Fractures following thyroidectomy in women: a population-based cohort study

Bone Vol. 27, No. 5 November 2000:695–700

Fractures Following Thyroidectomy in Women: A Population-based Cohort Study L. J. MELTON, III,1,2 E. ARDILA,2* C. S. CROWSON,1 W. M. O’FALLON,1 and S. KHOSLA2 1

Department of Health Sciences Research and 2Division of Endocrinology, Metabolism, Nutrition, and Internal Medicine, Mayo Clinic and Mayo Foundation, Rochester, MN, USA

onstrated a substantial increase in hip fracture risk among men following thyroidectomy27,28 or hyperthyroidism.28 Our group and others1,7,9,36 have also reported a strong association between a history of hyperthyroidism and hip fractures in women. However, fractures following thyroidectomy in women have not been adequately assessed. This is relevant because thyroidectomy could be associated with bone loss through a number of different mechanisms: endogenous excess of thyroxine prior to surgery; overenthusiastic thyroid replacement therapy following the operation; deregulation of bone resorption as a consequence of calcitonin deficiency caused by thyroidectomy; or some combination of these factors.10,14,26,35 Others have suggested that thyroid disease might be associated with an increased risk of falling,1 the main precipitating event for age-related fractures.16 To determine if thyroidectomy represents an important risk factor for fractures in women, and therefore should be subjected to closer scrutiny, we used the medical records linkage system in the community to identify all Rochester, MN women who underwent thyroidectomy between 1950 and 1974. The objective of this investigation was to estimate the risk of subsequent fractures among these women, relative to the incidence in the general population, and to evaluate several factors that might be associated with that risk.

Hip fracture risk has been associated with hyperthyroidism and thyroidectomy in men and with hyperthyroidism in women, but the influence of thyroidectomy on fracture risk in women has not been adequately addressed. The 630 Rochester, MN women who underwent thyroidectomy in 1950 – 1974 were followed subsequently for 12,804 person-years (retrospective cohort study) during which 601 fractures were observed. Relative to incidence rates in the community, there was no increase in overall fracture risk (standardized incidence ratio [SIR] 0.9; 95% confidence interval [CI] 0.8 – 1.00). No increase was seen in limb fractures generally or in distal forearm fractures specifically (SIR 1.1, 95% CI 0.8 – 1.4). There was a modest but statistically significant increase in the risk of hip fractures following thyroidectomy (SIR 1.3, 95% CI 1.01–1.8), but much greater increases were apparent in the risk of subsequent fractures of the ribs, spine, and pelvis. There was almost a threefold increase in vertebral fractures (SIR 2.8, 95% CI 2.3–3.3), but the excess was mostly observed long after the original operation and may be attributable to ascertainment bias. Fracture risk was associated with advancing age and with the presence of one or more of the diseases that have been associated with secondary osteoporosis but not with a history of hyperthyroidism, extent of thyroid surgery, or subsequent use of thyroid replacement therapy. Thus, with the exception of some fractures of the axial skeleton, which might have been more completely diagnosed among affected women, there was no increase in fracture risk among women following thyroidectomy performed mainly for adenoma or goiter. (Bone 27: 695–700; 2000) © 2000 by Elsevier Science Inc. All rights reserved.

Materials and Methods Population-based epidemiologic research can be conducted in Rochester because medical records for the entire population are available from almost all providers of care. Most surgical, endocrinological, and trauma care is provided by the Mayo Clinic, which has maintained a common medical record with its two large affiliated hospitals in the community (St. Marys and Rochester Methodist) for over 90 years. This dossier-type record, therefore, contains both inpatient and outpatient data, and the diagnoses and surgical procedures recorded in these records are entered into a computerized index.17 Medical records of the other providers who serve the local population, most notably the Olmsted Medical Group and its affiliated Olmsted Community Hospital, are indexed into the same system (the Rochester Epidemiology Project) and are also available for use in approved studies.17 Following approval by Mayo’s Institutional Review Board, we used this unique database to identify all 630 Rochester women who underwent thyroidectomy in the 25 year period of 1950 –1974. It was then possible to follow these women forward in time through their linked community medical records (retrospective cohort study) until death or the most recent clinical contact and to ascertain the risk of hip and other age-related fractures relative to the incidence in the general population.

