This retrospective study was approved by the local ethics committees of Nagoya University Hospital (Nagoya, Japan) (2020–0052) and Hokko Memorial Hospital (Sapporo, Japan) (I-03). Overall, 185 adults (male 25, female 160; average age 46.7 years) at two institutes in Japan were included: 114 who lived in the Aichi or Mie Prefecture and underwent RAI at Nagoya University Hospital, Nagoya, between April 2017 and March 2022 (group A), and 71 from the Hokkaido Prefecture who underwent RAI at Hokko Memorial Hospital, Sapporo, between December 2012 and May 2017 (group H). None of the patients dropped out in the year after RAI. All patients were pretreated with ATDs (MMI, PTU, and/or KI) before RAI. The patient inclusion criteria were as follows: (1) biochemically confirmed GH with positive thyrotropin receptor antibody (TRAb); (2) an estimated thyroid weight < 100 g; and (3) follow-up duration of thyroid function of > 1 year after RAI. These criteria ensured no bias between the two hospitals and that patients would be treated with the first RAI. All patients provided written informed consent to undergo RAI and associated procedures, including adherence to the LID. The sociophysiological characteristics of the patients are summarised in Table 1.

Two preparatory regimens were used (Fig. 1). In group A, pretreatment with KI was withdrawn 18 days before RAI, whereas ATDs were discontinued 7 days before RAI. In group H, KI was withdrawn 14 days before RAI, while ATDs were discontinued 3–4 days before RAI. After nutritional counselling with a dietitian or physician, patients in group A underwent LID from 18 days before RAI to 4 days after RAI. Patients in group H underwent LID from 7 days before to 3 days after RAI. We developed a brochure for patients who were followed up for a LID with RAI [17]. This was used for counselling in group H. In group A, commercially available materials were used for counselling.

Protocol for radioactive iodine treatment at the two facilities. Aichi group: Patients in Nagoya University Hospital living in Aichi. ATD, antithyroid drug; FT3: free triiodothyronine; FT4: free thyroxine; Hokkaido Group: Hokko Memorial Hospital, Hokkaido; KI: potassium iodine; p.o.: per os; RU: radioiodine uptake; TRAb: thyrotropin receptor antibody; UIC: urinary iodine concentration

The dose of 131I was determined according to Marinelli’s formula [18]. In this formula, the weight of the thyroid gland (TW), 24-h radioiodine uptake (24hRU), and the effective half-life of radioiodine in the thyroid (T1/2eff) served as the variables for the administered dose of 131I. In group A, radioiodine uptake (RU) was measured using 131I and an uptake probe (thyroid uptake system: AZ-800, Anzai Medical, Tokyo, Japan). Exactly 3.7 MBq of 131I was administrated 7 days before RAI. RU was sequentially measured at 3, 24, and 96 h to determine the T1/2eff. In group H, RU was determined using 123I and a gamma camera (Millennium MG, GE Healthcare, Milwaukee, USA), and 7.4 MBq of 123I was administered the day before RAI to perform thyroid scintigraphy and determine the 24hRU. 24hRU determined using 123I based on a standard method, a region of interest (ROI) was drawn around the margins of the thyroid gland and ROI in the supraclavicular area was used for background subtraction. The whole-body count was measured using a scintillation survey meter at 24, 72, 96, and 120 h after RAI in seven patients. The average value of 5.7 days from these patients was applied to the other patients to calculate the dosage of 131I. In group A, TW was measured using computed tomography (CT) (Aquilion 64 or Aquilion PRIME SP; Canon Medical Systems, Tokyo, Japan). The patients underwent non-enhanced neck CT examinations for thyroid volumetry. Contour extraction of the thyroid from CT images (5-mm slice thickness) was manually performed by a single radiologist (nuclear medicine physician). The sum of thyroid volumes obtained from each slice was denoted as the total thyroid volume. In group H, TW was measured using a Aquilion PRIME CT scanner (Canon Medical Systems) or by ultrasonography (US) (Xario; Canon Medical Systems). Thyroid volume was obtained by US according to the following formula:

\(\mathrm\pi/6\times\mathrm\times\mathrm\times\mathrm\;\mathrm\) [4]

Thyroid function (free tri-iodothyronine [FT3], free thyroxine [FT4], and thyroid-stimulating hormone [TSH]) and TRAb levels were measured on the day of RAI. The UIC was measured twice using spot urine samples before commencing the LID (UIC1) and on the day of RAI (UIC2). In group A, thyroid function was determined using a CLEA kit (Abbott Diagnostic Medical, Tokyo, Japan; normal range: FT3 1.68–3.67 pg/mL, FT4 0.70–1.48 ng/dL, TSH 0.35–4.94 μIU/mL). In group H, thyroid function was evaluated using the ECLEA (ECLEA) Kit (Roche Diagnostic Medical, Tokyo, Japan; normal range: FT3 2.3 − 4.0 pg/mL, FT4 0.90–1.7 ng/dL, TSH 0.50–5.00 μIU/mL). TRAb was determined using an ECLES Kit (Roche Diagnostic; normal range: < 2.0 IU/L in both institutes). UIC was measured by visible absorptiometry as previously described [19]. Thyroid function was monitored every 1–3 months after RAI. RAI was defined as successful when euthyroidism, subclinical hypothyroidism, or hypothyroidism was achieved within 1 year after RAI.

We investigated the changes in UIC before and after LID. We also investigated the relationship between UIC2 vs. 24hRU (group A and group H) and UIC2 vs. T1/2eff (group A only). Additionally, we investigated whether lower UIC2 enhances RAI success rate (SR). Commercially available KI in Japan contains 38.2 mg of iodine per 50 mg of the pill. Because 76 of 185 patients had been taking 1 − 6 pills of KI, the patient groups were further classified into four subgroups based on the pretreatment use of KI: A1 (group A pretreated with ATDs without KI, n = 74); A2 (group A pretreated with ATDs combined with KI or KI alone, n = 40); H1 (group H pretreated with ATDs without KI, n = 35); and H2 (group H pretreated with ATDs combined with KI, n = 36). The 185 patients underwent further stratification into quartiles based on UIC2, regardless of the pretreatment use of KI. The Wilcoxon signed-rank test was used to investigate changes in UIC before and after LID initiation. Spearman’s rank correlation coefficient was used to investigate the relationship between UIC2 vs. 24hRU and UIC2 vs. T1/2eff (group A only). This was a non-parametric correlation assessment of the two groups. FT3, FT4, TRAb, TW, UIC2 and 24hRU may be factors affecting the efficacy of RAI, and this hypothesis was tested using Fisher’s exact test. We examined the factors influencing the treatment effects by correcting for the effects of each variable. Similarly, T1/2eff might be a success factor for RAI, and we tested this hypothesis using the Fisher’s exact test. The Fisher’s exact test was used for nominal variables and independent comparisons between groups. Statistical significance was set at p < 0.05. All statistical analyses were performed using EZR (Saitama Medical Center, Jichi Medical University, Saitama, Japan), a graphical user interface for R (The R Foundation for Statistical Computing, Vienna, Austria) [20].