| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Departments of Medicine (M.R.R., J.P.B., D.J.M., T.J., E.Sh., E.Si., J.U., S.J.S.) and Pharmacology (J.P.B.), Columbia University, College of Physicians and Surgeons, New York, New York 10032
Address all correspondence and requests for reprints to: Shonni J. Silverberg, M.D., Department of Medicine, Columbia University, College of Physicians and Surgeons, 630 West 168th Street, New York, New York 10032. E-mail: sjs5{at}columbia.edu.
 |
Abstract |
|---|
 Top Abstract IntroductionPatients and MethodsResultsDiscussionReferences |
|---|
Objective: The objective of the study was to provide additional insights in patients with and without surgery for up to 15 yr.
Design: The study had an observational design.
Setting: The setting was a referral center.
Patients: Patients included 116 patients (25 men, 91 women); 99 (85%) were asymptomatic.
Intervention: Fifty-nine patients (51%) underwent parathyroidectomy and 57 patients were followed up without surgery.
Main Outcome Measure: BMD was measured.
Results: Lumbar spine BMD remained stable for 15 yr. However, BMD started to fall at cortical sites even before 10 yr, ultimately decreasing by 10 ± 3% (mean ± SEM; P < 0.05) at the femoral neck, and 35 ± 5%; P < 0.05 at the distal radius, in the few patients observed for 15 yr. Thirty-seven percent of asymptomatic patients showed disease progression (one or more new guidelines for surgery) at any time point over the 15 yr. Meeting surgical criteria at baseline did not predict who would have progressive disease. BMD increases in patients who underwent surgery were sustained for the entire 15 yr.
Conclusions: Parathyroidectomy led to normalization of biochemical indices and sustained increases in BMD. Without surgery, PHPT progressed in one third of individuals over 15 yr; meeting surgical criteria at the outset did not predict this progression. Cortical bone density decreased in the majority of subjects with additional observation time points and long-term follow-up. These results raise questions regarding how long patients with PHPT should be followed up without intervention.
|
Introduction |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
In 1984 we initiated a prospective study to define the natural history, pathophysiology, densitometric, and other skeletal abnormalities of primary hyperparathyroidism (PHPT) in the multichannel screening era (3). In terms of that natural history component of this study, we previously reported on this cohort at the conclusion of the first 10 yr of observation. In this report, mixed model analysis (see Statistical analysis below) allows a more statistically efficient attempt at detecting systematic patterns than the methods we previously used. This has permitted an analysis of all the data, which now includes more time points within the 10 yr and extends to 15 yr of prospective longitudinal follow-up. The longer follow-up period has led to new observations that have implications for decision-making and for management of this disorder.
|
Patients and Methods |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
The study, which began in 1984, enrolled 116 subjects with PHPT who had at least one follow-up bone mineral density test (Fig. 1
). The protocol was approved by the Institutional Review Board of Columbia University Medical Center, and all patients gave written informed consent.
|
600 mg/d). Intake of no more than 1000 mg (without calcium supplementation) was generally recommended for patients who did not undergo surgery; compliance with this recommendation was not specifically monitored. Study protocol
All patients underwent biochemical studies at baseline and every 4 months if they were followed up with no intervention, or every 6 months if they had parathyroid surgery Serum calcium, phosphorus, and alkaline phosphatase activity were measured by automated techniques (Technicon Instruments, Tarrytown, NY). Serum PTH was measured by immunoradiometric assay (interassay coefficient of variation, 5.6%) (6), urinary calcium by atomic absorption spectrophotometry, and serum 25-hydroxyvitamin D and 1,25-dihydroxyvitamin D as previously described (interassay coefficients of variation, 9.6 and 9.8%, respectively) (7).
Bone mineral density (BMD) of the lumbar spine, femoral neck, and distal third of the nondominant radius was measured at baseline and yearly thereafter. The short-term in vivo precision error (root mean square SD) was 0.026 g/cm2 for L1-L4, 0.032 g/cm2 for the total hip, 0.041 g/cm2 for the femoral neck, and 0.033 g/cm2 for the forearm. The data on bone density are reported both as absolute measurements and as Z scores (8).
