Introduction

In primary hyperparathyroidism (PHPT) with renal stone disease, it is currently recommended to perform parathyroidectomy (PTX) [1, 2]. In these patients, the prevalence of stone disease is higher [3] than in general population [4]. However, the presence of kidney stones has not always been an indication for PTX, as shown in guidelines from the third international workshop in 2009 [5]. Similarly, the measure of 24-h calciuria was no longer recommended in the preoperative management of this disease at this time. These observations underline that the link between stone disease and PHPT is not as univocal as it may appear. Interestingly, the risk of renal stones recurrence following PTX remains a major problem in patients with PHPT [6,7,8].

The reports of persistent hypercalciuria [8,9,10] and/or hypophosphatemia [11] in PTX patients suggests that this population may be different from PHPT patients with no stone disease [12]. This highlights the issue of some additional underlying calcium phosphate disorders in PHPT patients following PTX.

The aim of our study was to evaluate the effect of PTX both on calcium phosphate homeostasis, stone risk factors and stone recurrence among a well-characterized cohort of PHPT renal stone formers patients.

Methods

Study population

Over a decade, 1448 hypercalciuric stone patients were referred in our department for a calcium load test. Thirty patients were included in our study on the following criteria: patients with a final diagnosis of PHPT after a calcium load test, who underwent a PTX and performed a new load test within 1 year after PTX. PHPT was defined as the presence of hypercalcemia following the calcium load test with an inadequate serum parathyroid hormone (PTH) value (above 25 pg/ml). All 30 patients were explored before surgery by a parathyroid scintigraphy using methoxy isobutyl isonitrile labeled with Technetium-99 m (MIBI) for the localization of the hyperfunctioning parathyroid glands. Parathyroid ultrasounds were also performed in all cases to locate adenomas [13, 14].

Data collection

Clinical and biological data were collected from the department of physiology database (data collection was approved by the “Commission Nationale de l’Informatique et des Libertés” according to French legislation, n°2065902v0). Patients were referred to our unit for an oral calcium load following a 2-day calcium-free diet [15]. Hypercalciuria was defined as a daily urinary calcium excretion > 6.0 mmol/day and > 7.5 mmol/day for women and men, respectively. A fasting blood sample was analyzed for total and ionized calcium (iCa), phosphate, magnesium, creatinine, uric acid, PTH, 25(OH)-D3, 1,25(OH)-D3 and FGF-23. Among bone-remodeling biomarkers, bone alkaline phosphatase (BALP), cross-laps (CTX) and urinary deoxypyridinolin (DXP) were measured. A fasting urine sample over a 30-min period was also collected to assess phosphate reabsorption rate, TmPO4/GFR and creatinine clearance. An oral calcium load (calcium carbonate 1 g PO) was given [15], and following a 90-min delay, a 2-h urine collection was performed with simultaneous blood sample analysis for total and ionized calcium, serum phosphate, creatinine, and PTH (for this latter, one data was missing following calcium load). Biological data also included 24-h urine collections, performed under a normal diet. A diet survey aiming at evaluating daily calcium intake was executed during the out-day clinic stay. Urinary samples were collected after a 2-day calcium free diet. Of note, no medication was given during the calcium load test. A phone call survey was performed to assess stone recurrence follow-up after PTX among 28 out of 30 patients. Stone activity in our population was assessed as the ratio between the number of new events every year/number of patients. New events were defined as new renal colic episode, or surgery including shock wave lithotripsy, ureteroscopy or percutaneous nephrolithotomy.

