2025 年 72 巻 5 号 p. 637-643
Serum thyroglobulin (Tg) levels are highly sensitive and specific tumor markers in patients with thyroid cancer who have undergone total thyroidectomy and radioiodine (RAI) therapy. Recently, we reported a case wherein serum Tg levels fluctuated according to hemoglobin A1c (HbA1c) levels, demonstrating a strong correlation between serum Tg and HbA1c levels. However, whether this association exists broadly in other patients with thyroid cancer remains unclear. Therefore, we retrospectively investigated this association in six patients with thyroid cancer and diabetes who underwent total thyroidectomy and RAI therapy at our institution. Two patients exhibited a significant correlation between serum Tg and HbA1c levels (r = 0.53, p < 0.01 and r = 0.66, p = 0.01). In these patients, a gradual decrease in serum Tg levels was observed, along with improved glycemic control. Two other patients showed a non-significant correlation between serum Tg and HbA1c levels (r = 0.72, p = 0.11 and r = 0.54, p = 0.17). In these patients, a rapid increase in serum Tg levels was observed, with HbA1c levels showing only small fluctuations. In the remaining two patients, no correlation was found; in these patients, the fluctuations in serum Tg levels were only small, despite fluctuations in HbA1c levels. In conclusion, serum Tg levels may be associated with HbA1c levels in some patients with thyroid cancer and diabetes. The correlation between serum Tg and HbA1c levels was more evident in patients with gradual fluctuations in Tg levels. Future studies with larger cohorts are necessary to clarify the underlying mechanism by which glycemic control influences Tg levels and establish appropriate monitoring methods.
Thyroglobulin (Tg) is a glycoprotein synthesized by thyroid follicular cells. Serum Tg is believed to be derived from thyroid cancer cells after total thyroidectomy and radioiodine (RAI) therapy, making it a highly sensitive and specific tumor marker [1]. Generally, serum Tg levels correlate with tumor size during the natural course of thyroid cancer but are also influenced by serum TSH levels.
We recently reported a case of RAI-refractory thyroid cancer associated with type 2 diabetes [2]. Serum Tg levels of the patient fluctuated according to hemoglobin A1c (HbA1c) levels, with a strong correlation between the two. This suggests that serum Tg levels may also be influenced by glycemic control. However, it remains unclear whether there is a similar association between serum Tg levels and glycemic control among other patients with thyroid cancer. Therefore, we investigated the association between serum Tg levels and glycemic control in patients with thyroid cancer and diabetes who underwent total thyroidectomy and RAI therapy.
We reviewed the medical records of 159 consecutive patients with differentiated thyroid cancer who underwent total thyroidectomy and RAI therapy at the Kobe City Medical Center General Hospital, Japan, between April 2011 and March 2022. Among them, 29 had diabetes during the study period. We excluded 12 patients whose serum Tg levels were below the detection limit, two patients who were positive for Tg antibodies, and nine patients with insufficient data. We retrospectively analyzed the remaining six patients, along with our recently reported case (case A) [2]. The survey period was defined as the time during which serum Tg levels were measured six or more times, and no new therapeutic interventions, such as surgery or RAI therapy, were undertaken. Only serum Tg levels obtained when the TSH levels were <0.5 mIU/L were included in the analysis. We analyzed the relationship between serum Tg and HbA1c levels that were measured concurrently (within 2 weeks).
A summary of these cases follows.
Case seriesIn case A, a 25-year-old woman underwent total thyroidectomy for papillary thyroid carcinoma in 2003. She subsequently received 11 sessions of RAI therapy for lung metastases but was diagnosed as RAI-refractory. In 2017, at the age of 39 years, she was diagnosed with type 2 diabetes with an HbA1c level of 10.8% and was administered a diabetic drug. More detailed information on this patient has been previously described [2].
Case 1 involved a 49-year-old man who underwent total thyroidectomy for papillary thyroid carcinoma in 2013 and received two sessions of RAI therapy. As his serum Tg levels remained detectable with no evidence of structural disease, he was diagnosed with biochemical persistent disease (BPD). At the time of surgery, he was also diagnosed with type 2 diabetes with an HbA1c level of 7.0% and was administered a diabetic drug. His serum Tg levels were undetectable after the third RAI therapy session; therefore, the period after the third RAI therapy session was excluded from the analysis.
Case 2 involved a 51-year-old man who underwent total thyroidectomy and RAI therapy for papillary thyroid carcinoma in 2017. His serum Tg level remained detectable with no evidence of structural disease; thus, he was diagnosed with BPD. He was diagnosed with type 1 diabetes in 2003 at the age of 37 years and had been receiving insulin therapy. In 2021, the patient was diagnosed with rectal cancer, which interrupted his follow-up.
