4 mmol/mmol) values [37, 38].

Furosemide

Several (early) studies indicated furosemide for chronic lung disease to be the most important aetiological factor in the development of NC in preterm neonates [1, 7, 39]. Reabsorption of calcium in the loop of Henle is primarily passive, driven by the gradient created by sodium chloride (NaCl) transport. Inhibition by furosemide of the Na+ K+-2Cl− carrier in the apical membrane leads to a parallel reduction in the reabsorption of calcium and, hence, to hypercalciuria [40]. In preterm neonates calciuresis may be prolonged due to slower plasma clearance [41, 42]. However, development of NC has also been described in preterm neonates without furosemide therapy [8, 43].

Glucocorticoids

Many preterm neonates receive glucocorticoids for treatment or prevention of chronic lung disease [44–46]. High doses of glucocorticoids can lead to osteopenia, hypercalciuria and nephrolithiasis. The pathogenesis relates to an imbalance between resorption and formation of bone. An association between dexamethasone treatment, high calcium excretion and NC is, indeed, found in preterm neonates [13, 30, 47, 48].

Methylxanthines

Caffeine and theophylline, both methylxanthines, are frequently prescribed to prevent apnoea in preterm neonates and are also known for their hypercalciuric effect. Urinary calcium excretion rose two- to ten-fold in preterm neonates treated with methylxanthines in comparison with a control group [49, 50]. In keeping with this concept, a significant correlation was noted between theophylline prescription and NC in preterm neonates [30]. Suggested mechanisms to explain the hypercalciuria are increased diuresis and natriuresis, increased prostaglandin synthesis, and antagonism of adenosine-mediated effects with a change in renal blood flow and glomerular filtration rate [49, 51].

Acidosis

Preterm neonates may experience periods of respiratory and/or metabolic acidosis. Acidosis results in a number of changes in adults that increase the risk of stone formation: increased urinary calcium as a result of bone buffering, decreased urinary citrate by increased reabsorption in the proximal tubule and decreased urinary pH [52]. Significantly more preterm neonates with NC show a tendency toward metabolic acidosis than those without NC [15].

Calcium intake

As 80% of calcium and phosphorus accumulate in the fetus between the 25th post-conceptional week and full term, the majority of bone accumulation in preterm neonates takes place after birth [53]. Consequently, to prevent rickets of prematurity, the recommended intake of calcium (100–160 mg/kg per day) is high in comparison with that for full-term neonates [54]. High calcium intake is confirmed as a risk factor for NC in preterm neonates [30].

Vitamin D intake

Vitamin D excess can result in hypercalcaemia and hypercalciuria. In order to prevent rickets, the intake of vitamin D in preterm neonates is relatively high. Recent guidelines advise an intake of 800–1,000 IU [54]. In spite of this, vitamin D has not been found to be a risk factor for NC. Vitamin D intake and vitamin 1,25(OH)2-vitamin D3 plasma levels were equal in preterm neonates compared with those without NC [30].

Phosphorus intake

Recommended phosphorus intake for preterm neonates is 60–90 mg/kg per day and is like calcium intake, high in comparison with that for term neonates [54]. Phosphorus intake can influence the risk of developing NC in two ways.

Low intake of phosphorus can result in hypophosphataemia. Hypophosphataemia leads to a rise in phosphorus reabsorption in the proximal tubules, an increase in calcium and phosphorus absorption from the bone, and an increase in renal production of 1,25(OH)2-vitamin D3. Vitamin D stimulates the intestinal absorption and resorption of calcium and phosphorus from the bone. The consequent increase in serum calcium induces suppression of parathormone release, which results in a further decrease in urine phosphate and an increase in calcium excretion in adults [55]. In line with this hypothesis, Hein et al. confirmed that significantly more preterm neonates with NC had transient hypophosphataemia and hypercalciuria than did preterm neonates without NC [15]. In contrast Schell-Feith et al. did not find a difference in serum phosphate between preterm neonates with and without NC [30].

In contrast, a high intake of phosphorus is also associated with NC in preterm neonates [11, 30]. In addition, a high intake of phosphorus can lead to secondary hyperoxaluria, as will be explained later.

Sodium intake

In children and adults sodium intake and excretion are linked to urinary calcium excretion and can result in hypercalciuria. Similarly, sodium excretion in preterm neonates is positively correlated with urinary calcium excretion [56]. Nonetheless, the correlation between sodium excretion and occurrence of NC in preterm neonates has not yet been studied.

