Changes in both size and cytological features of thyroid nodule after levothyroxine treatment

We prospectively evaluated the effects of 12 months thyrotropin suppressive levo-thyroxine (L-T4) therapy in terms of changes in both thyroid nodule size and cytological features and considered whether thyroid nodule size changes actually resulted in (or were the result of) cytological changes.

We studied 142 consecutive patients with benign 'cold' solitary thyroid nodules. For the purposes of the study, we divided our patients into three groups according to their initial cytological nodule classification as follows: group 1, including 88 colloid nodules (CN); group 2, including 30 hyperplastic nodules (HN); and group 3, including 24 adenomatous nodules (AN).

The effects of TSH suppressive L-T4 treatment on both nodule volume and cytological characteristics were evaluated by ultrasonography and fine-needle aspiration (FNA) before and 12 months after the onset of therapy.

Twelve months of TSH suppressive L-T4 treatment were effective in shrinking about one-third to one-quarter (31.8% CN, 26.7% HN and 25% AN) of thyroid nodules, irrespective of their initial cytological characteristics. Similarly, there was no difference in the prevalence of 'non-responders' (increasing nodules) to L-T4, which accounted for about one-fifth or less (20.5% CN, 13.3% HN and 20.8% AN) of all nodules. We found also that cytological features changed in 33.8% nodules after 12 months of L-T4 treatment. These changes were most commonly observed in HN and AN and consisted largely ( approximately 80%) of cytological characteristics resembling colloid features, not only in shrinking but also in stable nodules, thus indicating that cytology changes might be the very first indicator of nodule sensitivity to L-T4 therapy even in the continued absence of nodule shrinkage. When nodules were divided into three subgroups according to variations in size (shrinking, stable and increasing nodules), we observed that the distribution of the three cytological types showed a trend towards colloid lesions in shrinking nodules (chi2 3.8, P < 0.005) and towards an augmentation of hypercellular, adenomatous and suspicious characteristics in increasing nodules (chi2 3.6, P < 0.005).

The frequency of shrinking nodules was not different between colloid, hyperplastic and adenomatous nodules. Repeat FNA should be advisable for thyroid nodules which increase in volume despite L-T4 therapy and might also provide useful information about nodule sensitivity to L-T4 treatment for both HN and AN, even where nodule size remains stable.