Hypercalciuria May Persist After Successful Parathyroid Surgery and It Is Associated With Parathyroid Hyperplasia

Transcription

1 ORIGINAL ARTICLE Hypercalciuria May Persist After Successful Parathyroid Surgery and It Is Associated With Parathyroid Hyperplasia Serena Palmieri, Cristina Eller-Vainicher, Elisa Cairoli, Valentina Morelli, Volha V. Zhukouskaya, Uberta Verga, Marcello Filopanti, Leonardo Vicentini, Stefano Ferrero, Anna Spada, and Iacopo Chiodini Unit of Endocrinology and Metabolic Diseases (S.P., C.E.-V., E.C., V.M., A.S., I.C.), Endocrine Surgery Unit (L.V.), and Division of Pathology (S.F.), Fondazione IRCCS Ca Granda-Ospedale Maggiore Policlinico, Milan, Italy; and Department of Clinical Sciences and Community Health (S.P., E.C., V.M., V.V.Z., U.V., M.F., A.S.) and Department of Biomedical, Surgical and Dental Sciences (S.F.), University of Milan, Milan, Italy Context: Hypercalciuria is frequently found in primary hyperparathyroidism (1HPT) and, although it generally normalizes after successful parathyroidectomy, may persist in some patients. The factors associated with persistent calcium renal leak (crl) have not been clarified. Objective: The purpose of this study was to determine the prevalence of crl in our 1HPT population and investigate crl-related factors. Design: This was a retrospective longitudinal study. Setting: The study was conducted in an outpatient setting. Patients/Intervention: The participants were 95 patients with 1HPT successfully operated on who had a normal estimated glomerular filtration rate. Main Outcome Measures: The biochemical parameters of calcium metabolism and bone mineral density (BMD) measured by dual-x-ray absorptiometry before and 24 months after surgery were assessed. All histological findings were recorded. Results: The prevalence of hypercalciuria before and after surgery was 74% and 32%, respectively. Before, surgerypatientswithcrlshowedlowercalciumandhigherphosphatelevelsthanthosewithoutcrl( vs mg/dl [ vs mmol/l], P.01 and vs mg/dl [ vs mmol/l], P.04, respectively), whereas 24-h calciuria levels and the prevalence of 1HPT complications (osteoporosis, renal stones, and hypertension) were comparable. After surgery, serum calcium, phosphate, and PTH levels were comparable between patients with and without crl. The prevalence of the histological finding of parathyroid hyperplasia was higher in patients with crl (50%) than in patients without crl (22%) (P.01). The presence of crl was independently associated with presurgery hypercalciuria(oddsratio,4.71;95%confidenceinterval, ;P.03)andparathyroidhyperplasia(oddsratio, 3.52;95%confidenceinterval, ;P.01).OnlypatientswithoutcRLhadimprovedBMDatthespine (P.04), total femur (P.01), and femoral neck (P.01). Conclusions: crl is present in 30% of patients with 1HPT after successful surgery, and it is associated with parathyroid hyperplasia before surgery and the lack of improvement in BMD after surgery. (J Clin Endocrinol Metab 100: , 2015) ISSN Print X ISSN Online Printed in USA Copyright 2015 by the Endocrine Society Received December 28, Accepted May 4, First Published Online May 8, 2015 Abbreviations: BMD, bone mineral density; BMI, body mass index; crl, calcium renal leak; ES, elliptically symmetric distribution; FN, femoral neck; FT, total femur; 1HPT, primary hyperparathyroidism; LS, lumbar spine; 25OHVitD, 25-hydroxyvitamin D; ROC, receiver operating characteristic; SN, sensitivity; SP, specificity press.endocrine.org/journal/jcem J Clin Endocrinol Metab, July 2015, 100(7): doi: /jc

2 doi: /jc press.endocrine.org/journal/jcem 2735 Primary hyperparathyroidism (1HPT) is the third most common endocrine disease, affecting at least 1 in 1000 persons (1, 2). The incidence of 1HPT rises with age and peaks in the seventh decade (1, 2). Most cases occur in women, with a 5-fold excess in those older than 75 years of age, although the incidence is similar between sexes before 45 years of age (1, 2). 1HPT is due mainly to a single gland adenoma (80% 85%), less frequently to hyperplasia (10% 15%), and rarely to parathyroid carcinoma ( 1%) (1, 2). Hypercalciuria is frequently found in patients with 1HPT and is generally considered a risk factor for renal stone formation, even if the precise relationship between 1HPT and nephrolithiasis is only incompletely understood (3 5). Hypercalciuria may come from enhanced intestinal calcium absorption and/or increased mobilization of skeletal calcium resulting from the PTH excess and generally normalizes after successful parathyroidectomy. However, previous studies reported that hypercalciuria may persist after successful surgery (6 8). Unfortunately, these studies are flawed by the low number of patients enrolled and by some selection bias (overrepresentation of patients with advanced disease or stone former patients). Overall, the available data did not clarify the real prevalence of persistent calcium renal leak (crl) after successful parathyroid surgery for 1HPT. On the other hand, hypercalciuria may also be the cause (rather than only the consequence) of hyperparathyroidism and hypercalcemia (9). Indeed, in patients with idiopathic hypercalciuria, the increased urinary calcium excretion represents a stimulus for PTH secretion (10). Therefore, long-standing idiopathic hypercalciuria may lead to chronic parathyroid tissue stimulation, which in turn may predispose to parathyroid hyperplasia and adenomas and, eventually, to the development of a