Key Words: Thyroidectomy; Osteoporosis; Hip fracture; Vertebral fracture; Thyroid replacement therapy. Introduction A number of medical conditions are important contributors to bone loss and fracture risk.14 For example, we previously demAddress for correspondence and reprints: Dr. L. Joseph Melton, III, Section of Clinical Epidemiology, Department of Health Sciences Research, Mayo Clinic, 200 First Street SW, Rochester, MN 55905. *Present address: Faculty of Medicine, Department of Internal Medicine, Endocrinology Unit, National University of Columbia, Bogota, Columbia. © 2000 by Elsevier Science Inc. All rights reserved.

695

8756-3282/00/$20.00 PII S8756-3282(00)00379-3

696

L. J. Melton, III et al. Fractures following thyroidectomy

For each subject, all inpatient and outpatient records at any local provider of medical care were searched for the occurrence of fractures. Mayo Clinic records, for example, contain the details of every inpatient hospitalization at its two hospitals, every outpatient or office visit at the Clinic, and every emergency room or nursing home encounter, as well as all radiographic and pathology reports, including autopsies. Emphasis was on subsequent fractures at the skeletal sites usually associated with osteoporosis. Because we had access to an average of 17 years of documented medical history prior to the time of operation, it was possible to distinguish old from new fractures. The records contained the clinical history and the radiologist’s report of each fracture, but the original radiographs were not available for review. Thus, the diagnosis of vertebral fracture was accepted on the basis of a radiologist’s report of compression or collapse of one or more thoracic or lumbar vertebrae. All fractures were classified according to the circumstances of the injury: By convention, falls from standing height or less were considered moderate trauma, whereas motor vehicle accidents and falls from greater heights were deemed severe trauma. Fracture ascertainment is believed to be complete except for vertebral fractures, some of which are never diagnosed.3 The influence of thyroidectomy on fracture risk was evaluated using three basic methods of analysis. In the primary analysis, we calculated standardized incidence ratios (SIR), comparing the number of fractures that were observed at each skeletal site (based on the first fracture of a given type per person) to the number expected in this cohort during their follow-up in the community. One hundred eight women were deleted from this analysis who had no follow-up after age 35 years. Expected numbers of fractures were derived by applying gender- and age-specific incidence rates for these fractures to the age-specific person-years of follow-up in the thyroidectomy cohort. Incidence rates from the general population of Rochester women were available for fractures of the proximal femur,21 vertebrae,3 distal forearm,19 proximal humerus,31 pelvis,25 and all other sites.22 Ninety-five percent confidence intervals (95% CI) for the SIRs were calculated assuming that the expected rates are fixed and the observed fractures follow a Poisson distribution.5 In the second method of analysis, the cumulative incidence of new fractures (1 minus survival-free-of-fracture) was projected for up to 40 years following the date of surgery using productlimit life-table methods.12 Life-table methods were also used to assess survival, with expected death rates derived from Minnesota data. Cumulative incidence estimates, as well as the survival curves, were compared using the log-rank test statistic.11 Finally, in a third approach, Cox proportional hazards models,6 which do not incorporate the population expected rates, were used to assess the impact of various covariates on subsequent fracture risk. Bivariate relationships between the risk of specific fractures and each clinical characteristic under consideration were first assessed. Stepwise methods with forward selection and backward elimination were then used to choose independent variables for the final models. The dependent variable was time until the first new fracture at various skeletal sites, and the independent variables were the clinical characteristics. For the final multiple models, as well as for the bivariate models, the assumption of proportional hazards was examined and was not violated for the variables considered. Results Six hundred thirty Rochester, MN women underwent thyroidectomy in the 25-year time period of 1950 –1974. Patients ranged in age from 4 to 86 years, with a median age at surgery of 42.5