Case 3 involved a 60-year-old woman who underwent total thyroidectomy for papillary thyroid carcinoma in 2016. RAI therapy was administered for lung metastases; however, no uptake was observed, leading to a diagnosis of RAI-refractory disease. In 2017, she experienced recurrence in the cervical lymph nodes, and cervical lymph node dissection was performed. She was diagnosed with type 2 diabetes in 2013 at the age of 57 years with an HbA1c level of 10.2% and was administered a diabetic drug. However, her glycemic control improved in 2018, and the diabetes medication was discontinued.
Case 4 involved a 55-year-old man who underwent total thyroidectomy in 2014 for papillary thyroid carcinoma and two sessions of RAI therapy. As his serum Tg levels were measurable with no evidence of structural disease, he was diagnosed with BPD. At the time of surgery, he was also diagnosed with type 2 diabetes, with an HbA1c level of 8.9%; however, his glycemic control improved with dietary therapy. In 2020, the patient was transferred to another hospital, leading to an interruption in his follow-up.
Case 5 involved a 59-year-old woman who underwent a total thyroidectomy in 2017 for papillary thyroid carcinoma. In 2018, the patient underwent cervical lymph node dissection for recurrence. Subsequently, she received two sessions of RAI therapy. As her serum Tg levels were measurable with no evidence of structural disease, she was diagnosed with BPD. She was diagnosed with type 2 diabetes in 2009 at the age of 51 years and was administered diabetic drugs. In 2022, she died of cervical cancer.
Case 6 involved a 66-year-old woman who underwent total thyroidectomy for papillary thyroid carcinoma in 2013 and received two sessions of RAI therapy. Her serum Tg level was measurable with no evidence of structural disease; thus, she was diagnosed with BPD. In 2019, at the age of 72 years, she was diagnosed with type 2 diabetes with an HbA1c level of 7.0% and was administered a diabetic drug. Table 1 shows the patient characteristics at the start of the survey period.
| Patient No. | |||||||
|---|---|---|---|---|---|---|---|
| A | 1 | 2 | 3 | 4 | 5 | 6 | |
| Sex | F | M | M | F | M | F | F |
| Age (Y) | 39 | 51 | 52 | 63 | 57 | 61 | 72 |
| Age at surgery (Y) | 25 | 49 | 51 | 60 | 55 | 59 | 66 |
| Pathology of thyroid cancer | PTC | PTC | PTC | PTC | PTC | PTC | PTC |
| Body mass index (kg/m2) | 29.7 | 31.4 | 25.5 | 29.0 | 25.0 | 30.6 | 24.0 |
| eGFR (mL/min/1.73m2) | 98 | 89 | 79 | 54 | 67 | 73 | 73 |
| Metastatic sites | Lung | Lung, cLN | |||||
| Serum Tg level (ng/mL) | 192.7 | 4.1 | 0.7 | 36 | 2.5 | 10.5 | 2.4 |
| Cumulative RAI dose (mCi) | 1,100 | 200 | 100 | 40 | 200 | 200 | 200 |
| Type of diabetes | Type 2 | Type 2 | Type 1 | Type 2 | Type 2 | Type 2 | Type 2 |
| Duration of diabetes (Y) | 0 | 2 | 15 | 6 | 2 | 10 | 0 |
| HbA1c (%) | 10.8 | 8.3 | 7.4 | 6.0 | 6.1 | 6.7 | 7.0 |
F, female; M, male; PTC, papillary thyroid carcinoma; eGFR, estimated glomerular filtration rate; cLN, cervical lymph nodes; Tg, thyroglobulin; RAI, radioactive iodine; HbA1c, hemoglobin A1c
Serum Tg levels were measured using an electrochemiluminescence immunoassay kit (ECLusys Tg II kit; Roche Diagnostics, Basel, Switzerland).
Calculation of the Tg-doubling rate (Tg-DR)The Tg-DR is an index that indicates the proliferation rate of thyroid cancer cells [3]. Positive Tg-DR values indicate an increase in serum Tg levels, while negative values indicate a decrease; larger values indicate a faster increase. The Tg-DR was calculated using the Doubling Time, Doubling Rate & Progression Calculator (available at https://www.kuma-h.or.jp/kumapedia/kuma-medical/detail/?id=290).
Ethics approval and consent to participateThis study was approved by the Research Ethics Committee of Kobe City Medical Center General Hospital (approval no. zn231107) and was conducted in accordance with the principles of the Declaration of Helsinki. The requirement for informed consent was waived by the Research Ethics Committee of Kobe City Medical Center General Hospital due to the retrospective nature of this study.