Assays

Serum and urinary creatinine levels were measured by enzymatic method on a Konelab 20 analyzer (Thermo Fisher Scientific). Uric acid levels were determined with the Konelab analyzer. Total CO2 in blood, ionized calcium, sodium, and potassium levels were measured with an ABL 815 (Radiometer). Calcium and magnesium serum and urinary levels were measured using atomic absorption spectrometer (PerkinElmer 3300). PTH, 25(OH)-D3 and 1,25(OH)-D3 were analyzed by radioimmunoassay kits (Cisbio International and Immunodiagnostics Systems Ltd, Codolet, France). FGF-23 was measured by determination of human FGF-23 C-Term ELISA (Immutopics International). Urinary deoxypyridinolin was determined by the RIA method from Immunodiagnostics Systems Ltd. Bone alkaline phosphatase (BALP) serum levels were assessed by Enzyme immunoassay (Immunodiagnostics Systems Ltd). Crystallization indexes were calculated from 24-h urine collections with Equil 2 Software [16]. Altogether, Tiselius index (APCaOx) defined as 3.8 × Ca^0.71 × Ox × Mg^(− 0.14) × Citrate^(− 0.1) × 24-h Urine volume^(− 1.2), Coe index defined as 0.0253 Ca − 0.00548 Citrate + 0.224 (mg/l) for women and 0.01489 Ca − 0.00336 Cit − 0.391 (mg/g de creatinine) for men, calcium phosphate (SSBr), calcium oxalate (SSCaOx) and uric acid indexes [17,18,19] were used to determine the risk of stone formation.

Statistical analyses

Statistical analysis was performed using Xlstat© software (Addinsoft Paris©). Quantitative values are reported as means ± SD or median [Interquartile Range]. Comparison of quantitative data before and after PTX was assessed using a non-parametric Wilcoxon signed rank. The χ2 test was used to compare qualitative variables. A p value < 0.05 was considered statistically significant.

Results

Patient’s profile before PTX

Among the 30 renal stone formers, the mean age was 38 years for the first stone episode and 49 years for PTX surgery. Sixty percent of patients were women. Stone activity was 0.27 ± 0.45/year. Lithotripsy or renal ureteroscopy (RUS) were performed in 87% of patients (and in 37% of cases within 1 year prior to PTX). Body mass index (BMI) was 25.3 [23.0–27.3] kg/m2. Median systolic and diastolic blood pressure were 124 [115–137] and 77 [64–85] mmHg, respectively. Before surgery, no patient demonstrated chronic renal failure [i.e., estimated glomerular filtration rate (GFR) below 60 ml/min/1.73 m2].

We collected 22 stones from 17 out 30 patients. A large majority of stones from the cohort were calcium-dependent mainly composed of calcium oxalate dihydrate (COD) and calcium phosphate (CA) (Fig. 1). We also detected eight stones with calcium oxalate monohydrate (COM) as a first component. In these cases, the second component was still COD or CA. On the remaining stones, brushite (Br) was the main component (9% among all of cases). All patients were explored by MIBI. Twenty-eight patients experienced a single parathyroidectomy due to an adenoma, and 2 parathyroid glands were removed in two patients having hyperplasia (and no familial form of hyperparathyroidism).

Fig. 1
figure 1

Characteristics of stones among our population. Prevalence of first component and second component. COM calcium oxalate monohydrate, COD calcium oxalate dihydrate, CA carbapatite, Br brushite

Full size image

Following a 2-day calcium-free diet, median fasting ionized calcium (iCa) was higher than the physiological value (1.37 mmol/l as compared with normal value < 1.31 mmol/l), and associated with a high mean serum PTH value (Table 1). We also detected a mean low serum phosphate. Of note, fasting iCa was normal in 8/30 patients (26%) and among them four also had a normal PTH (Fig. 2a). After calcium load, all patients exhibited high iCa values associated with an inadequate PTH (Fig. 2b), confirming the diagnosis of primary hyperparathyroidism. Hypophosphatemia (< 0.85 mmol/l) was encountered in 73% of cases, and renal phosphate leak was confirmed by maximal tubular reabsorption threshold (Tm/GFR < 0.77 mM/l in all cases) [20]. 25(OH)-D3 was low (under 20 ng/ml) in 67% of cases with a high serum calcitriol in 57% of cases. Bone biomarkers such as BALP and DXP were high in 50% and 57% of cases, respectively.