Bone Vol. 27, No. 5 November 2000:695–700

years. All but 17 patients were white, reflecting the racial composition of the community (99% white in 1970). Thyroidectomy was “total” in 7 patients (1%), whereas 74 women (12%) had a subtotal lobectomy and the remainder had a partial lobectomy. At the time of surgery, 85 women were hyperthyroid, but most had hyperfunctioning nodular thyroid disease; only 11 had Graves disease. Three other women were hypothyroid at operation, of whom 2 had Hashimoto thyroiditis. Most of the women were euthyroid and were operated for adenoma (381, 70%) or goiter (17, 3%). Forty-seven women were found to have a malignancy, which was papillary adenocarcinoma in 37, follicular cell adenocarcinoma in 9, and anaplastic adenocarcinoma in 1. Miscellaneous indications accounted for the remaining cases. Survival was not impaired in this cohort of women (at 40 years an estimated 37% were still alive compared with an expected 39%; p ⫽ 0.299), who were observed for a total of 12,804 person-years (median 20.9 years per subject). During this period of observation, 261 subjects experienced 601 different fractures (Table 1). Of these, 122 (20%) were due to severe trauma (31 from motor vehicle accidents, 35 due to falling from a height, and 56 from miscellaneous other causes), but the majority of fractures (387, 64%) were attributed to moderate or minimal trauma. Most limb fractures were caused by falling from a standing height or less, whereas most of the vertebral fractures occurred “spontaneously” during the activities of daily living including lifting heavy objects. Twenty-eight fractures were due to a specific pathological process (e.g., metastatic malignancy), and the final 64 fractures could not be associated with a specific etiology. The cumulative incidence of any fracture (at 40 years, an estimated 72%) or any fracture due to moderate trauma (at 40 years, an estimated 33%) is shown in Figure 1. In either instance, fracture risk seemed steady, with a constant increase in the cumulative incidence over time. The number of first fractures observed at each skeletal site, the number expected, and the corresponding SIRs are shown in Table 2. The observed fractures differ somewhat from the numbers in Table 1 because only the first fracture of each type was counted and because person-years and fractures that occurred before age 35 years were excluded from this analysis, which was based on expected incidence rates from the general population of Rochester women age 35 years and above. Among women followed beyond 35 years of age, 255 (49%) experienced at least one fracture, which was not statistically significantly different from 289.7 expected on the basis of overall fracture rates in the community (SIR 0.9, 95% CI 0.8 –1.00). There were, however, significant increases in the risk of vertebral (SIR 2.8, 95% CI 2.3–3.3) and rib fractures (SIR 2.3, 95% CI 1.8 –2.8). The increase in vertebral fracture risk mainly occurred long after the original thyroid surgery (Figure 2A). At 40 years, the cumulative incidence of a vertebral fracture was 44% compared with an expected 19% (p ⱕ 0.001). There were also significant increases in fractures of the pelvis (SIR 3.0, 95% CI 2.1– 4.2) and proximal femur (SIR 1.3, 95% CI 1.01–1.8). After 40 years, the cumulative incidence of a hip fracture was 20% compared with an expected 19% (p ⫽ 0.067) as shown in Figure 2B. These increases were offset somewhat by a significant decrease in fractures of the tibia and fibula (SIR 0.5, 95% CI 0.3– 0.7). There was no increase in distal forearm fractures in this cohort (SIR 1.1, 95% CI 0.8 –1.4); by 40 years after operation, the cumulative incidence was 16% compared with an expected 16% (Figure 2C). Altogether, there was a 20% increase in fractures of the hip, spine, and distal forearm, the fractures that have traditionally been associated with osteoporosis (SIR 1.2, 95% CI 1.02–1.4). The only independent predictors of fracture risk overall or of spine, hip, or distal forearm fractures specifically were age and the presence of one or more of the diseases that have been

Bone Vol. 27, No. 5 November 2000:695–700

L. J. Melton, III et al. Fractures following thyroidectomy

697

Table 1. Distribution of subsequent fractures by skeletal site and cause among Rochester, MN women following thyroidectomy, 1950 –1974 Fracture cause Fall from standing

Severe trauma a

Skeletal site

n

(%)

n

(%)

