Parathyroid Hormone Related Peptide (PTHrP) Revisited: A Comparison of Levels in Malignant and Non-malignant Conditions.

Abstract

Elevated parathyroid hormone-related peptide (PTHrP) levels are often presumed to indicate malignancy-associated hypercalcemia, yet PTHrP may be elevated in nonmalignant conditions. We compared PTHrP levels in patients with known malignancies with patients with elevations of PTHrP who had defined nonmalignant diagnoses to determine which levels of PTHrP were more indicative of malignancy. A retrospective review of electronic medical records in a large midwestern hospital system was conducted using search terms hypercalcemia with elevated PTHrP (> 2.3 pmol/L (men) and > 3.4 pmol/L (women). Binary logistic regression analyses assessed the positive predictive value (PPV, probability) of cancer. Among 236 patients with hypercalcemia and elevated PTHrP, 58 had known cancer. The others had defined nonmalignant diagnoses including primary hyperparathyroidism (n = 87), absorptive hypercalcemia (n = 28), and "other" causes (n = 63). PTHrP was highest in the cancer group (PTHrP = 36.9, 95% CI 19.8, 53.9 pmol/L, p < 0.001), however, overlap occurred. Binary logistic regression analyses showed a positive association of PTHrP levels (p < 0.001), while PTH levels were negatively associated (p < 0.001) with the presence of cancer. When analyzed together, PTH on PTHrP had independent effects (p < 0.001). At PTH ≥ 30 pg/mL, the PPV of cancer was < 5% when PTHrP < 7.0 pmol/L and < 10% at a PTHrP < 9.0 pmol/L. In the evaluation of hypercalcemia with elevated PTHrP, simultaneous levels of PTH measurements improve estimation of cancer probability. Mild PTHrP elevations with PTH > 30 pg/mL are more consistent with nonmalignant diagnoses, although higher PTHrP levels may still warrant evaluation for occult malignancy.