We retrospectively reviewed data from patients who underwent US-guided FNAC for thyroid nodules from January 2005 to December 2010 at a tertiary referral hospital. This search identified 289 individuals with a cytological diagnosis of lymphocytic thyroiditis. Exclusion criteria were (1) current thyroid hormone replacement therapy (n = 62), (2) absence of an initial TFT (n = 37), and (3) no regular visit with follow-up of TFT or thyroid US (n = 81). Thus, we initially included 109 patients in the present study.

During follow-up period, seven patients among 109 were diagnosed with papillary thyroid cancer and were excluded from the study due to the difficulty in differentiation between true Hashimoto’s thyroiditis and peritumoral lymphocyte infiltration around papillary thyroid carcinoma.

The follow-up interval of thyroid US was 12 months. However, the follow-up interval of TFT varied from 3 to 6 months. If there was a significant change in the TFT, follow-up was performed within 3 months at the discretion of the clinician. Thus, the follow-up intervals were 6 to 12 month, and the patients underwent TFT and thyroid US at least three times.

The TFT was performed. Also, thyroid peroxidase antibody (TPOAb) and thyroglobulin antibody (TgAb) were measured by radioimmunoassay. Our institution’s uses the following reference ranges TSH 0.17 to 4.05 mIU/mL, free thyroxine (T4) 0.85 to 1.86 ng/dL, triiodothyronine (T3) 0.78 to 1.82 ng/mL, TPOAb 0 to 80 IU/mL, and TgAb 0 to 70 IU/mL. Patients were assigned to group E (euthyroidism) if they had normal values for serum TSH and free T4 [6]. Subclinical hypothyroidism was defined as a serum TSH concentration that was above the upper limit of the reference range, but with normal free T4 levels [5]. Overt hypothyroidism was generally defined as serum TSH concentrations above the reference range and free T4 levels below the reference range. However, some individuals with clinical symptoms of hypothyroidism, high TSH levels (> 10 mIU/L), and low but normal free T4 levels were classified as having overt hypothyroidism [7]. High initial TSH levels and TPOAb positivity are risk factors for overt hypothyroidism. Therefore, in this study, the clinical hypothyroidism group (H group) consisted of patients with both overt and subclinical hypothyroidism with TSH levels ≥ 10 mIU/L. Clinical hypothyroidism was confirmed on repeat TFT at minimum 2 months after that initial assessment. After repeat TFT, if a patient presented high TSH and free T4 levels within the reference range, the clinician made decision according to patients’ clinical presentation.

The FNAC was performed using a 22 G Franseen needle under US guidance and the specimens were smeared onto glass slides and placed in 95% alcohol for Papanicolaou staining. The classical cytological features of lymphocytic thyroiditis were accompanied by some plasma cells and macrophages. The lymphocytes have close contact with the thyrocytes and mediate their destruction. The cytological findings of lymphocytic thyroiditis are a mixed group of follicular and Hurthle cells with low-to-moderate colloid and lymphocytic infiltration in the epithelium [8,9]. A pathologist assessed the Hurthle cell changes, as well as the increased numbers of mature and transformed lymphocytes impinging on follicular cells [8]. Lymphocytic thyroiditis was diagnosed when the FNAC revealed a mixture of follicular cells and Hurthle cells in a background of mature lymphocytes (Supplementary Fig. 1). Two pathologists interpreted the cytological findings.

The US findings for thyroid nodules that were evaluated using FNAC were also examined to determine their echogenicity, margin, shape, homogeneity, calcification, and size, based on previously reported methods [10]. Size was defined based on the maximum nodule diameter. Nodules with confirmed lymphocytic thyroiditis based on FNAC were categorized according to the grading system that was suggested by Sostre and Reyes [11], with grade 1 assigned to an enlarged gland with a normoechoic pattern, grade 2 nodules exhibiting multiple hypoechoic foci or patches scattered throughout the normoechoic gland (i.e., focal rather than diffuse involvement), grade 3 defined as an enlarged gland with diffuse but mild hypoechogenicity, and grade 4 nodules exhibiting marked hypoechogenicity (Fig. 1) [12,13].

All statistical analyses were performed using SPSS software version 14 (SPSS Inc., Chicago, IL, USA). The distributions of all continuous variables were examined using the Shapiro-Wilk test. The non-normally distributed variables were follow-up duration and TFT results (levels of TSH, free T4, and T3), which were reported as median and interquartile range. Normally distributed continuous variables were reported as mean ± standard deviation. Number and percentage were used for all categorical variables (sex, background thyroiditis, TPOAb positivity, and TgAb positivity). The characteristics of groups E and H were compared using the independent t test (continuous variables) or the chi-square and Fisher exact test (categorical variables). The outcomes for groups E and H were evaluated without adjusting for the time until the detection of hypothyroidism, and times to events were calculated based on the start of levothyroxine therapy. Cox proportional hazards regression analysis was adjusted for potential confounders (p < 0.05 in the univariate analyses) and used to evaluate the associations for the TFT and thyroid US findings. The results were reported as the relative risk (RR), 95% confidence interval (CI), and corresponding p value. The cut-off values for nodule size (17 mm) and TSH levels (2.02 mIU/L) were determined using receiver operating characteristic curve analysis. Differences were considered statistically significant at a two-tailed p value of < 0.05.

