Differences in radioimmunoassay results for thyroglobulin that affect management of patients with thyroid cancer

We compared two commercial assays for measurement of serum thyroglobulin [Nuclear Medical Systems (NMS) and "CIS" (Damon Diagnostics)] with each other and with one developed at Stanford (J Clin Endocrinol Metab 49:557-564, 1979). The NMS assay is a competitive-binding RIA, the CIS and Stanford assays are two-site immunoradiometric assays. The kit standards varied in thyroglobulin concentration. The NMS standards differed in immunoreactivity from thyroglobulin in clinical specimens and from the other standards. Also, nonparallelism between standards and patients' sera in the NMS assay suggested a less-specific antiserum. Results by the CIS and Stanford assays correlated well (n = 120, r = 0.964), those by the NMS assay less strongly (n = 101, r = 0.855 vs CIS, r = 0.888 vs Stanford). Clinical evaluation in 50 patients treated for differentiated thyroid carcinoma (10 with metastases and 40 currently disease-free) indicated good agreement for positive results by the three assays. The CIS and the Stanford assay both gave high results (greater than or equal to 25 micrograms/L) in all 10 cases with metastases; the NMS RIA identified eight of these patients (thyroglobulin greater than or equal to 30 micrograms/L), but excluded two with anti-thyroglobulin autoantibodies. In subjects without disease, however, the percentage of undetectable thyroglobulin (negative result), as opposed to low measurable thyroglobulin (inconclusive result) varied considerably: 85% by CIS, 30% by NMS, and 75% by the Stanford assay.