Endocrine Abstracts

Aim: A key diagnostic step to determine the cause of hyperthyroidism is the measurement of circulating thyroid-stimulating immunoglobulin (TSI). High levels are associated with Graves’ disease (GD). A thyroid scintigraphy (TRS) is recommended when TSI is absent or its measurement unavailable, as TRS is capable of differentiating between GD (diffuse uptake) and toxic goitre (TG) (patchy uptake). Nevertheless, and while rarely necessary, histopathology remain the best method for obtaining an accurate aetiology. Our purpose was to compare TSI vs. TRS sensitivity and specificity in the differential diagnosis of thyrotoxicosis.

Methods: We retrospectively studied 235 outpatients with hyperthyroidism (overt or subclinical) admitted in our Centre from 2006 to 2016 from whom TSI and TRS were performed at the time of diagnosis. Subsequently, we selected a sample of thyroidectomised patients in whom a histological specimen was available. A pathologist reviewed the samples. SPSS 23.0 was used for statistical analysis. Pearson’s correlation was used to quantify the relationship between the diagnostic tests.

Results: A total of 45 patients met the inclusion criteria. We divided the sample according to the histopathology diagnosis of GD: Group A: 14 patients (31.1%) consistent with GD; Group B: 31 patients (68.9%) without features of GD. In Group A, one patient (7.1%) had negative TSI and two patients (14.3%) did not have a TRS with a high diffuse uptake. In Group B, 10 patients (32.2%) had positive TSI and 6 patients (19.3%) had a TRS with a high diffuse uptake. In comparison with histopathology, the TRS yield offered better diagnostic precision than TSI in both accuracy (82.2% vs. 75.5%) and specificity (80.6% vs. 67.7%) although inferior sensitivity (85.7% vs. 92.8%). The Positive Predictive Value (PPV) for TSI was 56.5% and the Negative Predictive Value (NPV) was 95.45%, whereas the PPV for TRS was 66.7% and the NPV was 83.3%. Pearson’s correlation between TSI and histology was 0.561 (P<0,001) versus 0.556 (P<0.001) between TRS and histology.

Conclusions: TRS has higher accuracy and specificity when compared to TSI in the differential diagnosis of thyrotoxicosis. These results suggest that the differential diagnosis of thyrotoxicosis cannot rely initially or solely on TSI, as this approach may result in misdiagnosis. Our results lend support to the value of both TRS and TSI as the first step in differential diagnosis, owing to the fact that TRS may detect 7.1% of mild GD with negative TSI and TSI alone may leave 32.3% of patients misdiagnosed.