tertiary form of hyperparathyroidism (11). In this case, after the removal of the enlarged parathyroid glands and the recovery from hyperparathyroidism, hypercalciuria may persist. Hence, given this pathophysiology, one night expect the histological diagnosis of parathyroid hyperplasia to be more common in patients with crl after successful parathyroid surgery. However, no data are available regarding the relationship between crl after successful surgery for 1HPT and the histological findings. Finally, from a clinical point of view, the presence of persistent crl may have a detrimental effect on the recovery of bone density and promote the formation or recurrence of renal stones. Thus, the aims of our study were to determine the prevalence of crl in 1HPT after surgery, the association between parathyroid hyperplasia and persistent crl, and the effect of this latter on bone density changes. Subjects and Methods Subjects We retrospectively analyzed data from 126 consecutive patients with 1HPT operated on between 2008 and 2012 in our hospital. We selected only patients who had a complete metabolic profile with at least 2 24-hour urine collections and bone evaluation with bone mineral density (BMD) measurements and spinal radiographs before surgery and after 24 months (120 patients). From this group, we excluded 3 patients because of persistent 1HPT after surgery, 11 patients because of the use of drugs interfering with calcium metabolism, 10 patients because of possible renal failure (estimated glomerular filtration rate [egfr] of 60 ml/min, calculated by the Cockcroft-Gault equation) (12), and 1 patient because of a previously unrecognized familial benign hypocalciuric hypercalcemia, diagnosed by a calcium clearance/creatinine clearance ratio of 0.01 mg/dl (13) and subsequently genetically diagnosed. Eventually, 95 patients (69 postmenopausal women, 10 premenopausal women, and 16 men) were selected for further analysis. Methods In all patients, before and 24 months after surgery, serum and 24-hour urinary calcium, serum albumin, phosphate, PTH, and 25-hydroxyvitamin D (25OHVitD) levels were assessed. Serum and urinary calcium, albumin, and phosphate were measured by standard colorimetric techniques. Total calcium was corrected for serum albumin (Caalb adj) according to the formula Caalb adj (milligrams per deciliter) total calcium [(4.4 albumin in grams per deciliter) 0.8] (reference interval mg/dl [ mmol/l]) (14). Serum intact PTH was measured by electrochemiluminescence immunoassay (reference interval pg/ml [15 65 ng/l]). The serum 25OHVitD concentration was measured by chemiluminescent immunoassay (reference interval ng/ml [ nmol/l]). Before and 24 months after surgery, a dual-energy X-ray absorptiometry scan was carried out to measure BMD (Hologic Discovery) at the lumbar spine (LS) (in vivo precision of 1.0%), total femur (FT) (in vivo precision of 1.7%), and femoral neck (FN) (in vivo precision of 1.8%). Fractured vertebrae were excluded from BMD measurement. Individual BMD values are expressed as SD units (Z-values) in relation to our reference population. Conventional spinal radiographs in the lateral (T4 to L4) and anteroposterior projections were obtained in all patients by a standardized technique. Vertebral fractures were diagnosed on visual inspection using semiquantitative visual assessment and defined as a reduction of 20% in anterior, middle, or posterior vertebral height (15). Before surgery, all patients were screened for renal stones by abdominal ultrasound. In patients with persistent crl and patients with renal stones at the preoperative evaluation the abdominal ultrasound was repeated during the follow-up. All patients with hypovitaminosis D were given supplements of standard doses of cholecalciferol per os according to the Italian national guidelines (16) to achieve 25OHVitD levels of 30 ng/ml (75 nmol/l). Hypercalciuria was defined as urinary calcium excretion of 4 mg/kg body weight in both sexes with consumption of a normal calcium diet (17). In all patients, bilateral neck exploration with visualization of all parathyroid glands was performed before excision of any abnormal gland(s). Intraoperative PTH monitoring was not

3 2736 Palmieri et al Hypercalciuria After Successful Parathyroidectomy J Clin Endocrinol Metab, July 2015, 100(7): used. Sixty-two (65%), 10 (11%), and 16 (17%) patients had 1, 2, and 3 parathyroid glands removed, respectively. Seven patients (7%) were treated with subtotal parathyroidectomy; in these patients, parathyroid glands were resected with 1 2 of the most normal-appearing parathyroid gland preserved in situ on its vascular pedicle. Among the patients with adenomas, 55, 6, 4, and 1 had 1, 2, and 3 parathyroid glands removed, respectively. In 1 patient, 2 adenomas were found. Immediately after surgery, the excided parathyroid glands were measured by caliper on fresh tissue in our pathology department. The volume of each pathological parathyroid gland was assessed by the ellipsoid formula ( /6 a b c, where a, b, and c represent the diameters of the gland in 3 dimensions). For each patient, the sum of the volumes of the pathological parathyroid glands (total parathyroid volume) was calculated. The volume of the histologically normal glands was omitted in the calculation of the total parathyroid volume. One trained pathologist with expertise in parathyroid glands and blinded to the hormonal and biochemical data reviewed all