Fracture Risk After Bilateral Oophorectomy in Elderly Women*
L JOSEPH MELTON III,1,2SUNDEEP KHOSLA,2GEORGE D MALKASIAN,3SARA J ACHENBACH,4
ANN L OBERG,4and B LAWRENCE RIGGS2
Elderly women with the lowest serum estrogen levels are at the greatest risk of bone loss and fractures, but
it is controversial whether the ovaries contribute to estrogen production after menopause, and therefore,
whether bilateral oophorectomy in postmenopausal women might have adverse skeletal effects. To address
this potential problem, we estimated long-term fracture risk among 340 postmenopausal Olmsted County,
MN, women who underwent bilateral oophorectomy for a benign ovarian condition in 1950–1987. In over
5632 person-years of follow-up (median, 16 years per subject), 194 women experienced 516 fractures (72%
from moderate trauma). Compared with expected rates, there was a significant increase in the risk of any
osteoporotic fracture (moderate trauma fractures of the hip, spine, or distal forearm; standardized incidence
ratio [SIR], 1.54; 95% CI, 1.29–1.82) but almost as large an increase in fractures at other sites (SIR, 1.35; 95%
CI, 1.13–1.59). In multivariate analyses, the independent predictors of overall fracture risk were age,
anticonvulsant or anticoagulant use for >6 months, and a history of alcoholism or prior osteoporotic fracture;
obesity was protective. Estrogen replacement therapy was associated with a 10% reduction in overall fracture
risk (hazard ratio [HR], 0.90; 95% CI, 0.64–1.28) and a 20% reduction in osteoporotic fractures (HR, 0.80;
95% CI, 0.52–1.23), but neither was statistically significant. The increase in fracture risk among women who
underwent bilateral oophorectomy after natural menopause is consistent with the hypothesis that androgens
produced by the postmenopausal ovary are important for endogenous estrogen production that protects
against fractures. (J Bone Miner Res 2003;18:900–905)
Key words: epidemiology, hip fracture, oophorectomy, osteoporosis, vertebral fracture
by estrogen replacement therapy.(1)More recently, it be-
came evident that this rapid phase of bone loss, which can
last for up to 8–10 years, is followed by a slower phase of
age-related bone loss that continues indefinitely.(2)Convinc-
ing evidence has now emerged that this continuing slow
phase of bone loss is caused in large part by estrogen
deficiency, through effects on the gut and kidney that impair
intestinal calcium absorption and renal calcium conserva-
tion.(3)Because testosterone and androstenedione produced
by the ovary can serve as substrate for extragonadal endog-
T WAS SHOWN OVER 50 years ago that menopause is asso-
ciated with a period of rapid bone loss that is preventable
enous estrogen production after menopause,(3)there could
be effects on the skeleton of bilateral oophorectomy later in
life, even after the rapid phase of bone loss has ceased.
Little attention has been given to this potential problem, and
there is controversy of whether the postmenopausal ovary
is(4–8)or is not(9)an important source of these androgens.
However, population-based studies have shown that even
slightly lower levels of circulating estrogens are associated
with increased bone loss and fracture risk in postmeno-
pausal women.(10–15)To the extent that ovarian androgens
make a contribution to endogenous estrogen production
after menopause, there may be unexpected adverse conse-
quences of oophorectomy in elderly women. The purpose of
this study was to assess the practical significance of this
potential problem by estimating the long-term risk of frac-
tures among an inception cohort of women who were al-
ready past menopause when they underwent bilateral oo-
phorectomy for a benign ovarian condition.
*Presented in part at the 24th Annual Meeting of the American
Society of Bone and Mineral Research, San Antonio, TX, Septem-
ber 20–24, 2002.
The authors have no conflict of interest.
1Division of Epidemiology, Department of Health Sciences Research, Mayo Clinic and Mayo Foundation, Rochester, Minnesota, USA.
2Division of Endocrinology, Metabolism and Nutrition, Department of Internal Medicine, Mayo Clinic and Mayo Foundation,
Rochester, Minnesota, USA.
3Department of Obstetrics and Gynecology, Mayo Clinic and Mayo Foundation, Rochester, Minnesota, USA.
4Division of Biostatistics, Department of Health Science Research, Mayo Clinic and Mayo Foundation, Rochester, Minnesota, USA.
