Introduction

Graves’ disease (GD) is the most common cause of thyrotoxicosis [1]. Thyrotoxicosis leads to increased bone turnover, loss of bone mineral density (BMD), and an increased fracture risk in the first years following diagnosis [2, 3]. BMD improves with establishment of euthyroidism [4,5,6], but studies have reported conflicting results on the restitution of bone density compared to controls [7, 8]. A recent study on patients with previous Graves’ thyrotoxicosis found a similar risk of fracture but a higher risk of a diagnosis of osteoporosis compared to controls [9]. Evaluation of bone microarchitecture adds important information on bone quality and fracture risk compared to osteodensitometry [10]. High resolution peripheral quantitative computerized tomography (HRpQCT) scanning is a non-invasive procedure that provides information on microarchitecture and estimated bone strength [11]. Knowledge on the microarchitectural changes in bone during thyrotoxicosis and its treatment is sparse. Only one study investigated bone changes in GD using this technique, showing changes in the cortical compartment of non-weightbearing extremities [12].

Vitamin D is crucial to skeletal health and aiming at a replete vitamin D status by supplementation with vitamin D is standard of care in conditions with a low BMD. In healthy adults, vitamin D supplementation does not improve BMD [13]. However, the effect of vitamin D in patients recovering from a high bone turnover is unknown. A recent longitudinal study demonstrated a decline in the vitamin D level during treatment of GD thyrotoxicosis [14]. The benefit of vitamin D supplementation on bone convalescence in GD is unknown. We previously demonstrated an unexpected adverse effect of vitamin D supplementation on muscle strength recovery in GD [15].

Purpose

In a randomized clinical trial, we tested whether supplementation with vitamin D versus placebo affects bone density and quality, as assessed by areal and volumetric BMD, microarchitecture and estimated bone strength, in Graves’ disease thyrotoxicosis. Furthermore, we evaluated the effect of antithyroid drugs (ATD) on bone.

Methods

The DAGMAR study was an investigator-initiated double-blinded clinical trial, randomizing 278 patients with GD thyrotoxicosis to a daily supplement of cholecalciferol (vitamin D3) 70 μg or placebo as add-on to ATD treatment. Findings on the primary endpoint of the study have previously been reported showing no beneficial effect on the risk of relapse [16].

Study population

Here, we report the effect on bone indices in a subgroup (n = 86) of participants in the DAGMAR trial [16] who completed an extensive examination program on bone indices. From this subgroup, we have previously reported the effect of vitamin D on arterial stiffness (no effect) and muscle strength (adverse effect) [15, 17]. Participants were recruited from endocrine outpatient clinics in Central Denmark Region between March 2015 and December 2017. Follow-up visits were completed by September 2018. All investigations were performed at the same study site.

Study participants were adults (18–80 years) with newly diagnosed GD’s thyrotoxicosis defined as a suppressed level of thyrotropin (TSH) with levels of total or free triiodothyronine (T3) and/or thyroxine (T4) above the upper limit of the reference range, and a positive thyroid stimulating hormone receptor antibody (TRAb). Furthermore, participants were supposed to have planned or initiated antithyroid treatment.

Exclusion criteria were intake of vitamin D supplement (ergo- or cholecalciferol) of more than 10 μg/day that could not be discontinued, a previous diagnosis of hyperthyroidism, planned treatment with radioactive iodine or thyroid surgery, hypercalcemia (ionized plasma calcium > 1.40 mmol/l), reduced kidney function (eGFR < 45 ml/min), treatment with immune-modulating drugs, malignant disease, or pregnancy. A questionnaire on dietary habits was used to assess the daily calcium intake by the formular: 350 mg + (glasses of milk × 150 mg) + (cheese × 200 mg) + (yogurt × 250 mg).

Study intervention

The trial was performed with parallel groups and an equal allocation to vitamin D3 70 μg (2800 IU/day) or similar placebo provided as one daily tablet at mealtime. The intervention was given as add-on to standard of care with ATD. Details on treatment allocation has previously been reported [15].

Dual energy X-ray absorptiometry (DXA)

At baseline and at 9 months, we performed DXA scans (using a Hologic Discovery, Hologic, Inc., Waltham, MA, USA) at the lumbar spine (L1- L4), the hip, the distal right forearm (or the opposite in the case of previous fracture) and whole body. We report areal BMD (aBMD [g/cm2]) and Z-scores. From the lumbar spine scans, a trabecular bone score (TBS) was extracted using TBS iNsight software, version 2.1.0.0 (Medimaps, Mérignac France).