Table 1 Biological data before and after PTX
Full size table
Fig. 2
figure 2

iCa and PTH values in patients before (a, b) and after PTX (c, d). a, c Pre-calcium load fasting values. b, d Post-calcium load values

Full size image

Patient’s profile after PTX

Calcium load measurement was performed 4.7 [3.3–6.5] months before, and then 4.9 [3.4–8.0] after PTX, respectively. As expected, PTX markedly decreased iCa and serum PTH (Table 1). No patient was displaying fasting hypercalcemia and four patients experienced a mild fasting hypocalcemia (Fig. 2c). Six patients demonstrated a high serum fasting PTH before calcium load. Among them, four normalized iCa and serum PTH after calcium load (Fig. 2d). Further, a mild hypercalcemia was detected in four other patients after calcium load, including three patients with PTH values lower to the normal range (≤ 30 pg/ml). Mean serum phosphate and Tm/GFR values significantly increased, supporting a correction of the phosphate leak. Nevertheless, 36% (11/30) of patients were still displaying hypophosphatemia (0.72 ± 0.9). No changes were observed in FGF-23 or vitamin D metabolites.

Following PTX, we observed a decrease of bone biomarkers (Table 1). In accordance, fasting urine calcium/creatinine ratio, a marker of calcium bone efflux, significantly decreased after PTX (p < 0.0001). Similarly, variation of urine calcium/creatinine before and after calcium load (also called the delta urine calcium/creatinine ratio (deltaUCa/Cr) that is a marker of intestinal calcium absorption), significantly decreased after PTX (p = 0.004). We did not detect difference for other stone risk factors such as diuresis, urine oxalate, urate, citrate or pH. Prevalence of hypercalciuria (assessed by 24-h urine collection on a normal diet) decreased only from 73 to 47% of cases following surgery (as shown in Table 1). A calcium diet survey and 24-h urine urea and sodium values allow to exclude potential co-funding factors like diet difference for calcium, proteins or salt intakes (Table 1). Nevertheless, assessment of stone risk factors by crystallization indexes (Coe, SSCaOx and SSBr) was significantly decreased after PTX (Fig. 3). A non-significant trend was also observed for APCaOx index (0.86 vs 0.68, respectively, p = 0.11), and no difference was detected for uric acid risk factors (0.53 [0.21–1.26] vs 0.58 [0.31–1.36], p = 0.28).

Fig. 3
figure 3

Comparison of different urine saturation indexes before and after PTX (24-h urine collection). SSCaOx Calcium oxalate supersaturation, SSBr Brushite supersaturation, APCaOx index Tiselius calcium oxalate index. Crystallization threshold values are indicated in dot lines

Full size image

Finally, a median follow-up of 4 [2–6] years was performed by a phone call survey. Six to three years before PTX, prevalence of stone events was observed between 18 and 32%. The prevalence was very high (50–75% events/year) within the 2 years preceding PTX (Fig. 4) and notably decreased thereafter. Nevertheless, a marked stone activity was still observed after PTX (5–15%), with nine patients who experienced clinical recurrence within the 6 years after surgery. Of note, stone recurrence was similar during the 3–6 years before PTX and following PTX with 5–15% of stone recurrence per year after PTX.

Fig. 4
figure 4

Renal stone events among patients at different time points before and after PTX (− 6 years to + 6 years). *Indicate a p value < 0.01 when comparing prevalence of stone events each year after PTX to the prevalence of stone events the year before surgery

Full size image

Discussion

Our study shows that in a well-characterized PHPT population with calcium-dependent renal stones, PTX is associated with a significant decrease of stone risk factor as well as recurrence and normalization of calcemia and PTH. Nevertheless, a fraction of patients still experienced stone activity indicating that PTX is not sufficient to fully abolish stone activity.

PHPT biological phenotype in renal stone disease patients is frequently diagnosed by the detection of increased serum calcium and high level of PTH, associated with hypercalciuria. Nevertheless, a new phenotype has been established over the past decade, which is characterized by a normal serum calcium with elevated level of PTH [21]. Interestingly, our data show a high prevalence of normocalcemia (26% of cases, i.e., 8 patients). In this work, performing a calcium load test under a calcium-restricted diet before PTX allowed the diagnosis of PHPT in up to 5% of our hypercalciuric renal stone population. This is a rather high prevalence than series reported in the literature [22,23,24]. Thus, the calcium load test allowed an early diagnosis in some patients and also provided a better understanding of calcium homeostasis in PHPT stone formers. Indeed, among eight patients with normal fasting iCa, four patients had also PTH values within normal range and thus could have been misdiagnosed. However, after calcium load, all patients displayed hypercalcemia associated with inadequate serum PTH values, thus assessing the diagnosis of PHPT. Notably, high serum PTH was encountered in 70% (21/30) and 38% (11/29) of patients before and after calcium load test, respectively. Thus, our study highlights the need to measure simultaneously iCa and PTH to assess PHPT diagnosis.

