Abstract
KEY POINTS
This meta-analysis found no significant survival or disease-control benefit from thyrotropin (TSH) suppression in intermediate- and high-risk differentiated thyroid cancer, while a pooled signal pointed to real cardiovascular and skeletal harm.
The underlying evidence for TSH suppression remains weak and heterogeneous, with suppression thresholds, assays, and treatment eras differing widely across studies; taken together, these findings should temper enthusiasm for aggressive TSH suppression in most patients.
These results were among the evidence cited in the 2025 American Thyroid Association (ATA) guidelines, contributing to a shift away from fixed, ATA risk-based suppression targets toward response-to-therapy–driven, largely normal-range TSH goals.
SUMMARY
Background
Thyrotropin (TSH) suppression with supraphysiological doses of levothyroxine has long been a mainstay of intermediate- and high-risk differentiated thyroid cancer (DTC) management, predicated on the observation that DTC cells express the TSH receptor and that TSH may exert a trophic effect promoting tumor growth. 1 The 2015 American Thyroid Association (ATA) guidelines recommend a serum TSH below 0.1 mIU/L for high-risk DTC (strong recommendation, moderate-quality evidence) and 0.1–0.5 mIU/L for intermediate-risk DTC (weak recommendation, low-quality evidence). 2 The supporting evidence is of low to moderate quality, and more recent cohort studies have questioned whether aggressive suppression improves survival. 1 This systematic review and meta-analysis was performed to assess the association between TSH suppression and both survival outcomes and treatment-related adverse events in intermediate- and high-risk DTC.
Methods
The authors performed a systematic review and meta-analysis of studies evaluating TSH suppression in adults aged 18 years or older with intermediate- and high-risk DTC who were treated with guideline-based initial therapy and followed for 5 years, using the 2015 ATA criteria for risk stratification. TSH suppression was defined as below the normal reference range or, when known, <0.5 mIU/L; the nonsuppressed comparator was an average TSH of 0.5–2 mIU/L (low-normal) or >2 mIU/L. Randomized controlled trials, cohort studies, and case–control studies were eligible for inclusion. The primary outcomes were a composite of progression-free survival (PFS), disease-free survival (DFS), and relapse-free survival (RLFS) and a composite of disease-specific survival (DSS) and overall survival (OS). The secondary outcome was a composite of cardiovascular and skeletal adverse events. Risk of bias was assessed with the National Heart, Lung, and Blood Institute Study Quality Assessment Tool, with each study rated good, fair, or poor. Pooled hazard ratios (HRs) with 95% confidence intervals (CIs) were calculated using a random-effects model, with heterogeneity quantified by the I2 statistic.
Results
Of 6369 reports screened, 9 studies published between 1996 and 2019 met the inclusion criteria: 5 retrospective cohort studies, 1 prospective cohort study, 1 case–control study, and 2 randomized controlled trials. On risk-of-bias assessment, four studies were rated good, four fair, and one poor. The definition of TSH suppression varied widely across studies. Based on seven studies (3591 patients), the composite of PFS, DFS, and RLFS did not differ significantly between the suppression and nonsuppression groups (HR, 0.75; 95% CI, 0.48–1.17; I2 = 76%). The composite of DSS and OS, based on four studies (3616 patients), likewise did not favor suppression (HR, 0.69; 95% CI, 0.31–1.52). Restricting the primary composite outcome to good-quality studies alone (3109 patients) again showed no benefit (HR, 0.98; 95% CI, 0.64–1.5). In contrast, the secondary composite of cardiovascular and skeletal adverse events, based on two studies (1294 patients), was significantly higher with TSH suppression (HR, 1.82; 95% CI, 1.30–2.55).
Conclusions
In intermediate- and high-risk DTC, TSH suppression was not associated with improved progression or survival outcomes but was associated with a higher risk of cardiovascular and skeletal adverse events. The authors caution that limited evidence and substantial study heterogeneity warrant cautious interpretation of these findings.
COMMENTARY
The headline of this meta-analysis is striking: in intermediate- and high-risk DTC, suppressing TSH conferred no measurable survival or disease-control advantage, while the pooled data pointed toward genuine cardiovascular and skeletal harm. Just as important, in my view, is what the study exposed about the evidence base itself. Despite TSH suppression being entrenched in DTC management, the authors' comprehensive search, formal risk-of-bias assessment, and good-quality sensitivity analysis expose how limited the evidence supporting TSH suppression really is, with no benefit emerging even among the highest-quality studies.
The nine included studies spanned more than two decades and defined suppression in vastly different ways, from a TSH below the reference range to undetectable values measured on older, less sensitive assays, and heterogeneity for the primary outcomes was considerable. The one hint of a survival benefit of TSH suppression came from a single study by Carhill et al., drawn from a large national thyroid cancer registry, 3 and only when its patient groups were broken down by tumor stage. Even then, the apparent benefit came entirely from comparing patients on modest suppression against those whose TSH was high (above the upper limit of normal), so that signal more likely reflects the harm of a high TSH than the benefit of a low one. The harm signal secondary to TSH suppression was a more straightforward story. Only two studies assessed the composite of cardiovascular and skeletal events, but they agreed closely with one another, and both pointed to greater harm with suppression.
This meta-analysis helped form the body of literature leading to a paradigm shift in the ATA 2025 guidelines regarding TSH suppression. Rather than fixing a TSH target by initial risk category for a defined number of years, the 2025 ATA guidelines tie the target to the response to therapy. The decision to suppress is individualized, weighed against comorbidities such as atrial fibrillation and osteoporosis, and reevaluated over time rather than maintained indefinitely. 4
This study sits within a largely concordant body of recent meta-analyses, most of which document cardiovascular and skeletal harm 5 and no clear survival benefit, 6 a marked shift from the prosuppression signal reported two decades ago. 7 One contrasting 2025 meta-analysis found that the overall risk of recurrence was unrelated to the degree of TSH suppression, except in the subgroup with distant metastases, in which a TSH ≥ 0.1 mIU/L carried a higher recurrence risk, suggesting that suppression may still benefit that specific high-risk group. 8 That nuance is precisely why the guidelines preserve a subnormal target for structurally incomplete disease.
My own practice has changed on the strength of this meta-analysis and the 2025 ATA guideline recommendations. I am more alert to the harms of TSH suppression and more skeptical of its presumed survival and recurrence benefits. Applying the new cutoffs to newly diagnosed patients may be easier than adjusting targets for patients diagnosed prior to guideline changes. Anecdotally, I have faced a few challenges when attempting to de-escalate TSH targets to within the reference range in patients already established on suppression. The first is psychological: Many patients come to regard their levothyroxine as their cancer treatment, the part of their care they can control, so a higher TSH target can cause fear of recurrence. The second is physiological: some patients accustomed to a suppressed TSH seem to feel hypothyroid as TSH is normalized. We will need to learn how best to conduct de-escalation conversations, and more studies are needed in this area, including work that captures the patient’s own perspective.