histological examinations. For each sample, 3 sections, 2 min thickness, were cut and stained with hematoxylin and eosin. We considered parathyroid hyperplasia to be a symmetric enlargement of multiple parathyroid glands as a results of a proliferation of chief cells and oncocytes and transitional cell hyperplasia without an evident rim of normal gland (18, 19). The diagnosis of adenoma was established on histological examination by the presence of a single enlarged gland with a proliferation of parenchymal cells and an evident rim of normal gland and in the presence of normal histological findings in the other parathyroid glands. In the parathyroid adenoma, the parenchymal cell arrangement was found generally in a follicular and/or acinar pattern. In the presence of difficulties in the histological picture (ie, asymmetrical enlargement of multiple parathyroid glands), the diagnosis of hyperplasia was established in the presence of 2 or more enlarged glands together with a proliferation of chief cells with a reduction of adipocyte cells (20). A parathyroid gland was defined as enlarged when it measured 7 mm in length, 4mm in width, and 2 mm in thickness (21). Written informed consent was obtained from all subjects, and the study was approved by our ethics committee. Table 1. Statistical analysis Statistical analysis was performed by SPSS version 21.0 (SPSS Inc). The results are expressed as means SD or medians standard error (SE) in the case of non-gaussian distribution. The normality of distribution was tested by the Kolmogorov-Smirnov test. The comparison of continuous variables was performed using the Student t test or Mann- Whitney U test as appropriate. Categorical variables were compared by the 2 test. The comparison of the continuous variables in patients with and without persistent crl before and after surgery was performed by one-way ANOVA. Bonferroni post hoc analysis was performed as appropriate. Logistic regression analysis was used to assess the possible independent associations between persistent crl and the variables that were significantly different between patients with and without persistent crl. Receiver operating characteristic (ROC) curve analysis was performed to assess the best cutoff value(s) of the postoperative urinary calcium excretion (expressed as milligrams per kilogram of body weight) for predicting the presence of parathyroid hyperplasia on histological examination. Furthermore, the sensitivity (SN), specificity (SP), and overall accuracy of the best cutoff(s) of postoperative urinary calcium excretion for predicting parathyroid hyperplasia were assessed. P values of.05 were considered significant. Results The clinical and biochemical characteristics of patients with 1HPT before and after surgery are reported in Table 1. As for the inclusion criteria, all patients had normalized PTH and calcium levels after surgery. Consequently, after surgery, patients showed lower calcium, PTH, and 24-h calciuria levels and higher phosphate and 25OHVitD levels and LS BMD, FT BMD, and FN BMD than before surgery. No difference in body mass index (BMI) was found between baseline and the end of the follow-up. The Presurgical and Postsurgical Clinical and Biochemical Characteristics of Patients With 1HPT Presurgery (n 95) Postsurgery (n 95) P Age, y (16 to 81) (18 to 83).27 BMI, kg/m (20 to 37) (21 to 39).49 Calcium, mg/dl (10.4 to 13.8) (8.4 to 10).0001 PTH, pg/ml (60 to 931) (6.8 to 65) h calciuria, mg/24 h (120 to 980) (52 to 622) h calciuria, mg/kg/24 h (2.0 to 13.0) (1.0 to 7.0).0001 Patients with 24-h calciuria 70 (74) 30 (32) mg/kg body weight Phosphate, mg/dl (1.4 to 3.5) (2.6 to 4.5) OHVitD, ng/ml (12 to 91) (30 to 102).0001 egfr, ml/min (60 to 176) (60 to 175).79 LS BMD (Z-score) ( 4.0 to 4.0) ( 3.8 to 3.8).03 FT BMD (Z-score) ( 3.4 to 2.2) ( 3.2 to 2.0).01 FN BMD (Z-score) ( 3.3 to 2.2) ( 3.2 to 2.0).0001 Data are expressed as means SD (range) or absolute number (percentage). PTH is expressed as median SE (range) because of its non-gaussian distribution. SI conversion factors: calcium, 0.25; PTH, 1.0; 24-h calciuria, 0.02; phosphate, 0.32; 25OHVitD, 2.49.

4 doi: /jc press.endocrine.org/journal/jcem 2737 Table 2. Presurgical and Postsurgical Clinical and Biochemical Characteristics of Patients With 1HPT With Persistence or Normalization of Urinary Calcium Levels After Recovery From 1HPT and of the Subgroup of Patients With 1HPT Without Hypercalciuria Before Surgery Normalized Hypercalciuria (n 65) Persistent Hypercalciuria (n 30) No Hypercalciuria (n 25) Presurgery Postsurgery P Presurgery Postsurgery P Presurgery Postsurgery P Age, y (16 to 81) (18 to 83) (31 to 74) (33 to 76) (16 to 77) (18 to 79) BMI, kg/m (20 to 35) (21 to 37) (21 to 37) (21 to 39) (21 to 37) (21 to 39) Calcium, mg/dl (10.4 to 13.8) (8.5 to 10) (10.4 to 12.5) a (8.4 to 10) (10.4 to 12.6) (8.6 to 10) PTH, pg/ml (73.4 to 931) (15 to 65.5) (60 to 275.9) (6.8 to 64) (60 to 475) (15 to 62) 24-h calciuria, mg/24 h (120 to 980) (52 to 276) (163.7 to 600) (199 to 622) b (120 to 328) (52 to 472) 24-h calciuria, mg/kg/24 h (2 to 13) (1 to 3.9) (3 to 9.9) (4 to 7) b (2 to 4.0) (1 to 4) Phosphate, mg/dl (1.4 to 3.4) (2.6 to 4.5) (1.6 to 3.5) c (2.6 to 4.5) (1.6 to 3.5) (2.6 to 4.4) 25OHVitD, ng/ml (12 to 91) (30 to 78) (12 to 85) (31 to 102) (12 to 91) (30 to 70) egfr, ml/min (60 to 176) (60 to 175) (61 to 139) (60 to 137) (60 to 178) (60 to 176) LS BMD (Z-score) ( 4.0 to 4.0) ( 3.8 to 3.8) ( 3.1 