| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Article |
Section of General Internal Medicine (M.R.C., P.J.C.) and Division of Endocrinology (M.R.C.), Albany Medical School and VA Healthcare Network Upstate New York, Albany, New York 12208; and Division of Endocrinology and Metabolism (J.C.M.), Mayo Clinic, Rochester, Minnesota 55905
Address all correspondence and requests for reprints to: M. Regina Castro, M.D., Stratton VA Medical Center, 113 Holland Avenue, Albany, New York 12208. E-mail: . Maria.Castro2{at}med.va.gov
Abstract
The effectiveness of thyroid hormone suppressive therapy in reducing the volume of benign thyroid nodules is controversial. It is important to clarify this therapeutic effect of thyroid hormone, because its prolonged use needs to be carefully weighed against its potential deleterious effects in the skeletal and cardiovascular systems. To evaluate the best available evidence, we conducted a systematic review and meta-analysis of the randomized controlled trials that fulfill the following inclusion criteria: single thyroid nodules proven benign by fine needle aspiration, treatment, and follow-up of at least 6 months; documented suppression of TSH; measurement of thyroid nodule volume by ultrasound; and response to therapy defined as more than 50% volume reduction from baseline. Six randomized clinical trials published between 1987 and 1999, with 346 patients, were included in the meta-analysis. Ninety percent of the participants were female. Using a random effects model, the overall effect size showed a relative risk of 1.9 (95% confidence interval, 0.95–3.81) favoring a treatment effect. A sensitivity analysis showed significant changes in the results.
Suppressive thyroid hormone therapy for longer than 6 months is associated with a trend toward a reduction of more than 50% in volume of benign thyroid nodules, without achieving statistical significance. The results are highly sensitive to changes in the statistical analysis, especially if the method used ignores heterogeneity among the effect sizes. More studies are needed before this therapy can be widely recommended.
ALTHOUGH BENIGN NODULAR thyroid disease is a very common disorder, significant controversy still exists regarding its management. Since the original publication by Astwood et al. in 1960 (1) reporting a reduction in nodule size with the use of thyroid hormone suppressive therapy in an uncontrolled trial, multiple reports have either confirmed (2, 3) or refuted these observations (4, 5, 6, 7). However, the majority of these reports were either uncontrolled trials or observational studies (8, 9). Furthermore, many of these studies included a heterogeneous group of patients with single benign thyroid nodules, colloid or cystic nodules, multinodular or simple adenomatous goiters; different treatment doses and length of therapy, often without clear documentation of TSH suppression with therapy (2, 3); and variable definitions of what constitutes a response to therapy, as well as different methods of assessment of nodule size [palpation vs. ultrasound (US)], all of which may explain the large variability in study results. Because prolonged thyroid hormone suppressive therapy has been associated with adverse effects on the skeletal and cardiovascular systems (10, 11, 12, 13), the expected benefits of this treatment must be carefully weighed against the potential adverse effects, and careful review of the current evidence is needed to settle this controversy.
The aim of this study is to systematically review the available evidence in the medical literature about the effectiveness of thyroid hormone suppressive therapy in patients with benign, solitary thyroid nodules and to perform a meta-analysis of all randomized, controlled, clinical trials (RCT).
Subjects and Methods
Search and study selection
We conducted a comprehensive, computerized search of the literature in MEDLINE from January 1985 until April 2001 and performed a hand search in the references of pertinent publications. The terms used in the search included: thyroid nodules, nodular thyroid disease, solitary thyroid nodules, goiter, nontoxic goiter, cold nodules combined with levothyroxine (L-T4) suppression, thyroid hormone suppression, suppressive therapy, and thyroxin. We searched for all publications reporting original data on the effect of thyroid hormone suppressive therapy on benign thyroid nodular disease. Studies that met the following criteria were considered for inclusion in the present report: RCT, with or without placebo; single thyroid nodule (STN) by palpation, proven benign by fine needle aspiration biopsy; documented suppression of TSH with thyroid hormone therapy; measurement of thyroid nodule volume by US; response to suppressive therapy defined as greater than 50% reduction in nodule volume from baseline; and treatment duration and follow-up of at least 6 months. Two publications (6, 14) that reported data on the same patient population were considered as one. Another study (15) reported results of suppressive therapy in patients with benign single nodules and multinodular goiter (MNG), but only results pertaining to those patients with STN were considered in the current report.