During the first 5 yr of the study, densitometric data were obtained using single- and dual-photon absorptiometry techniques (SP2 and DP3; Lunar Radiation Corp., Madison, WI), whereas data for virtually the entire past 20 yr were obtained using dual-energy x-ray absorptiometry (DXA). Twenty-six subjects who did not undergo parathyroidectomy switched densitometers (mean time from study entry was 3 yr; range 1–6 yr). Of the 11 patients followed up beyond 10 yr, eight switched densitometers (mean time of switch from study entry was 4 yr). When this transition occurred, the different machines were cross-calibrated by scanning 60 people on both machines within a few weeks. Subjects were selected to represent the range of bone densities that represented our cohort. Regression modeling was used to create the predictive equations that allowed interconversion of the data to the newer instrumentation. Data obtained on the cohorts followed for 20 yr with the DXA technology were essentially identical with the data that spanned the transition to DXA, giving assurance that an artifact of measurement was not introduced into the monitoring process.
Quality control included reading phantoms daily that were analyzed by Shewhart criteria for process stability (9, 10).
Statistical analysis
Independent t tests were used to compare baseline characteristics of the medical and surgical groups. All data were transformed to percent change from baseline before analysis and descriptive statistics examined for distributional characteristics. No measure exceeded a range of 3 orders of magnitude, and all mean and median percent changes from baseline were within 30% of each other on all measures. Whereas our 10-yr report used ANOVA for comparison among surgical and nonsurgical cohorts, and paired t tests to measure changes over time, a more precise tool was chosen for this report. Linear mixed model analysis for repeated measures was used to estimate within-subject changes in BMD and biochemical parameters. Identical models with a fixed effect of time (annual years of follow-up), random effects of subject, time and error, and an empirically determined block diagonal autoregressive [1] covariance structure were created for each bone measurement site and biochemical parameter (11). Estimated within-subject change in outcome between specific time points was taken from the differences between model estimated means for each time, after adjustment of the degrees of freedom, to test the statistical significance of the comparison (12). Adjustments for multiple comparisons were not used because the mixed model adjusted degrees of freedom for the calculation of the confidence interval for the difference. All models were run under restricted maximum likelihood.
Separate models were performed for medically managed subjects from the time of enrollment, with and without patients who were treated with antiresorptive medications and with and without patients who completed the 15 yr of the study and for surgical subjects starting with their last available presurgical evaluation. The proportion of subjects meeting surgical criteria at baseline that went on to have surgery was assessed with Fisher’s exact test. Data are presented as mean ± SEM.
|
Results |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
The cohort of 116 patients with PHPT included 66 postmenopausal women (56%), 25 premenopausal women (22%), and 25 men (22%). Although 17 patients (15%) had a history of kidney stones, none had osteitis fibrosa cystica or classical neuromuscular symptoms. Patients were followed up according to their treatment category (with or without surgery) and according to the presence or absence of PHPT symptoms (Fig. 1
and Table 1). Patients who underwent parathyroidectomy (n = 59) had higher serum calcium and PTH concentrations than those (n = 57) who did not undergo surgery (Table 1). Patients who underwent surgery also had significantly lower BMD Z-scores at the lumbar spine and femoral neck but not at the radius. In both groups, a typical pattern of skeletal involvement was seen with greater reductions at the distal one third radius (cortical site) than the lumbar spine (cancellous site). Urinary calcium excretion, alkaline phosphatase activity, and vitamin D concentrations were similar between the surgical and nonsurgical groups.