Skull/face Hands/fingers Distal forearm Shaft/proximal forearm Shaft/distal humerus Proximal humerus Clavicle/scapula/sternum Ribs Thoracic/lumbar vertebrae Other vertebrae Pelvis Proximal femur Shaft/distal femur Patella Tibia/fibula Feet/toes All sites

4 17 9 2 0 7 3 23 7

(44.4) (53.1) (13.9) (20.0) (0.0) (30.4) (23.1) (25.8) (4.2)

3 12 55 7 2 16 4 14 24

2 7 4 0 1 7 29 122

(50.0) (18.9) (6.5) (0.0) (10.0) (29.2) (58.0) (20.3)

2 23 50 6 9 12 7 246

Spontaneous a

a

Pathological a

Uncertain

All causes n

(%)b

(22.2) (9.4) (1.5) (10.0) (0.0) (0.0) (30.8) (36.0) (1.2)

9 32 65 10 2 23 13 89 165

(1.5) (5.3) (10.8) (1.7) (0.3) (3.8) (2.2) (14.8) (27.4)

(0.0) (8.1) (4.8) (0.0) (0.0) (12.5) (20.0) (10.7)

4 37 62 6 10 24 50 601

(0.7) (6.2) (10.3) (1.0) (1.7) (4.0) (8.3) (100.0)

n

(%)

n

(%)

n

(%)

(33.3) (37.5) (84.6) (70.0) (100.0) (69.6) (30.8) (15.7) (14.6)

0 0 0 0 0 0 0 9 120

(0.0) (0.0) (0.0) (0.0) (0.0) (0.0) (0.0) (10.1) (72.7)

0 0 0 0 0 0 2 11 12

(0.0) (0.0) (0.0) (0.0) (0.0) (0.0) (15.4) (12.4) (7.3)

2 3 1 1 0 0 4 32 2

(50.0) (62.2) (80.7) (100.0) (90.0) (50.0) (14.0) (40.9)

0 2 4 0 0 2 4 141

(0.0) (5.4) (6.5) (0.0) (0.0) (8.3) (8.0) (23.5)

0 2 1 0 0 0 0 28

(0.0) (5.4) (1.6) (0.0) (0.0) (0.0) (0.0) (4.7)

0 3 3 0 0 3 10 64

a

a

Percentage (%) of each type of fracture. Percentage (%) of total.

b

associated with secondary osteoporosis, including hyperparathyroidism (n ⫽ 17), Addison’s disease (n ⫽ 1), osteogenesis imperfecta (n ⫽ 1), peptic ulcer disease (n ⫽ 43), gastrectomy (n ⫽ 12), malabsorption syndrome (n ⫽ 8), chronic obstructive lung disease (n ⫽ 42), renal failure (n ⫽ 12), rheumatoid arthritis (n ⫽ 25), hemiplegia/hemiparesis (n ⫽ 16), parkinsonism (n ⫽ 9), and multiple myeloma (n ⫽ 2). For any fracture, the risk rose 60% per decade of age (HR per 10 year increase 1.61, 95% CI 1.46 –1.77) and by a comparable amount in the presence of other diseases (HR 1.56, 95% CI 1.21–2.01). Extent of surgery (total/ subtotal vs. partial) was a predictor of overall fracture risk in univariate (HR 1.76, 95% CI 1.10 –2.82) but not multivariate analyses. None of the other surgery-specific variables was a predictor of overall fracture risk. After adjusting for age, overall fracture risk was not significantly increased in the 85 women (13%) who were hyperthyroid at the time of surgery (HR 1.08, 95% CI 0.72–1.60), nor significantly decreased among the 3 women (0.5%) who were hypothyroid (HR 0.95, 95% CI 0.13– 6.84). Postoperative thyroid replacement therapy was needed by

Figure 1. Cumulative incidence of any new fracture or any fracture due to moderate trauma, among Rochester, Minnesota women following thyroidectomy, 1950 –1974.

