The Diagnosis and Management of Subclinical Hypothyroidism Is Assay-Dependent

Implications for Clinical Practice

Tejas Kalaria; Anna Sanders; Jonathan Fenn; Helen L. Ashby; Pervaz Mohammed; Harit N. Buch; Clare Ford; Rousseau Gama


Clin Endocrinol. 2021;94(6):1012-1016. 

In This Article

Patients and Methods

Two neighbouring hospital laboratories in our regional pathology network participated in an institution approved and registered cross-sectional diagnostic test study comparing the potential impact of Roche and Abbott TSH and fT4 assays on the diagnosis of SCH and its management. TSH and fT4 were measured on a Roche cobas e801 (Roche Diagnostics, GmbH) in one laboratory and on an Abbott Architect i2000 SR (Abbott Laboratories) in the other laboratory.

For the purposes of this study, we identified samples from patients in primary care with biochemical SCH who had a raised TSH <10 mIU/L and a normal fT4 on the parent assays as levothyroxine replacement would not be indicated based solely on these biochemical results. Consecutive serum samples with SCH were exchanged until at least 40 samples from each site were identified. Samples from patients with pituitary disease, pregnant women and children (<18 years) were excluded as were from patients on levothyroxine and patients taking anti-thyroid drugs. Following identification, samples were stored at 2 to 8°C until analysed at the other site within 24 hours. TSH and fT4 were measured in duplicate on each sample at both the sites.

Inter-assay coefficient of variation (CV) for Abbott TSH was 6.8% at 3.4 mIU/L and 6.6% at 17.8 mIU/L and for the Roche TSH was 2.0% at 3.4 mIU/L and 2.2% at 26.7 mIU/L. Inter-assay CV for Abbott fT4 was 4.7% at 9.5 pmol/L and 4.6% at 20.1 pmol/L and for the Roche fT4 was 3.5% at 11.4 pmol/L and 3.1% at 24.4 pmol/L. Both assays had satisfactory internal quality control and external quality assurance performance during the study period.

Manufacturer-provided reference ranges for the Abbott TSH and fT4 assays were 0.35 to 4.94 mIU/L and 9.0 to 19.0 pmol/L, respectively. Reference ranges for Roche TSH and fT4 assays were 0.27 to 4.2 mIU/L and 12.0 to 22.0 pmol/L, respectively. Both laboratories use manufacturer-provided assay-specific reference ranges, and these were applied for the interpretation of results.

Data were tabulated in Excel (Microsoft corporation), and statistical analysis was performed using IBM SPSS Statistics for Windows, version 25 (IBM Corp). Since the data were non-parametric (Shapiro-Wilk test), Spearman rank correlation was used to measure the degree of association between results on both the analytical platforms. Wilcoxon signed-rank test was used to assess the significance of difference between paired Abbott and Roche data. The threshold for statistical significance was 5%. Data are expressed as medians with inter-quartile ranges (IQR).