Calcified Tissue International

, Volume 77, Issue 3, pp 139–144

Influence of Hyper- and Hypothyroidism, and the Effects of Treatment with Antithyroid Drugs and Levothyroxine on Fracture Risk

Authors

    • Department of Endocrinology and Metabolism C Aarhus University Hospital
    • The Osteoporosis Clinic
  • L. Rejnmark
    • Department of Endocrinology and Metabolism C Aarhus University Hospital
  • L. Mosekilde
    • Department of Endocrinology and Metabolism C Aarhus University Hospital
Article

DOI: 10.1007/s00223-005-0068-x

Cite this article as:
Vestergaard, P., Rejnmark, L. & Mosekilde, L. Calcif Tissue Int (2005) 77: 139. doi:10.1007/s00223-005-0068-x

Abstract

The purpose of this study was to assess alterations in fracture risk in patients with hyper- and hypothyroidism, and the effects of antithyroid drugs and levothyroxine on fracture risk. The study was designed as a case–control study. All patients with a fracture (n = 124,655) in the year 2000 in Denmark served as cases. For each case, three age- and gender-matched controls were randomly drawn from the general population (n = 373,962). Exposure variables were a diagnosis of hyperthyroidism, hypothyroidism, ever use of antithyroid drugs, and ever use of levothyroxine. Adjustments were made for time since diagnosis, surgery for hyperthyroidism, thyroid cancer, prior fracture, alcoholism, and corticosteroid use. There was an increase in the risk of any fracture within the first 5 years after a diagnosis of hyperthyroidism, and the first 10 years after a diagnosis of hypothyroidism. Use of antithyroid drugs was associated with a significantly reduced fracture risk independently of the dose used (odds ratio [OR] 0.79–0.84, 2P < 0.05). No effect of levothyroxine on fracture risk was present. Thyroid surgery for hyperthyroidism and thyroid cancer were not associated with fracture risk. A prior fracture, alcoholism, and corticosteroid use were significant risk factors for fractures. Fracture risk is increased at the time of diagnosis of both hyper- and hypothyroidism, and this increase seems reversible. Low-dose levothyroxine was not associated with fracture risk. The effect of antithyroid drugs may be caused by the reduction in thyroid hormone levels.

Keywords

FractureHyperthyroidismHypothyroidismLevothyroxineAntithyroid drugs

Thyroid hormones have profound effects on bone turnover [1, 2]. Patients with hyperthyroidism have an increased bone turnover [1], leading to a negative bone balance, mainly because of an expansion of the remodeling space, a decreased cortical thickness [3], and an increased risk of trabecular perforations [4]. The clinical consequences are decreased bone mineral density (BMD) and increased fracture risk [5]. In patients with untreated hypothyroidism, bone turnover is decreased in both trabecular and cortical bone [6, 7]. However, in addition to the well-known effects of thyroid hormones on bone turnover mediated by thyroid hormone receptors (TRs) [8], thyroid stimulating hormone (TSH) may also influence bone turnover directly, with high levels of TSH inhibiting bone resorption [9].

Upon treatment for hyperthyroidism, bone turnover is normalized, the reduced BMD increases, and fracture risk decreases [5]. In patients with hypothyroidism, a transient loss of bone mineral [10, 11] and a parallel increase in fracture risk [12] is observed within the first 2 years after initiation of levothyroxine substitution therapy, probably because of a transient increase in bone turnover with remodeling of accumulated old bone, and perhaps an increased risk of falls. If TSH is suppressed by levothyroxine supplementation, a decreased BMD is observed [13]. The latter situation is often encountered in patients with thyroid cancer, in which TSH suppression is induced to stop tumor growth.