JOURNAL OF BONE AND MINERAL RESEARCH
Volume 18, Number 5, 2003
© 2003 American Society for Bone and Mineral Research
MATERIALS AND METHODS
This study was possible because medical care for the
residents of Rochester, Minnesota, is delivered almost ex-
clusively by the Mayo Clinic and a small number of other
providers. Most providers use a dossier (or unit) medical
record system, wherein all data on each patient are con-
tained in a single file. The unit records at Mayo, for exam-
ple, contain the information recorded for outpatient office or
clinic consultations, emergency room and nursing home
care, inpatient hospitalizations, autopsy, and death certifi-
cation. These records are accessible because Mayo Clinic
maintains a master index to all diagnoses and surgical
procedures recorded among its patients. The Rochester Ep-
idemiology Project supports a similar index to the records of
the other medical care providers serving the community.(16)
The original medical records have been preserved and are
easily retrieved for review. The result is documentation of
the medical care provided to Rochester residents by Mayo
Clinic or its two large affiliated hospitals (Saint Marys and
Rochester Methodist), the Olmsted Medical Group and its
affiliated Olmsted Community Hospital (Olmsted Medical
Center), and other hospitals and practitioners in the vicinity.
The potential of this data system for population-based stud-
ies has been described previously.(16)
After approval by Mayo’s Institutional Review Board, all
Rochester women who underwent an oophorectomy during
the 38-year period, 1950–1987, were identified using the
indexes described.(17)Oophorectomy was defined as com-
plete removal of the ovary and was classified as unilateral or
bilateral (including a second unilateral oophorectomy).
Hysterectomy was recorded, if done, but women who had
only a hysterectomy and no oophorectomy were excluded.
The indication for oophorectomy was determined by review
of each woman’s complete inpatient and outpatient medical
record, including the surgeon’s narrative report of the op-
eration. We previously reported on subsequent fracture risk
among the 463 Rochester women who were still menstru-
ating when they had their bilateral oophorectomy.(18)This
study focuses on 545 postmenopausal women who under-
went bilateral oophorectomy; 195 of them had ovarian
cancer. This analysis was restricted to the 350 women who
were already postmenopausal when they underwent bilat-
eral (or second unilateral) oophorectomy for a benign ovar-
ian condition. However, 10 of these women were excluded
from further study because they had not provided an autho-
rization for review of their medical records for research in
accordance with Minnesota State law.(19)
After additional approval by Mayo’s Institutional Review
Board, the remaining 340 women were followed forward in
time through their linked medical records in the community
(retrospective cohort study). Each subject’s complete inpa-
tient and outpatient medical record at each local provider of
medical care was searched by trained nurse abstractors for
the occurrence of any fracture. Follow-up continued until
death or the most recent clinical contact documented in
community medical records. The records contained the clin-
ical history and the radiologist’s report of each fracture, but
the original roentgenograms were not available for review.
Consequently, the diagnosis of vertebral fracture was ac-
cepted based on a radiologist’s report of compression or
collapse of one or more thoracic or lumbar vertebra.(20)
Fracture ascertainment is believed to be complete except for
vertebral fractures, some of which are never diagnosed.
Fractures were also classified according to the circum-
stances 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.
The influence of bilateral oophorectomy on fracture risk
was evaluated using three basic methods of analysis. In the
primary analysis, we calculated standardized incidence ra-
tios (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 during
follow-up in the community. Expected numbers of fractures
were derived by applying age- and sex-specific incidence
rates of these fractures in the Rochester population(20–24)to
the age-specific person-years of follow-up among the oo-
phorectomy cases. Ninety-five percent confidence intervals
(95% CI) for the SIRs were calculated with the assumption
that the expected rates are fixed and the observed fractures
follow a Poisson distribution.(25)
In the second method of analysis, the cumulative inci-
dence of a new fracture (1 ? survival-free-of-fracture) was
projected for up to 30 years after the index date, using
product-limit life table methods.(26)Two approaches were
used in this analysis. The customary analysis censored
women who died and therefore estimated cumulative frac-
ture incidence among survivors at each point in time. Al-
ternatively, we estimated fracture risk with death as a com-
peting event, thereby providing a cumulative incidence
figure more concordant with that actually observed by the
treating physician. The one-sample log-rank test was used to
compare observed and expected cumulative fracture inci-
dence curves.(27)Life-table methods were used to assess
survival, and the observed and expected survival curves
were also compared using the log-rank test statistic.
