Abstract
KEY POINTS
Thyrotropin receptor antibody (TRAb) concentration, measured both at diagnosis and at cessation of antithyroid drug therapy, independently predicts the risk of relapse over a 10-year horizon in first-episode Graves' thyrotoxicosis.
Most relapses occur within 2 years of drug withdrawal, with recurrence beyond 5 years being uncommon, allowing clinicians to reassure patients who remain euthyroid in the early post-withdrawal period.
Roughly half of the patients who complete a course of antithyroid drugs avoid any further thyroid-directed therapy long-term, an outcome valued by patients but not reliably identifiable in advance even with TRAb.
SUMMARY
Background
Antithyroid drugs (ATDs) reliably restore euthyroidism in Graves' thyrotoxicosis, but relapse after a planned course is common, affecting roughly half of the patients in Europe and up to 80% in North America. 1 Thyrotropin receptor antibody (TRAb) is elevated at diagnosis in the large majority of patients with Graves' disease 2 and has been linked to both relapse risk after ATD withdrawal and radioiodine treatment failure.3,4 Other reported risk factors—younger age, larger goiter, male sex, orbitopathy and smoking—show considerable inconsistency across the literature. 5 Much of the existing evidence is limited by short follow-up (typically 1–2 years) and by heterogeneity in whether patients had completed a physician-directed course with planned cessation. This study aimed to characterize the natural history of relapse across a full decade and to define the risk factors that matter most.
Methods
This was a single-centre retrospective study from Edinburgh. 1 Two hundred and ninety consecutive patients with a first presentation of Graves' thyrotoxicosis and available baseline TRAb, who completed a physician-directed ATD course between 2006 and 2011, were identified, with serial thyroid function followed for at least 10 years after withdrawal. Standard practice was an 18-month tapering course of carbimazole (typical starting dose 40 mg). Propylthiouracil was reserved for planned pregnancy or carbimazole intolerance, and block-and-replace was not used. TRAb was measured at baseline in all patients and before cessation in 144 patients. Relapse was defined by the first date of biochemical thyrotoxicosis. Continuous data were compared using Wilcoxon rank-sum tests, relapse was analyzed using Kaplan–Meier methods with log-rank comparison, and Cox proportional hazards models identified independent predictors.
Results
Of 290 patients, 191 completed 10-year follow-up. Relapse occurred in 103 of 191 (54%) within 10 years, with 54% of relapses within 1 year, 72% within 2 years, and 93% within 5 years. Patients who relapsed were younger (41 vs. 47 years, p = 0.011) and had higher TRAb at diagnosis (8.8 vs. 6.0 IU/L, p = 0.002) and at cessation (1.3 vs. 1.0 IU/L, p < 0.001), with longer times to normalization of TSH and fT4. Recurrence reached 74% in those with diagnostic TRAb > 12 IU/L versus 44% in those with TRAb < 5 IU/L. On multivariate analysis, diagnostic TRAb > 12 IU/L (HR 2.4), detectable TRAb at cessation, and younger age were independent predictors. At a median of 141 months, around half of the patients required no thyroid-directed therapy, irrespective of TRAb category.
Conclusions
TRAb at diagnosis and at cessation, together with age, predicts long-term relapse risk; most recurrences occur early, and approximately half of the patients completing ATD avoid further therapy.
COMMENTARY
This study addresses a question central to almost every Graves’ thyrotoxicosis consultation: after completing a course of antithyroid drugs, how likely is this patient to relapse, and when? Its major strength lies in the duration and completeness of follow-up. Most previous studies have reported outcomes over 1–2 years, despite relapse risk being unevenly distributed across time. 5 By following a well-characterized cohort for a full decade after physician-directed ATD therapy, Tun and colleagues define the natural history of recurrence with a clarity that shorter studies cannot achieve.
The temporal pattern of relapse is perhaps the study’s most clinically useful finding. With 72% of recurrences occurring within 2 years and 93% within 5 years, relapse is shown to be overwhelmingly an early event. This carries a practical implication for patient counseling: individuals who remain euthyroid several years after treatment cessation have a low and progressively diminishing likelihood of recurrence. Conversely, the early post-withdrawal period warrants the closest biochemical surveillance, as this is when most relapses occur.
The reaffirmation of TRAb as a prognostic marker, rather than merely a diagnostic one, is also valuable, though perhaps less novel. 3 Most clinicians—ourselves included—already incorporate TRAb into our prognostic thinking, commonly deferring withdrawal of carbimazole until TRAb becomes undetectable. This study adds durability, showing that the association persists, even up to 10 years, a period in which you might expect the baseline value to lose any discriminatory value. A diagnostic TRAb above 12 IU/L was associated with a recurrence rate of 74%, compared with 44% among those below 5 IU/L. These findings position TRAb as a meaningful tool for risk stratification and support its incorporation into early discussions regarding treatment strategy. However, the authors are appropriately candid about the limits of stratification. Composite tools such as the Graves’ Recurrent Events after Therapy (GREAT) score, which combine age, goiter size, free T4 and TRAb, 6 and which has been externally validated in a large multicenter cohort, 5 improve prognostic performance but do not identify an individual patient’s outcome with the confidence with which we would wish, and the modest sensitivity and specificity of TRAb for predicting 10-year recurrence in the present study reinforces this limitation.
The findings are particularly relevant to evolving practice surrounding longer-term ATD therapy. The most compelling evidence comes from the randomized trial by Azizi et al., in which patients continuing low-dose methimazole for a total of 60–120 months had a relapse rate roughly one-third that of those stopping at 18–24 months (around 15% vs. 53%), with no excess of adverse effects over the extended period. 7 Findings of this kind reframe long-term, low-dose ATD as a legitimate maintenance strategy in its own right rather than a holding measure before definitive therapy—a meaningful change in emphasis from the historical assumption that drug treatment is a time-limited attempt at remission, with mandatory radioiodine or surgery if unsuccessful. The observation that patients with the highest TRAb concentrations tend to relapse within the first 2 years raises the possibility that treatment duration could be individualized rather than fixed at 12–18 months.
Importantly, approximately half of the patients completing an ATD course required no thyroid-directed therapy at long-term follow-up, irrespective of TRAb category. For many patients, the possibility of avoiding lifelong treatment remains a major consideration, 8 and these data can inform that preference with a credible quantitative framework. TRAb therefore serves not to dictate management, but to support shared decision-making by identifying patients at higher relapse risk while preserving medical therapy as a reasonable option for those willing to accept uncertainty. As a single-centre retrospective study lacking access to variables such as goiter size, orbitopathy, and smoking status, these findings warrant the usual caution, and the optimal duration and selection criteria for extended ATD remain to be defined prospectively. Nonetheless, the study provides clinicians with a robust long-term evidence base for a conversation that has too often relied upon short-term data.