In patients with hyperthyroidism, treatment with radioactive iodine has been associated with an increased risk of fractures, whereas use of antithyroid drugs has been associated with a decreased risk of fractures [14]. It is likely that the antithyroid drugs quickly reverse the increased bone turnover [1517], and thereby help restore normal BMD and integrity, whereas radioactive iodine may have a more delayed effect on the increased thyroid hormone levels [14, 18]. However, it remains unclear whether the use of levothyroxine as supplementation to methiazole and propylthiouracil treatment (block-replace treatment) for Graves’ disease may modify the fracture risk in patients with hyperthyroidism by interacting with the balance between TSH and thyroid hormones (T3 and T4). The increase in thyroid hormone level induced by levothyroxine may, therefore, increase fracture risk by suppressing TSH. Previous studies have not taken this potential interaction into account and have not adjusted for the normalization of thyroid function over time [14]. The aim of the present study was to assess the effect of hyper- and hypothyroidism, antithyroid treatment, and thyroid hormone substitution therapy on fracture risk in a nationwide case–control study.

Subjects and Methods

Study Design

The study was designed as a case–control study. All subjects who sustained a fracture during the year 2000 in Denmark were included as cases (n = 124,655), and for each case three subjects of the same age (same birth year) and gender were randomly selected from the background population as controls (n = 373,962). The controls were selected using the incidence–density sampling technique [19], that is, the controls had to be alive and at risk for fracture diagnosis at the time the corresponding case was diagnosed. The follow-up time was time until fracture in cases and a corresponding dummy date among the controls. The cases occurred only once in the analyses, with the first occurrence of an incident fracture during the year 2000.

Endpoints

The study endpoints were occurrence of any fracture (ICD10 codes: S02.0–S02.9, S07.0–S07.9, S12.0–S12.9, S22.0–S22.9, S32.0–S32.8, S42.0–S42.9, S52.0–S52.9, S62.0–S62.9, S72.0–S72.9, S82.0–S82.9, and S92.0–S92.9) between January 1, 2000 and December 31, 2000. In Denmark almost all patients with fractures are managed in the hospital system (including the emergency rooms) [20]; even fractures sustained abroad are registered upon return for insurance reasons. The capture of fractures is therefore very high.

Exposure Variables

The exposure variables were (1) presence of prior hyperthyroidism (ICD8: 242.00, 242.01, 242.08, 242.09, 242.19, 242.20, 242.28, and 242.29; ICD10: E05.0–E05.5, E05.8, and E05.9) from 1977 to the date of censoring from the National Hospital Discharge Register, (2) presence of prior hypothyroidism (ICD8: 243.99, 244.00–244.03, 244.08, and 244.09; ICD 10: E00.1, E03.0–E03.5, E03.8, and E03.9) from 1977 to the date of censoring from the National Hospital Discharge Register, (3) prior use of antithyroid drugs (propylthiouracil [ATC: H03BA02], thiamazole [ATC: H03BB02], and carbimazole [ATC: H03BB01]) from 1996 to the date of censoring from the National Pharmacoepidemiological Database, (4) prior use of levothyroxine [ATC: H03AA01] from 1996 to the date of censoring from the National Pharmacoepidemiological Database, (5) prior surgery to the thyroid for hyperthyroidism [0800, 0802, 0804, 0806, 0808, 0810, 0812, 0814, 0816, 0818, 0820, 0822, 0834, 08000, 08020, 08040, 08060, 08080, 08100, 08120, 08140, 08160, 08180, 08200, 08220, 08240, 08340, KBAA00, KBAA05, KBAA10, KBAA20, KBAA25, KBAA30, KBAA40, KBAA50, KBAA60, KBAA99] from 1977 to the date of censoring from the National Hospital Discharge Register, and (6) prior thyroid cancer [ICD8: 193.99; ICD10: C73.9] from 1977 to the date of censoring from the National Hospital Discharge Register. The confounders were (1) ever use of corticosteroids, (2) presence of a prior fracture [21], and (3) prior alcoholism (a prior contact with a diagnosis of alcoholism from 1977 to the date of censoring from the National Hospital Discharge Register, or from 1968 to the date of censoring from the Psychiatric Central Register, or at least one prescription of disulfiram from 1996 to the date of censoring from the National Pharmacoepidemiological Database) [22].