Finally, Cox proportional hazards models(28)were used to
assess the impact of various covariates on the subsequent
risk of fracture after bilateral oophorectomy. Univariate
relationships between the risk of specific fractures and each
clinical characteristic under consideration were first as-
sessed. Stepwise methods with forward selection and back-
ward elimination were then used to choose independent
variables for the final models. The dependent variable was
time until the first new fracture, and the independent vari-
ables were the clinical characteristics. For the final multiple
models, as well as for the univariate models, the assumption
of proportional hazards was examined and was not violated
for the variables considered.
In the 38-year period, 1950–1987, 340 Rochester, Min-
nesota, women had a bilateral oophorectomy, which was
unrelated to primary ovarian cancer (median age at surgery,
62 years; range, 37–86 years), after natural menopause
(median age at menopause, 50 years). All but four of them
were white, reflecting the racial composition of the com-
901 LATE OOPHORECTOMY AND FRACTURES
munity (98% white in 1980). The majority of operations,
292 (86%), were incidental in the course of total abdominal
hysterectomy. Fifteen others (4%) were elective procedures
with other types of surgery, whereas 31 women (9%) were
operated for a variety of benign conditions (e.g., endome-
triosis). The final two women were castrated for manage-
ment of breast cancer. Altogether, 314 (92%) of the women
underwent hysterectomy at the time of bilateral oophorec-
tomy, whereas 6 others had a hysterectomy previously and
4 had one afterward. The hysterectomy status of 16 women
These 340 women were followed subsequently for 5632
person-years (range, 1 day to 42 years per subject). Survival
was unimpaired in this cohort (at 30 years, 30% remained
alive compared with an expected 27%; p ? 0.826), and
follow-up was complete to death in 59% of the subjects
(median, 14 years of follow-up per subject). Among survi-
vors, the median duration of follow-up was 17 years. During
this entire period of observation, 194 subjects experienced
516 different fractures (Table 1). Eighty-two women had a
single fracture, whereas 39 had two different fractures, 25
had three fractures, and 48 had four or more fractures. After
30 years of follow-up, the cumulative incidence of a new
fracture was 67% compared with an expected 60% (p ?
0.001). Among 30-year survivors, an actuarially estimated
86% of the women had experienced at least one new frac-
ture compared with 76% expected (p ? 0.001). Eighty-two
of the fractures (16%) were caused by severe trauma (26
motor vehicle accidents, 24 falls from a height, and 32
miscellaneous causes, e.g., hand crushed in car door), but
the majority of fractures (369, 72%) were attributed to
minimal or moderate trauma (Table 1). Excluding hands and
feet, 77% of the limb fractures were caused by a fall from
standing height or less, which accounted for 44% of all
fractures observed. Three-fourths of all vertebral fractures
occurred “spontaneously” in the course of everyday activi-
ties. Twenty-two fractures (4%) were the result of a specific
pathological lesion (e.g., metastatic malignancy), and the
etiology of the remaining 43 fractures was uncertain.
Skeletal site-specific data are delineated in Table 2. Com-
pared with expected rates among community women gen-
erally, statistically significant increases were seen for frac-
tures at most sites in the axial skeleton, such as the vertebrae
(SIR, 3.01; 95% CI, 2.46–3.64) and ribs (SIR, 2.24; 95%
TABLE 1. DISTRIBUTION OF SUBSEQUENT FRACTURES AMONG 340 ROCHESTER, MINNESOTA, WOMEN WHO UNDERWENT BILATERAL (OR SECOND
UNILATERAL) OOPHORECTOMY AFTER NATURAL MENOPAUSE IN 1950–1987, BY FRACTURE SITE AND CAUSE
standing Spontaneous Pathological Uncertain All causes
* Percentage of each type of fracture.
†Percentage of total.
TABLE 2. NUMBER OF SUBSEQUENT FRACTURES OBSERVED (OBS)
AND EXPECTED (EXP), WITH STANDARDIZED INCIDENCE RATIOS
(SIRS) AND 95% CONFIDENCE INTERVALS (CI) AROUND THE RATIOS,
AMONG 340 ROCHESTER, MINNESOTA, WOMEN WHO UNDERWENT
BILATERAL (OR SECOND UNILATERAL) OOPHORECTOMY AFTER
NATURAL MENOPAUSE IN 1950–1987
Skeletal siteOBS*EXP SIR (95% CI)
* The number of fractures observed at these specific skeletal sites differs
from those reported in Table 1 because only the first fracture of each type
per patient was counted in this analysis.