to 2.2) ( 3.0 to 3.2) ( 3.1 to 2.2) ( 2.6 to 3.2) FT BMD (Z-score) ( 3.4 to 1.7) ( 3.2 to 1.8) ( 2.5 to 2.2) ( 2.1 to 2.0) ( 2.1 to 2.2) ( 1.5 to 2.0) FN BMD (Z-score) ( 3.3 to 1.8) ( 3.2 to 1.8) ( 2.3 to 2.2) ( 2.0 to 2.0) ( 2.1 to 2.2) ( 1.6 to 2.0) Hyperplasia/ 14/51 (23/77) 15/15 (50/50) d 6/19 (24/76) adenoma at histology Total parathyroid (0.2 to 7.3) c (0.1 to 8.0) (0.2 to 7.3) volume, ml Data are expressed as means SD (range) or absolute number (percentage). PTH is expressed as median SE (range) because of its non-gaussian distribution. Total parathyroid volume represents the sum of the volumes of the pathological parathyroid glands. SI conversion factors: calcium, 0.25; PTH, 1.0; 24-h calciuria, 0.02; phosphate, 0.32; 25OHVitD, a P.01 vs presurgery in patients with 1HPT with normalized hypercalciuria. b P.0001 vs post-surgery in patients with 1HPT with normalized hypercalciuria. c P.04 vs presurgery in patients with 1HPT with normalized hypercalciuria. d P.008 vs patients with 1HPT with normalized hypercalciuria. prevalence of vertebral fractures before surgery was 42%. No patient experienced a new vertebral fracture after surgery. The prevalence of hypercalciuria before and 24 months after surgery was 74% and 32%, respectively. The comparison of the clinical and biochemical features before and after surgery of patients with 1HPT with and without persistent crl is shown in Table 2. At baseline (presurgery), calcium levels were lower and phosphate levels were higher in patients with persistent crl than in patients without persistent crl. No differences were found in the other clinical and biochemical features. In particular, no difference in preoperative 24-h calciuria levels was seen. After surgery, with the exception of urinary calcium levels, no differences were found in the other biochemical features between patients with crl and those without crl. No difference was seen in the prevalence of the possible complications related to 1HPT (osteoporosis and/or vertebral fractures, renal stones, and hypertension) between patients with and without persistent crl presurgery and postsurgery. Interestingly, in patients with persistent crl, the histological finding of parathyroid hyperplasia was significantly more frequent. In addition, the total parathyroid volume was higher in patients with persistent crl than in those without persistent crl. Kidney stones did develop or worsen and nephrocalcinosis did not occur during the follow-up in any patient with persistent crl. In patients without persistent crl, BMD at both the spine and femur improved significantly after surgery, whereas no significant improvement was seen in patients with persistent crl. Urinary calcium levels were comparable before surgery between patients with adenoma (5 2.1 mg/ kg/24 h) and hyperplasia (6 1.6 mg/kg/h, P.51). After surgery, patients with hyperplasia showed higher urinary calcium levels than patients with adenoma (4 1.7 vs mg/kg/24 h, respectively, P.02). After surgery, the urinary calcium levels significantly decreased even in the subgroup of 25 patients without hypercalciuria before surgery (Table 2)

5 2738 Palmieri et al Hypercalciuria After Successful Parathyroidectomy J Clin Endocrinol Metab, July 2015, 100(7): Table 3. Association Between the Persistence of Hypercalciuria and Age, BMI, Basal Hypercalciuria, and Histological Findings by Logistic Regression Analysis Odds Ratio P 95% CI Age (1-y increase) BMI (1-kg/m 2 increase) Basal hypercalciuria (absence vs presence) Hyperplasia (presence vs absence) Abbreviation: CI, confidence interval. The logistic regression analysis assessed the association between the persistence of hypercalciuria postsurgery (dependent variable, expressed as a categorical variable) and independent variables: age, BMI, basal hypercalciuria (presence of hypercalciuria presurgery) and histological examination. The logistic regression analysis showed that persistent crl was positively and independently associated with presurgery hypercalciuriaandwithparathyroidhyperplasiaonhistological examination after adjustment for age and BMI (Table 3). The comparison of the clinical and biochemical features of patients with 1HPT with or without persistent crl and with the histological finding of parathyroid hyperplasia or adenoma is shown in Table 4. Patients with persistent crl and with the histological finding of parathyroid adenoma showed reduced calcium levels compared with those of patients with normalized hypercalciuria with either parathyroid adenoma or hyperplasia and lower PTH and higher phosphate levels than patients with normalized hypercalciuria and adenoma. In addition, in patients with persistent crl and adenoma, the number of subjects who had 1 parathyroid gland removed was higher than for patients with normalized hypercalciuria and adenoma. Finally, among patients with either normalized hypercalciuria or persistent crl, the subjects with parathyroid hyperplasia on histological examination showed a higher total parathyroid volume than the subjects with the histological findings of adenoma. Table 4. Baseline Biochemical Characteristics of Patients With 1HPT With Persistence or Normalization of Urinary Calcium Levels After Recovery From 1HPT and With Hyperplasia or Adenoma at Histology Normalized Hypercalciuria (n 65) Persistent Hypercalciuria (n 30) Adenoma (n 51) Hyperplasia (n 14) P Adenoma (n 15) Hyperplasia (n 15) P Age, y (16 to 81) (38 to 77) (31 to 72) (31 