Information from the studies was tabulated by one reviewer (M.R.C.) and subsequently confirmed by a second reviewer (P.J.C.). The following data were collected from each study: 1) publication data, first author’s last name, year of publication, and country of the population studied; 2) study design; 3) duration of treatment in months; 4) characteristics of the exposed and unexposed groups, i.e. number of subjects in each group, diagnosis, gender, and age; and 5) treatment dose of L-T4 and documentation of TSH suppression.
To assess the methodological quality of the studies, we performed an open review of all the studies initially identified by the search of the literature. Studies that did not meet predetermined inclusion criteria were excluded from the final review.
Analysis
We performed a meta-analysis on the effect of L-T4 suppressive therapy in patients with benign STNs. We considered only the studies that met all of the above-mentioned inclusion criteria, to avoid significant methodological differences in study design and treatment protocols. We calculated the effect size in each study using relative risk (RR) with their 95% confidence intervals (CI). The RR was calculated using the number of patients assigned to treatment vs. control group whose nodule volume decreased by more than 50%, as determined by US measurement. The effect sizes were pooled on the basis of a random effects model (16) that incorporates the statistical heterogeneity, if present, and provides a more conservative estimate of the pooled effect size (wider CI). If heterogeneity is not present, the fixed effect and random effects model provide similar results (17). Using this methodology we minimized the probability of an 
error (to conclude that the treatment is effective when, in fact, it is not), although it may result in an increased probability of a ß error (concluding that the treatment is not effective, when in reality it is). We chose this method because of concern about the potential adverse effects of long-term suppressive thyroid hormone therapy. Heterogeneity of study results was evaluated by visual examination of a plot of the effect sizes (RR and 95% CI), where the degree of homogeneity is judged on the basis of the degree of overlapping of the CI. We also used a statistical test for homogeneity, which assesses the probability that the difference between the result of the studies is due to chance (the smaller the P value, the less homogeneity present among study results). To evaluate the impact of each selected study on the overall results of the meta-analysis, we performed a one-way (one study excluded at the time) and two-way sensitivity analysis (two studies excluded at a time). We recalculated the overall RR and 95% CI using a fixed effect model.
Results
Thirteen original clinical trials assessing the effect of L-T4 suppressive therapy on benign STNs were identified (4, 5, 6, 7, 8, 9, 14, 15, 18, 19, 20, 21, 22). Of these, two studies (6, 14) reported data on the same population, and only one of them (6) was included for this analysis. Of the remainder, six other studies were excluded (Table 1
) for not meeting predefined criteria with regard to: 1) study design, lack of randomization (8, 9, 15, 22) or of control groups (8, 9); 2) nodule size not assessed by US (20); and 3) definition of response to treatment (21).
|
). The dose of L-T4 used varied from 1.5–3 µg/kg, and in every study appropriate suppression of TSH was documented, either by measurement of TSH by second or third generation assay or by TRH stimulation test. The studies were performed in the United States (4), Italy (6, 18), Spain (5), Brazil (19), and Iran (7). The publication dates of these studies ranged between 1987 and 1999.
|
); however, its degree was judged to justify the use of a random effects model (
2 = 7.42; P = 0.19). The meta-analysis resulted in an overall RR of 1.9 favoring the treatment group, which did not reach statistical significance (95% CI, 0.95–3.81).