This study complied with the ethical standards of the Helsinki Declaration and was approved by the Catholic University of Korea, Catholic Medical Center, Seoul St, Mary's Hospital Institutional Review Board (IRB approval No. KCT18RESI0831).

This study was exempted from the consent of the IRB, because there is no reason to estimate the refusal of consent of the subject and the risk to the subject is extremely low even if the consent is waived.

The patients’ baseline clinical characteristics and TFT findings are shown in Table 1. The median follow-up interval was 10.2 months (range, 3 to 18). The median number of nodules observed on thyroid US was 1.0 (range, 1.0 to 2.0) and the mean nodule size as 17.9 mm (range, 9.1 to 26.7). Thyroid US findings for nodules that were evaluated by FNAC were also analyzed again according to echogenicity, calcification, margin, and shape. Hypoechogenicity was detected for 126 of the 178 nodules (70.8%), and 30 of the nodules (16.9%) were accompanied by calcification. Lobulated and poorly defined margins were observed for 71 nodules (39.9%). The orientation was non-parallel (taller than wide) for 58 nodules (32.6%). According to the Korean Thyroid Imaging Reporting and Data System (K-TIRADS) [14], 35 of the 178 nodules (19.7%) were considered highly suspicious (K-TIRADS 5), 82 nodules (46.1%) had an intermediate suspicion (K-TIRADS 4), 47 nodules (26.4%) had a low suspicion (K-TIRADS 3), and 14 nodules (7.9%) were considered benign (K-TIRADS 2). No significant differences were observed in the nodule characteristics when we compared groups E and H (data not shown).

The US findings of background thyroiditis were classified as grade 1 to 2 for 82 of the 109 patients (75.3%) and as grade 3 to 4 for 27 patients (24.8%). Fifty-one patients (46.8%) were positive for TPOAb and 43 patients (39.4%) were positive for TgAb. Table 2 shows the clinical characteristics and TFT findings for groups E and H. No significant intergroup differences were observed for age, sex, follow-up duration, number of nodules, and initial TFT findings (levels of TSH, free T4, and T3). However, group H had significantly larger nodules, compared to group E (27.5 mm vs. 15.1 mm, p < 0.001). Grade 3 to 4 background thyroiditis was observed for 10 of the 14 patients in group H (71.4%) and in 10 of the 88 patients in group E (11.4%). Group H had a higher proportion of TPOAb positivity, compared to group E (78.5% vs. 43.2%, p < 0.001). The difference in TgAb positivity was not statistically significant (57.1% vs. 25%, p = 0.068).

We evaluated the outcomes of 109 patients with lymphocytic thyroiditis based on FNAC. Eighty-eight patients did not exhibit any symptoms or TFT abnormalities until the last follow-up visit. However, 14 of the 109 patients (12.8%) progressed to clinical hypothyroidism during a mean follow-up of 58.7 months (range, 23 to 94). The detailed characteristics of these cases are shown in Table 3. Their median time to the onset of hypothyroidism was 16 months (range, 3 to 88), and nine of the 14 patients (64.3%) developed clinical hypothyroidism within 1 year. Eleven patients exhibited progression to overt hypothyroidism and received levothyroxine therapy (25 to 75 μg). The remaining three patients underwent TFT every 3 to 6 months, but were found to remain in a subclinical hypothyroid state

We evaluated potential clinical parameters at the time of FNA to find out predictable risk factors for future clinical hypothyroidism events (Table 4). In univariate analysis, the variables associated with clinical hypothyroidism were background thyroiditis on US (RR, 8.10; p < 0.001; 95% CI, 2.503 to 26.232), TPOAb positivity (RR, 7.34; p = 0.009; 95% CI, 1.640 to 32.844), initial TSH (RR, 1.55; p < 0.001; 95% CI, 1.531 to 1.580), nodule size (RR, 1.47; p = 0.004; 95% CI, 1.415 to 1.479), and TgAb positivity (RR, 2.95; p = 0.054; 95% CI, 1.640 to 32.844). In multivariate analysis, background thyroiditis on US (RR, 9.78; p = 0.004; 95% CI, 2.077 to 46.124), TPOAb positivity (RR, 9.90; p = 0.003; 95% CI, 1.012 to 96.887), nodule size (RR, 1.24; p < 0.001; 95% CI, 1.096 to 1.394), and initial TSH (RR, 1.47; p = 0.009; 95% CI, 1.011 to 1.484) were found to be independent risk factors for predicting clinical hypothyroidism.