Unit of Endocrinology-Diabetology, Department of Medical Sciences, University of Milan, Fondazione Ospedale Maggiore Policlinico, Mangiagalli e Regina Elena, IRCCS, Milan, Italy
Address correspondence to: Cristina Eller-Vainicher, MD, Unit of Endocrinology-Diabetology, Department of Medical Sciences, University of Milan, Foundation Polyclinic, Mangiagalli e Regina Elena, IRCCS, Milan 2012, Italy
The authors state that they have no conflicts of interest
Published online on March 16, 2009
Primary hyperparathyroidism (PHPT) is a common endocrine disease that is associated with multiple endocrine neoplasia type 1 (MEN1) in ∼2% of PHPT cases. Lack of a family history and other specific expressions may lead to underestimated MEN1 prevalence in PHPT. The aim of this study was to identify clinical or biochemical features predictive of MEN1 and to compare the severity of the disease in MEN1-related versus sporadic PHPT (sPHPT). We performed a 36-mo cross-sectional observational study in three tertiary referral centers on an outpatient basis on 469 consecutive patients with sporadic PHPT and 64 with MEN1-related PHPT. Serum calcium, phosphate, PTH, 25(OH)D3, and creatinine clearance were measured, and ultrasound examination of the urinary tract/urography was performed in all patients. In 432 patients, BMD was measured at the lumbar spine (LS) and femoral neck (FN). MEN1 patients showed lower BMD Z-scores at the LS (−1.33 ± 1.23 versus −0.74 ± 1.4, p = 0.008) and FN (−1.13 ± 0.96 versus −0.6 ± 1.07, p = 0.002) and lower phosphate (2.38 ± 0.52 versus 2.56 ± 0.45 mg/dl, p = 0.003) and PTH (113.8 ± 69.5 versus 173.7 ± 135 pg/ml, p = 0.001) levels than sPHPT patients. Considering probands only, the presence of MEN1 was more frequently associated with PTH values in the normal range (OR, 3.01; 95% CI, 1.07–8.50; p = 0.037) and younger age (OR, 1.61; 95% CI, 1.28–2.02; p = 0.0001). A combination of PTH values in the normal range plus age <50 yr was strongly associated with MEN1 presence (OR, 13.51; 95% CI, 3.62–50.00; p = 0.0001). In conclusion, MEN1-related PHPT patients show more severe bone but similar kidney involvement despite a milder biochemical presentation compared with their sPHPT counterparts. Normal PTH levels and young age are associated with MEN1 presence.
Primary hyperparathyroidism (PHPT) is considered a common endocrine disease with a prevalence of three per thousand in the general population(1) and of 2.5–3% in women >65 yr of age.(2,3) Multiple endocrine neoplasia type 1 (MEN1) is an autosomal dominant inherited disease (OMIM #131100) characterized by the occurrence of several endocrine tumors(4) and a prevalence of 2–10 per 100,000 in the general population. About 2% of all cases of PHPT are associated with MEN1, the MEN1 frequency in PHPT being estimated at 1–18%.(5,6) PHPT is the most highly penetrant expression of this syndrome, occurring in nearly 100% of all MEN1 patients by the age of 50 yr, and the earliest endocrine expression in most cases. In up to the 30% of mutation-positive MEN1 cases, however, a family history is lacking.(7) Thus, the lack of a family history or other expressions of the syndrome may lead to an underestimate of MEN1 prevalence in PHPT.(8)
Nevertheless, some clinical features differentiating PHPT in MEN1 from the sporadic form (sPHPT) are well known. In MEN1, the average age of onset of PHPT is typically ∼30 yr earlier than in sPHPT (second and third decades versus fifth and sixth decades in MEN1 and sPHPT, respectively). Moreover, the female/male ratio is even in MEN1-related PHPT but is closer to 3:1 in sPHPT. Finally, sPHPT is generally caused by a single adenoma, whereas in MEN1, all the parathyroid tissue is typically involved.(9) This difference is usually given considerable clinical weight. Indeed, in the absence of family history or at least one of the other two main MEN1-related endocrine neoplasias (e.g., pituitary adenoma and enteropancreatic neuroendocrine tumors), the search for MEN1 gene mutations is considered mandatory only in PHPT patients <30 yr of age in the presence of parathyroid hyperplasia or disease recurrence.(4)
To date, there are not large studies aimed to identify features highly suggestive of a MEN1-related form of PHPT. To get further insight into PHPT in MEN1, we therefore compared clinical and biochemical parameters of PHPT in a group of sPHPT and PHPT MEN1-related patients. The endpoints of our study were to identify differences in the severity of the disease and in the biochemical presentation of PHPT suggestive of MEN1.