|
To evaluate the robustness of the results of this meta-analysis, and due to the significant controversy surrounding this therapy, we performed a one-way sensitivity analysis. In this analysis, the overall homogeneity and effect size were calculated, removing one study at a time (Table 3). The RR achieves statistical significance only when the study by Gharib et al. (4) is excluded (RR, 2.35; 95% CI, 1.21–4.57). On the other hand, the homogeneity improved in a similar way when the study by either Gharib et al. (4) or La Rosa et al. (18) was removed from the analysis (2 = 4.39, P = 0.355; and 2 = 4.19, P = 0.380, respectively). In a two-way sensitivity analysis, when both of these studies that introduce significant heterogeneity are excluded, the remainder of the studies become homogeneous (2 = 1.81; P = 0.6127), and the RR (2.08; 95% CI, 1.09–3.06) approaches that seen when a fixed effects model is used including all the studies (RR, 2.2; 95% CI, 1.3–3.72). However, exclusion of these studies is arbitrary and may introduce unintentional bias.
|
The efficacy of L-T4 therapy to reduce thyroid nodule size has been a matter of controversy for many years. Because of growing concern about potential adverse effects of long-term suppressive doses of L-T4 on the cardiovascular and skeletal systems, it becomes imperative to analyze the current evidence in a more systematic way.
This study systematically reviewed the six best RCT that evaluate the effectiveness of L-T4 suppressive therapy given for more than 6 months in decreasing volume of solitary nodules by more than 50%. We found an overall treatment response that did not achieve statistical significance (RR, 1.9; 95% CI, 0.95–3.81). This result is very sensitive to the small statistical heterogeneity of the effect sizes of the studies included in the meta-analysis. If we assume homogeneity (fixed effect model) or exclude outliers to improve homogeneity, the results become statistically significant. The clinical and methodological similarities of the studies make it difficult to assume these changes without jeopardizing the integrity of the systematic review and the meta-analysis.
Although five of the six studies included in the meta-analysis showed no significant reduction in nodule volume with L-T4 suppressive therapy, most of them did show a trend in favor of L-T4 treatment (5, 6, 7, 19). The proportion of responders was greater in the L-T4 (mean, 22%; range, 14–39%) than in the placebo or control groups (mean, 10%; range, 0–20%), but statistical significance was achieved in a single study (18). Only one study (4) was clearly negative, with a RR less than 1. Although most studies included patients with exclusively (18) or predominantly solid nodules (5, 6, 19), which are known to be more responsive to L-T4 suppressive therapy, in the study by Gharib et al. (4) up to 38% of patients had predominantly cystic nodules (14% in the L-T4 group and 24% in the control group), which may at least partially explain the greater proportion of patients who failed to respond to therapy, when compared with other studies. A placebo-controlled study by McCowen et al. (24) demonstrated lack of efficacy of L-T4 suppressive therapy in preventing recurrence of cystic nodules after aspiration. The study by Larijani et al. (7) also included a significant proportion of patients with mixed or predominantly cystic nodules, although the relative proportion of nodules in each of these two categories is not clearly defined.
The results of this meta-analysis are in contrast to those reported by Zelmanovitz et al. (19). However, some significant differences between these meta-analyses need to be pointed out, which may explain the contradictory results of the two studies. Zelmanovitz et al. included seven trials in their meta-analysis (4, 5, 6, 15, 18, 19, 22), five of which are also included in our current report (4, 5, 6, 18, 19). We excluded the studies by Mainini et al. (22) and Lima et al. (15) because they did not meet one or more of our inclusion criteria. Specifically, on detailed review of these studies, the trial reported by Mainini et al. (22) was not a randomized study, and therefore, potential for selection bias existed. The second study, by Lima et al. (15), was also a nonrandomized trial in which the control group had specific contraindications to L-T4 therapy. In this study, which included 101 patients, only 54 of them (rather than 74, as stated in that meta-analysis) had a diagnosis of STN, the rest being patients with MNG. Finally, our study includes an additional trial by Larijani et al. (7) published in 1999, not included in the previous meta-analysis.