†p ? 0.05.
902MELTON ET AL.
CI, 1.69–2.92), as well as the proximal femur (SIR, 1.54;
95% CI, 1.17–1.99) and distal forearm (SIR, 1.47; 95% CI,
1.06–1.98). As shown in Fig. 1, the increase in vertebral
fracture risk seemed to accelerate over follow-up, but the
ratio of observed to expected cumulative incidence did not
actually change. In contrast, the excess of proximal femur
and distal forearm fractures was relatively constant over
time (Fig. 1). Overall, there was a significant increase in the
risk of any of the traditional osteoporotic fractures (moder-
ate trauma fractures of the hip, spine, or distal forearm ?35
years of age: SIR, 1.54; 95% CI, 1.29–1.82), but almost as
large an increase was seen in the likelihood of a fracture at
any of the remaining skeletal sites (SIR, 1.35; 95% CI,
In multivariate analyses, advancing age was the strongest
risk factor for any fracture (hazard ratio [HR] per 10-year
increase, 1.60; 95% CI, 1.35–1.89) or for any osteoporotic
fracture (HR, 1.85; 95% CI, 1.48–2.31). There was no
suggestion of an association with earlier age or shorter
duration since menopause nor any significant relation to the
indication for oophorectomy (ovarian pathology vs. inci-
dental), or in the latter group, to the indication for concom-
itant hysterectomy (uterine descensus vs. endometrial can-
cer vs. other vaginal bleeding vs. fibroids vs. other
indications). The other independent predictors of any sub-
sequent fracture were use of anticonvulsants or anticoagu-
lants for more than 6 months and a history of alcoholism or
prior osteoporotic fractures, while obesity was protective
(Table 3). As also shown in Table 3, the other risk factors
for any osteoporotic fracture included anticonvulsant use, a
history of prior osteoporotic fracture or kyphosis, and more
recent year of oophorectomy, while thiazide use was pro-
tective. Fifty-one women (15%) had been on hormone re-
placement therapy (80% of them for at least 6 months)
before the oophorectomy and over one-half continued this
treatment after surgery. An additional 61 women (14%)
were put on hormone replacement for the first time after
their oophorectomy, but the median delay to initiating treat-
ment was 5.7 years; only 13 women (4%) were started
within 6 months of surgery. In a univariate model, hormone
replacement therapy, when handled as a time-dependent
variable, was associated with a 10% reduction in overall
fracture risk (HR, 0.90; 95% CI, 0.64–1.28) and a 20%
reduction in osteoporotic fractures (HR, 0.80; 95% CI,
0.52–1.23), but neither difference was statistically signifi-
In this population-based study, women who had a bilat-
eral oophorectomy on average 14 years after natural meno-
pause experienced a 32% increase in subsequent overall
fracture risk and a 54% increase in the fractures traditionally
associated with osteoporosis. This finding is consistent with
the notion that postmenopausal women experience reduc-
tions in circulating testosterone and androstenedione levels
after the removal of their ovaries.(4–8)The androgen reduc-
tions themselves could have an independent adverse effect
on bone,(3)but more importantly, they are aromatized to
estrogens systemically in fat and locally in bone tissue.(29,30)
To the extent that endogenous estrogen production is re-
duced even slightly,(31)postmenopausal bone turnover
might be exacerbated and fracture risk increased.(32,33)Al-
though most authors have concluded that circulating estro-
gen levels are not lowered after oophorectomy, in fact, small
reductions are typically observed, though the differences may
not be statistically significant.(5–8,34)Similarly, in an age-
stratified sample of 213 postmenopausal Rochester women,
we found that those who had undergone bilateral oophorec-
tomy but were not on estrogen replacement therapy, when
compared with untreated women with intact ovaries, had
lower serum total estradiol (13.6 vs. 24.1 pg/ml) and estrone
(25.9 vs. 32.7 pg/ml) levels (LJ Melton, III, unpublished
data, 2002). These reductions were not statistically signifi-
cant given the relatively small number of subjects, but
differences of this magnitude in otherwise low circulating
estrogen levels seem to adversely influence bone metabo-
lism in most,(10–15)if not all,(35,36)recent studies of post-
fractures of the vertebrae (p ? 0.001), proximal femur (p ? 0.001), and
distal forearm (p ? 0.012) among 340 Rochester, Minnesota, women
who underwent bilateral (or second unilateral) oophorectomy after
natural menopause in 1950–1987. Death was considered a competing
event in this analysis.