Parenteral nutrition

In preterm neonates on parenteral nutrition urinary calcium/citrate ratio increased in the first 10 days of life, while in human-milk-fed infants this ratio decreased, which indicates a higher risk of renal calcifications in preterm neonates fed parenterally [57]. Narendra et al. confirmed that NC occurs more frequently in preterm neonates with longer duration of parenteral nutrition [13], while Schell-Feith et al. could not find such a correlation [30].

Hyperoxaluria

As NC in preterm neonates can consist of calcium oxalate precipitates, high urinary excretion of oxalate might be an important factor in the development of NC [1, 29]. Hoppe et al. demonstrated that preterm neonates excrete more oxalate in the first months of life than do full-term neonates [58]. An association with higher urine oxalate/creatinine ratio was found in preterm neonates with NC as opposed to those without NC by some [13, 59], but not by others [30].

Precursors of oxalate

Ascorbic acid and glycine are precursors of oxalate. This could be the reason why preterm neonates receiving parenteral nutrition that contains ascorbic acid and glycine excrete higher urinary oxalate than do neonates receiving a glucose and electrolyte solution in the first week of life [60]. Formula-fed preterm neonates have a higher excretion of oxalate than do human milk-fed preterm neonates. This can be explained either by the fact that in formula for preterm neonates there is a higher concentration of ascorbic acid than in human milk, or by the mechanism of secondary hyperoxaluria, described below [61].

Secondary hyperoxaluria

If calcium is bound to a compound in the gut, less calcium will be available for oxalate, and, therefore, oxalate will form complexes with sodium. In the intestines sodium–oxalate complexes are more readily absorbed than calcium–oxalate complexes. Thus, enteral oxalate absorption will increase and higher concentrations of oxalate will be excreted in the urine. One of these calcium-binding compounds in the gut is fat. Because fatty acid absorption is more effective in human milk-fed infants than in formula-fed infants, relative fat malabsorption can induce secondary hyperoxaluria in formula-fed neonates [61]. Moreover, phosphate can also form complexes with calcium in the gut and lead to secondary hyperoxaluria. In patients with hypophosphataemic rickets, high intake of phosphate has been described to lead to hyperoxaluria and NC, even in a normocalciuric state [62].

Parenteral nutrition seems to increase not only urinary calcium/citrate ratio (see above), but also urinary oxalate excretion in preterm neonates. Therefore, parenteral nutrition may be an additional factor in the pathogenesis of nephrocalcinosis [57, 60].

Miscellaneous

Demographic characteristics

NC occurs more often in Caucasians, male gender, and preterm neonates with a positive family history of kidney stones [10, 13].

Nephrotoxic medication, e.g. gentamicin, has also been found to be a risk factor for NC in preterm infants [13].

Factors inhibiting nephrocalcinosis

Factors inhibiting nephrocalcinosis are shown in Table 1.

Citrate

Citrate is an important chelator of calcium stones, because it forms complexes with calcium that are more soluble than calcium oxalate and calcium phosphate. Preterm neonates on mechanical ventilation had a low urinary citrate excretion when compared with a control group [63]. Schell et al. confirmed a lower urine citrate/calcium ratio in preterm neonates with NC as opposed to those without NC [30]. Likewise, hypocitraturia was a major risk factor for NC in infants with very low birth weight, especially in those < 1,000 g [64]. In contrast, White et al. found similar values for urinary citrate in preterm neonates and healthy term babies. Furthermore, they found no association between urinary citrate and NC in preterm neonates [65]. A possible explanation for the lower excretion of citrate in some preterm neonates might be tubular reabsorption of citrate to compensate for episodes of respiratory and metabolic acidosis.

Magnesium

Magnesium is another low molecular weight chelator, which forms complexes with oxalate that are more soluble than calcium oxalate. The role of urinary magnesium excretion in the development of NC has not been studied in preterm neonates.

Stone-inhibiting macromolecules

Urine contains macromolecules, such as osteopontin, nephrocalcin and Tamm-Horsfall protein, which can inhibit the processes of nucleation, agglomeration and growth in vitro [66]. Osteopontin, a glycoprotein originally identified as a component of bone, has been associated with a variety of functions, including mineralization, signalling and cell adhesion. The ability of osteopontin to bind and coat calcium oxalate crystals in renal tubules suggests that osteopontin may play a role in the prevention of renal calcium oxalate accumulation [38]. Urinary osteopontin concentration is significantly lower in premature neonates than in adults [38]. However, levels do not differ from those of term infants; therefore, interpretation of these lower values in the aetiology of NC is still under debate, and additional studies are warranted. On the other hand, expression of crystal-binding molecules (hyaluronan and osteopontin) at the luminal surface of the distal tubular cells in preterm neonates precedes crystal retention at the distal nephron epithelium, the first step to NC [67]. Nephrocalcin and Tamm-Horsfall protein have not been studied in preterm neonates.

