Comparative efficacy and thyroid function effects of non-surgical treatments for benign thyroid nodules: a Bayesian network meta-analysis

OBJECTIVE: Non-surgical treatments for benign thyroid nodules (BTN) have shown variable effects across different outcome domains, but randomized comparative evidence remains fragmented, particularly regarding both structural efficacy and thyroid function effects. This study aimed to compare available non-surgical interventions for nodule volume and thyroid function parameters using a Bayesian network meta-analysis. METHODS: PubMed, Embase, Web of Science, and Cochrane Library were searched through August 17, 2025. RCTs evaluating levothyroxine (L-T4), radiofrequency ablation (RFA), microwave ablation (MWA), percutaneous ethanol injection (PEI), or percutaneous laser ablation (PLA) in adults with BTN were included. Outcomes were nodule volume, TSH, FT3, and FT4. Bayesian random-effects network meta-analysis used R Studio (gemtc) and STATA; rankings employed SUCRA. RESULTS: Fourteen RCTs (545 participants) were included. For nodule volume, PLA (WMD≈-5.23, 95% CI: -8.26 to -2.15) and RFA (WMD≈-4.11, 95% CI: -6.19 to -2.38) were superior to L-T4 and placebo. L-T4 showed strongest TSH suppression (vs. MWA: WMD≈-1.59, 95% CI: -2.42 to -0.79). No significant differences were observed for FT3 or FT4. CONCLUSIONS: Within the available randomized evidence base, non-surgical treatments for BTN showed outcome-specific differences. Thermal ablation techniques, particularly PLA and RFA, were associated with greater reductions in nodule volume, whereas levothyroxine showed the strongest effect on TSH suppression. However, because the network was largely built on historical comparators and did not include several key patient-centered clinical outcomes, the applicability of these findings to current practice remains limited. Further head-to-head trials comparing contemporary minimally invasive treatments are needed.