Expanding the net: The re-evaluation of the multidimensional nomogram calculating the upper limit of normal PTH (maxPTH) in the setting of secondary hyperparathyroidism and the development of the MultIdimensional Predictive hyperparaTHyroid model (Mi-PTH)

Background The multidimensional nomogram calculating the upper limit of normal PTH (maxPTH) model identifies a personalized upper limit of normal parathyroid hormone (PTH) and successfully predicts classical primary hyperparathyroidism (PHP). We aimed to assess whether maxPTH can distinguish normocalcemic PHP (NCPHP) from secondary hyperparathyroidism (SHP), including subjects who underwent bariatric surgery (BrS). Methods A total of 172 subjects with 359 complete datasets of serum calcium (Ca), 25-OH vitamin D, and intact PTH from Oregon were analyzed: 123 subjects (212 datasets) with PHP and 47 (143) with SHP, including 28 (100) with previous BrS. An improved prediction model, MultIdimensional evaluation for Primary hyperparaTHyroidism (Mi-PTH), was created with the same variables as maxPTH by the use of a combined cohort (995 subjects) including participants from previous studies. Results In the Oregon cohort, maxPTH's sensitivity was 100% for classical PHP and 89% for NCPHP, but only 50% for normohormonal PHP (NHPHP) and 40% specific for SHP. In comparison, although sensitivity for NCPHP was similar (89%), Mi-PTH vastly improved SHP specificity (85%). In the combined cohort, Mi-PTH had better sensitivity of 98.5% (vs 95%) and specificity 97% (vs 85%). Conclusion MaxPTH was sensitive in detecting PHP; however, there was low specificity for SHP, especially in patients who underwent BrS. The creation of Mi-PTH provided improved performance measures but requires further prospective evaluation.