High-resolution peripheral quantitative computed tomography (HRpQCT)

At baseline and 9 months, HRpQCT scans, using a first generation Xtreme CT scanner (Scanco Medical AG, Brüttisellen, Switzerland), were performed at the radius (right hand) and tibia (right side) estimating bone microarchitecture, volumetric BMD (vBMD) (mgHA/cm3), geometry, and strength. The manufacturer’s standard operating procedures were followed including daily and weekly phantom quality scans. The scans started 9.5 and 22.5 mm proximately from the endplate of the radius and tibia, respectively, and proceeded 9.02 mm in the proximal direction comprising 110 slices. At the baseline scan, a scout-view with the reference-line set, defining the starting point of the scan, was saved for comparison and manually positioning the region of interest at follow-up measurement. During the scanning, the limb was immobilized in a carbon cast to avoid movement artefacts. Images were assessed and rated immediately for quality and artifacts, and we accepted images graded between 1 and 3 (good) whereas images graded 4 or 5 were rescanned. The separation of the trabecular and the cortical areas was automated using evaluation software, but manually checked and corrected if needed. Using standard analytic tools, the trabecular structures were assessed as trabecular number, trabecular thickness, trabecular separation, and trabecular volume to total volume. The cortical compartment was evaluated by cortical thickness and cortical porosity. Finite element analysis (Scanco FEA software version 1.13) was performed to estimate bone strength.

Biochemistry

Blood samples were drawn in the fasting condition at baseline, three and 9 months. The bone turnover markers C-terminal telopeptide of type I collagen (CTX), procollagen I N-terminal peptide (P1NP), and osteocalcin were measured by electrochemiluminescence immunoassay (Cobas 8000 E; Roche Diagnostics GmbH) and the intra-assay CV was 10% (CTX), 8.8% (P1NP) and 6% (osteocalcin) for all levels. Bone-specific alkaline phosphatase was determined by immunoassay (METRA BAP EIA kit; Quidel Corporation) with an intra-assay CV at 20% at all levels. Total plasma level of 25 hydroxy vitamin D (25(OH)D) (a combined measurement of 25(OH)D2 + 25(OH)D3) and calcitriol was analyzed by isotope dilution liquid chromatography tandem mass spectrometry. TRAb, TSH, free T3 and free T4 were measured by electrochemiluminescence immunoassays using the Cobas 6000 analyser (Roche, Denmark). Analyses of 25(OH)D, parathyroid hormone (PTH), bone turnover markers, TRAb, and free T3 and free T4 were performed in batches.

Outcomes measures

The primary outcome was the between-group difference in lumbar spine aBMD at 9 months. Secondary outcomes were between-group differences of aBMD at the hip and forearm, tibia- and radius vBMD, cortical porosity, estimated failure load, indices of microarchitecture, and differences in the level of bone turnover markers (BTM). We also analyzed the changes in all outcomes in response to 9 months of ATD treatment. Finally, we explored the association between changes in BMT and thyroid hormones, and the development in areal and volumetric BMD and cortical porosity with the changes in thyroid hormones or BTM. All analyses were tested for interaction with baseline vitamin D status (insufficient (plasma 25(OH)D < 50 nmol/l) vs. replete).

Statistical analyses

Sample size was calculated with the intent to detect a clinically relevant difference of a lumbar aBMD of 2%. Based on the assumption of a mean lumbar spine aBMD of 0.920 g/cm2 with SD 0.03 g/cm2, we calculated a sample size of 36 participants in each group. To allow for dropouts, 86 patients were included.

The intention to treat principle was applied in all analyses. Comparison of randomization groups was performed using t-test or Wilcoxon ranksum test, as appropriate. Linear mixed modelling was used in all analysis of outcome with time and randomization group and their interaction as fixed factors and participants as random effects. The unequal standard errors and correlations in the two groups were taken into account in the analysis, as appropriate. Model validation was performed by inspection of QQ-plots of the residuals. The between-group differences were assessed by comparing the difference in change from baseline to 9 months between groups. When analyzing the effect of ATD treatment, linear mixed modelling was applied as described above, the interaction term of time and group was kept in the model, and the main effect of time is reported. In the analysis of associations between changes in bone indices and BTM and thyroid hormones and between BTM and thyroid hormones, we assessed the change in the dependent variable per standard deviation (SD) increase in the independent variable which was included as a covariate in the analysis. We used the SD on the mean change in the independent variable, e.g., T3 from baseline to 9 months. In all analyses using mixed modelling, outcomes were logarithmically transformed to obtain relative effect sizes and the results are presented as percentage with a 95% confidence interval (95%CI). We explored any interaction between the intervention and smoking status (current smoker or not), menopausal status, and sex, and we analyzed whether the effect of ATD treatment (effect of time) differed according to these grouping. In sensitivity analyses of the effect of vitamin D, we excluded participants with a suppressed TSH level at 9 months.

