Our study showed that as much as 30% of the patients seen in pediatric nephrology practice had renal Mg wasting. Hypomagnesemia was seen in 13% and all had renal Mg wasting. In adults with chronic kidney disease (CKD), hypomagnesemia was the most common electrolyte abnormality and seen in 14.7%, similar to our results [2]. The frequency of renal Mg wasting without hypomagnesemia in adults with CKD is unknown but assumed to be higher than that of hypomagnesemia.
As a marker of renal Mg wasting, FEMg is more commonly used than urine Mg/Cr because one does not have to compare with reference values. FEMg has been reported to be elevated in patients with tubulointerstitial fibrosis in adults [1, 10] and to increase along with CKD progression [11]. FEMg, however, is affected by SMg as well as eGFR, and was previously suggested to be useful only in patients with eGFR > 30 ml/ml/1.73 m2 in adults [10]. In children, the cut-off of value FEMg >4% is also used [3]. In the presence of hypomagnesemia, the cut-off value of FEMg is set at 2% in view of the increased Mg reabsorption along the tubules [4]. In the present study all patients with hypomagnesemia had FEMg >2%. Urine Mg/Cr, however, was elevated in only 4, in agreement with the increased Mg reabsorption. Since there was no patient with FEMg <2% in patients with hypomagnesemia, we could not determine whether this value is the best. In normomagnesemic patients with normal eGFR, 32% of the patients with FEMg >4% had normal urine Mg/Cr. Only one patient with FEMg <4% had elevated urine Mg/Cr. The ROC curve analysis gave the cut-off value of 4.1% with high accuracy (93% sensitivity and 90% specificity). The sensitivity was the same with the commonly used cut-off value of 4.0%, therefore it is reasonable to use FEMg >4% for the screening of Mg wasting in patients with normal SMg and eGFR. The confirmation by urine Mg/Cr, however, is necessary.
In patients with normal SMg and low eGFR, FEMg proved not to be useful in detecting Mg wasting. In these patients urine Mg/Cr should be used from the beginning.
Taken together, for hypomagnesemic patients, FEMg >2% is thought to be an appropriate cut-off. For patients with normal SMg and normal eGFR, FEMg >4% can be used as a screening test, but Mg wasting should be confirmed by increased urine Mg/Cr because the specificity is not high enough. For patients with normal SMg and reduced eGFR, urine Mg/Cr not FEMg should be used (Fig. 2).
The underlying disease conditions of Mg wasting included low eGFR, CAKUT, hematuria, urolithiasis, Fanconi syndrome, ciliopathy, and use of calcinuerin inhibitors. As mentioned above, hypomagnesemia is common in adults with CKD [2]. Although in Fultrakul’s study, tubulointerstitial fibrosis correlated with renal FEMg [1], the frequency of hypomagnesemia was not different among different CKD staging in Oka’s study [2]. The frequency and characteristics of Mg wasting with normomagnesemia in adults need to be studied in the future. In the present study, urine ß2m/Cr and NAG/Cr were associated with Mg wasting to a certain extent. Noiri et al also showed a correlation between FEMg and NAG/Cr in adults [10].
In Oka’s study in adults, 46.7% of the patients had diabetes, and diabetes was a risk factor for hypomagnesemia [2]. Urine Mg excretion has been associated with metabolic factors such as body mass index, fat mass, and fasting insulin levels [12]. TRPM6, which encodes an epithelial Mg channel, has been shown to be influenced by metabolic genes [12]. In our patients there were no diabetics, which are not very common in pediatric nephrology practice.
Twenty-six percent of patients with CAKUT had Mg wasting. Although the CAKUT gene HNF1B is known to cause Mg wasting [13], there was only one patient with HNF1B abnormality in the present study. The reason for Mg wasting in CAKUT in the present study is unknown but could be tubulointerstitial fibrosis.
It was unexpected that 25% of patients with hematuria had Mg wasting. Several possibilities are considered. Uromodulin has been reported to be associated with Mg wasting through epithelial Mg channel TRPM6 [14], although there has been no association between hematuria and either uromodulin or TRPM6. Hypercalciuria and hyperuricosuria are known to cause hematuria. Urine Mg, however, works to prevent crystallization, and Mg ammonium phosphate crystallization is rare in our country.
The high frequency of Mg wasting in Fanconi syndrome is well expected in view of the fact that 20–25% of filtered Mg is reabsorbed at proximal tubules. Fanconi syndrome is a well-known differential diagnosis of hypomagnesemia [15]. Compared with other solutes, however, Mg wasting in Fanconi syndrome tends to be understated. This study underscores the importance of searching for Mg wasting routinely in Fanconi syndrome.
Mg wasting was also seen in patients with ciliopathy. In autosomal dominant and recessive polycystic kidney disease, increased FEMg has been reported [16, 17]. Further examination in other ciliopathies is warranted.
Mg wasting by calcineurin inhibitors is reported to be linked to the downregulation of epidermal growth factor and of Mg transport proteins including TRPM6 at distal tubules [18, 19].
In conclusion, we have established an algorithm to diagnose Mg wasting in pediatric nephrology practice (Figure. 2). Mg wasting is not uncommon and patients should be routinely investigated.