Registers Used

We retrieved the information on fracture occurrence and occurrence of other diseases, prior fractures, diagnosis of alcoholism from the National Hospital Discharge Register between 1977 and the date of censoring [23], and the Psychiatric Central Register for the period from 1968 to the date of censoring [24]. We acquired information on the use of prescribed drugs from the Danish Medicines Agency in the period from 1996 to 2000. The average daily dose of drug was calculated as the cumulated dose divided by the time from the date of first prescription to the date of censoring.

The National Hospital Discharge Register was founded in 1977 [23]. It covers all inpatient contacts from 1977 to 1994, and from 1995 also all outpatient visits to hospitals, outpatient clinics, and emergency rooms [23]. Upon discharge, the physician codes the reason for the contact using the ICD system. The code used is at the discretion of the individual physician. The register has a nationwide coverage and an almost 100% capture of contacts [23]. In general, the validity of registrations is high [25], especially for fractures, for which a precision of 97% has been reported for fractures treated both on an inpatient basis and on an outpatient basis via emergency rooms (e.g., a forearm fracture) [26].

The Psychiatric Central Register was founded in 1968 and covers all in- and outpatient contacts to Danish mental hospitals [24]. It has a nationwide coverage, and a high validity of diagnoses has been reported [20]. This register also uses the ICD system for coding contacts.

The Danish Medicines Agency keeps a nationwide register of all drugs sold at pharmacies throughout the country from 1996 and onward (The National Pharmacological Database run by the Danish Medicines Agency—http://www.dkma.dk). Any drugs bought are registered with an ATC code, dosage sold, and date of sale for the period from January 1, 1996 to December 31, 2000. Because all sales are registered to the individual who redeemed the prescription, the capture and validity are high.

We linked these sources of information through the Central Person Register Number, which is a unique registration code given to every inhabitant—to some degree similar to the American Social Security Number — that allows registration on an individual basis.

The project was approved by the National Board of Health and subject to control by the National Data Protection Agency.

Statistical Analyses

Mean and standard deviation (SD) were used as descriptive statistics. Crude and adjusted odds ratios (ORs), and 95% confidence intervals (CIs) were calculated. A conditional logistic regression analysis was used to assess the association between any fracture and the exposure variable.

Crude and multiply adjusted ORs were calculated. The addition of interaction terms between levothyroxine and antithyroid drugs (to analyze the effects of block-replacement therapy) or between levothyroxine and thyroid cancer (to adjust for levothyroxine used to suppress TSH), and introduction of additional comorbidiry (Charlson index) did not change the results (data not shown).

Analyses were performed using STATA 8.1 (STATA Corp., College Station, TX, USA) and SPSS 10.1.0 (SPSS Inc., Chicago IL, USA), both in the UNIX version.

Results

Table 1 shows baseline characteristics of the participants. The cases and controls were well matched concerning age and gender. More cases than controls had sustained a prior fracture, had ever used glucocorticoids, or had a history of alcoholism. More cases than controls had a prior history of hyper- or hypothyroidism or had used levothyroxine, whereas the use of antithyroid drugs, prevalence of thyroid cancer, and surgery for hyperthyroidism did not differ.
Table 1

Characteristics of patients and controls for all fracture types.

Variable

Cases (n = 124,655)

Controls (n = 373,962)

P

Age (y)

43.44 ± 27.39

43.44 ± 27.39

Gender

   

  Men

60,107 (48.2%)

180,321 (48,2%)

  Women

64,548 (51.8%)

193,641 (51.8%)

 

Previous fracture

41,315 (33.1%)

56,200 (15.0%)

< 0.01

Alcoholism

8,863 (7.1%)

9,473 (2.5%)

< 0.01

Ever use of any glucocorticoid

67,695 (54.3%)

189,636 (50.7%)

<0.01

Antithyroid drugs

1,221 (1.0%)

3,567 (1.0%)

< 0.42

Levothyroxine

2,750 (2.2%)

7,356 (2.0%)

< 0.01

Hyperthyroidism

1,517 (1.2%)