to 74) 1.00 BMI, kg/m (20 to 35) (20 to 30) (21 to 35) (21 to 37).73 Calcium, mg/dl (10.4 to 13.8) (10.4 to 12.1) (10.4 to 12.4) a (10.4 to 12.5).44 PTH, pg/ml (78 to 931) (73.4 to 202) (60 to 276) b (60 to 232) c h calciuria, mg/24 h (120 to 980) (149 to 563) (164 to 580) (248 to 600) h calciuria, mg/kg/24 h (2 to 13) (2.7 to 7.9) (3 to 9.6) (3.5 to 9.9).78 Phosphate, mg/dl (1.4 to 3.4) (1.7 to 3.3) (1.7 to 3.5) b (1.6 to 3.4).5 25OHVitD, ng/ml (12 to 91) (25 to 72) (16 to 85) (12 to 82).12 egfr, ml/min (60 to 176) (60 to 178) (61 to 139) (60 to 127).56 LS BMD (Z-score) ( 4.0 to 4.0) ( 1.6 to 0.2) ( 3.1 to 2.2) ( 3.0 to 2.1).95 FT BMD (Z-score) ( 4.0 to 1.5) ( 3.1 to 0.8) ( 3.4 to 0.3) ( 2.4 to 0.1).96 FN BMD (Z-score) ( 4.0 to 1.5) ( 3.4 to 1.4) ( 3.5 to 0.3) ( 2.7 to 0.8).88 No. of patients with 1 or 2 or 3 45/4/1/1 (88/8/2/2) 3/3/5/3 (21/21/37/21) /2/3/0 (67/13/20/0) f 4/1/7/3 (27/7/46/20).009 removed parathyroid glands d or subtotal parathyroidectomy e No. of patients with 1 6 (12) 11 (79) (33) f 11 (73).0001 removed parathyroid glands No. of patients with 1 or 2 or 3 50/1/0 (98/2/0/0) 4/3/4/3 (29/21/29/21) /0/0/0 (100/0/0/0) 4/2/6/3 (27/13/40/20).0001 or 3 pathological parathyroid glands Total parathyroid volume, ml (0.2 to 4.5) (1.5 to 7.3) (0.1 to 3.2) g (1.0 to 8.0) h.0001 Data are expressed as means SD (range) or absolute number (percentage). PTH is expressed as median SE (range) because of its non-gaussian distribution. Total parathyroid volume represents the sum of the volumes of the pathological parathyroid glands. SI conversion factors: calcium, 0.25; PTH, 1.0; 24-h calciuria, 0.02; phosphate, 0.32; 25OHVitD, a P.007 vs patients with 1HPT with normalized hypercalciuria and adenoma at histology. b P.004 vs patients with 1HPT with normalized hypercalciuria and adenoma at histology. c P.04 vs patients with 1HPT with normalized hypercalciuria and hyperplasia at histology. d Parathyroid glands that were removed and/or histologically examined during surgery. e Subtotal parathyroidectomy: removal of 3 parathyroid glands and half of the fourth. f P.04 vs patients with 1HPT with normalized hypercalciuria and adenoma at histology. g P.0001 vs patients with 1HPT with normalized hypercalciuria and hyperplasia at histology. h P.0001 vs patients with 1HPT with normalized hypercalciuria and adenoma at histology.

6 doi: /jc press.endocrine.org/journal/jcem 2739 In the histological sample from patients with parathyroid hyperplasia and adenoma, the mean SD percentages of adipocyte cells were % and %, respectively, of chief cells were % and %, respectively, of water cells were % and %, respectively, of oncocytic cells were % and %, respectively, and of transitional cells were % and 3 3.7%. In the 11 patients with parathyroid adenomas who had more than 1 parathyroid gland removed, the pathologist analyzed a total of 28 parathyroid glands and found 16 normal parathyroid glands. In the histological sample of these glands, the mean SD percentage of adipocyte cells was %, of chief cells was %, of water cells was 0 0%, of oncocytic cells was 8 8.8%, and of transitional cells was 0 0%. Consistent with the literature data (22), the quantitative histological results for the normal parathyroid glands showed more fat cells and fewer chief cells than the quantitative histological results for adenoma and hyperplasia. Figure 1 depicts the relation between the total parathyroid volume and the urinary calcium levels after surgery in each patient and for each group. In the whole sample of patients before surgery, no correlation was found between the total parathyroid volume and the urinary calcium levels (data not shown), whereas after surgery this correlation was present (Figure 1). The ROC curve analysis showed that the cutoff of postoperative urinary calcium excretion (expressed as milligrams per kilogram of body weight) with the best compromise between SN and SP for predicting the presence of parathyroid hyperplasia on histological examination was 3.5 mg/kg/24 h (SN of 63.6%, SP of 62.1%, and overall accuracy of 66.3%, P.009). The same analysis showed that the cutoff of 4 mg/kg/24 h (used in the present study for classifying crl) had the best overall accuracy (69.4%, P.005) and showed a higher SP (77.3%) but a lower SN (51.7%) than the cutoff of 3.5 mg/kg/24 h for individuating the presence of hyperplasia at histology. As expected, lowering the cutoffs of urinary calcium excretion led to an increase in SN at the expense of a decrease in SP and vice versa. Discussion The aims of this study were to assess the prevalence of persistent crl in a large cohort of surgically treated patients with 1HPT, to determine the possible relation between persistent crl and the histological findings, and to evaluate the possible detrimental effect of persistent crl Figure 1. Correlation between the postoperative urinary calcium levels and the sum of the volumes of the pathological parathyroid glands (total parathyroid volume) in patients with a histological diagnosis of either parathyroid adenoma or hyperplasia and persistence or normalization of urinary calcium levels after surgery. The postoperative urinary calcium levels directly correlate (R 0.25, P.013) with the sum of the volumes of the pathological parathyroid glands (total parathyroid volume). The 80% of patients with a histologic diagnosis of adenoma had a total parathyroid volume of 2 ml along with a normal urinary calcium level after surgery. The