MATERIALS AND METHODS
The study was performed in the following three Italian centers: “Fondazione Policlinico,” IRCCS, Milan, “Casa Sollievo della Sofferenza,” IRCCS, and San Giovanni Rotondo, “Policlinico Umberto I,” University of Rome.
From January 2005 to January 2008, 531 consecutive patients with PHPT were recruited. In all patients, the presence of MEN1 syndrome was evaluated by taking a full personal and family history, conducting abdominal/renal ultrasonography, and measuring serum prolactin and gastrin. In patients with hyperprolactinemia, the presence of drug interference or macroprolactinemia was ruled out. Patients with true hyperprolactinemia and those with signs or symptoms suggestive of pituitary disease underwent pituitary nuclear magnetic resonance. In patients with hypergastrinemia drug interference, atrophic gastritis, gastric outlet obstruction, previous vagotomy, chronic renal failure, or short bowel syndrome were excluded. Patients with true hypergastrinemia underwent gastric endoscopy, computer tomography (CT) of the abdomen, and the determination of gastrin levels after secretin stimulation. In patients with evidence of pancreatic lesions at abdomen ultrasonography, abdomen CT and pentetreotide scintigraphy were performed. Genomic DNA from all patients with a clinical diagnosis of MEN1 or familial PHPT was subjected to mutation screening of the MEN1 gene.
MEN1 syndrome was diagnosed in patients presenting at least two of three major MEN1 lesions—hyperparathyroidism, endocrine enteropancreatic, or anterior pituitary tumors—including both those with and without MEN1 gene germline mutation.(4) MEN1 syndrome was also diagnosed in patients with any single major MEN1 syndrome component and the presence of an MEN1 gene germline mutation.(4) All relatives of MEN1 probands who consented were screened for the presence of MEN1 syndrome.
Among the whole sample of 531 patients, we excluded 13 patients (2.4%) with familial hypocalciuric hypercalcemia (FHH), 3 (0.5%) with a diagnosis of MEN2, 4 (0.7%) with a diagnosis of familial PHPT without MEN1 gene mutations, and 4 (0.7%) with a histological diagnosis of parathyroid carcinoma (0.7%).
The presence of FHH has been assessed by calcium clearance. Subjects with calcium clearance <0.02 and familial PHPT without MEN1 mutations underwent calcium sensing receptor gene mutation analysis. We decided to use a calcium clearance cut-off of <0.2 to increase sensitivity for diagnosing FHH.(10)
Among the remaining 505 patients, 469 were classified as affected by apparently sPHPT and 36 by MEN1-related PHPT. Among the latter, an MEN1 germline mutation was identified in 31 subjects. In the remaining five subjects, the MEN1 diagnosis was substantiated by the presence of the three major MEN1 tumors or of two of three of the major MEN1 neoplasias in the presence of a positive family history.
Twenty-eight patients with MEN1-related PHPT, identified among relatives of patients diagnosed as having MEN1, were also enrolled in the study. Because no differences were found between MEN1 probands and their 28 relatives in respect to clinical characteristics (Table 1), these two groups were also considered together. Subsequent analyses were thus performed on a total of 533 patients (469 sPHPT and 64 MEN1-related PHPT subjects).
Table Table 1.. Clinical Characteristics of MEN1 Probands and Relatives
The study was approved by the local ethics committees, and a genetic study was made after informed consent was obtained from of each patient.
In all patients, serum total and ionized calcium (Ca2+), albumin, phosphate (P), intact PTH, 25(OH)D3, creatinine, prolactin (PRL), gastrin, and calcitonin levels were measured on a blood venous sample obtained after overnight fasting under a free diet, and urinary calcium (CaU), creatinine clearance (CrCl), and calcium clearance (CaCl) were measured in 24-h urine collections.
Serum calcium, albumin, and creatinine and urinary calcium and creatinine were measured by standard colorimetric techniques. Total calcium was corrected for serum albumin (Caalb adj) according to the formula: Caab adj (mg/dl) = total calcium + (4 − albumin mg/dl) × 0.8.(11) Plasma Ca2+ was measured by a potentiometric method (ABL System 625; Radiometer, Copenhagen, Denmark) on heparinized blood samples within 30 min of blood collection, with a reference interval of 1.15–1.29 mM. Serum intact PTH was measured by a chemiluminescent method (Nichols Institute Diagnostic, San Juan Capistrano, CA, USA), with intra- and interassay CVs of 5.1% and 8.2%, respectively (reference interval: 10–60 pg/ml or 10–60 ng/liter), and 25(OH)D3 by radioimmunoassay (RIA; DiaSorin) as previously described.(12) Reference intervals for Caab adj, Ca2+, and PTH are derived from a normal population recruited in our laboratory.