3,753 (1.0%)

< 0.01

Hypothyroidism

939 (0.8%)

1,904 (0.5%)

< 0.01

Surgery for hyperthyroidism

226 (0.2%)

618 (0.2%)

0.23

Thyroid cancer

51 (0.04%)

143 (0.04%)

0.68

Table 2 shows the adjusted fracture risk associated with prior hyper- and hypothyroidism, surgery for hyperthyroidism, thyroid cancer, and ever use of antithyroid drugs or levothyroxine for selected fracture sites adjusted for prior fracture and alcoholism. There was an increased risk of all fracture types except Colles’ fractures in patients with hypothyroidism, whereas patients with hyperthyroidism had an increased risk of any fracture and hip fractures. Use of antithyroid drugs was associated with a decreased risk of any fracture and hip fractures. Levothyroxine and surgery for hyperthyroidism were not associated with fracture risk and the same was the case for most fracture types for thyroid cancer. Corticosteroid use was associated with an increased fracture risk. A prior fracture and alcoholism were both significant risk factors for fractures.
Table 2

Effects of hyper- and hypothyrodism, Surgery, thyroid cancer, and use of levothyroxine and antithyroid drugs on fracture risk

Variable

Any fracture

Hip fracture

Colles’ fracture

Spine fracture

Hypothyroidism (yes/no)

1.32 (1.20–1.44)

1.72 (1.40–2.11)

1.10 (0.88–1.39)

1.63 (1.03–2.58)

Hyperthyroidism (yes/no)

1.21 (1.12–1.31)

1.54 (1.31–1.82)

1.14 (0.94–1.38)

1.11 (0.75–1.65)

Surgery for hyperthyroidism (yes/no)

0.92 (0.77–1.09)

0.78 (0.51–1.21)

0.88 (0.58–1.33)

1.72 (0.65–4.60)

Thyroid cancer (yes/no)

1.02 (0.73–1.42)

0.80 (0.29–2.19)

0.23 (0.05–0.97)

2.23 (0.54–9.30)

Ever use of antithyroid drugs (yes/no)

0.87 (0.80–0.94)

0.83 (0.70–0.99)

0.88 (0.73–1.07)

1.23 (0.82–1.83)

Ever use of levothyroxine (yes/no)

0.95 (0.90–1.00)

0.90 (0.79–1.02)

0.97 (0.85–1.10)

0.75 (0.55–1.02)

Ever use of corticosteroids

1.13 (1.11–1.14)

1.06 (1.01–1.11)

1.06 (1.03–1.10)

1.30 (1.20–1.41)

Prior alcoholism (yes/no)

2.49 (2.42–2.57)

4.12 (3.66–4.65)

2.36 (2.15–2.59)

2.76 (2.33–3.27)

Prior fracture (yes/no)

2.69 (2.65–2.73)

2.47 (2.36–2.59)

2.05 (1.97–2.13)

2.69 (2.46–2.94)

The ORs are adjusted for the other variables in the table.

Adjusted odds ration (OR) and 95% confidence intervals (CI)

Table 3 and 4 details the analyses by adding the time since diagnosis and the daily dose of drugs used.
Table 3

Crude and adjusted odds ratio (OR) and 95% confidence intervals (CI) for any fracture

Variable

Crude OR

Adjusted OR

Time since diagnosis of hyperthyroidism

  

  < 5 years

1.37 (1.25–1.50)

1.44 (1.28–1.62)

  5–9.9 years

1.12 (0.98–1.27)

1.08 (0.94–1.25)

  ≥ 10 years

1.13 (1.02–1.24)

1.09 (0.97–1.22)

Time since diagnosis of hypothyroidism

  

  < 5 years

1.66 (1.48–1.86)

1.51 (1.33–1.71)

  5–9.9 years

1.48 (1.26–1.74)

1.32 (1.11–1.57)

  ≥ 10 years

1.26 (1.09–1.46)

1.12 (0.96–1.31)

Daily dosage of antithyroid drugs

  