remaining 20% of patients with adenoma at histology had a total parathyroid volume of 2 ml along with an increased urinary calcium level after surgery. At variance, all patients with parathyroid hyperplasia showed a total parathyroid volume of 2 ml. The 50% of patients with persistent hypercalciuria after surgery had a histological diagnosis of hyperplasia, and the remaining 50% had a histological diagnosis of adenoma. The solid line represents the upper limit of the normal urinary calcium levels. Three patients showed a parathyroid adenoma total volume of 0.5 ml. In these subjects, the urinary calcium levels were 1.7, 2.7, and 6.1 mg/kg/24 h and the diameters of the adenomas were cm (volume 0.5 ml), cm (volume 0.32 ml), and cm (volume 0.45 ml), respectively. on the recovery of bone density after successful parathyroid surgery. We found that hypercalciuria persisted in 30% of patients with 1HPT even after successful parathyroidectomy. Moreover, for the first time, we report that persistent crl is associated with the histological finding of parathyroid hyperplasia. Finally, we found that patients with persistent crl showed no significant improvement in BMD Z-score at either the spine or femur 24 months after surgery. The possibility of crl after successful parathyroidectomy has already been described with a prevalence ranging from 34% to 60% (6 8). The low persistent crl prevalence found in the present study is probably related to the different design, sample size, and inclusion criteria compared with those in previous studies. Indeed, in the studies of Farias et al (6, 7), only 10 patients with histologically confirmed parathyroid adenomas were recruited, persistent crl was defined with nonconventional criteria, and patients had complicated and, probably,

7 2740 Palmieri et al Hypercalciuria After Successful Parathyroidectomy J Clin Endocrinol Metab, July 2015, 100(7): long-standing disease (6 patients with osteitis fibrosa cystica and 4 with recurrent renal stone disease). In the study of Spivacow et al (8), stone former patients were probably overrepresented. As suggested by some authors, chronic hypercalcemic nephropathy may explain all the alterations in renal hemodynamics and distal calcium reabsorption in patients with persistent crl (7). Moreover, this hypothesis might also be plausible in patients with 1HPT with advanced disease and with long-standing hypercalcemic hyperparathyroidism, as those enrolled in the previous studies (6, 7). In the present study, data regarding renal hemodynamics are lacking. However, we excluded patients with decreased renal function, and the presence of persistent crl was independent of the egfr. Therefore, at least in our sample, the presence of persistent crl does not seem to be related to chronic hypercalcemic nephropathy. In keeping with this result, a previous study did not find different renal function between patients with and without persistent crl after successful surgery for 1HPT (8). Overall these findings suggest that the tubular defect could have existed before the 1HPT or be an acquired defect due to sustained hypercalcemia (8). The noteworthy and novel finding of our study is that parathyroid hyperplasia was significantly more frequent in patients with persistent crl after successful parathyroid surgery. Moreover, before surgery, patients with persistent crl showed lower serum calcium and higher serum phosphorus levels than patients without persistent crl. Although entirely speculative, these findings may suggest that the pathogenesis of 1HPT is different in patients with persistent crl than in those without persistent crl. It may be hypothesized that in patients with persistent crl, a preexistent long-standing hypercalciuria had exerted a chronic stimulatory effect on the parathyroid glands that led to the development of parathyroid hyperplasia and eventually to hyperparathyroidism and hypercalcemia. This hypothesis may explain the finding that patients with persistent crl had a higher total parathyroid volume than patients with normalized hypercalciuria after surgery and that the total parathyroid volume directly correlated with the urinary calcium excretion after surgery. Finally, among patients with either normalized hypercalciuria or persistent crl, the subjects with parathyroid hyperplasia on histological examination showed a higher total parathyroid volume than the subjects with the histological findings of adenoma. This may be surprising, even if it is known that in primary parathyroid hyperplasia the total weight may range from 150 mg to 20 g (even though it is usually between 1 and 3 g), whereas the weight of a parathyroid adenoma is variable (between 300 mg and several grams) (21). Unfortunately, data regarding the weight of the excided parathyroid glands are not available, and, therefore, we could not compare our finding with literature data. Finally, an unrecognized bias in overestimating the size of lesser glands cannot be excluded. In keeping with our findings, in 129 patients with 1HPT treated surgically, D Angelo et al (10) found that hyperplasia prevailed in stone former and not in non-stone former patients and that stone former patients had lower levels of serum calcium and clinical evidence of recurrent hyperparathyroidism. Indeed, the presence of persistent crl could explain, at least in part, the sustained risk of stone disease in some patients for many years after surgery, even though normocalcemia had