299 women (48%), but was not associated with any increase in fracture risk (HR 0.88, 95% CI, 0.69 –1.12). Of the 47 women operated for thyroid malignancy, 43 were put on suppressive doses of thyroid (usually 0.15– 0.20 mg levothyroxine per day). Even at these relatively high doses, thyroid replacement was not associated with an increase in age-adjusted fracture risk (HR 1.42, 95% CI 0.84 –2.41). The only independent predictor of vertebral fracture was age (HR per 10 year increase 2.14, 95% CI 1.82–2.52), although secondary osteoporosis was a risk factor in the univariate analysis (HR 1.62, 95% CI 1.13–2.33). Both age (HR per 10 year increase 3.19, 95% CI 2.39 – 4.27) and secondary osteoporosis (HR 2.37, 95% CI 1.38 – 4.05) were independent Table 2. Observed (Obs) fractures following thyroidectomy 1950 –1974 with the expected numbers (Exp) and standardized incidence ratios (SIRs) among Rochester, MN women

Skeletal site

Obs

Exp

SIR

95% confidence interval

Skull/face Hands/fingers Distal forearm Shaft/proximal forearm Shaft/distal humerus Proximal humerus Clavicle/scapula/sternum Ribs Thoracic/lumbar vertebrae Other vertebrae Pelvis Proximal femur Shaft/distal femur Patella Tibia/fibula Feet/toes Any site

9 26 52 9 2 21 10 74 124

7.17 26.04 47.23 13.3 5.25 21.35 9.77 32.76 44.76

1.3 1.0 1.1 0.7 0.4 1.0 1.0 2.3 2.8

0.6–2.4 0.7–1.5 0.8–1.4 0.3–1.3 0.1–1.4 0.6–1.5 0.5–1.9 1.8–2.8a 2.3–3.3a

4 35 54 6 10 23 43 255

2.53 11.62 40.25 7.64 7.94 47.10 47.37 289.7

1.6 3.0 1.3 0.8 1.3 0.5 0.9 0.9

0.4–4.1 2.1–4.2a 1.01–1.8a 0.3–1.7 0.6–2.3 0.3–0.7a 0.7–1.2 0.8–1.00

a

p ⬍ 0.05.

698

L. J. Melton, III et al. Fractures following thyroidectomy

Figure 2. Observed and expected cumulative incidence of vertebral fracture (A), proximal femur fracture (B), and distal forearm fracture (C) among Rochester, Minnesota women following thyroidectomy, 1950 – 1974.

predictors of hip fracture. Distal forearm fractures were associated only with age (HR per 10 year increase 1.51, 95% CI 1.22–1.85). Discussion In this population-based study, the large inception cohort comprised of all Rochester women who underwent thyroidectomy from 1950 to 1974 was followed for nearly 13,000 person-years, on average more than 20 years per subject. Over this long period of observation, there was no overall increase in fracture incidence compared with expected rates from the general population of Rochester. The cumulative incidence of fractures in this cohort