  < 0.4 DDD/day

1.04 (0.94–1.15)

0.84 (0.75–0.95)

  0.4–0.659 DDD/day

0.99 (0.87–1.13)

0.79 (0.69–0.91)

  > 0.66 DDD/day

1.04 (0.93–1.17)

0.84 (0.73–0.95)

Daily dosage of levothyroxine

  

  < 0.4 DDD/day

1.16 (1.07–1.26)

1.01 (0.92–1.10)

  0.4–0.659 DDD/day

1.12 (1.03–1.22)

0.91 (0.83–0.99)

  ≥ 0.66 DDD/day

1.10 (1.03–1.18)

0.94 (0.87–1.02)

Surgery for hyperthyroidism

1.10 (0.94–1.28)

0.98 (0.82–1.17)

Thyroid cancer

1.07 (0.78–1.47)

1.02 (0.73–1.43)

Ever use of corticosteroids

1.16 (1.14–1.17)

1.13 (1.11–1.14)

Prior fracture

2.69 (2.44–2.97)

2.69 (2.65–2.73)

Alcoholism

2.95 (2.86–3.03)

2.49 (2.42–2.57)

CI, Confidence interval; DDD, defined daily dose,

The ORs are adjusted for the other variables in the table.

The drugs are ever versus never use.

Table 4

Selected fracture sites. The drugs are ever versus never use. Crude and adjusted OR (95% CI)—after adjustment for the other covariates.

Variable

Hip

Colles

Spine

Time since diagnosis of hyperthyroidism

   

  < 5 years

1.93 (1.61–2.46)

1.32 (0.99–1.75)

0.90 (0.51–1.60)

  5–9.9 years

1.23 (0.92–1.65)

1.17 (0.84–1.65)

0.90 (0.44–1.85)

  ≥ 10 years

1.47 (1.17–1.85)

0.98 (0.74–1.29)

1.47 (0.85–2.54)

Time since diagnosis of hypothyroidism

   

  < 5 years

2.14 (1.64–2.80)

1.44 (1.04–1.99)

2.08 (1.17–3.70)

  5–9.9 years

1.19 (0.81–1.75)

0.92 (0.59–1.44)

0.88 (0.34–2.32)

  ≥ 10 years

1.73 (1.23–2.43)

0.97 (0.67–1.41)

1.72 (0.79–3.73)

Daily dosage of antithyroid drugs

   

  < 0.4 DDD/day

0.66 (0.51–0.85)

0.83 (0.63–1.11)

1.53 (0.88–2.65)

  0.4–0.659 DDD/day

0.88 (0.67–1.15)

0.87 (0.63–1.22)

0.95 (0.47–1.92)

  ≥ 0.66 DDD/day

0.87 (0.67–1.17)

0.81 (0.60–1.10)

1.38 (0.72–2.63)

Daily dosage of levothyroxine

   

  < 0.4 DDD/day

0.98 (0.80–1.19)

1.08 (0.88–1.32)

0.84 (0.53–1.33)

  0.4–0.659 DDD/day

0.87 (0.71–1.06)

0.92 (0.75–1.14)

0.77 (0.47–1.27)

  ≥ 0.66 DDD/day

0.84 (0.69–1.02)

0.91 (0.76–1.10)

0.65 (0.42–1.01)

Surgery for hyperthyroidism

0.81 (0.52–1.27)

0.97 (0.62–1.50)

1.56 (0.58–4.22)

Thyroid cancer

0.81 (0.30–2.24)

0.24 (0.05–1.01)

2.35 (0.55–10.0)

Ever use of corticosteroids

1.06 (1.01–1.11)

1.06(1.03-1.10)

1.30 (1.20–1.41)

Prior fracture

2.47 (2.36–2.59)

2.05 (1.97–2.13)

2.79 (2.47–2.94)

Alcoholism

4.13 (3.66–4.66)

2.36 (2.15–2.59)

2.76 (2.32–3.27)

CI, confidence interval; DDD, defined daily dose.

The ORs are adjusted for the other variables in the table.