We analyzed data using STATA version 15 (STATA Corp LP, College Station, TX, USA). The statistical significance level was set at 0.05.

Ethical considerations

The trial was approved by the regional committee on biomedical Research Ethics and all patients gave written informed consent. The trial was prospectively registered (Clinicaltrials.gov identifier: NCT02384668) and trial reporting is in adherence with the CONSORT guidelines [18]. Orkla Health a/s provided study medication free of charge but had no part in planning, performing, interpreting, or reporting of the trial.

Results

Out of 348 eligible GD patients, 86 were included of which 44 received vitamin D supplementation and 42 received placebo. A flow chart of in- and exclusion has previously been published [17]. In the vitamin D group, 43 and 37 completed three and 9 months of follow-up and the corresponding numbers in the placebo group were 39 and 37, respectively. No adverse events occurred. All participants received the intended treatment and were included in the analyses.

Randomization was balanced (Table 1). Participants were (mean ± SD) 41 ± 14 years old and 86% were females. There was no between-group difference in baseline 25(OH)D level (67 ± 30 nmol/l) and no difference in the proportion of participants with vitamin D insufficiency (< 50 nmol/l) (vitamin D vs placebo: 35 vs 31%, p = 0.70). Only six participants had a 25(OH)D level below 30 nmol/l. The participants had a sufficient daily calcium intake (median 950; interquartile range (IQR): 850 to 1100 mg/d) and a comparable proportion was postmenopausal (23%) or smokers (17%). Participants had similar BMD with a baseline total hip Z-score of − 0.2 ± 0.8.

Table 1 Baseline characteristics according to randomization
Full size table

At 9 months, a proportion of 26% (n = 19) had not fully recovered from thyrotoxicosis as they continued to have a suppressed TSH level and six (three in each group) continued to have elevated thyroid hormones.

Effect of vitamin D supplementation

Levels of 25(OH)D increased significantly in response to the intervention as the 25(OH)D level increased by 44.8 (IQR: 32.7 to 56.9) nmol/l in the vitamin D group and decreased in the placebo group by 14.3 (IQR: − 23.9 to − 4.4) nmol/l (p < 0.0001). The intervention did not affect the course of disease or the level of TRAb, T3, or T4 [16].

At the hip, spine, forearm, and whole body, aBMD did not change significantly with vitamin D supplementation (Table 2). The effects of vitamin D supplementation on bone geometry, vBMD, microarchitecture and estimated bone strength in tibia and radius are shown in Table 3. Vitamin D did not improve bone microarchitecture in any parameter. In response to vitamin D supplementation, the increase in PTH level observed with euthyroidism was attenuated with a between-group difference of − 17% (95%CI: − 31 to − 1) (Table 4 and Fig. 1). Vitamin D did not affect the change over time in any BTM (Table 4). The effect of supplemental vitamin D did not differ between the groups with a baseline insufficient (< 50 nmol/l) or replete 25(OH)D level (data not shown). Also, there was no interaction between the intervention and smoking status, menopausal status, or sex (data not shown). In sensitivity analyses, results were similar when excluding participants with persistent thyrotoxicosis (n = 19) at 9 months (data not shown).

Table 2 Areal bone density at the lumbar spine, hip and forearm in Graves’ disease hyperthyroidism in response to nine months of vitamin D supplementation and treatment with antithyroid drugs
Full size table
Table 3 HRpqCT-derived measures of bone geometry, volumetric bone density, microarchitecture, and estimated bone strength in the tibia and the radius in patients with Graves’ disease thyrotoxicosis in response to nine months of vitamin D supplementation or placebo as add-on to treatment with antithyroid drugs
Full size table
Table 4 The effects of nine months vitamin D supplementation and treatment with antithyroid drugs on bone turnover markers in Graves’ disease thyrotoxicosis
Full size table
Fig. 1
figure 1

Changes in parathyroid hormone level in response to vitamin D as add-on to nine months of antithyroid treatment of Graves’ disease. Group means with standard deviation. The change in parathyroid hormone in response to vitamin D 70 µg per day or placebo was evaluated using linear mixed modelling. # indicates a significant between-group difference from baseline to three or nine months

