Trabecular bone deterioration in differentiated thyroid cancer: Impact of long‐term TSH suppressive therapy

Abstract Background Conflicting results has been reported regard osteoporosis and fractures in patients with Differentiated Thyroid Cancer (DTC). Our objective was to evaluate the long‐term effects of TSH suppression therapy with Levothyroxine (LT4) on trabecular bone score (TBS) and bone mineral density (BMD) in females with DTC after thyroidectomy. Methods About 145 women with resected DTC and receiving long‐term TSH therapy, were stratified according to the degree of TSH suppression. Mean duration of follow‐up was 12.3 ± 6.1 years. BMD and TBS, were assessed using dual‐energy X‐ray absorptiometry (DXA) and TBS iNsight (Med‐Imaps), at baseline (1‐3 months after surgery) and at the final study visit. Results In patients stratified by duration of TSH suppression therapy (Group I, 5‐10 years; Group II, >10 years), slight increases from baseline TSH levels were observed. Significant decreases in LS‐BMD and FN‐BMD were seen in patients after >10 years. TBS values were lower in Groups I (1.289 ± 0.122) and II (1.259 ± 0.129) compared with baseline values (P = .0001, both groups). Regarding the degree of TSH suppression, TBS was significantly reduced in those with TSH < 0.1 µU/mL (P = .0086), and not in patients with TSH suppression of 0.1.‐0.5 or >0.5 µU/mL. Conclusions We found deterioration of trabecular structure in patients with DTC and TSH suppression therapy below 0.1 µU/mL and after 5‐10 years of follow‐up. Significant changes in BMD according to TSH levels were not observed. Trabecular Bone Score is a useful technique for identifying thyroid cancer patients with risk of bone deterioration.