Plasma gastrin levels were measured using commercially available kits (Gamma Dab Gastrin [125I] RIA Kit (DiaSorin, Stillwater, MN, USA). Serum calcitonin levels were measured by a chemiluminescence assay (Nichols Advantage Calcitonin Chemiluminescence assay). The intra- and interassay CVs were 3.5% and 4.5%, respectively, and the upper limit of normal range was 10 pg/ml (10 ng/liter).
In 432 patients (388 sPHPT and 44 MEN1 patients), BMD was measured by DXA (Hologic, Waltham, MA, USA) at the spine (LS; DXA L2–L4, in vivo precision 1.0%) and femoral neck (FN; in vivo precision 2.3%). Individual BMD values were expressed as Z-score (age- and sex-matched comparison in SD units).
All subjects were routinely assessed for ultrasound examination of the urinary tract or urography, and a history of renal stones was specifically asked for. We defined patients as affected by nephrolithiasis if they had a history positive for kidney stones or in the presence of kidney stones or microstones at abdomen ultrasound or urography.
Statistical analysis was performed using the SPSS version 12.0 statistical package (SPSS, Chicago, IL, USA). The results are expressed as mean ± SD.
Comparison of continuous variables among the different groups was performed using Student t-test, Mann-Whitney U test, one-way ANOVA. General linear modeling (GLM) was used to compare PTH levels from MEN1 and sPHPT patients after adjusting for age and sex. Categorical variables were compared by χ2 test.
The bivariate associations between age, parameters of disease activity (Caalb adj, Ca2+, P, PTH, CaU, CaCl), and BMD were tested by Pearson product moment association or Spearman correlation as appropriate.
In PHPT probands (MEN1-related probands plus sPHPT patients), logistic regression analysis assessed the association between the presence of MEN1 syndrome (dependent variable, expressed as categorical variable) and the following independent variables: sex, age (expressed both as continuous—i.e., decades—or categorical—i.e., below or above 50 yr), and PTH (expressed both as continuous or categorical—i.e., below or above the upper limit of normal range—variable); the same analysis assessed the association between the presence of MEN1 syndrome and sex and the combination of age <50 yr plus PTH below the upper limit of normal range.
Logistic regression analysis was used to test the association between the presence of nephrolithiasis (dependent variable, expressed as categorical variable) and the following independent variables of age and log-transformed PTH levels or Caalb adj or Ca2+ (the latter three variables taken separately). p < 0.05 was considered significant.
At study entry, no patient had had parathyroidectomy before being enrolled. During the subsequent follow-up, 250 patients underwent parathyroidectomy (25 in MEN1-related PHPT, 225 in sPHPT) and 10 had reoperations (4 in MEN1-related PHPT, 6 in sPHPT). In 180 cases, the pathological finding was a single adenoma (0 patients in MEN1-related PHPT, 180 in sPHPT); in 5, multiple adenomas were found (0 in MEN1-related PHPT, 5 in sPHPT); and in 65, parathyroid hyperplasia was diagnosed (25 patients in MEN1-related PHPT, 40 in sPHPT).
Among MEN1 patients, no subject was affected by Cushing's disease, Cushing's syndrome, or hypopituitarism. Ten MEN1-related PHPT subjects (seven females, three males) were hyperprolactinemic, but no one was hypogonadal.
In MEN1 probands and their relatives, clinical characteristics were not different (Table 1). However, comparison of all MEN1-related and sporadic PHPT patients showed several important differences (Table 2). MEN1 patients were younger, showed lower serum P and PTH levels (the latter even after adjusting for age by GLM), and lower BMD at both the LS and FN (Figs. 1 and 2). However, urinary calcium excretion parameters and the prevalence rates of nephrolithiasis were not different in the two groups.
Table Table 2.. Clinical Characteristics of Patients With MEN1 and sPHPT
In MEN1 subjects, PTH and calcium levels were directly associated with age but not with urinary calcium excretion parameters and BMD (Table 3). Surprisingly, serum PTH was also not associated with plasma calcium levels. In the sPHPT group, in contrast, PTH levels were not correlated with age but were directly associated with serum calcium levels, urinary calcium excretion parameters, and inversely with BMD both at the LS and FN. Serum calcium levels were also found to be directly associated with urinary calcium excretion parameters and inversely with age and both spinal and femoral neck BMD (Table 3).