|
Fifty-seven subjects with PHPT were followed without intervention for up to 15 yr. The vast majority (n = 49; 86%) were asymptomatic. During this period of time, 11 patients died. Four of the 11 subjects who died had increased risk factors for cardiovascular disease at study entry, such as diabetes mellitus and hypertension. The causes of death included myocardial infarction (n = 2), complications of diabetes mellitus (n = 2), congestive heart failure (n = 1), stroke (n = 1), aortic aneurysm rupture (n = 1), gall bladder cancer (n = 1), and unknown causes (n = 3). The subjects who died did not have higher serum calcium levels at baseline (10.3 ± 0.1 mg/dl) than the remainder of the asymptomatic medical subjects but did have higher initial PTH levels (161 ± 25 vs. 107 ± 8 pg/ml, P < 0.05). Eight subjects were unable to comply with regular study visits and five were eventually lost to follow-up. None of the demographic variables available to predict duration of follow (age at study enrollment, sex, and race) achieved statistical significance. Nearly half of the nonintervention group (20 of 57) ultimately had successful surgery. Of these, one was symptomatic at baseline but had initially refused surgery, six met surgical criteria but deferred intervention initially, six had disease progression, and seven no longer wished to be followed up without intervention.
Biochemical changes
Of the 49 asymptomatic subjects, 29 (59%) did not meet surgical criteria at baseline, whereas 20 (41%) did meet criteria but refused intervention. There were no significant changes in PTH, total alkaline phosphatase activity, or urinary calcium excretion (Table 2) among the asymptomatic patients who did not have surgery, for as long as 15 yr of follow-up. However, in the subjects who were followed up conservatively for up to 15 yr, serum calcium concentrations rose to levels, beginning at yr 13, that were slightly but significantly higher than their individual baseline values. In the 11 patients who completed 10 yr of follow-up, neither the serum calcium nor the PTH changed significantly over yr 0, 5, and 10 (serum calcium: 10.5 ± 0.1, 10.7 ± 0.1, 10.8 ± 0.2 mg/dl; PTH: 109 ± 13, 102 ± 13, 110 ± 14 pg/ml). In the six patients who completed 15 yr of follow-up, the serum calcium changed significantly only at yr 15, compared with baseline (yr 0, 5, 10, and 15 were 10.6 ± 0.2, 10.7 ± 0.2, 11.1 ± 0.3, and 11.1 ± 0.2 mg/dl, respectively). PTH did not change significantly (yr 0, 5, 10, and 15 were 98 ± 12, 81 ± 12, 74 ± 14, and 89 ± 12 pg/ml, respectively). Longitudinal biochemical profiles were similar among those asymptomatic subjects who met or did not meet surgical criteria.
|
BMD did not change at any measurement site during the first 8 yr of follow-up (Fig. 2) in the 49 asymptomatic patients. The lumbar spine BMD was stable for the entire 15 yr of follow-up, in both the overall cohort and the 28 postmenopausal women. However, even before the 10-yr time point, the femoral neck, a site containing a substantial amount of cortical bone, and the distal one third radius, a site containing a predominance of cortical bone, showed significant reductions in BMD. Overall, 29 of the 49 asymptomatic patients (59%) had more than a 10% decline in BMD at one or more sites over the 15-yr period. Eleven of the 29 subjects were patients who had been previously found to have declined within 10 yr (3); eight other patients, with the availability of additional time points, were now found to also have declined within 10 yr, whereas 10 other patients declined after 10 yr of observation.
|
The six asymptomatic subjects who completed the 15-yr period of monitoring did not differ from the rest of the asymptomatic subjects (n = 43) in terms of age, baseline biochemistries (serum calcium, PTH, urinary calcium excretion, total alkaline phosphatase, 25-hydroxyvitamin D, or 1,25-dihydroxyvitamin D), or baseline Z-scores. However, even before year 10, BMD at the distal one third radius site showed accelerated losses. Linear mixed-model analyses for repeated measures were performed, comparing BMD changes in the first 10 yr of study among those followed up for less than 10 yr, and those followed up for a longer period. At the lumbar spine or femoral neck, the rates of change between these groups did not differ (P = 0.85, lumbar spine; P = 0.09, femoral neck). However, after yr 10, BMD declined significantly at the distal one third radius (P < 0.001). The slope of BMD loss therefore differed at the distal radius (there was an accelerated rate of loss after 10 yr) but not at the femoral neck (there was the same rate of loss continuing from before 10 yr).
