Suppressive therapy, in which thyroid hormone is administered to reduce the concentration of thyroid-stimulating hormone (TSH) to a level below normal (as opposed to replacement levothyroxine therapy, which normalizes the TSH level), had been used for years in the nonsurgical management of nodular thyroid disease [2-4]. However, this treatment is no longer popular: Recent studies failed to show efficacy [5, 6]; other studies showed an increased risk for osteoporosis as a consequence of thyrotoxicosis [7-9]; and fine-needle aspiration biopsy can better ascertain malignancy than can a therapeutic trial [10].

Although many patients do not need suppressive therapy, there are four main concerns about how the lack of suppressive therapy for cytologically benign disease may affect thyroid lesions or the patient's prognosis: First, in the past the clinician was alerted that a nodule might be malignant if it grew during treatment with levothyroxine. Enlargement of an untreated nodule is less informative. Second, because sampling errors from fine-needle aspiration biopsy may occur and negative or inconclusive cytologic findings do not exclude malignancy, it is uncomfortably open-ended to not treat patients who have "negative" cytologic findings and some clinical concern about malignancy but not enough data to warrant surgery. Third, previously, the emergence of a nodule in a goiter during therapy with levothyroxine attracted attention to the possibility of neoplasm. Obtaining a biopsy of all "dominant" nodules in a goiter has been suggested, but a cancer could be in the nodule next to the one that was sampled. This Herculean task might be made more difficult without levothyroxine. Fourth, it is unknown which, if any, nonsuppressed nodules or lobes remaining after partial thyroidectomy for benign nodular disease will grow and cause obstruction of the thoracic inlet. I recall that 30 years ago, before suppressive therapy was common, there were relatively more patients than we see now with large, obstructive nodules and distorted, hypertrophied contralateral postoperative lobes that posed diagnostic and management problems.

New reports suggest a tendency [11] or show [12] that levothyroxine can reduce the volume of some nodules. In this issue, La Rosa and colleagues [13] describe patients who had solitary cold thyroid nodules that were benign on fine-needle aspiration biopsy. As ascertained by an ultrasonographer who was blinded to treatment, nodule size decreased in 9 of 23 patients who received levothyroxine for 1 year without producing thyrotoxicosis, in 5 of 25 patients who received low-dose intermittent potassium iodine, and in none of 22 patients who received no treatment. During the 1-year study, nodules resumed growth when patients no longer received therapy, and untreated control nodules continued to grow. Unfortunately, the trial was not completely double-blinded, and the numbers, although significant, are small.

Opposing, strong opinions held by well-informed people about medical matters are usually accompanied by a lack of critical data, few participants in otherwise well-designed studies, or a lack of appropriate controls. Assuming that some euthyroid nodules get smaller and that further growth of some nodules is arrested with suppressive therapy and perhaps after treatment with iodide, we need to explore several questions.

1) Are all palpable nodules similar? No. Only some palpable nodules are truly solitary masses in an otherwise normal gland. More than 90% of these are benign and belong to a pathologically heterogeneous group of disorders that includes nodular thyroid, colloid nodules, adenomas, thyroiditis, and simple cysts. Even the thyroid follicular cells are heterogeneous [14]. In many cases, a clinically palpable nodule is actually a dominant nodule in a sea of smaller nonpalpable nodules, which is called a nodular goiter. Most dominant nodules in a goiter are not separate lesions but rather are benign and part of the goiter; only a few are malignant. The likelihood of detecting a response to levothyroxine or iodide in any small study group depends on the chance mix of pathologic findings in the series and other variables.

2) Is TSH the only thyroid growth factor? No. Other growth factors include immunoglobulins, systemic and paracrine growth factors [15], and mutations in the TSH receptor [16].

3) How do these factors affect a therapeutic trial? Nodules that have grown in response to excessive stimulation by TSH or iodine deficiency usually get smaller when TSH is suppressed or when the mineral is replenished. In contrast, autonomous nodules will not respond to suppression of TSH. Furthermore, the size of tissue that has enlarged in response to an abnormal immunologic or other growth factor will not be reduced. Finally, normal thyroid tissue is partially free of TSH control. One study showed low-level secretion of thyroid hormone that does not depend on TSH [17]; growth could also be independent of TSH. A therapeutic trial with thyroid hormone is the only way to determine whether suppression of TSH will arrest thyroid growth.

4) Does a decrease in nodule size when TSH is suppressed exclude malignancy? No. Atrophy of surrounding normal thyroid gland or reabsorption of fluid may deceive palpation, and although a few thyroid malignancies are said to respond, most do not.

5) How is reduction in size defined? We need to agree on criteria. Reduction in the size of the solid portion by 50% has been suggested. Documentation of change in size based on objective measurements such as sonography is essential [11, 18].

6) How safe is levothyroxine? When administered in quantities sufficient to produce thyrotoxicosis, it increases risk for adverse cardiovascular effects and osteoporosis but not for bone fracture [9]. It seems prudent to avoid iatrogenic thyrotoxicosis for benign disorders.

7) How is suppressive therapy defined? A reliable third-generation assay is essential for measuring the necessary low levels of TSH. Is reducing TSH into the low-to-normal range adequate for benign disease, or must TSH be below normal? The answer is unclear. However, there seems to be no advantage to suppressing TSH below 0.1 mU/L [19] or 0.4 mU/L [20].

8) Is iodide therapy effective? The efficacy of iodide in arresting the growth of thyroid nodules when the patient's diet contains enough or too much iodide needs to be investigated further.

9) How safe is iodide? It is unsafe in pregnancy. Hypothyroidism may result when the Graves-Hashimoto diathesis is present. Thyrotoxicosis may develop with an autonomous nodule or toxic nodular goiter and poses a risk for elderly patients or those with heart disease. Supplemental iodide is essential when dietary iodine is deficient, but it may temporarily precipitate thyrotoxicosis.

We must refine the answers to these questions and enhance our efforts to ascertain biochemical, cytologic, or receptor-related criteria that identify the types of thyroid tissue that are likely to enlarge if untreated but that will respond to suppression. Morita and colleagues [12] showed one such marker. Thyroglobulin levels decreased in 18 of 49 patients with cold benign nodules that responded to levothyroxine but not in the other patients whose TSH was equally suppressed.

Fortunately, most nodules are benign; even the malignant ones grow slowly and rarely cause death. But some nodules do cause a problem, and a few result in death. Whether the patient has a cytologically benign solitary nodule, a nodule in a subclinical or clinical goiter, or a lobe that remains after surgery, there is a small possibility that a neoplasm will develop or that the tissue may have the propensity to grow and result in obstruction of the thoracic inlet. Life-long observation is necessary. The clinical data must be interpreted with mature judgment and an understanding of the disease process, a specific person's needs, and available resources. In many cases, suppressive therapy is not needed; in others, the patient's medical condition or advanced age precludes this treatment; in still others, treatment with levothyroxine that avoids thyrotoxicosis seems a reasonable choice.

The debate about suppressive therapy continues because the answers are important and clinically relevant but not yet complete.