Table Table 3.. Bivariate Correlations Between PTH and Plasma Calcium Levels With Age, Urinary Calcium Excretion Parameters, and BMD in MEN1 and sPHPT Patients
Logistic regression analysis showed that, in sPHPT patients, but not in MEN1 patients, serum Caalb adj, Ca2+, and PTH levels were associated to the presence of nephrolithiasis even after adjustment for age (OR, 1.27; 95% CI, 1.02–1.57, p = 0.03; OR, 10.80, 95% CI, 2.28–51.20, p = 0.003; and OR, 3.25, 95% CI, 1.57–1.80, p = 0.002, respectively).
Comparing MEN1 probands and sPHPT patients, PTH levels and age were independently inversely associated with the presence of MEN1 syndrome regardless of sex (OR, 4.37, 95% CI, 1.03–18.52, p = 0.045; and OR, 1.58, 95% CI, 1.26–1.99, p = 0.0001, respectively). Moreover, one quarter of all MEN1 patients had PTH levels within the normal range (PTH < 60 pg/ml; Table 2). Serum PTH concentrations in the normal range were significantly associated with the presence of MEN1 syndrome, regardless of age and sex (Table 4).
Table Table 4.. OR for Detecting MEN1 for Potential Risk Factors Using the Multivariable Logistic Regression Model Analyzing Only Probands (n = 36)
In PHPT patients (sPHPT subjects and MEN1 probands), logistic regression analysis showed significant association between age <50 and presence of MEN1 syndrome regardless of PTH and sex (Table 5).
Table Table 5.. OR for Detecting MEN1 for Potential Risk Factors Using the Multivariable Logistic Regression Model Analyzing Only Probands (n = 36)
Finally, in PHPT patients (sPHPT subjects and MEN1 probands), the combination of normal PTH levels plus age <50 yr had a >13.5-fold higher risk of being diagnosed with MEN1 regardless of sex (Table 6).
Table Table 6.. OR for Detecting MEN1 for Potential Risk Factors Using the Multivariable Logistic Regression Model Analyzing Only Probands (n = 36)
Applying this combination as an adjunctive criterion for MEN1 gene analysis, 10 subjects <50 yr of age and showing PTH levels in the normal range of 505 patients (36 MEN1 probands plus 469 sPHPT patients) would have been screened for MEN1 gene mutations. Among these, five patients had already been diagnosed as MEN1 probands and five as affected apparently by sPHPT. On the basis of these findings, two patients among these latter have been screened for MEN1 gene mutations, and in one patient, the MEN1 gene mutation has been identified.
Our study indicates that PHPT MEN1-related patients have higher severity of bone involvement but the same prevalence of nephrolithiasis along with lower PTH and phosphate levels compared with sPHPT patients. In MEN1-related PHPT, serum calcium and PTH levels correlate with age, whereas urinary parameters of calcium excretion do not, and they do not correlate with kidney stones or bone involvement. In MEN1 patients, normal, but inappropriate, PTH levels are often present; this finding applied to nearly 40% of MEN1 cases <50 yr of age (Table 2). PTH levels in the normal range are associated with MEN1 syndrome, regardless of age and sex. Furthermore, a PHPT subject in our study <50 yr of age and having PTH levels in the normal range has a 13.5-fold higher likelihood of MEN1 syndrome than one >50 yr of age and/or having an elevated PTH.
This is the largest study aimed at identifying features suggestive of a MEN1-related form of PHPT. In comparison with previous studies regarding the same topic, this study included a much higher number of unrelated probands.(13,14)
The finding of lower BMD levels in MEN1-related PHPT patients is in keeping with a previous study(15) showing that PHPT is usually associated with osteopenia present by age 20–30 yr among women with MEN1. At variance with our findings, that study found an inverse correlation between PTH levels and BMD. However, we were able to enroll 36 MEN1 patients from 36 unrelated families, whereas the earlier study looked at women from a single large kindred.
We also found no difference in the prevalence of kidney stones between MEN1-related PHPT and sPHPT patients. Until now, very few data have been available about nephrolithiasis in MEN1-related PHPT.(16) In one older study, early urolithiasis was suggested to be common in MEN1 patients.(17) Our finding of a similar prevalence of urolithiasis in MEN1-related PHPT patients compared with sPHPT patients, despite younger age, suggests higher stone disease activity in MEN1-related PHPT. It is also important to consider that, because some previous studies have used different criteria for diagnosing MEN1,(12,13) the results obtained may be not fully comparable with those of this one.