Thirteen asymptomatic subjects who were followed up without surgery received antiresorptive therapy at some time point in the study (off protocol). The median duration of antiresorptive drug use was 4 yr (range 1–11 yr), whereas the median observation time in the remaining 36 asymptomatic subjects was 5 yr (range 1–15 yr). However, linear mixed-model analyses for repeated measures, comparing BMD changes in those who did and did not receive antiresorptive therapy, showed no differences in terms of the densitometric measurements over the 15 yr at any site.
Disease progression
Most of the 49 asymptomatic patients did not develop densitometric or other surgical guidelines. However, 18 patients (37%) did develop new surgical criteria at any time point over the 15 yr of observation. Meeting surgical criteria at study baseline did not predict who would have progressive disease. Of the 20 asymptomatic subjects who met surgical criteria at baseline (5 yr, n = 10; 10 yr, n = 5; 15 yr, n = 3), seven had disease progression (35%); likewise, of the 29 asymptomatic subjects who did not meet surgical criteria at baseline (5 yr, n = 15; 10 yr, n = 12; 15 yr, n = 3), 11 had disease progression (38%), a similar proportion. Changes in bone density over time were also similar among the asymptomatic subjects, regardless of whether they met surgical criteria at baseline.
All eight of the patients who had nephrolithiasis, and were therefore symptomatic, but refused surgery, showed progression by at least one criterion: six had recurrent kidney stones, one sustained a fracture, and one developed marked hypercalcemia. Despite these further complications, only one of these patients agreed to undergo parathyroidectomy.
Clinical course over 15 yr in subjects who underwent parathyroidectomy
All 59 patients had successful surgery as evidenced by normal serum calcium and PTH values postoperatively. Only nine of the 59 subjects (15%) who underwent surgery were symptomatic with kidney stones. At 15 yr, serum calcium was 9.8 ± 0.2 mg/dl (2.5 ± 0.1 mmol/liter), PTH 34 ± 21 pg/ml (3.6 ± 2 pmol/liter), and urinary calcium excretion 138 ± 36 mg/dl (3.5 ± 0.9 mmol/liter), all significantly lower (P < 0.01) in comparison with the individual subjects’ baseline values and all well within normal limits. Postoperative increases in BMD were sustained with BMD remaining significantly above baseline for the entire 15 yr of follow-up at all three skeletal sites (Fig. 3). Pre- and postmenopausal women and men, and those with and without symptoms, all had similar increases in BMD. Assessment of postoperative changes in BMD in those subjects who were initially followed up without intervention, but ultimately underwent parathyroidectomy (n = 20), demonstrated similar increases in BMD to those who had parathyroidectomy at the outset (at yr 1, 5, and 10, the lumbar spine increased 9, 6, and 12%; the femoral neck 1, 7, and 10%; the distal radius 4, 8, and 7%). There were no recurrent kidney stones in any of the nine symptomatic patients with kidney stones who underwent parathyroid surgery.
|
|
Discussion |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
In those who did not undergo surgery, lumbar spine BMD continued to be relatively stable during yr 10–15. Although progressive osteoarthritic changes at this site could have contributed to the stable lumbar spine findings, it is also consistent with our prior observations that the lumbar spine is a site that is relatively protected from bone loss in PHPT. Thus, despite histomorphometric evidence for increased bone turnover in PHPT (13), cancellous bone appears to be relatively well preserved, even after many years. Potential mechanisms for preservation of cancellous bone mass are suggested by quantitative histomorphometry of bone biopsies from patients with PHPT (13, 14, 15). Although activation frequency is increased, resorption pits are relatively shallow such that increased osteoblast activity is likely to ensure reasonably well-balanced remodeling units. As a result, the expected age-associated loss of trabecular plates is not regularly seen in PHPT (16).
In contrast to the stability at the lumbar spine, declines in BMD became apparent before yr 10 at the femoral neck and distal one third radius sites in the few remaining patients who were followed up without surgery. The detection of statistically significant declines in cortical bone density before yr 10 is a result of both the additional observation points in the time line as well as the more powerful statistical mixed model approach, in that variance attributable to within-subject covariance among repeated measures no longer inflates the error term (3).