Thiazides

Thiazides, diuretics that inhibit sodium reabsorption in the distal tubules by blocking the thiazide sensitive Na+Cl− co-transporter, are frequently administered to preterm neonates with chronic lung disease. Thiazides decrease urinary excretion of calcium. This hypocalciuric effect can either be caused by an increase in proximal reabsorption of sodium and calcium in reaction to distal natriuretic effect of thiazides, or by a direct stimulatory effect of thiazides on the distal reabsorption of calcium. Nevertheless, Toffolo et al. have described preterm neonates with chronic lung disease who were treated with thiazides and developed NC anyway [43].

Animal model for nephrocalcinosis

In an animal model NC occurred within a few days of extremely high (40 mg/kg) furosemide administration in weanling rats. Interestingly, NC in this model was not age dependent but reflected a property of the loop diuretic itself [68].

Natural course and long-term effects of nephrocalcinosis

Nephrolithiasis

Rarely, urolithiasis is found in former preterms with severe NC. Downing et al. describes two patients who needed nephrolithotomy for ureteral obstruction [69]. That urolithiasis is not a frequent finding in NC is also suggested by a large prospective study, where none of the patients with NC developed urolithiasis during a 2-year follow-up period [70]. Haematuria is not a characteristic of patients with NC, unless they have urolithiasis [70].

Persistence of nephrocalcinosis

In the majority of patients spontaneous resolution of NC occurs in the first years of life. However, persistence for several years has been described [11, 14, 69, 71–74]. In the largest ongoing study of preterm neonates, NC persisted in 34%, 15% and 10% after 15 months, 30 months and 7.4 (± 1.0) years, respectively (Fig. 2) [70, 75].

Fig. 2
figure 2

Persistence of nephrocalcinosis (NC) with time, n = 70 (continuous line) (95% confidence interval dotted line) [70]

Full size image

Long-term consequences

Long-term consequences are shown in Table 2.

Blood pressure

Although there is evidence that low birth weight is an important factor in the development of hypertension and the metabolic syndrome in adults, prematurity did, but birth weight standard deviation score (SDS) did not, predict hypertension in a large cohort of 19-year-old former preterms (Table 2) [76, 77]. Likewise, blood pressure of children born preterm with and without NC did not differ when they were a mean age of 7.4 years. However, it was significantly higher than expected for healthy children, although only a minority of former preterms with (3/42) and without (2/31) NC in fact had a systolic blood pressure >95th percentile [75] (Table 2).

Table 2 Possible long-term effects of nephrocalcinosis and prematurity per se (TRP tubular reabsorption of phosphate)
Full size table

Renal growth

Decreased renal growth at the age of 20 years was noted in former preterms (gestational age < 32 weeks) compared to full-term born controls [5]. Likewise, the kidneys of patients with and without neonatal NC were significantly smaller than expected for healthy children of the same height [75]. It is hypothesized that preterm neonates are specifically at risk of renal growth impairment because nephrogenesis peaks at 32 weeks and continues until 36 weeks. So far as limited evidence suggests, NC does not further impair renal growth.

Glomerular function

Undoubtedly, the most important question is: does NC in very prematurely born children affect renal function in the long term? Several (early) studies comprising a limited number of selected children with preterm NC demonstrated reduced glomerular filtration rate (GFR) when they were aged between 1 year and 5 years [69, 71, 72]. In a prospective follow-up study, after 7.4 ± 1 years, significantly more children with NC (6/40, 15%) had low GFR (< 85 ml/min per 1.73 m2 body surface area) than did healthy children, this in contrast to children without neonatal NC (2/32, 6%), Fig. 3. However, there was no significant difference in GFR or microalbuminuria between the groups. Furthermore, no correlation between persistence of NC and low GFR was found, but the number former preterms with persisting NC (n = 4) was small [75]. This is in contrast with results of a study by Saarela et al., who found no significant difference in GFR at a mean age of 4.7 years between 20 former preterms with and 20 without NC in the neonatal period [73]. Hoppe et al. and Porter et al. also observed normal GFR after 3–6 years and 5.8–7.7 years, respectively, in 12 and 14 prematurely born children with NC [14, 74].

Fig. 3
figure 3

Estimated GFR of former preterm infants with (n = 42) and without (n = 32) neonatal nephrocalcinosis at a mean age of 7.4 years. +NC with nephrocalcinosis, −NC without nephrocalcinosis. The asterisk indicates that significantly more children with neonatal nephrocalcinosis have (mild) chronic renal insufficiency than do healthy children, P < 0.0001. The dashed line shows glomerular filtration rate < 85 ml/min per 1.73 m2 body surface area [75]

Full size image

In conclusion, long-term follow-up of preterm neonates with NC demonstrates normal renal function in most patients. However, an unfavourable effect on renal function is seen in a small number of children.