Table 3 shows the crude and adjusted ORs for any fracture. After adjustment, an increase in relative fracture risk was only present less than 5 after a first diagnosis of hyperthyroidism was made, whereas an increase in relative fracture risk was present up to 10 years after a first diagnosis of hypothyroidism was made.

In the crude analysis, use of antithyroid drugs was not associated with fracture risk, whereas a reduced relative risk was observed after adjustment. There was no dose–response relationship between fracture risk and average daily dose of antithyroid drug. In the crude analysis, a limited excess fracture risk was present in users of levothyroxine, whereas no increase was present after adjustment. Surgical treatment of hyperthyroidism and thyroid cancer were not associated with fracture risk. Corticosteroid use was associated with an increased fracture risk. Both a prior fracture and alcoholism were significantly associated with an increased fracture risk.

Table 4 shows the fracture risk in the hip, forearm, and spine associated with thyroid disease. No trend with time since a diagnosis of hyperthyroidism was present except for hip fractures. For hypothyroidism, an increased fracture risk was seen with less than 5 years since diagnosis in the hip, spine, and the forearm. An increase in relative fracture risk was seen more than 10 years after diagnosis in the hip. Antithyroid drugs and levothyroxine were not associated with fracture risk, except for a decrease in hip fracture risk with the use of less than 0.4 defined daily dosages (DDDs) per day of antithyroid drugs. Surgical treatment and thyroid cancer were not associated with fracture risk. Corticosteroid use was associated with increased fracture risk. Alcoholism and a prior fracture were both significant risk factors for fractures.

Discussion

In this nationwide study, we have shown an increased fracture risk within 10 years after diagnosis of hypothyroidism and within 5 years after diagnosis of hyperthyroidism. Substitution with low-dose levothyroxine did not seem to alter fracture risk, whereas users of antithyroid drugs had a reduced fracture risk.

We observed a borderline significant reduction in overall fracture risk with the use of antithyroid drugs, but no dose–response relationship was observed. The absence of a dose–response relationship may be due to the fact that antithyroid drugs per se have little or no effect on BMD but act through the reduction in thyroid hormone level. These levels are again used to guide the dose of antithyroid drugs according to individual need. The risk of fracture decreased with the time since diagnosis, pointing to an effect of normalization of thyroid hormone levels. The antithyroid drugs lower thyroid hormone levels, and therefore, reduce fracture risk—the observed reduction in relative risk in this study may thus be the effect of residual confounding. The residual confounding results from the fact that patients on thyroid hormones have a more rapid reduction in fracture risk than those treated with, for example, radioactive iodine. This is probably also the reason that no independent effect of thyroid surgery for hyperthyroidism was present — thyroid surgery was not offered as first-line therapy and was first introduced for long-term management of hyperthyroidism after the hyperthyroidism had been controlled through the use of, for example, antithyroid drugs. This is supported by the observation that risk reduction was observed even in users of low-dose antithyroid drugs. The risk reduction in users of antithyroid drugs in the present study (OR approximately 0.8–0.85) was smaller than in a prior study [14] (OR 0.4). This may be due to the fact that the prior study did not adjust for time since diagnosis, thus ascribing the entire risk reduction to the antithyroid drugs rather than to the reduction in thyroid hormone levels as a result of the treatment.

Use of levothyroxine was not associated with any change in overall fracture risk. Levothyroxine is not used in Denmark in suppressive doses to treat benign thyroid disorders such as atoxic goiter. The doses used were low and probably did not result in suppression of TSH. The absence of a change in fracture risk may therefore, be in accordance with the metaanalysis of Uzzan et al. [13], who did not observe any changes in BMD in patients with nonsuppressed TSH in contrast to a decrease in BMD among patients with suppressed TSH.