This observation suggests that the catabolic effects of PTH at the femoral neck and distal radius, which contain more cortical bone than the lumbar spine, might begin to emerge with prolonged exposure to PTH excess. Cortical bone is more vulnerable than cancellous bone to the catabolic effects of PTH (17). The possibility that cortical bone density is eventually compromised in PHPT has been suggested by some (18, 19) but not all studies (20, 21). Why so much time is required before these changes manifest themselves may relate to the low turnover rate of cortical bone, requiring years of exposure to demonstrate the negative effect. Alternatively, the process by which excess PTH leads to resorption of cortical bone may be compensated initially but not sustained over time.
The declines in cortical bone density were observed regardless of whether the patients taking antiresorptive therapy were excluded. Including those on antiresorptive therapy might have potentially masked an even greater bone loss than what was observed. However, the important point is that use of antiresorptives did not alter the natural history of the eventual cortical reductions after 10 yr or the prolonged maintenance of bone density at the lumbar spine.
Concepts of bone loss in mild PHPT may need to be re-examined in light of these new observations. It is currently hypothesized (22, 23) that cortical BMD is lost early in the disease before hypercalcemia brings the disorder to clinical attention. Although some patients may have had PHPT for years before diagnosis, for most patients, as demonstrated in this study, BMD remains stable for several years after coming to medical attention. The new steady state may be sustained for an even longer period at the cancellous lumbar spine but may not be at sites in which cortical bone is more prominent. The loss of bone mass in the hip and distal one third radius suggests that cortical bone might not remain stable over time.
Over the 15-yr follow-up period, 37% of asymptomatic patients showed evidence for disease progression at some time point, a substantial increase over the 25% progression during the first 10 yr. Meeting surgical criteria at study entry did not predict who would have worsening disease. In asymptomatic patients who did not undergo parathyroidectomy, there were no reliable predictors of bone loss, the development of symptomatic complications, or other surgical criteria. This point, along with the eventual decline in cortical bone density, highlights the importance of regular monitoring in asymptomatic patients with PHPT who do not undergo parathyroid surgery.
Meeting surgical guidelines at the time of diagnosis did not predict a worse outcome in asymptomatic patients who were followed up without surgery, raising questions as to whether the guidelines for surgery in asymptomatic patients with PHPT are sufficient to support appropriate decision making (24). Since this study began, surgical criteria were revised after a National Institutes of Health-sponsored workshop on asymptomatic PHPT in 2002 (25). Surgical recommendations were expanded to include a serum calcium concentration greater than 1.0 mg/dl (0.25 mmol/liter) above normal or a marked reduction in BMD at any site (T score below –2.5). Using these new more stringent criteria would, in all likelihood, have uncovered evidence for even greater disease progression. In addition, a larger group of less severely affected individuals would have met surgical criteria at baseline.
Fifteen years after successful parathyroid surgery, patients sustained their postoperative increases in BMD both at cancellous and cortical sites despite the expectation of age-related bone loss. Most likely, the early increase in cancellous BMD results from mineralization of the expanded remodeling space that is characteristic of PHPT (26, 27). Although the increase in BMD at the more highly cortical distal one third radius was not as great as that at cancellous sites, these data suggest that cortical bone loss in patients with PHPT is at least partially reversible. This new observation contrasts with data from previous reports (19, 28) and is most likely explained by the longer period of observation.
A limitation of this study is the small number of asymptomatic participants who remained in the nonsurgical group at the conclusion of the 15 yr of observation, a byproduct of this type of very long-term natural history study. Nevertheless, the six asymptomatic subjects who completed the 15-yr period of monitoring did not differ from the rest of the asymptomatic subjects in terms of age, baseline biochemistries, or baseline BMD Z-scores. An additional limitation was the change in densitometers in some of the patients in the early years of the study. However, the contrast between the increase in BMD in the patients who underwent parathyroidectomy (many of whom were also studied before and after change in equipment) and the lack of change in patients who did not argues against any widespread systematic error.