| INTRODUCTION
The incidence of thyroid cancer has been described as increasing worldwide in the last decades. 1 The mainstay of treatment of patients with differentiated thyroid cancer (DTC) is surgery. Thereafter, TSH suppression with levothyroxine (LT4) is recommended as a main therapeutic option for patients with DTC, to prevent tumor recurrence and increase survival. 2 American Thyroid Association (ATA) 2016 guidelines for the use of thyroid hormone therapy in DCT recommend TSH level targets based on a patient's risk of recurrence: 0.1 µU/mL for high-risk patients; 0.1-0.5 µU/L for intermediate-risk patients; and 0.5-2 µU/mL for low-risk patients, who have undergone remnant ablation and have undetectable serum thyroglobulin levels. 3 Although meta-analysis has confirmed that patients with DTC and TSH suppression showed a significant reduced risk of disease progression, recurrence, and death (relative risk [RR] = 0.73, 95% CI = 0.60-0.88, P < .05), 4 recent studies have found no significant benefit t4regarding disease-specific or disease-free survival, in DTC patients with undetectable serum TSH levels versus subnormal TSH levels, when also taking in consideration the degree of low-risk of thyroid cancer. 5,6 Further, because DTC usually is an indolent tumor, and its mortality rate is very low, TSH suppression with LT4 can be controversial, given that treatment can induce a state of iatrogenic subclinical hyperthyroidism, which can be associated with bone and cardiovascular adverse effects.
There is controversy about the association between fracture risk and subclinical hyperthyroidism, either of endogenous cause or related to thyroxine treatment. Patients with LT4 treatment and TSH suppression levels, followed for 8 years, were not associated to any increase fractured rate; however, there was an increased risk of ischemic heart disease. 7 In the first Cardiovascular Health Study (CHS) no association between endogenous subclinical hyperthyroidism and hip fracture in women was found. 8 Data from the CHS has been recently enlarged, confirming no association between endogenous subclinical hyperthyroidism and an increased risk of hip fracture or lower BMD at the spine or hip in a study with 5888 elderly subjects. 9 In the contrary, an observational study showed that patients taking thyroid medication, with high TSH and those with a suppressed TSH, were both at increased risk of fracture, suggesting that hyperthyroxinemia together with suppressed TSH could detrimental to bone. 10 A meta-analysis of 13 prospective cohorts' studies showed that endogenous subclinical hyperthyroidism was associated with HRs of 1.36 (95% CI, 1.13-1.64) for hip fracture, and 1.28 (95% CI, 1.06-1.53) for any fracture, while in the comparison between participants treated with LT4 versus untreated patients, therapy with LT4 was not associated with any fracture outcomes 0.98 (95% CI, 0.82-1.17). 11 A recently meta-analysis with 24 studies, confirm this results, and extends this effect on fracture risk at various sites and to lower distal and ultradistal BMD. 12 In particular, long-term TSH suppression therapy has been called into question due to its association with increased bone loss and fracture incidence. 13 Recently, in a large study, compared with controls, osteoporosis, but not fractures was more frequent in patients with thyroid cancer (OR 1.33; 95%CI 1.18-1.49). 14 Although several studies have not found an association between bone loss and TSH suppressive therapy in men and premenopausal women with DTC, a substantial number of other studies in postmenopausal women with DTC have identified this adverse effect. 15 The discrepancy in these studies could be explained by the heterogeneity of the patient populations included, the degree and duration of TSH suppression, or the methods used to measure bone density and quality. Also, because the number of thyroid cancer survivors is growing, more patients may experience the long-term effect of TSH suppression therapy, and is mandatory to develop methods for the accurately identification of patients at risk of osteoporosis and fractures.
Dual-energy X-ray absorptiometry (DXA) for measurement of bone mineral density (BMD), can be now performed together with Trabecular Bone Scores (TBS) analysis. TBS is a gray-level texture measure derived from lumbar spine DXA images. It is an indirect measure of trabecular microarchitecture, which gives additional information regarding bone quality, that can be useful for patients with risk factors for bone loss, including those under TSH suppression therapy. 10 Longitudinal studies have shown that TBS predicts fracture risk in women, even after adjusting for BMD. 16 The use of both measurements, TBS plus BMD, also improves fracture discrimination. 17 The aim of this study was to use TBS and DXA to assess the effect of the degree and duration of long-term TSH suppression therapy, on bone microarchitecture and BMD in female patients with DTC after total thyroidectomy who were treated in our Thyroid Cancer Unit.

| Data and study population
In this study, inclusion criteria were women with DTC who received total thyroidectomy with 131 I ablation, when necessary, and who received long-term TSH suppressive therapy with LT4 according to guidelines. 2 Patients were required to initiate TSH suppression immediately after surgery, have blood extraction and a DXA scan 1-3 months after surgery, and have a follow-up period of ≥5 years. The exclusion criteria were as follow: (a) the use of medications that might affect bone metabolism including estrogen/progestin, glucocorticoids, bisphosphonate, calcitonin, selective estrogen receptor modulators, denosumab, teriparatide, and lithium; (b) malabsorption syndrome; (c) diuretics; (d) diseases affecting bone metabolism (eg, Paget's bone disease, renal osteodystrophy), malignant neoplasms, hyperparathyroidism, primary and postsurgical hypoparathyroidism and hyperthyroidism. Patients without a complete set of data were also excluded. Men were not included in the analysis.
To study the effect of the long-term TSH suppression therapy, the included cohort was stratified at the final visit according to the degree of suppression of TSH: suppressed (<0.1 µU/mL), moderately suppressed (0.1-0.5 µU/mL) and nonsuppressed (>0.5 µU/mL). Patients were also surveyed to assess risk factors for low bone mass, as smoking, daily calcium intake, and physical activity. Other clinical data were retrieved using the information collected in patient files. Body mass index (BMI) was calculated as the weight in kilograms divided by the height in meters squared. Time of follow-up was calculated from the start of TSH suppression therapy, which was initiated immediately after total thyroidectomy, to the final study visit. Diagnosis, surgery, and follow-up of all patients occurred at the Thyroid Cancer Unit of our Hospital. Ethical approval for this study was obtained from our Institutional Review Board prior to beginning this study. Informed consent was obtained from all enrolled patients.