Bone Vol. 27, No. 5 November 2000:695–700

increased steadily over time and was most closely associated with increasing age, a well-described risk factor for fractures.16 Like other investigators,13,15,34 we saw no increase in distal forearm fractures among women following thyroidectomy. On the other hand, the risk of a subsequent hip fracture was increased by 30%, a difference that was statistically significant. The only other study that evaluated the effect of thyroidectomy found no association with hip fracture risk,36 but detailed data were not reported and the number of affected subjects was small. This figure contrasts with the fivefold excess risk of hip fracture attributable to thyroidectomy that we previously observed among Rochester men.27,28 Most32–34 but not all4 retrospective studies, including our own,20 have found no association between thyroid disease and vertebral fractures. In this large cohort study, by contrast, we observed an almost threefold increase in the likelihood of a subsequent vertebral fracture following thyroidectomy. There were also significant increases in the risk of rib and pelvis fractures. Substantial increases in rib, spine, and pelvis fractures, but not forearm or hip fractures, could conceivably be explained by ascertainment bias if women were subjected to closer surveillance following thyroidectomy. Although we have no evidence that this is so, the potential exists because the incidence rates used in these calculations were based on clinically recognized vertebral fractures. These rates provide a low estimate of vertebral fracture occurrence in the population3 and might cause the SIR to be overstated if women with thyroidectomy were followed more closely for vertebral fractures than community women in general. An alternative approach is to use incidence rates derived from prevalence data,24 but these overestimate vertebral fracture incidence in the general population because some vertebral deformities do not represent actual fractures. If the latter rates are substituted in the analysis, however, the relative risk of vertebral fracture is no longer elevated (SIR 1.0, 95% CI 0.8 –1.2). Fracture risk is reportedly increased among patients with hyperthyroidism in most studies.1,4,7,9,28,36 The reduction in bone density associated with hyperthyroidism is largely reversed by treatment,10 but transient changes could be due to increased bone turnover,26 which is an important independent risk factor for vertebral fractures.23,30 In this study, however, the increase in vertebral fracture risk was not seen around the time of the thyroidectomy but only became evident decades later. Increased bone turnover would also exacerbate bone loss from other causes,29 and this is more likely to be a problem among women. In fact, we saw no increase in vertebral fracture risk among Rochester men following thyroidectomy.27 Others have suggested that the increased fracture risk in women with hyperthyroidism is independent of bone density2,7 and possibly related to impaired neuromuscular function, inducing a greater likelihood of falling.1 Falls were, indeed, the precipitating cause of most limb fractures in this population, but the overall risk of limb fractures was not increased following thyroidectomy. Moreover, only 15% of the vertebral fractures resulted from falls; most of them occurred during the activities of daily living as would be expected.3 Instead, fracture risk was associated with underlying diseases (other than hyperthyroidism) that have been linked to secondary osteoporosis.14 This raises the possibility of confounding—that is, that fracture risk is related to these other conditions which are also associated with thyroid disease. The present study has a number of strengths. This large population-based inception cohort represents the complete spectrum of women undergoing thyroidectomy in the community. During the extensive follow-up, many fractures were observed, and these were documented in detailed medical records that spanned each subject’s entire period of residency in the commu-

Bone Vol. 27, No. 5 November 2000:695–700

nity. Because the vast majority of fractures come to medical attention,22 ascertainment should be nearly complete, with the possible exception of vertebral fractures.3 There are also corresponding limitations of this retrospective approach. Risk factors for falls could not be thoroughly assessed and special measurements, such as bone densitometry, were not routinely performed. Also, our results are not generalizable to nonwhites, because the Olmsted County population is largely white and slightly younger than United States whites in general,17 although age-adjusted hip fracture incidence rates from Rochester are quite comparable to those for United States whites.18 These considerations had little influence on our main conclusion, that the long-range impact of thyroidectomy on overall fracture risk in women is negligible. The majority of thyroidectomies were for adenoma and goiter, however, and we cannot exclude the possibility of adverse effects following thyroidectomy for other causes, although no independent influence of hyperthyroidism on fracture risk was seen in our multivariate analysis. Thyroid replacement therapy was not an independent predictor of fracture risk either, even though 48% of the women were exposed postoperatively. This observation is consistent with the recent review of a large and somewhat discrepant literature, which concluded that there was little evidence for adverse skeletal effects of thyroid replacement but that thyroid suppression might be associated with moderate bone loss in postmenopausal women.10 There was no significant increase in fracture risk among those on suppressive doses of thyroid hormone in this study, but the number of affected women was small. In a retrospective study such as this, we were unable to directly address pathogenetic mechanisms, but the main influence of thyroid disease may be to alter bone turnover and thus to increase (hyperthyroidism) or decrease (hypothyroidism) any underlying imbalance in bone formation and resorption.8 Among the women who undergo thyroidectomy, this process appears to be neither severe enough nor long lasting enough to have a great influence on fracture risk later in life.

L. J. Melton, III et al. Fractures following thyroidectomy

8. 9.

10. 11. 12. 13.

14.

15.

16.

17. 18. 19.

20.

21. 22.

Acknowledgments: The authors thank Judy Bruen, Judy Stancl, and Celia Wright for their help in data collection, and Mary Roberts for assistance in preparing the manuscript. This project was supported in part by Grants AG-04875 and AR-30582 from the National Institutes of Health, United States Public Health Service.