In contrast to sPHPT, in MEN1-related PHPT, serum calcium and PTH levels do not correlate with urinary parameters of calcium excretion and are not predictive of kidney stones and bone involvement. Notwithstanding the small sample size of the MEN1 group, this finding may reflect a different behavior of PHPT in MEN1 syndrome. Although entirely speculative, this difference could be caused by higher PTH bioactivity. This would also explain our finding of the simultaneous presence of lower phosphorous and PTH levels in MEN1-related PHPT patients compared with sPHPT subjects. On the other hand, reduced downregulation of the calcium sensing receptor in the parathyroids of MEN1 patients can not be excluded. Further studies regarding these topics are warranted.
In our MEN1 group, age is directly associated with serum calcium and PTH levels, whereas sPHPT subjects showed a decrease of calcium levels with increasing age. The direct association between age and both calcium and PTH in MEN1 patients has been already reported in two previous studies looking at a Japanese MEN1 population.(13,14) This different behavior could be explained by the progressive proliferation of parathyroid tissue in MEN1-related PHPT patients.
In PHPT, the age of 50 is a crucial threshold, because, at this point, almost all MEN1 patients have developed PHPT.(4) Moreover, most cases of PHPT develop around the fifth decade,(18) and an age <50 yr has been shown to be an independent risk factor for developing PHPT complications.(11)
The novel conclusion of this study is the finding that the presence of normal serum PTH concentrations in patients <50 yr of age is associated with a 13.5-fold increased risk of having MEN1 syndrome. This is important as a clinical aid to avoid missing the diagnosis of MEN1 syndrome in patients affected with PHPT. Indeed, in young MEN1-related PHPT patients, the family history is often negative and unhelpful, and the lack of other MEN1 syndrome manifestations is frequent. However, the presence of hypercalcemia, normal PTH level, young age, and family history of hypercalcemia are also typical features of FHH that should always be taken into account.
Currently, in the absence of features and family history suggestive of MEN1, the determination of MEN1 gene mutations is considered mandatory only in PHPT patients <30 yr of age with parathyroid hyperplasia or primary hyperparathyroidism relapse.(4) The introduction of presurgical localization technologies (i.e., Tc99m-sestamibi scintigraphy, neck ultrasound) and intraoperative PTH sampling has led to the diffusion of “limited” surgical approaches, including focal (single gland) and unilateral (single side) explorations. As clearly pointed out by a recent study, this sort of surgical approach has failed to identify multiglandular disease in one of seven (16%) patients,(19) risking future recurrence. Therefore, it is likely that these surgical approaches, which seem to underestimate multiglandular disease, may lead away from the consideration of MEN1 as the underlying cause of PHPT. In this light, the combination we identified of normal PTH and age <50 should be a sensitive rule-of-thumb for identifying patients at high risk of an MEN1 syndrome before surgery and, thereby pointing to the need for genetic testing and selection of appropriate surgical treatment.
Put another way, we consider it advisable that, before performing parathyroidectomy, MEN1 gene mutations analysis should be undertaken on all apparently sporadic PHPT patients <50 yr who have PTH within the normal range. This approach seems unlikely to be cost ineffective because, in our study, only 1.9% of all subjects met these combined criteria and only 0.95% would receive a negative genetic analysis that possibly would have been uninformative. Our finding of MEN1 gene mutation in one of two subjects previously diagnosed as sPHPT but who underwent the genetic analysis on the basis of the combined criteria strengthens our conclusions.
Our study has several limitations. Because it was conducted in referral centers, the relative prevalence of MEN1 syndrome and sPHPT may not reflect the prevalence of the two diseases in the general population. Second, also enrolling relatives of MEN1 probands in the analysis may have introduced a bias in the patient selection. However, comparing only MEN1-related PHPT probands with sPHPT patients showed essentially the same results.
In conclusion, our study showed that, in PHPT, (1) patients affected with MEN1 syndrome show more severe bone and similar renal chronic involvement despite a milder biochemical presentation compared with those affected with the sporadic form; and (2) a combination of normal PTH levels and an age <50 yr is a strong and clinically useful indicator of MEN1 syndrome that should warrant genetic testing before selection of surgical treatment options.
This work was supported by grants from Ministero della Salute of Italy (Ricerca Corrente 2001, 2002 to A.S. and 2007, 2008 to L.D.).