Contradicting results of multiple previous reports on this topic are likely the result of extensive and significant methodological differences in study design. Many of these studies were uncontrolled (8, 9), often conducted in endemic goitrogenous areas (8) in which the therapeutic effect of TSH suppression may be enhanced; an otherwise well designed study was flawed by lack of an objective method of assessment of nodule volume (20). It has been demonstrated that US measurement is far more sensitive and accurate for estimating number and size of thyroid nodules than palpation (25, 26).
Our study attempts to overcome these flaws by establishing very well defined inclusion criteria to minimize, in as much as possible, these methodological differences and ensure a certain degree of homogeneity in the design of the studies included in our report. However, despite our efforts to minimize these methodological differences, a test for homogeneity demonstrated that there remained between-study heterogeneity that cannot be accounted for by the above criteria. Also, results of the meta-analysis differed depending on the model used to analyze the result, suggesting meaningful discrepancies (heterogeneity) among the study-specific effect estimates.
Although the results of our meta-analysis do not demonstrate a statistically significant difference in nodule size after 6–12 months of L-T4 suppressive therapy, when compared with control group, there is a clear trend favoring the treatment group, and a larger proportion of patients in the L-T4 group responded with more than 50% reduction in nodule volume. Also, in one study, L-T4 has been shown to significantly decrease the number of new nodules forming in the gland (14). It is possible, therefore, that a subgroup of patients, not easily identifiable on the basis of pretreatment characteristics, may, in fact, benefit from this therapy. Certain nodule features may predict a better response to L-T4 therapy, such as small nodular volume (<10 ml) and abundant colloid on fine needle aspiration (23). Patients with exclusively or predominantly solid nodules may also be more responsive to this therapy than those with a major cystic component, because thyroid cysts have been shown to recur after aspiration despite treatment with L-T4 suppression (24). Although, on the basis of the results of our study, L-T4 suppressive therapy cannot be generally recommended for treatment of single benign thyroid nodules, with the expectation of achieving significant reduction in their size, it may still be reasonable to offer this option to carefully selected patients whose risk for development of potential adverse effects is low, and consider discontinuation of treatment if no objective response is seen after a predetermined period of time. Further carefully designed studies including a larger number of patients are needed to provide a clear answer to this important clinical question.
Acknowledgments
Footnotes
Abbreviations: CI, Confidence interval; L-T4, levothyroxine; MNG, multinodular goiter; RCT, randomized, controlled, clinical trials; RR, relative risk; STN, single thyroid nodule; US, ultrasound.
Received November 2, 2001.
Accepted June 7, 2002.
References
This article has been cited by other articles:
 |
|
 |
|
  H Gharib, E Papini, and R Paschke Thyroid nodules: a review of current guidelines, practices, and prospects Eur. J. Endocrinol., November 1, 2008; 159(5): 493 - 505. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. R. Castro and H. Gharib Continuing Controversies in the Management of Thyroid Nodules Ann Intern Med, June 7, 2005; 142(11): 926 - 931. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. J. Mandel A 64-Year-Old Woman With a Thyroid Nodule JAMA, December 1, 2004; 292(21): 2632 - 2642. [Full Text] [PDF]
|
|
|
|
 |
|
 L. Hegedus The Thyroid Nodule N. Engl. J. Med., October 21, 2004; 351(17): 1764 - 1771. [Full Text] [PDF]
|
|
|
|
 |
|
 G. Brenta, M. Schnitman, O. Fretes, E. Facco, M. Gurfinkel, S. Damilano, N. Pacenza, A. Blanco, E. Gonzalez, and M. A. Pisarev Comparative Efficacy and Side Effects of the Treatment of Euthyroid Goiter with Levo-Thyroxine or Triiodothyroacetic Acid J. Clin. Endocrinol. Metab., November 1, 2003; 88(11): 5287 - 5292. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Hegedus, S. J. Bonnema, and F. N. Bennedbaek Management of Simple Nodular Goiter: Current Status and Future Perspectives Endocr. Rev., February 1, 2003; 24(1): 102 - 132. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