The increase in fracture risk in both hyper- and hypothyroidism seemed reversible, except for hip fractures, in which an increase may persist. The reason for this discrepancy is unclear. The reason that both hyper- and hypothyroidism were associated with an increase in fracture risk despite opposing effects on bone turnover may be multifaceted. First of all does untreated hyperthyroidism lead to an increased bone turnover [1] with a decrease in BMD, which leads to an increase in fracture risk [5]. Upon normalization of the hyperthyroid state, bone turnover is reduced [17], BMD increase, and fracture risk declines [5]. In addition, falls are more often reported in patients with hyperthyroidism, probably because of impairment of muscular function by the hyperthyroid state [27, 28].

In patients with hypothyroidism in the untreated state bone turnover is decreased with an increased cortical thickness in histomorphometric studies [6, 7]. However, upon initiation of levothyroxine replacement there is an increased remodeling activity that exceeds that of normal individuals with an increased cortical porosity and an increased trabecular resorption activity [6], which may lead to trabecular perforations and thus an increased risk of fractures. Concomitantly a decrease in BMD of 5.4% in the spine and 7% in the hip is seen within the first year of levothyroxine replacement therapy [10]. However, if one extends the observation period, the decreased BMD returns to normal, probably signaling a normalization of bone turnover and therefore of fracture risk [11]. This is probably one of the explanations for the temporary increase in fracture risk after a diagnosis of hypothyroidism and thus initiation of levothyroxine therapy. Another explanation is the increased risk of falls seen in patients using levothyroxine, although only significant in men [27]. This may be the result of an increased physical activity at a time at which the patients still may have weak muscles and frail bones.

Strengths and Limitations of the Study

The main strengths of our study are that it is population-based, and there is a high capture rate for fractures, a large study size, a uniformly organized health care system allowing a population-based design, and the use of data on exposure and confounders that are collected before the date of hospitalization for hip fracture. Therefore, recall bias did not influence data collection.

The major weakness is the inability to adjust for use of radioactive iodine. The other weaknesses include potential selection bias, for example, the use of routine hospital discharge diagnoses of fractures coded by hospital doctors to ascertain case status. Some coding errors have probably occurred. However, misclassification of case status is unlikely to be related to prescription of drugs before hospitalization, and any nondifferential misclassification will lead to underestimation of our risk estimates. Therefore, selection biases are probably of minor importance in our study.

Other major weaknesses are the absence of information about concomitant diseases such as liver disease, malnutrition, renal disease, vitamin D deficiency, and hyperparathyroidism. However, these diseases are all relatively rare in the Danish population in general, and the effect on the risk estimates is therefore limited.

Furthermore, we did not address whether the fractures resulted from traumas or were spontaneous. However, previous studies have failed to show a difference in the proportion of trauma-related fractures in patients with hyper- or hypothyroidsm compared to controls [12, 14].

We cannot exclude that information biases may have influenced our results, for example, we had no information on patient compliance in our study, since redeeming a prescription was used as a proxy for actual use of a drug. However, the patients had paid for part of the cost of the drug, which increases the likelihood of compliance. Furthermore, data on drugs administered during hospitalization are not registered in the prescription database and, therefore, are not included in our study. Both of these uncertainties could have led to misclassification of the exposure. However, in Denmark hospitalization time for hyperthyroidism is short, and the amounts of antithyroid drugs administered in this way is therefore limited. Patients with hypothyroidism are seldom hospitalized initially for their hypothyroidism, and the amount of drugs administered in hospital is therefore also limited.

Although our analyses included many potential confounders, the design of our study did not allow for adjustment for clinical differences between users and nonusers. Most importantly, we were not able to adjust for differences in body weight and smoking [29].

In conclusion do both hyper- and hypothyroidism seem associated with a temporary increase in fracture risk that is reversed upon normalization of the thyroid disorder. Antithyroid drugs may lower fracture risk through reversal of the hyperthyroidism. Low-dose levothyroxine treatment is not associated with an increase in fracture risk.

Acknowledgement

The National Bureau of Statistics (Statistics Denmark) is acknowledged for their help, without which this project would not have been possible.

Financial support was provided by the Danish Medical Research Council (Grant number 22-04-0495).

Copyright information

© Springer Science+Business Media, Inc. 2005