In accordance with the literature [25], our data show that PTX was effective, inducing a normalization of fasting iCa, PTH and bone remodelling biomarkers. Prevalence of hypophosphatemia also decreased significantly after PTX [26], probably explained by a decreased serum PTH level (as no FGF-23 decrease or calcitriol increase were detected). In addition, a decrease of urine CaP crystallization indexes was observed in line with a calcium urine reduction in all patients, emphasizing a lower risk to develop stones following PTX. As a matter of fact, the majority of stones components were COD and CaP (i.e., calcium-dependent stones). Conversely, other stone risk factors such as urate, pH, oxalate or citrate were not different before and after surgery and thus were not influenced by PTX.

Noteworthy, some metabolic disorders were persistent following PTX. Indeed, four patients experienced secondary hyperparathyroidism (fasting high PTH values with iCa < 1.20 mmol/l and a normal iCa and PTH after calcium load) after PTX. This could be explained by the 2-day calcium-restricted diet, associated in most cases with a low 25(OH)-D3 plasma level. Moreover, after calcium load, hypercalcemia associated with serum PTH values ≤ 30 pg/ml was detected in three patients. This raises the issue of a residual calcium-sensing receptor or a mild remnant PHPT following PTX. Interestingly, a persistent hypercalciuria was detected in 47% of cases after PTX. This was independent from dietary confounding factors [27], as no significant difference in calcium intake, sodium or urea excretion was observed after PTX. This might be due to underlying idiopathic hypercalciuria, as previously reported [10, 28,29,30,31]. Accordingly, Frøkjaer et al. [10] observed that the decrease of calciuria after PTX was significantly less marked in stone formers as compared with non-stone formers. Their work further suggests that an underlying PHPT-independent hypercalciuria is responsible for stone activity.

Our data also show that a remaining 0.05–0.15 stone/year activity persists after PTX during 6-year follow-up. Of note, an incidence of 0.20–0.30 stone/year was also recorded 3–6 years before PTX with a notable increase (up to > 0.5 stone/year) during the 2 years before surgery. Accordingly, Mollerup et al. [25] in a series of 167 renal stone patients also reported a higher frequency of stone events within the 4 years before PTX (with a peak in risk immediately before diagnosis). This indicates that the disease started several years before diagnosis. Indeed, Fig. 4 suggests that parathyroid gland growth occurred within a short time frame, between 2 and 3 years before surgery in most cases. In our series, PTX exerted a beneficial effect on stone recurrence as soon as the first year, but a remaining stone activity was detectable long term after surgery in a fraction of patients.

Study limitation

Our observational study suffers several limitations due to potential inclusion bias. Some patients with a severe hypercalcemia were not evaluated by a calcium load test and thus were not included. Moreover, free stone status was unknown at the time of PTX in most patients as renal echography or CT-scans were not regularly performed. Thus, some patients may not have been stone free, due to small residual stone after PTX (which did not require surgery before PTX). The increase of stone size and numbers during follow-up would have require additional sequential CT scans, which were not available in our cohort. However, stone recurrence was based only on clinical events both before and after PTX (see Fig. 4). This rules out this latter bias and provides a rather reliable estimate of stone activity within a noticeable 6-year period before and after PTX. Last, following PTX, we show a persistent hypercalciuria in nearly half of patients, suggesting a remaining calcium stone process at play. Unfortunately, data about stone composition following PTX were not available to confirm this speculation.

Conclusion

PTX in calcium-dependent renal stone patients with hypercalciuria significantly decreases stone recurrence, urine CaP saturation indexes, and normalize serum calcium, PTH and bone remodeling. However, following surgery, an idiopathic hypercalciuria persisted in half of patients with a noticeable remaining 0.05–0.15 stone/year activity, reflecting the need for a regular follow-up to prevent stone recurrence.