Full size image

Effect of antithyroid medication

Microarchitecture, volumetric bone density and estimated bone strength

Marked changes were observed in bone microarchitecture with 9 months of ATD treatment in both the cortical and the trabecular compartment (Table 3). In tibia and radius, cortical porosity decreased − 7% (95%CI: − 12 to − 2) and − 14% (95%CI: − 24 to − 3). The trabecular compartment was affected in both tibia and radius with an increase in trabecular thickness and spacing and a decrease in trabecular numbers. There was a non-significant trend for improvement of estimated bone strength in tibia (estimated failure load 2% (95%CI: 0;4), but not in the radius (Table 3). In sensitivity analyses, results were similar when excluding participants with a persistently suppressed TSH level (n = 19) at 9 months, and the changes observed in cortical and trabecular bone did not differ significantly from the euthyroid group (data not shown).

Areal bone density

Over the course of 9 months of ATD treatment, aBMD of the hip increased significantly, except at the neck region, with estimates ranging from 2 to 3% (Table 2). aBMD at the lumbar spine did not increase. Of notice, we observed a continued decrease in aBMD at the forearm over the 9 months of ATD treatment. Smoking status, menopausal status, baseline vitamin D status, or sex did not affect the changes in aBMD with ATD treatment.

Bone turnover markers

Marked changes in BTMs and calcium-phosphate homeostasis were observed with antithyroid drug treatment (Fig. 2). As expected, with decreasing thyroid hormones, we observed decreased plasma levels of calcium and phosphate as well as increased levels of PTH and calcitriol. A pronounced reduction of ionized calcium and an increased PTH level were observed already at 3 months with no further significant changes at 9 months, whereas the normalization of bone formation and resorption markers continued beyond 3 months reflecting that bone recovery from high turnover takes time (Fig. 2).

Fig. 2
figure 2

Changes in bone turnover markers and calcium-phosphate homeostasis in response to nine months of antithyroid treatment of Graves’ disease. Group mean with standard deviation. The change in bone turnover markers and calcium-phosphate homeostasis was evaluated using linear mixed modelling. There was no effect of vitamin D supplementation, except for parathyroid hormone, see Fig. 1. * indicates significant changes from baseline; # indicates significant changes from three to nine months. The dashed lines indicate the upper limit of the reference range for adults or premenopausal women. The limit of the normal range if not indicated was: parathyroid hormone 2.0–8.5 pmol/l, ionized calcium 1.18–1.32 mmol/l, phosphate 0.76–1.41 mmol/l, calcitriol 60–180 pmol/l. *p < 0.01 **p < 0.001. #p < 0.05, ###p < 0.001. CTX C-terminal telopeptide of type I collagen; P1NP procollagen I N-terminal peptide

Full size image

The changes in BTMs were associated with the changes in thyroid hormones during ATD treatment, even when adjusted for the changes in PTH level, except for bone specific alkaline phosphatase. Thus, the percentages change (95%CI) in CTX, P1NP, and osteocalcin per SD decrease in T3 was − 10% (− 5 to − 16), 8% (3 to 13), and − 7% (− 3 to − 11). This pattern was similar for T4 (data now shown) and the associations remained statistically significant when adjusting for PTH. We tested and found no association between BTM and TRAb (data now shown). As expected, bone measures showed an inverse association with the development of BTM (Table 5). Thus, aBMD, cortical vBMD, and estimated failure load and stiffness increased and cortical porosity declined when BTM levels decreased. The changes in bone microarchitecture were more strongly associated to changes in P1NP than CTX. Thus, a decline in P1NP was associated with increased total vBMD, cortical thickness, stiffness and failure load of both the radius and the tibia, whereas CTX was not. Changes in PTH were merely associated with changes in the trabecular compartment of radius, but not tibia or aBMD.

Table 5 The association of changes in areal BMD and volumetric BMD, microarchitecture and estimated bone strength with bone turnover markers in Graves’ disease hyperthyroidism during nine months of antithyroid treatment
Full size table

The change in the level of T3 was not associated with changes in bone microarchitecture in the tibia (data not shown), whereas in the radius, T3 was positively associated with trabecular vBMD and trabecular bone per tissue volume (increase pr SD increase in T3: 2.8% (95%CI: 0.7 to 2.9) p = 0.002 and 2.7% (95%CI:0.6 to 2.9), p = 0.003).

Discussion

This is the first study to investigate the effect of supplemental vitamin D on bone convalescence in GD thyrotoxicosis treated with antithyroid drugs. We found that 9 months supplementation with a daily dose of 70 µg vitamin D3 did not improve areal or volumetric BMD, bone microarchitecture or estimated bone strength in a population with an adequate dietary calcium intake. The only significant effect of vitamin D was a smaller increase in PTH level compared to placebo, whereas there was no impact on bone turnover markers. On the other hand, in response to treatment of thyrotoxicosis with ATD, we observed marked improvement in bone health with changes in the bone microarchitecture of radius and tibia especially a reduced cortical porosity. Moreover, our data shows that both the cortical and the trabecular compartments are affected and improve with treatment of thyrotoxicosis. As expected, treatment inferred a decrease in bone markers and a significant increase in hip aBMD, whereas we observed a continued decline in the aBMD of the forearm.

Vitamin D and bone

We found no other clinical studies investigating the effect of vitamin D on bone density or microarchitecture in hyperthyroidism.

After initiation of this trial, studies have questioned the bone protective effect of supplemental vitamin D on bone density unless the 25(OH)D level is < 30 nmol/L [19]. Whether a similar low threshold for a beneficial effect of vitamin D exists in hyperthyroidism is unknown. Very few of our participants had a baseline 25(OH)D level below this level (n = 6). In subanalysis, we found that the effect of supplemental vitamin D did not differ between those with a baseline replete vs insufficient (< 50 nmol/L) 25(OH)D level on any outcome, but it must be acknowledged that this subanalysis should be considered as underpowered due to the low number of patients with vitamin D insufficiency. Recently, Burt et al. [20] showed detrimental effects on bone microarchitecture with three years of high-dose vitamin D supplementation with 4000 and 10,000 IU/day. In the present study with a lower dose of 2800 IU/day and a duration of 9 months, no such adverse effects were seen, and treatment was safe regarding bone homeostasis. However, as previously reported, findings from our study did suggest an impaired muscle recovery with vitamin D supplementation [15].

During hyperthyroidism the increased bone remodeling causes an excess release of calcium to the blood stream which leads to a lower level of PTH that normalizes with achievement of euthyroidism. We observed that vitamin D reduced this expected increase in PTH level. This effect was not mediated by thyroid hormones as euthyroidism was archived in a similar manner in the placebo group [15, 16]. The effect of vitamin D on PTH level is in line with another study [20]. PTH induces bone resorption via activation of osteoclasts to tailor calcium levels in the blood, and a lower PTH level might have bone protective effect in the long run. However, in the study by Burt et al., high-dose vitamin D led to detrimental skeletal effects despite a pronounced decline in the PTH level. A recent meta-analysis showed some improvement in BTMs with vitamin D supplementation [21]. This was not the case during the first 9 months of ATD-treated Graves’ thyrotoxicosis.

Thyrotoxicosis and the bones

Our study confirms previous robust findings that hyperthyroidism affects aBMD and BTM, and that treatment of hyperthyroidism improves aBMD and lowers BTM substantially.

Thyrotoxicosis leads to accelerated bone turnover and the reduction in BMD with hyperthyroidism has been estimated to be as high as 12–20% in older studies, whereas the observed increase in BMD with restoration of euthyroidism at the hip was 6–7% in other studies [14, 22]. The observed changes in BMD with treatment of hyperthyroidism were smaller in the present study. One explanation could be differences in the duration and severity of thyrotoxicosis at diagnosis affecting the BMD loss prior to baseline investigation. The shorter follow-up period in the present study and the prolonged thyrotoxicosis in a subset of participants may also have contributed to a smaller BMD increase.

Changes in vBMD, microarchitecture, and bone strength with treatment of hyperthyroidism

Data on the impact of hyperthyroidism on bone microarchitecture is less consistent. Earlier histomorphometry studies have been divergent regarding the impact of thyrotoxicosis on the different bone compartments. Thus, Mosekilde and Melsen performed histomorphometry analysis of iliac crest biopsies before and after treatment of hyperthyroidism and demonstrated an initial cortical porosity that normalized with treatment, whereas the trabecular bone was unchanged [23]. Eriksen et al. showed predominantly trabecular impact [24]. Our data convincingly confirms that both the trabecular and the cortical compartment of the extremities are affected in GD hyperthyroidism. We found only one study of the changes in bone microarchitecture with treatment of hyperthyroidism using HRpQCT scans. In an observational study of patients with hyperthyroidism of mixed etiology (N = 62), treatment of thyrotoxicosis lead to an increase in bone strength and vBMD and a reduction of cortical porosity similar to our findings [12]. However, contrary to our data, no impact on the trabecular compartment was observed, presumably due to differences in disease severity between study populations as the majority of patients in the study by Nicolaisen et a.l had subclinical hyperthyroidism [12]. We believe the present data better illustrate the detrimental effects of thyrotoxicosis on bone and the early improvements with antithyroid treatment.

Bone turnover markers

Over the course of 9 months, T3 did not predict changes in bone indices, except changes in the trabecular compartment in the radius. Bone indices changed with BTM, and BTM were associated to changes in T3 even when adjusted for changes in PTH. The bone remodeling process is accelerated during thyrotoxicosis and as expected, we found a positive association between T3 and CTX. Bone remodeling is a linked process where bone resorption is followed by bone formation. This explains the increment in P1NP and osteocalcin with thyrotoxicosis and the decline observed with antithyroid treatment. However, in our data the association of T3 and P1NP was opposite that of CTX, suggesting that T3 does not only pace the bone resorption but may also negatively affect the bone formation process. Siddiqi et al. studied 17 patients with GD before and after 12 months of treatment and found similar results [22].

We also investigated the association of bone indices and BTM. One previous study explored the correlation of changes in bone microarchitecture with changes in BTM (CTX and P1NP) [25]. Similar to our results, Vinther et al. found that both CTX and P1NP were inversely correlated with cortical density and positively correlated with porosity of the tibia and the radius, whereas the authors did not find correlation between P1NP and failure load. The difference could be explained by milder disease in the study by Vinther et al. The marked changes in BTM reflect the impact of thyrotoxicosis on bone microarchitecture that is not fully captured by aBMD measurement.

Different impact of thyrotoxicosis at various bone sites?

We observed a continued decline in aBMD at the forearm despite treatment of thyrotoxicosis. Our observation confirms the finding of one observational study of hyperthyroid men but is contrary to others [8, 26]. An increased risk of forearm fracture among post-menopausal women with a history of hyperthyroidism was previously reported [27]. One contributing factor for our finding could be an incomplete bone recovery at this timepoint. Another factor might be a different recovery pace in the trabecular and the cortical bone compartment. Quantitative morphometric analysis based on iliac bone biopsies showed that the cortical porosity is normalized within 8 months of ATD treatment, whereas the changes in the trabecular compartment did not recover within this time frame [23]. Bone convalescence appeared different in the tibia and the radius in our study. In tibia the estimated bone strength increased and trabecular bone per tissue volume tended to increase, contrary to the radius where the trabecular vBMD and trabecular bone volume/tissue volume tended to decline despite recovery from thyrotoxicosis. A continued decline in the trabecular compartment presumably explains the ongoing aBMD reduction in the forearm, measured by DXA. The radius and the tibia also exhibited differences regarding the impact of BTM and PTH. We believe that the stimulus of weight on the tibia contributes to faster bone convalescence and counteract the effect of PTH and BTM.

Strengths and limitations

This was the largest longitudinal study on changes in BMD, BTM, and microarchitecture in GD patients, and the first study to assess the effect of supplemental vitamin D on bone indices during thyrotoxicosis.

The simultaneous bone scans using DXA and HRpQCT allowed a direct comparison on measurement of areal and volumetric BMD providing detailed information and a deeper understanding of the changes in bone microarchitecture and the impact of thyrotoxicosis that is not captures measuring osteodensitometry by conventional DXA, and our study demonstrates convincingly that both the trabecular and the cortical compartment is affected in hyperthyroidism and that different bone sites recover differently.

At 9 months, not all participants were euthyroid, delaying bone convalescence, and the study duration was inadequate to capture complete bone recovery. Whether a longer follow-up would unmask benefits from vitamin D is unknown, but the lack of effect on BTM during 9 months did not suggest it.

As participants had an overall sufficient 25(OH)D level, we cannot conclude on any effect on bone recovery in those with a deficient 25(OH)D level and future studies should focus on the beneficial effects of supplementation in this group of patients.

Conclusion

During the first 9 months of ATD treated Graves’ thyrotoxicosis, a daily high-dose vitamin D intake of 70 µg did not improve bone convalescence. We observed a different bone recovery in weight-bearing and non-weight-bearing bone sites and the association with BTM and PTH differed. HRpQCT scans demonstrated marked changes in bone microarchitecture revealing that both the trabecular and the cortical department are affected during thyrotoxicosis but also that the most pronounced change with treatment of thyrotoxicosis was a decrease in cortical porosity.