| Biochemical analysis
Blood extraction were obtained 1-3 months postoperatively and at the final study visit. Serum samples for biochemical analyses were obtained between 8 and 9 am after overnight fast and immediately kept frozen at −70°C until they were measured by auto analyzer (Modular P800 Chemistry Analyzer, Roche Diagnostic). Serum levels of creatinine, calcium (corrected for albumin binding), and phosphate were measured. Serum TSH (Architect TSH reagent; Abbot Laboratories) and free (T4) by electrochemiluminescence (ElecsysT4, Roche Diagnostic; functional sensitivity <0.01 µg/mL).
BMD values are expressed as grams per square centimeter (g/ cm 2 ) which is expressed from the expected peak young-adult mean BMD for the T scores. According to the WHO criteria, patients were classified as osteoporotic (T score equal or worse than −2.5), osteopenic (T score −1≥ and >−2.5), and normal (T score > −1). 18 Reference data corresponding to the Spanish population were obtained from a multicenter study with 2442 healthy subjects, aged 20-80 years. 19 TBS measurements were performed applying the TBS iNsight2.0 software (Med-Imaps, Geneva, Switzerland) to the LS DXA exams. Lumbar TBS was calculated as the mean value of individual measurements for vertebrae L1-L4. Weight and height of each patient are entered in the software program in each visit corresponding to the TBS study.

| Statistical analysis
All data were analyzed using SAS statistical package (version 9.3; SAS Institute). Continuous variables were expressed as the mean ± standard deviation (SD). Normality of data was confirm using the Kolmogorov-Smirnov test. Category variables were expressed by their absolute and relative percentage, and analyzed using contingence tables and Chi-square or Fischer test. The Wilcoxon nonparametric test or Kruskal-Wallis nonparametric test were used for the analysis of more than two parameters in the transversal study, and the Student´s t test for the longitudinal study. The Pearson test was used to evaluate the relationship between bone parameters and duration and TSH suppression level. Fisher's exact test was used to study the significance of the association observed in the categorical data. Multiple lineal regressions was performed to evaluate the dependence and influence between TBS and other variables. All analyzes were adjusted for follow-up time. A level of α = .05 was considered significant in all statistical procedures. The Bonferroni test was used in the correction of multiple comparison tests.

| Description of sample
A total of 145 Caucasian women (131 postmenopausaland 14 premenopausal) with DTC were included in this study. Clinical and bone densitometry data at baseline and at the final study visit are shown in Table 1. The mean follow-up with LT4 suppression therapy after total thyroidectomy was 12 years, with a range of 7-20 years, and the mean age of women at the end of the study was 64 ± 10.6 years. Mean BMI was higher in patients at the end of the study (28.45 ± 5.3 kg/ m 2 ), compared to baseline (27.27 ± 0.6 kg/m 2 ; P < .0001). At the end of the study, there was a significant decrease from baseline in prescribed LT4 doses from 2.29 ± 0.6 µg/kg to 1.70 ± 0.4 µg/kg (P = .0417) and a significant increase in TSH levels from 0.23 ± 0.4 µU/mL to 0.89 ± 0.1 µU/mL (P < .0001).
No significant changes were observed in areal BMD or T score at all skeletal sites. In contrast, TBS decreased from 1.346 ± 0.136 (normal range) to 1.273 ± 0.136 (partially degraded) (P < .0001). Estimated mean dietary calcium intake of DTC patients (575.94 ± 282 mg/d), was only collected at final study visit, and no differences were observed among patients with normal BMD (561 ± 195 mg/d), osteopenia (572 ± 301 mg/d), and osteoporosis (588 ± 294 mg/d). Table 2 shows the results of the final visit stratified by the number of years of follow-up: Group I (n = 69) includes patients with a follow-up duration of 5-10 years and Group II (n = 76), includes patients with >10 years of follow-up. As expected, there were also significant changes in BMI at the end of the study in both groups, compared to baseline (Group I, end of study 28.67 ± 5.9 kg/m 2 vs 27.77 ± 5.6 kg/ m 2 , P = .0042; Group II, end of study 28.41 ± 4.8 kg/m 2 vs 26.87 ± 4.0 kg/m 2 , P = .0009). Both groups also had increases in TSH levels (from 0.17 ± 0.36 µU/ml to 0.76 ± 1.6 µU/ mL in Group 1 [P = .0027], and from 0.27 ± 0.5 µU/mL to 1.12 ± 1.85 in Group II [P = .0002]). There was also a significant decrease in mean LT4 doses in both groups (Group I from 2.10 ± 0.5 µg/kg to 1.71 ± 0.3 µg/kg, P < .0001; Group II from 2.47 ± 0.62 µg/kg to 1.69 ± 0.50 µg/kg, P < .0001).