A further limitation is that the study was not randomized or controlled. The patients who underwent surgery were different at the outset from those who were managed conservatively. Despite this point, the presence within each group (surgical and nonsurgical) of patients who did or did not meet criteria for surgery has allowed us to gain some insights that would not have been possible with a more rigorous experimental design. Moreover, it is noteworthy that data from short-term studies in which identically matched PHPT subjects were randomized to a surgical or nonsurgical group are virtually indistinguishable from the findings we currently present (29, 30). Like the randomized trial of Rao et al. (30), many of our medical subjects eventually elected to undergo parathyroidectomy. Most recently, randomized 1- (31) and 2-yr (32) trials of surgery have also corroborated our early densitometric observations. These studies, together with ours, seem to raise the question as to whether long-term follow-up without intervention is the best option in PHPT.
The clinical significance of densitometric changes in PHPT needs to be considered along with other indices of bone strength (33). For example, the decreases in cortical BMD in the group followed up medically might be due to increased cortical diameter, which itself might counteract the apparent BMD reduction and thus reduce fracture risk by virtue of size-related improvements in biomechanical indices (34). It is also unclear whether postoperative improvements in cortical BMD are associated with improved bone strength. Assessment of bone geometry, microarchitecture, and material properties of bone (33, 35, 36), as well as fractures per se, is necessary for a complete description of fracture risk in PHPT (37, 38).
Whether mortality is increased in mild PHPT, as it clearly is in more severe disease, is controversial (39, 40). In our study, the number of deaths in the surgical group is unknown because many subjects were lost to follow-up, but 11 of the nonsurgical subjects (19%) died over the 15 yr of follow-up. The observations that the majority of these deaths were related to cardiovascular causes, and that baseline PTH levels in these subjects were significantly higher, is consistent with the observation that more severe PHPT and/or higher sustained PTH levels might be associated with increased cardiovascular mortality. Nevertheless, it is important to reiterate that this study was not powered to allow us to comment on this issue. Only epidemiological studies can address the issue of mortality in PHPT and whether there is a difference between those who undergo parathyroid surgery and those who do not.
These results indicate that there is evidence of progressive disease in individuals with mild PHPT, particularly with regard to declines in cortical bone density. Although the study was limited by its observational design and the small number of patients who completed 15 yr of observation, these new data raise the question as to whether conservative, nonsurgical management of PHPT, among those deemed not to be surgical candidates, should be limited to a decade of follow-up. Decisions about further conservative management depend on continued densitometric stability. These results also emphasize the need for continued long-term surveillance of biochemical indices and bone densitometry in all patients with mild disease who do not undergo parathyroid surgery.
|
Footnotes |
|---|
Disclosure Statement: M.R.R., J.P.B., D.J.M., T.J., E.Sh., E.Si., J.U., and S.J.S. have nothing to declare.
First Published Online June 10, 2008
Abbreviations: BMD, Bone mineral density; DXA, dual-energy x-ray absorptiometry; PHPT, primary hyperparathyroidism.
Received May 31, 2007.
Accepted May 29, 2008.
|
References |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  J. Bollerslev, T. Rosen, C. L. Mollerup, J. Nordenstrom, M. Baranowski, C. Franco, Y. Pernow, G. A. Isaksen, K. Godang, T. Ueland, et al. Effect of Surgery on Cardiovascular Risk Factors in Mild Primary Hyperparathyroidism J. Clin. Endocrinol. Metab., July 1, 2009; 94(7): 2255 - 2261. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. J. Silverberg, E. M. Lewiecki, L. Mosekilde, M. Peacock, and M. R. Rubin Presentation of Asymptomatic Primary Hyperparathyroidism: Proceedings of the Third International Workshop J. Clin. Endocrinol. Metab., February 1, 2009; 94(2): 351 - 365. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Khan, A. Grey, and D. Shoback Medical Management of Asymptomatic Primary Hyperparathyroidism: Proceedings of the Third International Workshop J. Clin. Endocrinol. Metab., February 1, 2009; 94(2): 373 - 381. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. El-Hajj Fuleihan Hyperparathyroidism: Time to Reconsider Current Clinical Decision Paradigms? J. Clin. Endocrinol. Metab., September 1, 2008; 93(9): 3302 - 3304. [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