Proximal tubular function

How does NC affect long-term tubular function? Jones et al. studied 11 former preterms when they were aged 4–5 years with neonatal NC. They found low median TmP/GFR (tubular maximum of phosphate reabsorption corrected for glomerular filtration rate) in comparison with reference values [71]. Likewise, Saarela et al. observed a significantly higher urine β2-microglobulin/creatinine ratio in 20 children with NC than in 20 children without NC as preterm neonates, when they were a mean age of 4.7 years. However, in their study, tubular reabsorption of phosphate (TRP) did not differ significantly in children with and without NC [73]. In contrast, Downing et al. found significantly lower TRP (84 ± 2% versus 93 ± 1%, normal value > 85%) in former preterms at the age of 1–2 years with NC than in former preterms without NC [69]. Kist-van Holthe et al. also noted a significantly lower TRP in children with NC (n = 39) than in those without NC (n = 32) at a mean age of 7.4 ± 1 years. However, plasma phosphate was within normal limits in all children [75]. Therefore, the implication of low TRP in these patients is debatable. Glucosuria was not found in any of these children [75].

Considering these data, there is no firm evidence for proximal tubular dysfunction caused by neonatal NC in former preterms, with the exception of low TRP found in some patients (Table 2).

Distal tubular function

Downing et al. observed a lower ability to excrete hydrogen ions in the distal tubule in preterms with NC (n = 10) than in those without NC (n = 14) [69]. Median plasma bicarbonate level was significantly lower in a larger study of children with (n = 42) and without (n = 32) neonatal NC. Urine anion gap of the children with low plasma bicarbonate levels was inappropriately high, indicating distal rather than proximal tubular dysfunction [75]. On the other hand, Hoppe et al. noted no acidosis after a follow-up period of 3–6 years in 12 former preterms with neonatal NC [14].

Most studies found no difference in early morning urine osmolality between former preterms with and without NC [14, 69, 74, 75]. On the other hand, two studies demonstrated early morning urine osmolality to be significantly lower in former preterms than in healthy children [6, 71]. However, test results indicating impaired concentrating ability after desmopressin was found in 4/30 children with neonatal NC at 1 year and in 2/25 at 2 years [70]. In conclusion, NC in preterm neonates can have long-term sequelae, mainly for distal tubular acidification. Furthermore, some tubular defects (e.g. reduced early morning osmolality) cannot solely be attributed to NC but are also seen in former preterms without neonatal NC.

Hypercalciuria

Hypercalciuria is frequently seen in former preterms with neonatal NC. Significantly more (9/41, 22%) children with neonatal NC had hypercalciuria after a mean age of 7.4 years than expected for healthy children, in contrast to former preterms without NC (2/32, 6%) [75]. This is in keeping with the findings of other smaller studies in which hypercalciuria with or without NC is described in former preterms (Table 2) [69, 71, 73, 74, 78].

Prevention and treatment of nephrocalcinosis in preterm neonates

In children with hypercalciuria, urinary calcium excretion decreases after administration of thiazides [79]. Nevertheless, a study in rats showed that, once established, NC caused by furosemide is not affected by thiazide therapy, in spite of its anti-calciuric effect [80]. The effect of thiazides on the natural course of NC in former preterm neonates has not yet been studied.

Children with primary hyperoxaluria have been treated successfully with sodium citrate [81]. Citrate supplementation in preterm neonates for prevention of NC from day 7 until term at a dose of 0.52 mmol/kg per day did decrease the urinary calcium/citrate ratio and was safe. Although a positive trend, no significant decrease in the prevalence of NC was found [82]. Prevention of NC with higher citrate dosage in preterm neonates deserves further study.

The balance between high intake of protein, phosphate, calcium and vitamin D for accretion of tissues on the one hand, and the risk of renal damage on the other hand, remains delicate, implying that, also in the newly developed guidelines for preterm neonates, attention to the development of NC is warranted.

Summary

Prevalence of NC of prematurity varied from 7–41% in the different populations studied. NC in preterm neonates has a multifactorial aetiology, consisting of low gestational age and birth weight, often in combination with severe respiratory disease, and occurs as result of an imbalance between stone-promoting and stone-inhibiting factors. Although spontaneous resolution of NC occurs in most children, some are at risk of renal damage later in life, validating the screening for NC of preterm neonates. Long-term follow-up of preterm neonates, especially with NC, is warranted. Further research pertaining to prevention of NC is necessary.