Statistical analysesPearson’s correlation tests were used to analyze the association between serum Tg and HbA1c levels. Statistical significance was set at p < 0.05. Statistical analyses were performed using the Statistical Package for the Social Sciences version 27.0 (IBM SPSS 27.0; IBM Corp., Armonk, NY, USA).
We present the time course of serum Tg and HbA1c levels for each case in Fig. 1, along with scatter plots for the survey period. Significant correlations were observed in cases A (r = 0.92, p < 0.01), 1 (r = 0.53, p < 0.01), and 2 (r = 0.66, p = 0.01). Although there appeared to be a correlation in cases 3 (r = 0.72, p = 0.11) and 4 (r = 0.54, p = 0.17), no significant differences were detected. There was no correlation in cases 5 (r = 0.10, p = 0.81) or 6 (r = 0.19, p = 0.47). To elucidate the relationship between serum Tg and HbA1c levels across all patients, we present scatter plots with regression lines depicting the percentage changes in serum Tg levels (%Tg) and the changes in HbA1c levels (ΔHbA1c) from the baseline of the survey period for each patient (Fig. 2). The %Tg was used instead of absolute Tg values because baseline Tg levels varied by more than 100-fold.
In the upper panel, red circles represent serum Tg levels, while blue diamonds represent HbA1c levels. The lower panel illustrates the time course of serum TSH levels. Only serum Tg levels recorded when TSH levels were <0.5 mIU/L are presented.
Tg, thyroglobulin; HbA1c, hemoglobin A1c; Ope, operation; RAI, radioiodine; MET, metformin; DPP4i, dipeptidyl peptidase-4 inhibitor; SGLT2i, sodium-glucose cotransporter-2 inhibitor; SU, sulfonylurea; GLP1 RA, glucagon-like peptide-1 receptor agonist
Scatter plots with regression lines depict the percentage changes in serum Tg levels (%Tg) and the changes in HbA1c levels (ΔHbA1c) from the baseline of the survey period for each patient.
Tg, thyroglobulin; HbA1c, hemoglobin A1c
Table 2 presents the clinical characteristics of participants during the survey period. The survey duration ranged from 22 to 75 months. In case A, the patient exhibited lung metastases and showed no change in tumor size during the survey period, while in case 3, the tumor size gradually increased. In case 6, the patient developed new lung metastases during the study period. In cases A, 1, and 2, the HbA1c levels were 10.8%, 8.3%, and 7.4%, respectively, at the baseline of the survey period and improved by 4.6%, 1.3%, and 0.4%, respectively, during the survey period; the Tg levels decreased by 71.9%, 78.0%, and 71.4%, respectively. Patients in cases 3 and 4 exhibited high Tg-DR levels, indicating a rapid increase in serum Tg levels. Additionally, the average HbA1c levels were 6.0% and 6.2%, respectively, suggesting good glycemic control with small fluctuations. Among the two cases with no observed correlation, the patient in case 5 exhibited small fluctuations in serum Tg levels, whereas the patient in case 6 showed small fluctuations in both serum Tg and HbA1c levels.
| Patient No. | |||||||
|---|---|---|---|---|---|---|---|
| A | 1 | 2 | 3 | 4 | 5 | 6 | |
| Correlation coefficient | 0.92 | 0.53 | 0.66 | 0.72 | 0.54 | 0.1 | 0.19 |
| p | <0.01 | <0.01 | 0.01 | 0.11 | 0.17 | 0.81 | 0.47 |
| Survey period (M) | 75 | 65 | 39 | 43 | 50 | 22 | 51 |
| Mean TSH level (mIU/L) | 0.020 | 0.023 | 0.165 | 0.007 | 0.004 | 0.023 | 0.072 |
| Final disease outcome | Lung (SD) | BPD | BPD | Lung (PD) | BPD | BPD | Lung (new lesion) |
| Tg-DR (/year) | –0.13 | –0.50 | –0.54 | 0.53 | 1.01 | –0.05 | 0.17 |
| %Tg (%) | –71.9 | –78.0 | –71.4 | 297.2 | 300.0 | –2.9 | 37.5 |
| Mean HbA1c (%) | 7.4 | 7.2 | 7.2 | 6 | 6.2 | 8.4 | 6.9 |
| ΔHbA1c (%) | –4.6 | –1.3 | –0.4 | 0.3 | 0.2 | 1.6 | –0.6 |
| Diabetic drugs | |||||||
| Metformin | 1 | 1 | 1 | ||||
| DPP4i | 1 | 1 | 1 | ||||
| GLP1 RA | 1 | ||||||
| SGLT2i | 1 | 1 | 1 | ||||
| Sulfonylurea | 1 | ||||||
| Insulin | 1 | ||||||
BPD, biochemically persistent disease; SD, stable disease; PD, progressive disease; Tg, thyroglobulin; Tg-DR, thyroglobulin-doubling rate; HbA1c, hemoglobin A1c; DPP4i, dipeptidyl peptidase-4 inhibitor; SGLT2i, sodium-glucose cotransporter-2 inhibitor; GLP1 RA, glucagon-like peptide-1 receptor agonist
Despite the limited sample size in this study, our findings suggest a correlation between serum Tg and HbA1c levels in some patients. This relationship was more evident in patients with elevated HbA1c levels at the baseline of the survey period and subsequently exhibited fluctuations. Conversely, this relationship was less pronounced in patients with low and relatively stable HbA1c levels. A similar trend was observed with Tg levels, where the correlation was more evident in patients with gradual fluctuations. In contrast, this relationship was less clear in patients with either excessive or minimal fluctuation in Tg levels. As most patients with differentiated thyroid cancer experience indolent disease progression, this observation implies that changes in glycemic control should be considered when interpreting serum Tg levels in patients with both thyroid cancer and diabetes.