| Stratification according to levels of TSH suppression
Clinical, thyroid hormone and bone parameters stratified by the level of TSH suppression at the final visit, are shown in Table 3. Although patients were treated in our Thyroid Cancer Unit by the same physician during follow-up, there were changes in the degree of TSH suppression from the beginning to the end of the study, either due to intentional decision or from factors including age or BMI.
During the course of follow-up, there was a reduction in the number of patients with suppressed TSH (<0.1 µU/mL) and moderately suppressed TSH (0.1-0.5 µU/mL), while T A B L E 2 Study of patients with differential thyroid carcinoma according to duration of follow-up

Years follow-up (n) 5-10 years Group I (n = 69) >(10 years Group II (n = 76)
Period of study  Figure 2). In the final visit the number of patients from the Group with TSH suppression <0.1 µU/mL, that continued with this suppressed values, was reduced to 26. This subgroup had low TBS values compared to initial suppressed group values (n = 75) (1.278 ± 0.130 vs, 1.362 ± 0.112, P = .0056). Taking in consideration the subgroup of 49 patients that change from total suppression, to moderate or no suppression, they also low values of TBS (1.272 ± 0.103) at the final visit study, as well as those 70 that continued to maintain the initial moderate or no suppression, with low TBS values (1.26 ± 0.14, P = .0062). According to the grade of risk of TBS values in the initial study there were: 23 patients (15.9%) with TBS < 1.23 (degraded); 45 (31%) with a TBS score between 1.23 and 1.35 (partial degraded), and 77 (53.1%) with TBS > 1.35 (normal); and the final study there were 35 patients (24.1%) with TBS < 1.23(degraded), 72 (49.65%) with a TBS score between 1.23 and 1.35 (partial degraded), and 38 (26.2%) with TBS > 1.35 (normal). At the final study, there was an important increase in the number of patients with degraded and partially degraded microarchitecture, while there was a decrease in more than a half, in patients with normal TBS scores and normal LS-BMD T scores (42% vs 13.3%)(Fisher´s exact test < 0.0001).
In the multivariate analysis, clinical parameters associated with TBS as a dependent variable were: levels of TSH suppression below 0.1 µU/mL (B = 0.2542, P < .0280); between 0.1 and 0.5 µU/mL (B = 0.2271, P = .0513), and >0.5 µU/mL (B = 0.2327, P = .0427). Adjustments for BMD were made to show that the association between TBS and duration of TSH suppression was independent from BMD.

| DISCUSSION
Our study provide information that total thyroidectomized female patients due to DTC who received long-term TSH suppressive therapy had lower vertebral TBS, after both 5-10 years and >10 years of follow-up. In contrast, LS-BMD F I G U R E 2 TBS scores of patients with DTC according to the degree of TSH suppression. Upper and lower lines indicates limits of normal, partially degraded and degrades TBS scores values