been established (23, 24). It should be emphasized that osteomalacia, rickets, and gastrointestinal malabsorption (bariatric surgery, celiac disease, and pancreatic disease) are well-known causes of secondary hyperparathyroidism, like renal failure (2, 25). In these patients, prolonged stimulation of the parathyroid glands by a low extracellular calcium concentration is described, and sometimes tertiary hyperparathyroidism can develop (2, 25). Moreover, in 97% of patients with tertiary hyperparathyroidism after successful kidney transplantation, the histological finding of multiple hyperplastic parathyroid glands has been reported (26). In keeping with these results, in the 50% of our patients with persistent crl, the histological diagnosis was consistent with parathyroid hyperplasia. This reinforces our hypothesis that in some patients, a primary metabolic defect in the tubular reabsorption of calcium may precede the development of autonomous hyperparathyroidism. It must be noted that the 50% of patients with persistent hypercalciuria after surgery, in fact, had a histological diagnosis of adenoma. This subgroup of patients showed significantly lower calcium and PTH levels and a higher prevalence of 1 parathyroid gland removed than patients with normalized hypercalciuria with parathyroid adenoma. This result suggests that, despite a better biochemical picture, the macroscopic presentation at surgery was more severe in patients with adenoma and persistence of hypercalciuria than in patients with adenoma and normalized hypercalciuria. Whether patients with parathyroid adenoma and persistence of hypercalciuria may represent a subgroup in whom a parathyroid adenoma has occurred in the setting of a preexistent long-standing hypercalciuria has to be determined. Finally, the finding that 14 subjects of the 65 patients with normalized hypercalciuria had a histological diagnosis of parathyroid hyperplasia may be explained by considering that hypercalciuria is not the sole chronic stimulus on parathyroid glands and that other factors may be involved in the pathogenesis of parathyroid hyperplasia (25).

8 doi: /jc press.endocrine.org/journal/jcem 2741 The third main finding of the present study is that only patients with normalized urinary calcium excretion had significantly improved BMD at both the spine and femur. This finding is in line with previous reports, showing that hypercalciuria is a well-known cause of secondary osteoporosis in both sexes (27) as well as being a risk factor for nephrolithiasis (5). However, it must be noted that in 1HPT the risk of fracture is only partially explained by the low BMD, because decreased bone quality may also play a role (28). Unfortunately, from the present study no information can be derived regarding the incidence of fractures in patients with and without persistent crl, as no incident fracture was recorded. Overall, the present findings are of clinical relevance. Indeed, the present data suggest that it is important to evaluate urinary calcium excretion after surgery for 1HPT, even in the presence of normal PTH and serum calcium levels, because 30% of patients may have persistent crl and remain, therefore, at risk of persistent osteoporosis. Our study has some limitations. First, because of its retrospective design, we cannot exclude the possibility that some unrecognized bias might have influenced the results. Second, specific studies of renal function are lacking, and, therefore, the possibility that persistent crl may be related to a chronic hypercalcemic nephropathy cannot be definitely ruled out. Third, there is a limit to the reliability of the differential diagnosis between parathyroid adenoma and hyperplasia, which in some patients may be difficult to establish (20). Unfortunately, we have not performed a clonality analysis to confirm the histological diagnosis of adenoma or hyperplasia. Classically, adenomas are defined as neoplastic clonal growths, whereas hyperplasias are considered to be reactive processes of polyclonal origin. However, previous and recent studies showed that even the clonality analysis cannot provide a clear distinction between these 2 entities, as classically diagnosed, and that a monoclonal expansion and polyclonal proliferation may coexist in parathyroid adenomas (29, 30). Finally, the cutoff for postoperative urinary calcium excretion (ie, 4 mg/kg/24 h) that we used, although derived from literature data (17), might be considered arbitrary. To check this point, we performed a ROC curve analysis for assessing the best cutoff(s) of postoperative urinary calcium excretion for predicting the presence of hyperplasia on histological examination. Although from the present data, the cutoff for urinary calcium excretion with the best compromise between sensitivity and specificity for detecting patients with parathyroid hyperplasia was set at 3.5 mg/kg/24 h, it must be noted that the commonly used cutoff of 4 mg/kg/24 h (17) had the best overall accuracy and higher specificity. Notwithstanding these limitations, the present finding of the association between persistent crl after successful surgery for 1HPT and parathyroid hyperplasia suggests that an idiopathic renal hypercalciuria may, in some cases, be the cause of later hyperparathyroidism. In conclusion, the present study shows that (1) persistent crl is present in 30% of patients with 1HPT successfully operated on (2) persistent crl is independently associated