It is well known that glycemic control influences the levels of tumor markers, such as carcinoembryonic antigen (CEA) and carbohydrate antigen 19-9 (CA19-9) [4, 5]. The mechanism underlying the association between these tumor markers and glycemic control likely involves increased synthesis of glycoproteins and carbohydrate chain antigens associated with hyperglycemia or decreased catabolism due to glycation [4, 5]. However, the mechanism linking serum Tg levels and glycemic control remains unclear. Tg is a glycoprotein synthesized in thyroid follicular cells, and its mechanism of association may be similar to that of CEA and CA19-9. Furthermore, thyroid cancer is associated with obesity and insulin resistance [6, 7]. In this study, all participants had a baseline overweight; however, an association could not be evaluated between serum Tg levels and factors influencing glycemic control, including body weight and insulin levels. Future studies with a larger cohort are necessary to clarify the underlying mechanism by which glycemic control influences Tg levels.
In case A, the possibility that metformin is involved in the decrease in Tg levels was suggested [2]. Similar to case A, in cases 1 and 5—wherein the patients were taking metformin—a decrease in serum Tg levels was observed during the survey period. These results align with a prospective interventional study involving patients with RAI-refractory thyroid cancer wherein all five patients treated with metformin showed decreased serum Tg levels [8]. Furthermore, in case A, a decrease in Tg was observed alongside a decrease in HbA1c [2]. Similarly, in cases 1 and 2—which demonstrated a significant correlation between serum Tg and HbA1c levels—a decrease in serum Tg levels was observed with improved glycemic control. Cancer cells undergo significant changes in their glucose metabolic programs and depend on glycolysis, even under aerobic conditions (commonly called the Warburg effect) [9]. Therefore, cancer cells require substantial amounts of glucose. Metformin mimics calorie restriction at both the cellular and systemic levels, making it a rational choice as an anticancer agent. Most RAI-refractory thyroid cancers progress slowly and require long-term treatment, and these patients desire cost-effective treatment options with minimal adverse effects. However, there are currently no effective treatments other than TSH suppression therapy. Targeting glucose metabolism may, therefore, be an important therapeutic option for these patients.
This study had several limitations. First, the small sample size restricts the generalizability of the findings. Thyroid cancer and diabetes are often managed separately, and factors such as treatment interventions and TSH levels can influence serum Tg levels, resulting in a limited number of patients and observation points. Second, the potential influence of TSH could not be entirely excluded. To minimize the influence of TSH, we included only patients with TSH levels <0.5. However, we acknowledge that TSH may still influence serum Tg levels. Third, as a retrospective observational study, we could not fully address biases and confounding factors, including the influence of previous treatments, tumor characteristics, effects of antidiabetic medications, and changes in body weight. Despite these limitations, the findings of this study still provide valuable insights for future research on thyroid cancer.
In conclusion, serum Tg levels may be associated with HbA1c levels in some patients with thyroid cancer and diabetes who have undergone total thyroidectomy and RAI therapy (Graphical Abstract). The correlation between serum Tg and HbA1c levels was more evident in patients with gradual fluctuations in Tg levels. Future studies involving larger numbers of patients may lead to the establishment of appropriate monitoring methods and effective treatment strategies.
We would like to thank Editage (www.editage.com) for the English language editing.
None of the authors have any potential conflicts of interest associated with this research.