23.

24.

References 1. Bauer, D. C., Black, D. M., Ashby, M. A., Cummings, S. R. for the Fracture Intervention Trial Research Group. Thyroid function and bone mass of the hip in postmenopausal women. In: Christiansen, C. and Riis, B., Eds. Proceedings of the Fourth International Symposium on Osteoporosis and Consensus Development Conference, Hong Kong. Aalborg, Denmark: Handelstrykkeriet Aalborg; 1993; 170 –171. 2. Bauer, D. C., Cummings, S. R., Tao, J. L., Browner, W. S., and the Study of Osteoporotic Fractures Research Group. Hyperthyroidism increases the risk of hip fractures. A prospective study. J Bone Miner Res 7(Suppl. 1):S121; 1992. 3. Cooper, C., Atkinson, E. J., O’Fallon, W. M., and Melton, L. J., III. Incidence of clinically diagnosed vertebral fractures: A population-based study in Rochester, Minnesota, 1985–1989. J Bone Miner Res 7:221–227; 1992. 4. Cooper, C., Shah, S., Hand, D. J., Adams, J., Compston, J., Davie, M., and Woolf, A. Screening for vertebral osteoporosis using individual risk factors. The Multicentre Vertebral Fracture Study Group. Osteopor Int 2:48 –53; 1991. 5. Cox, D. R. Some simple approximate tests for Poisson variates. Biometrika 40:354 –360; 1953. 6. Cox, D. R. Regression models and life-tables (with discussion). J R Stat Soc 34B:187–220; 1972. 7. Cummings, S. R., Nevitt, M. C., Browner, W. S., Stone, K., Fox, K. M.,

25. 26.

27.

28. 29. 30.

31. 32.

