A New, Vitamin D-Based, Multidimensional Nomogram for the Diagnosis of Primary Hyperparathyroidism

Adrian Harvey, MD; MengJun Hu, MS; Manjula Gupta, PhD; Robert Butler, MS; Jamie Mitchell, MD; Eren Berber, MD; Allan Siperstein, MD; Mira Milas, MD, FACS


Endocr Pract. 2012;18(2):124-131. 

In This Article


The results of this study support the notion that a diagnosis of 1°HPT can be strengthened by refining the interpretation of key biochemical variables relevant to this disease. We were able to identify clinical factors—total serum calcium, 25(OH)D, and age—that significantly affected PTH values. We developed a useful mathematical model for predicting normal PTH levels that is based on these multiple factors, not simply calcium values as is currently practiced. A unique aspect of our model is that the PTH nomogram generates a patient-specific upper limit of normal PTH, rather than providing generic and static PTH reference ranges.

The major motivation for developing such a model was the realization that the 2-dimensional, sigmoid relationship between total serum calcium and PTH levels, initially described in the 1970s, was insufficient to characterize fully all clinical scenarios of 1°HPT.[21–23] This particularly applies to those patients with borderline or atypical laboratory values, inasmuch as the diagnosis of 1°HPT in patients with elevated serum calcium and intact PTH concentrations and urinary calcium excretion is straightforward. With use of existing 2-dimensional models, which usually define the normal PTH reference range as <60 pg/mL, patients with atypical 1°HPT are impossible to distinguish from those with secondary hyperparathyroidism and sometimes even those with hypercalcemia of malignancy. Our expectation for this study was that multidimensional mathematical modeling of relevant biochemical factors may help support the 1°HPT diagnosis in patients with atypical or borderline laboratory values. In our patients with confirmed and treated 1°HPT, the PTH nomogram would, in fact, have correctly supported this diagnosis before surgical intervention in 84% of those with normocalcemic 1°HPT and 54% with hypercalcemia and inappropriately normal PTH values.

It is not surprising that about a third of our surgical cohort were in an atypical 1°HPT category. With the introduction of routine serum calcium measurements, the epidemiologic features of 1°HPT have shifted; the result is a greater proportion of cases of both asymptomatic hypercalcemia and mild serum calcium elevations being detected incidentally on routine blood studies.[1,2,24] Another demographic trend has been observed: unlike the 90% prevalence of single adenomas seen in previous decades, modern series of patients undergoing parathyroid surgical treatment identify up to 20% to 30% with multigland hyperplasia.[24–26]

Many patients with what is potentially "mild" or "early" 1°HPT have borderline laboratory results that pose diagnostic challenges. In a recent series of 60 consecutive patients with 1°HPT, Glendenning et al[27] found that 22% had normal serum calcium levels, with 8% having both serum calcium and PTH values in the reference range. Thus, there is a growing need and context for the application of our PTH nomogram or similar diagnostic aids. At our center, a substantial number of these patients were eventually offered surgical treatment and underwent neck exploration, with removal of pathologically confirmed abnormal parathyroid glands. Typically, these patients underwent follow-up for an extended period and had multiple laboratory measurements. Use of adjuncts such as ionized calcium and 25(OH)D or the presence of adverse consequences of the disease invariably contributed to the final decision. Thus, it is likely that our endocrinologists and endocrine surgeons used an informal, if not partially subconscious, version of this algorithm in deciding the course of management in these patients. As such, our study helps make explicit those factors that may enter into appropriate decision making in these challenging scenarios.

The traditional treatment regimen at our institution has been to correct vitamin D deficiency after a parathyroid surgical procedure. Because no patient underwent a negative surgical exploration, however, we do not believe that vitamin D treatment preoperatively would have influenced the ultimate treatment algorithm for these patients. Furthermore, it is important to emphasize again that the utility of the nomogram is to provide the expected PTH value for the given serum calcium and 25(OH)D measurements at a specific time. Treated secondary hyperparathyroidism should be reflected in improved calculated and measured PTH values.

One additional potential application of the PTH nomogram may be the assessment of eucalcemic elevation of PTH levels after parathyroid surgical treatment, a phenomenon noted in up to 40% of patients in some reported series.[28–32] In some patients, this finding represents an adaptive response to underlying vitamin D deficiency or renal insufficiency. Indeed, Beyer et al[33] found the incidence of postoperative eucalcemic PTH elevations to be significantly lower in patients who received oral vitamin D supplementation in comparison with those who did not (14% versus 39%; P<.04). A proportion of these patients, however, will have persistent or recurrent 1°HPT. If one assumes that parathyroid function and calcium metabolism return to a baseline state after a curative surgical procedure, then the generation of a patient-specific upper limit of normal for PTH may help to sort out those patients who require further investigation.

Our study has some limitations that warrant discussion. Even with the use of multiple variables in the predictive model, the explained variance (R2) is 0.18. This finding indicates that baseline "statistical noise" and additional factors beyond those we identified are likely influencing PTH levels. Despite this limitation, all factors used in our model were significantly correlated with PTH, and plotting the upper 95% confidence interval highlighted the difference between patients with 1°HPT and healthy control subjects. Other investigators such as Aloia et al[20] have identified body mass index (R 2 = 0.09), age (R 2 = 0.01), and 25(OH)D (R 2 = 0.008) as significant predictors of PTH in a model with overall R 2 of 0.11, and others have suggested that sex, race, and serum creatinine are potential factors.[15,16,18–20,34,35] The potential exists to modify or improve our PTH nomogram through a larger scale study, with examination of a broader range of predictive variables. Although promising also in its retrospective evaluation of our surgical series of patients with 1°HPT, this predictive equation requires validation on a prospective basis. We are currently undertaking such a study. We are additionally examining whether the use of the PTH nomogram leads to fewer repeated laboratory tests for patients with atypical 1°HPT or shorter time frames between screening, diagnosis, and surgical intervention. These variables were not possible to examine in our current study.

A final factor that bears mentioning is measurement of ionized calcium. In this study, we used total serum calcium for the diagnosis of 1°HPT, with the rationale as stated in the "Methods" section. Although high ionized calcium levels have been reported in patients with suspected 1°HPT and normal total serum calcium.[27,36] even then 18% of patients had laboratory values, either ionized calcium or PTH, within the normal range.[27] Thus, in future modifications of our PTH nomogram, we may wish to examine ionized calcium in an attempt to improve diagnostic accuracy further.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.