with the histological finding of parathyroid hyperplasia, and (3) patients with persistent crl shows no significant improvement in BMD postsurgery. Further studies are needed to ascertain whether a condition of chronic hypercalciuria may exert a causative role in 1HPT and whether the persistent crl has to be excluded after successful surgery for 1HPT for personalizing the therapy of choice in these patients. Acknowledgments Address all correspondence and requests for reprints to: Serena Palmieri, MD, Endocrinology and Diabetology Unit, Padiglione Granelli, Via Francesco Sforza 35, Milan, Italy. serena.palmieri@hotmail.it. Disclosure Summary: The authors have nothing to disclose. References 1. Marcocci C, Cetani F. Clinical practice. Primary hyperparathyroidism. N Engl J Med. 2011;365: Fraser WD. Hyperparathyroidism. Lancet. 2009;374: Silverberg SJ, Shane E, Jacobs TP, et al. Nephrolithiasis and bone involvement in primary hyperparathyroidism. Am J Med. 1990;89: Vezzoli G, Scillitani A, Corbetta S, et al. Polymorphisms at the regulatory regions of the CASR gene influence stone risk in primary hyperparathyroidism. Eur J Endocrinol. 2011;164: Corbetta S, Baccarelli A, Aroldi A, et al. Risk factors associated to kidney stones in primary hyperparathyroidism. J Endocrinol Invest. 2005;28: Farias ML, Delgado AG, Rosenthal D, et al. The cause of maintained hypercalciuria after the surgical cure of primary hyperparathyroidism is a defect in renal calcium reabsorption. J Endocrinol Invest. 1996;19: Farias ML, Delgado AG, Rosenthal D, et al. Changes in renal hemodynamics and tubular function of surgically cured primary hyperparathyroid patients are probably due to chronic hypercalcemic nephropathy. J Bone Miner Res. 1998;13: Spivacow FR, Negri AL, del Valle EE, Fradinger E, Martinez C, Polonsky A. Persistence of hypercalciuria after successful surgical treatment for primary hyperparathyroidism. Int Urol Nephrol. 2012;44: Maschio G, Vecchioni R, Tessitore N. Recurrence of autonomous hyperparathyroidism in calcium nephrolithiasis. Am J Med. 1980; 69: D Angelo A, Lodetti MG, Giannini S, et al. Hyperparathyroidism:

9 2742 Palmieri et al Hypercalciuria After Successful Parathyroidectomy J Clin Endocrinol Metab, July 2015, 100(7): cause or consequence of recurrent calcium nephrolithiasis? Miner Electrolyte Metab. 1992;18: Parfitt AM. Parathyroid growth: normal and abnormal. In: Bilezikian JP, ed. The Parathyroids. Basic and Clinical Concepts. New York, NY: Raven Press; 1994; Cockcroft DW, Gault MH. Prediction of creatinine clearance from serum creatinine. Nephron. 1976;16: Marx SJ, Attie MF, Levine MA, et al. The hypocalciuric or benign variant of familial hypercalcemia: clinical and biochemical features in fifteen kindreds. Medicine (Baltimore). 1981;60: Figge J, Jabor A, Kazda A, Fencl V. Anion gap and hypoalbuminemia. Crit Care Med. 1998;26: Genant HK, Engelke K, Fuerst T, et al. Noninvasive assessment of bone mineral and structure: state of the art. J Bone Miner Res. 1996; 11: Adami S, Romagnoli E, Carnevale V, et al. Guidelines on prevention and treatment of vitamin D deficiency. Italian Society for Osteoporosis, Mineral Metabolism and Bone Diseases (SIOMMMS). Reumatismo. 2011;9;63: Pak CY, Oata M, Lawrence EC, Snyder W. The hypercalciurias. Causes, parathyroid functions, and diagnostic criteria. J Clin Invest. 1974;54: Grimelius L, Bondeson L. Histopathological diagnosis of parathyroid diseases. Pathol Res Pract. 1995;191: DeLellis RA, Mazzaglia P, Mangray S. Primary hyperparathyroidism: a current perspective. Arch Pathol Lab Med. 2008;132: Grimeliu L, De Lellis RA, Bondeson L, et al. Parathyroid adenoma. In: De Lellis RA, Lloyd RV, Heitz PU, Heng C, eds. WHO Classification of Tumours. Tumours of Endocrine Organs. Lyon, France: IARC Press; 2004; Livolsi VA. Parathyroids morphology and pathology. In: Bilezikian JP, Marcus R, Levine MA, eds. The Parathyroids, Basic and Clinical Concepts, 2nd ed. San Diego, CA: Academic Press; 2001; Grimelius L, Johansson H. Pathology of parathyroid tumors. Semin Surg Oncol. 1997;13: Mollerup CL, Vestergaard P, Frøkjaer VG, Mosekilde L, Christiansen P, Blichert-Toft M. Risk of renal stone events in primary hyperparathyroidism before and after parathyroid surgery: controlled retrospective follow up study. Br Med J. 2002;12:325: Frøkjaer VG, Mollerup CL. Primary hyperparathyroidism: renal calcium excretion in patients with and without renal stone disease before and after parathyroidectomy. World J Surg. 2002;26: Cordellat IM. Hyperparathyroidism: primary or secondary disease? Reumatol Clin. 2012;8: Kebebew E, Duh QY, Clark OH. Tertiary hyperparathyroidism: histologic patterns of disease and results of parathyroidectomy. Arch Surg. 2004;139: Eller-Vainicher C, Cairoli E, Zhukouskaya VV, et al. Prevalence of subclinical contributors to low bone mineral density and/or fragility fracture. Eur J Endocrinol. 2013;169: Eller-Vainicher C, Filopanti M, Palmieri S, et al. Bone quality, as measured by trabecular bone score, in patients with primary hyperparathyroidism. Eur J Endocrinol. 2013;169: Sanjuan X, Bryant BR, Sobel ME, Merino MI. Clonality Analysis of benign parathyroid lesions by human androgen receptor (HU- MARA) gene assay. Endocr Pathol. 1998;9: Shi Y, Hogue J, Dixit D, Koh J, Olson JA Jr. Functional and genetic studies of isolated cells from parathyroid tumors reveal the complex pathogenesis of parathyroid neoplasia. Proc Natl Acad Sci USA. 2014;111: