Key words
thyroid cancer - osteoporosis - osteopenia - thyroid-stimulating hormone
Introduction
Vertebral fracture (VF) is one of a major osteoporotic fracture, which is associated
with a shorter life expectancy and risk of cardiovascular events [1]. Fractures are consistently ranked as one of
the leading causes of ADL (activities of daily living) decline and needs care. Their
prevention is important for a healthy life expectancy and health economics [2].
Thyroid cancer is the most frequent endocrine malignancy, with papillary thyroid
carcinoma (PTC) accounting for more than 90% of cases. Although the
frequency of PTC is increasing [3]
[4], the 10-year survival rate is more than
95% and the long-term prognosis is promising for most of the patients with
PTC [5].
Cases with a history of thyroid cancer exhibits the lower mean age of first fracture
than those with a history of other thyroid diseases [6]. Thyrotropin (TSH) suppression therapy by levothyroxine is commonly
applied for the patients with thyroid cancer; the levothyroxine treatment suppressed
TSH levels has shown to be the fracture risk [7]. Additionally, we have shown that PTC itself, before the TSH
suppression therapy, has a high frequency of severe osteoporosis [8]. However, whether the risk of VFs is higher
in PTC prior to TSH suppression therapy was not elucidated yet.
Subjects and Methods
Participants
We enrolled 66 patients after thyroidectomy due to thyroid cancer at Shimane
University Hospital between September 2012 and November 2019 ([Fig. 1]). The patients were evaluated
according to the UICC-TNM classification (8th edition). Some were excluded due
to using drugs affecting bone metabolism such as glucocorticoids or diagnosing
with bone metastasis and osteoporosis secondary to renal failure or
endocrinological abnormalities or having a history of traumatic fractures.
Forty-three patients whose bone metabolism, including the presence or absence of
VF, could be assessed were included in the PTC group. The control group
comprised a stratified random sampling of 43 age- and sex-matched healthy
individuals who underwent health screening for osteoporosis at a community
health center. In addition to age, sex, and BMI, we asked them whether they
smoked or drank alcohol, their age at menopause, and whether they had a family
history of proximal femur fractures. This study was approved by the ethics
review board of Shimane University Faculty of Medicine (No. 201809–1)
and complied with the Helsinki Declaration. Since this study was retrospective
design and collected data from our daily clinical practice, the formal written
informed consent was not required. Therefore, we opened this research project to
our hospital homepage for allowing participants to refuse research
participation.
Fig. 1 Flowchart of the study population. We enrolled 62 patients
who were diagnosed with PTC from September 2012 to November 2019 and 19
were excluded due to using drugs affecting bone metabolism or diagnosed
with secondary osteoporosis or having a history of traumatic fractures.
The control group comprised a stratified random sample of 43 age- and
gender-matched healthy individuals. PTC: Papillary thyroid carcinoma:
FTC: Follicular thyroid carcinoma.
Biochemical measurements
Blood and urine samples were collected just before TSH suppression therapy. Serum
concentrations of albumin, creatinine, calcium, phosphorus, intact parathyroid
hormone (PTH), urine type I collagen cross-linked N-telopeptide (NTX, as a
marker of bone resorption), were evaluated in both groups by automated
techniques at the central laboratory of our hospital. Intact PTH and urine NTX
were evaluated using ECLIA and CLEIA, respectively. Reference ranges of them
were: urine NTX 13.0–66.2 (male), 9.3–54.3 (premenopausal
female), 14.3–89.0 (postmenopausal female) nmolBCE/mmol·Cre,
intact PTH 10–65 pg/ml. Estimated glomerular filtration rate
(eGFR) was calculated using the equation proposed by the Modification of Diet in
Renal Disease Study with modified coefficients for the Japanese population.
Hemoglobin A1c (HbA1c) (International Federation of Clinical Chemistry) was
determined by high-performance liquid chromatography.
In the PTC group, preoperative free T3, free T4, TSH thyroglobulin (Tg), were
measured. Anti-thyroid peroxidase antibodies (TPO-Ab) and anti-thyroglobulin
antibodies (Tg-Ab) were evaluated using the two-step chemiluminescence enzyme
immunoassay kits “CL AIA-PAC TRAb”, “CL AIA-PAC
TPOAb” and “CL AIA-PACK TgAb”, respectively. Reference
ranges of the variables used to assess thyroid function were: Free T3
2.1–3.8 pg/ml, Free T4 0.8–1.5 ng/dl, TSH
0.45–4.5 μIU/ml, TPO-Ab≤3.2 IU/ml and Tg-Ab≤13.6
IU/ml.
Bone mineral density (BMD) measurements
BMD at the lumbar spine and femoral neck were measured by dual-energy X-ray
absorptiometry using the QDR-4500 system (Hologic, Waltham, MA). Automatically
calculated BMD based on the bone mineral content (in grams) and bone area (in
square centimeters) were shown as an absolute value in grams per square
centimeter. T-score, the number of standard deviations (SDs), indicates
difference from the mean for a normal young adult reference population.
Radiography
In the same week as serum collection, the lateral X-rays of the thoracic and
lumbar spine were performed. VF was diagnosed by semiquantitative assessment of
13 vertebral bodies from Th4-L4 [9].
Diagnosis of osteoporosis
Osteoporosis was diagnosed according to WHO criteria. The WHO defines normal BMD
as a T score of –1 and above, low bone mass as a T score between
–2.5 and –1, osteoporosis as a T score of –2.5 and
below, and severe osteoporosis as a T score of –2.5 and below with a
fragility fracture.
Definition of vertebral fracture
We defined VFs as grades 1 to 3 according to the classification by Genant et al.
[9]. Grade 1 corresponds to a
20% to 25% reduction in at least one height (anterior, middle,
or posterior) along the length of the same vertebra compared with the height of
the nearest uncompressed vertebral body. Grade 2 VF corresponds to a 25%
to 40% reduction in vertebral height, and grade 3 corresponds to a more
than 40% reduction in any height. We defined severe fracture as grade 2
or 3 VF. VFs were diagnosed by two investigators who were blinded to each
other’s readings and also blinded to PTC group and control group status.
Fractures were assessed at the same time, and if there was disagreement between
the two investigators, the findings were assessed by a third investigator. No
participants had any history of serious trauma.
Statistical analysis
Significant differences between groups were determined using the chi-square and
Mann–Whitney U-test, and the association with VF was analyzed by
logistic regression analysis. Statistical analysis was performed using SPSS
software (ver. 19; IBM Corporation, Tokyo, Japan) and statistical significance
was determined at p <0.05.
Results
Baseline characteristics of subjects
Forty-three healthy individuals and 43 patients with PTC (15 males and 28
females) were included in this study. Their ages ranged from 27 years at the
youngest to 82 at the oldest, with an average of 61 years old
(mean±standard deviation, 61±11.8 years). Characteristics at
baseline are shown in [Table 1]. The PTC
patients had a mean duration after operation of 4.0 months at the time of study
enrollment. The number of patients with the different PTC stages as diagnosed by
UICC 8th criteria were: Stage I: 4, Stage II: 3, Stage III: 14, and Stage IV:
22. Patients with PTC had a higher body mass index (BMI), phosphorus, and HbA1c,
as well as lower intact PTH (p <0.05 for all). There was no significant
difference in serum calcium, urinary NTX, femoral neck and lumbar spine BMD. On
the other hand, the prevalence of osteoporosis, severe osteoporosis, VFs, or
severe VFs was higher in the PTC group compared to the control group (27.8 vs.
14.0%, 25.0 vs. 7.0%, 44.1 vs. 16.3%, 14.0 vs.
4.7%, p <0.05 for all).
Table 1 Baseline characteristics of patients with
papillary thyroid carcinoma and controls.
|
Parameter
|
Papillary thyroid carcinoma (n=43)
|
Control subjects (n=43)
|
p
|
|
Male/Female (%)
|
15 (35)/28 (65)
|
15 (35)/28 (65)
|
–
|
|
Ages (years)
|
61.2±12.0
|
61.5±12.1
|
0.921
|
|
BMI (kg/m2)
|
24.7±3.60
|
22.8±2.30
|
0.019
|
|
Creatinine (mg/dl)
|
0.71±0.17
|
0.64±0.14
|
0.062
|
|
eGFR (ml/min/1.73 m2)
|
74.3±13.0
|
83.1±18.3
|
0.058
|
|
HbA1c (mmol/mol)
|
42.6±9.92
|
35.6±8.79
|
<0.001
|
|
Alb-corrected Ca (mg/dl)
|
9.10±0.53
|
9.20±0.36
|
0.575
|
|
Phosphorus (mg/dl)
|
3.80±0.51
|
3.25±0.49
|
<0.001
|
|
Intact PTH (pg/ml)
|
22.6±17.5
|
44.5±13.2
|
<0.001
|
|
U-NTX (nMBCE/mM·Cr)
|
46.7±41.2
|
51.5±107
|
0.451
|
|
Preoperative Free T3 (pg/ml)
|
2.66±0.390
|
–
|
|
|
Preoperative Free T4 (ng/dl)
|
1.14±0.19
|
–
|
|
|
Preoperative TSH (μU/ml)
|
2.05±1.64
|
–
|
|
|
Preoperative Tg (ng/ml)
|
180.5±581.8
|
–
|
|
|
Tumor size (cm)
|
1.93±1.09
|
–
|
|
|
Postoperative period (months)
|
4.0±3.0
|
–
|
|
|
L2–4 BMD (g/cm2)
|
0.975±0.198
|
0.942±0.184
|
0.570
|
|
T-score
|
–0.566±1.60
|
–0.740±1.51
|
0.68
|
|
Neck BMD (g/cm2)
|
0.693±0.139
|
0.717±0.14
|
0.668
|
|
T-score
|
–1.03±1.14
|
–0.858±1.01
|
0.760
|
|
Diabetes mellitus (%)
|
9 (20.9)
|
–
|
–
|
|
Osteoporosis (%)
|
10 (27.8)*
|
6 (14.0)
|
0.128
|
|
Severe osteoporosis (%)
|
9 (25.0)*
|
3 (7.0)
|
0.026
|
|
Vertebral fracture (%)
|
19 (44.1)
|
7 (16.3)
|
0.004
|
|
Severe vertebral fracture (%)
|
9 (14.0)
|
3 (4.7)
|
0.0015
|
Values are expressed as mean±SD.; BMI: Body mass index; eGFR:
Estimated glomerular filtration rate; HbA1c: Hemoglobin A1c; PTH:
Parathyroid hormone; U-NTX: Urine type I collagen cross-linked
N-telopeptide; TSH: Thyroid hormone stimulation hormone; L BMD: Bone
mineral density at the lumber spine 2–4; Neck BMD: Bone mineral
density at the femoral neck.; * BMD assessment was
performed on 36 patients with PTC group.
Vertebral Fracture and severity
Patients were excluded if they had traumatic fractures or known fractures
diagnosed prior to the study. All of VFs were morphometric fracture. In PTC
group, the site of VF was Th5 in two, Th6 in one, Th11 in three, Th12 in seven,
L1 in six, L2 in six, L3 in three ([Fig.
2a]). Of the 19 patients who had more than two VFs, 9 had severe VFs.
Forty percent of the PTC group in their 40 s, 40% in their
50 s, 47% in their 60 s, 33% in their
70 s, and 50% in their 80 s had VFs. Notably, fractures
were more prominent in the younger generation compared to the control group
([Fig. 2b]). In the control group,
the VFs were located mainly at the thoracolumbar transition: Three patients had
Th11 fractures, five had Th12 fractures, one had L1 fractures, and one had L3
fractures ([Fig. 2a]). Of the seven
patients with more than two fractures, three were severe fractures. Ten percent
of the patients in their 50 s, 11.8% in their 60 s,
25% in their 70 s, and 50% in their 80 s had
fractures. The fracture rate increased with older age ([Fig. 2b]). Previous non-VF s were found in
one patient in the PTC group (left forearm) and four patients in the control
group (toe, foot, ribs, and wrist). There was no statistically significant
difference between the two groups.
Fig. 2
a: Location and number of fractures. In the papillary carcinoma
group, two patients had fractures outside the thoracolumbar transition.
On the other hand, the location of the fracture was only at the
thoracolumbar transition. PTC: Papillary thyroid carcinoma; Th: Thoracic
vertebra; L: Lumber vertebra. b: Age of fracture victims and
number of fractures. In the papillary carcinoma group, fractures were
more prominent in the younger generation compared to the control group.
In the control group, the fracture rate increased with age. PTC:
Papillary thyroid carcinoma.
Comparison of parameters between patients with or without VFs in PTC
patients
We compared the demographic and biochemical parameters of PTC patients with or
without VFs ([Table 2]). The patients
with VFs showed significantly lower T-score of lumbar spine and femoral neck
compared to patients without VFs. Serum creatinine levels were also
significantly lower in patients without VFs (0.68 vs. 0.76, p <0.05).
Other variables such as BMI, eGFR, HbA1c, serum Ca, intact PTH, urinary NTX,
preoperative free T3, free T4, preoperative TSH and tumor size were not
significantly different between the PTC group with and without VF.
Table 2 Comparison of demographic and biochemical
parameters, as well as bone turnover markers, and bone mineral
density between subjects with and without vertebral fracture in
patients in papillary thyroid carcinoma.
|
Without vertebral fracture (n=24)
|
With vertebral fracture (n=1 9)
|
p
|
|
Number of subject (Male/Female)
|
24 (6/18)
|
19 (9/10)
|
0.082
|
|
Age (years)
|
59.6±12.9
|
63.2±10.7
|
0.287
|
|
Duration after operation (months)
|
3.8±2.2
|
4.7±3.6
|
0.717
|
|
BMI (kg/m2)
|
25.3±4.01
|
23.9±2.87
|
0.24
|
|
Creatinine (mg/dl)
|
0.68±0.17
|
0.76±0.15
|
0.046
|
|
eGFR (ml/min/1.73 m2)
|
77.6±14.3
|
70.0±9.83
|
0.074
|
|
HbA1c (mmol/mol)
|
43.1±11.4
|
42.2±7.77
|
0.784
|
|
Ca (mg/dl)
|
9.1±0.55
|
9.1±0.52
|
0.825
|
|
Intact PTH (pg/ml)
|
23.6±17.8
|
21.3±17.4
|
0.696
|
|
U-NTX (nMBCE/mM·Cr)
|
46.0±25.2
|
47.4±53.7
|
0.215
|
|
Preoperative Free T3 (pg/ml)
|
2.64±0.40
|
2.69±0.38
|
0.663
|
|
Preoperative Free T4 (ng/dl)
|
1.10±0.16
|
1.18±0.22
|
0.279
|
|
Preoperative TSH (μU/ml)
|
2.02±1.30
|
2.10±2.05
|
0.404
|
|
Preoperative Tg (ng/ml)
|
262±793
|
46.2±48.6
|
0.229
|
|
L BMD (g/cm2)
|
1.03±0.20
|
0.90±0.18
|
0.039
|
|
T-score
|
0.035±1.70
|
–1.37±1.04
|
0.01
|
|
Neck BMD (g/cm2)
|
0.75±0.12
|
0.63±0.14
|
0.018
|
|
T-score
|
–0.54±0.91
|
–1.64±1.12
|
0.007
|
|
Tumor size (cm)
|
1.92±1.30
|
1.94±0.84
|
0.447
|
|
Tg-Ab positive (%)
|
5 (20.8)
|
4 (21.1)
|
0.292
|
|
TPO-Ab positive (%)
|
3 (17.6)
|
4 (22.2)
|
0.309
|
|
Smoking (%)
|
5 (20.8)
|
9 (47.4)
|
0.05
|
|
Drinking (%)
|
5 (20.8)
|
8 (42.1)
|
0.09
|
Values are expressed as mean±SD.; BMI: Body mass index; eGFR:
Estimated glomerular filtration rate; HbA1c: Hemoglobin A1c; PTH:
Parathyroid hormone; U-NTX: Urine type I collagen cross-linked
N-telopeptide; TSH: Thyroid hormone stimulation hormone; L BMD: Bone
mineral density at the lumber spine 2–4; Neck BMD: Bone mineral
density at the femoral neck.
Association between PTC and presence of VF
Multivariate logistic regression analysis adjusted for age, sex, BMI revealed
that PTC was associated with the presence of VF (odds ratio 5.68: 95%
confidence interval 1.77–18.2; p <0.05). This relationship
remained significant after additional adjustment for HbA1c, lumbar and femoral
neck BMD ([Table 3]).
Table 3 Associations between papillary thyroid carcinoma
and vertebral fracture.
|
Adjusted variables
|
OR (95% CI)
|
p
|
|
None
|
4.07 (1.49–11.2)
|
0.006
|
|
Age
|
4.51 (1.58–12.8)
|
0.005
|
|
Age+gender
|
4.81 (1.64–14.1)
|
0.004
|
|
Age+gender+BMI
|
5.63 (1.82–17.5)
|
0.003
|
|
Age+gender+BMI+HbA1c
|
5.68 (1.77–18.2)
|
0.004
|
|
Age+gender+BMI+L2–4 BMD
|
6.37 (1.83–22.2)
|
0.004
|
|
Age+gender+BMI+neck BMD
|
4.31 (1.32–14.0)
|
0.015
|
BMI: Body mass index; HbA1c: Hemoglobin A1c; L2–4 BMD: Bone
mineral density at the lumber spine 2–4; neck BMD: Bone mineral
density at the femoral neck.
Discussion
Not only as the risk for severe osteoporosis [8], here, we report for first time that there is an association between
PTC and the risk of VF. In our cohort, most PTC patients displayed VFs before
receiving TSH suppression therapy and only five out of 43 patients with PTC had low
TSH level at diagnosis. Therefore, this study suggested that PTC itself was
associated with VFs. Our analysis demonstrated that evaluation of VFs is important
in PTC patients with indications for TSH suppressive therapy.
It is well known that suppressed TSH or hyperthyroidism is associated with a risk
of
osteoporotic fracture [7]
[10]. From population-based cohort, elderly
women with low TSH had a four times higher risk for VFs [11]. Even in subclinical hyperthyroidism, the
risk of fracture increased to 1.28 (95% CI: 1.06–1.53) [12]. From a study in patients with thyroid
cancer, randomized trial from Japan showed that BMD was predominantly decreased in
women over 50 years of age who received TSH suppression [13]. In patients at low to medium risk as
defined by the American Thyroid Association, TSH suppression therapy with
levothyroxine has shown to increase the risk of osteoporosis without altering to
life expectancy [14], and the management of
bone fragility is essential for cases with a promising long-term prognosis. However,
these studies evaluated fractures during TSH-suppressive therapy, and none has
examined whether PTC itself is associated with fractures. In the current study, the
duration of TSH suppression supposed to be too short to affect bone metabolism
leading to the fracture. Also, we found no significant differences in thyroid
hormones or TSH level when comparing patients with or without VFs. Taken together,
PTC could be a risk factor for VFs.
In our study, the prevalence of VF and osteoporosis were higher in PTC patients
compared to that of control subjects [8].
There are a few reports on thyroid cancer and osteoporosis. The population-based
cohort study in Taiwan showed that osteoporosis was significantly higher risk factor
for the development of thyroid cancer [15].
The authors discussed why osteoporosis is associated with the development of thyroid
cancer and described the involvement of potential selection bias. We previously
showed that patients with PTC had a higher prevalence of osteoporosis compared to
patients without PTC, and that patients with severe osteoporosis had significantly
higher rates of anti-TPO antibodies [8].
According to a study using a questionnaire, subjects with a history of thyroid
cancer have shown to display a lower mean age of first fracture when compared to
those with a history of other thyroid diseases [6]. The reason why patients with PTC exhibited an increased risk of VF
was not elucidated yet. Our study revealed that PTC was a risk factor for VF in age-
and sex-matched subjects, independent of BMI, glucose profile and BMD. It suggests
that there is an increased risk of fracture independent of BMD in the PTC group.
Because bone strength is defined by BMD and bone quality, evaluation of bone quality
is important. Unfortunately, we were not able to examine parameters to assess bone
quality such as TBS or bone metabolism markers in the present study. In order to
elucidate the bone fragility mechanism in PTC, accumulation of bone quality
parameters is a future issue.
The vitamin D deficiency is a known risk for osteoporosis [16]
[17]
and fractures [18]. Sahin et al. reported that
vitamin D3 levels were significantly lower and vitamin D deficiency was more
frequent in the patients with PTC [19]. In the
PTC group, 97% of the patients were vitamin D deficient [20], suggesting the potential contribution of
Vitamin D deficiency on fracture in PTC. In our cohort, 25-hydroxyvitamin D (25D)
was measured in only 34.9% (15/43 patients) of the PTC group and
86.0% (37/43 patients) of the control group. Even though with such
insufficient data set, 25D levels were significantly lower (p=0.037) in the
PTC group (15.5±5.75) when compared to that of the control group
(18.8±4.93). Among the participants with 25D measured, VF was present in
10/15 in PTC group and 7/37 in control group.
The oxidative stress contributes to the development of osteoporosis [21]
[22].
Ramli et al. showed that serum reactive oxygen species (ROS) levels in patients with
PTC were significantly higher than those in controls, suggesting that thyroid cancer
causes the generation of ROS and induces oxidative stress [23]. These reports suggest the
pathomechanistical role of the oxidative stress in the development of VF in PTC
patients.
Advancing age, as expected, is the most important risk factor for VF [24]
[25].
Although prevalence of VF was increasing with age in control subjects, that of the
PTC patients showed an age-independent distribution in this study. Furthermore, it
is known that VFs are generally more likely to occur at the thoracolumbar junction
[26]. In the control group, the
thoracolumbar junction was the most common site of VF, but VFs of PTC patients are
found in the upper thoracic and lower lumbar vertebrae in this study. These
differences suggest that PTC-related bone deterioration has a different mechanism
than age-related osteoporosis.
In our study, there was no difference in BMD between PTC patients and control;
however, the prevalence of VFs was significantly higher in PTC patients. Evaluating
the BMD in post-thyroidectomized PTC patients by both DXA and central quantitative
computed tomography (cQCT), the total prevalence of osteoporosis and osteopenia was
higher by cQCT (92.6%) compared to that by DXA (37.0%) [27]. The volumetric BMD calculated by cQCT is
much strongly associated with bone microarchitecture, the potential determinant of
bone strength independent of bone density [28]. Therefore, PTC patients would display the deterioration of bone
microarchitecture without altering the BMD levels. The interaction between PTC and
bone microarchitecture is important research topic and further investigation would
be required.
Inflammatory cytokines affect osteoporosis because high cytokine content promotes
bone resorption. Within the tumor microenvironment of PTC, the presence of cancer
proteins induces an inflammatory state that upregulates several chemokines [29]
[30].
The genes encoding chemokines are induced in PTC compared to normal thyroid,
suggesting that chemokines are associated with tumor-associated inflammation [30]. In this study, there was no association
between tumor size, lymph node metastasis, distant metastasis, or thyroglobulin
levels and prevalence of VFs in PTC patients. PTC stage and disease activity would
be not involved in the development of VFs. The present study included only either
intermediate- or higher-risk patients who were treated with intravenous radioiodine
therapy; therefore, further studies with lower-risk patients are needed.
Study limitation
Several limitations of this study must be clarified. First of all, the number of
subjects is small, and all subjects are Japanese. Second, this is a
retrospective study, which means that many people at high risk of fracture might
be included in the study (involving subjects who are interested in fractures
tend to undergo bone health screening). Third, we could not exclude all
influence of hospitalization/immobilization/inflammation caused
by the surgical treatment in the VFs occurrence. We need a prospective study to
measure each confounding factors before surgery including patients with
early-stage PTC who are not candidates for radiotherapy. Fourth, in this report,
we have not collected enough clinical indicators to adequately assess the
fracture mechanism of PTC. Fifth, 70% of the fractures were grade 1,
which may include some fractures that are usually overlooked or considered to be
just deformities. Careful consideration is still required regarding the
pathological significance of mild fractures. Finally, we did not fully evaluate
factors that may be involved in the pathophysiology, such as concentrations of
serum 25-hydroxyvitamin D and oxidative stress marker. Lastly, this study did
not include lower-risk PTC patients.
Conclusion
We found that PTC itself was a risk factor for VF independent of age, sex, BMI,
glucose metabolism, and BMD. Bone densitometry and fracture evaluation by chest and
lumbar X-ray could be recommended for those considering TSH suppression therapy.