699

Ensrud, K. E., Cauley, J., Black, D., and Vogt, T. M. Risk factors for hip fracture in white women. The Study of Osteoporotic Fractures Research Group. N Engl J Med 332:767–773; 1995. Eriksen, E. F., Mosekilde, L., and Melsen, F. Trabecular bone remodeling and bone balance in hyperthyroidism. Bone 6:421– 428; 1985. Gallagher, J. C., Melton, L. J., III, and Riggs, B. L. Examination of prevalence rates of possible risk factors in a population with a fracture of the proximal femur. Clin Orthop 153:158 –165; 1980. Greenspan, S. L. and Greenspan, F. S. The effect of thyroid hormone on skeletal integrity. Ann Intern Med 130:750 –758; 1999. Kalbfleisch, J. D. and Prentice, R. L. The Statistical Analysis of Failure Time Data. New York: Wiley; 1980; 1–321. Kaplan, E. L. and Meier, P. Non-parametric estimation from incomplete observations. J Am Stat Assoc 53:457– 481; 1958. Kelsey, J. L., Browner, W. S., Seeley, D. G., Nevitt, M. C., and Cummings, S. R. Risk factors for fractures of the distal forearm and proximal humerus. The Study of Osteoporotic Fractures Research Group. Am J Epidemiol 135:477– 489; 1992. Khosla, S. and Melton, L. J., III. Secondary osteoporosis. In: Riggs, B. L. and Melton, L. J., III, Eds. Osteoporosis: Etiology, Diagnosis, and Management, 2nd Ed. Philadelphia: Lippincott-Raven, 1995; 183–204. Mallmin, H., Ljunghall, S., Persson, I., and Bergstro¨m, R. Risk factors for fractures of the distal forearm: A population-based case-control study. Osteopor Int 4:298 –304; 1994. Melton, L. J., III. Epidemiology of fractures. In: Riggs, B. L. and Melton, L. J., III, Eds. Osteoporosis: Etiology, Diagnosis, and Management, 2nd Ed. Philadelphia: Lippincott-Raven; 1995; 225–247. Melton, L. J., III. History of the Rochester Epidemiology Project. Mayo Clin Proc 71:266 –274; 1996. Melton, L. J., III. Lifetime risk of a hip fracture. Am J Publ Health 80:500 –501; 1990. Melton, L. J., III, Amadio, P. C., Crowson, C. S., and O’Fallon, W. M. Long-term trends in the incidence of distal forearm fractures. Osteopor Int 8:341–348; 1998. Melton, L. J., III, Atkinson, E. J., Khosla, S., O’Fallon, W. M., and Riggs, B. L. Secondary osteoporosis and the risk of vertebral deformities in women. Bone 24:49 –55; 1999. Melton, L. J., III, Atkinson, E. J., and Madhok, R. Downturn in hip fracture incidence. Publ Health Rep 111:146 –150; 1996. Melton, L. J., Crowson, C. S., and O’Fallon, W. M. Fracture incidence in Olmsted County, Minnesota: Comparison of urban with rural rates and changes in urban rates over time. Osteopor Int 9:29 –37; 1999. Melton, L. J., III, Khosla, S., Atkinson, E. J., O’Fallon, W. M., and Riggs, B. L. Relationship of bone turnover to bone density and fractures. J Bone Miner Res 12:1083–1091; 1997. Melton, L. J., III, Lane, A. W., Cooper, C., Eastell, R., O’Fallon, W. M., and Riggs, B. L. Prevalence and incidence of vertebral deformities. Osteopor Int 3:113–119; 1993. Melton, L. J., III, Sampson, J. M., Morrey, B. F., and Ilstrup, D. M. Epidemiologic features of pelvic fractures. Clin Orthop 155:43– 47; 1981. Mosekilde, L., Eriksen, E. F., and Charles, P. Effects of thyroid hormones on bone and mineral metabolism. Endocrinol Metab Clin N Am 19:35– 63; 1990. Nguyen, T. T., Heath, H., III, Bryant, S. C., O’Fallon, W. M., and Melton, L. J., III. Fractures after thyroidectomy in men: A population-based cohort study. J Bone Miner Res 12:1092–1099; 1997. Poo´r, G., Atkinson, E. J., O’Fallon, W. M., and Melton, L. J., III. Predictors of hip fractures in elderly men. J Bone Miner Res 10:1900 –1907; 1995. Riggs, B. L. and Melton, L. J., III. Medical progress: Involutional osteoporosis. N Engl J Med 314:1676 –1686; 1986. Riggs, B. L., Melton, L. J., III, and O’Fallon, W. M. Drug therapy for vertebral fractures in osteoporosis: Evidence that decreases in bone turnover and increases in bone mass both determine antifracture efficacy. Bone 18(Suppl.): 197S–201S; 1996. Rose, S. H., Melton, L. J., III, Morrey, B. F., Ilstrup, D. M., and Riggs, B. L. Epidemiologic features of humeral fractures. Clin Orthop 168:24 –30; 1982. Santavirta, S., Konttinen, Y. T., Helio¨vaara, M., Knekt, P., Lu¨thje, P., and Aromaa, A. Determinants of osteoporotic thoracic vertebral fracture: Screening of 57,000 Finnish women and men. Acta Orthop Scand 63:198 –202; 1992.

700

L. J. Melton, III et al. Fractures following thyroidectomy

33. Seeman, E., Melton, L. J., III, O’Fallon, W. M., and Riggs, B. L. Risk factors for spinal osteoporosis in men. Am J Med 75:977–983; 1983. 34. Solomon, B. L., Wartofsky, L., and Burman, K. D. Prevalence of fractures in postmenopausal women with thyroid disease. Thyroid 3:17–23; 1993. 35. Suwanwalaikorn, S. and Baran, D. Thyroid hormone and the skeleton. In: Marcus, R., Feldman, D., and Kelsey, J., Eds. Osteoporosis. San Diego, CA: Academic Press; 1996; 855– 861. 36. Wejda, B., Hintze, G., Katschinski, B., Olbricht, T., and Benker, G. Hip

Bone Vol. 27, No. 5 November 2000:695–700 fractures and the thyroid: A case-control study. J Intern Med 237:241–247; 1995.

Date Received: February 28, 2000 Date Revised: June 29, 2000 Date Accepted: July 18, 2000