Observed and expected cumulative incidence of subsequent
TABLE 3. HAZARDS RATIOS (HR) FOR THE DEVELOPMENT OF ANY
SUBSEQUENT FRACTURE OR ANY OSTEOPOROTIC FRACTURE* AMONG
340 ROCHESTER, MINNESOTA, WOMEN WHO UNDERWENT BILATERAL
(OR SECOND UNILATERAL) OOPHORECTOMY AFTER NATURAL
MENOPAUSE IN 1950–1987
[HR (95% CI)]
Age (per 10 years)
Anticonvulsants ?6 months
Anticoagulants ?6 months
Thiazide diuretics ?6
Obesity (BMI ? 27.3 kg/m2)
Prior osteoporotic fracture
* Fracture of the hip, spine, or distal forearm caused by moderate trauma
at age 35 years or older.
903LATE OOPHORECTOMY AND FRACTURES
Conversely, 50% of the women with a history of bilateral
oophorectomy in the Rochester population sample were on
estrogen replacement compared with only 14% of the un-
operated women, an association observed previously,(37)
and estrogen levels among the treated women were practi-
cally identical. In the present study, 27% of the women were
on estrogen replacement therapy after bilateral oophorec-
tomy, and the effect on fracture risk was generally positive
(a 20% reduction in osteoporotic fractures and a 10% re-
duction in fractures overall), although not statistically sig-
nificant. Randomized controlled clinical trials, on the other
hand, have shown that bone loss can be slowed in elderly
women(38,39)and fracture risk reduced(40)by hormone re-
In addition to the well-known effects of age,(41)the other
risk factors identified here have been associated with frac-
ture risk in previous studies. Thus, a prior history of frac-
tures has been shown to be a strong predictor of future osteo-
porotic fracture risk,(42)while anticonvulsant use has been
associated with an increased risk of fractures generally,(43)
probably through a relationship with seizure disorders and
falling. Anticoagulants also had an effect on overall fracture
risk in this study, but other investigators have not found
associations except with vertebral and rib fractures.(44,45)
Alcoholism has been linked to an increase in all sorts of
fractures as we also observed.(46)The use of thiazide diuret-
ics was linked with a 40% reduction in osteoporotic frac-
tures, but we saw no independent protective effect of thia-
zides on specific fractures or on overall fracture risk in
accordance with most other work on the subject.(47)Obesity
was protective for fractures in general as shown in a host of
earlier studies,(47)and this may relate to the role of fat mass
as a site for aromatization of androgens to estrogens in
The present investigation has a number of strengths. The
population-based retrospective cohort design allowed us to
rapidly estimate long-term fracture outcomes among un-
selected community women undergoing bilateral oophorec-
tomy after the menopause. The clinical characteristics of
these women were recorded before any knowledge of re-
sultant fractures, and a large number of such fractures were
documented during extensive follow-up in the detailed med-
ical records that spanned each subject’s entire period of
residency in the community. Because the vast majority of
fractures come to medical attention,(24)ascertainment
should be nearly complete with the possible exception of
vertebral fractures.(20)Although our results are not gener-
alizable to non-whites, because the Rochester population is
largely white,(16)age-adjusted hip fracture incidence rates
from Rochester are comparable with those for United States
whites in general.(22)There are also corresponding limita-
tions of an observational study based on medical records.
Risk factors for falls could not be fully assessed, and mea-
surements of the variables of special interest (e.g., bone
density, circulating estrogen levels) were not routinely per-
formed. Moreover, these data do not allow direct assess-
ment of any pathophysiologic mechanism for the increase in
fracture risk, such as elevated bone resorption caused by
As in most populations in this country,(48)the great ma-
jority of bilateral oophorectomies in Rochester were per-
formed incidentally in the course of hysterectomy. Concur-
rent oophorectomy is technically simple, does not increase
short-term morbidity, and limits the risk of ovarian cancer,
although only 2.6 ovarian cancers would have been ex-
pected in this cohort, given the incidence rates in this
community.(49)Consequently, this practice is recommended
routinely for women undergoing hysterectomy after 50
years of age,(50)although some have argued that incidental
oophorectomy should be re-evaluated based on the obser-
vation that androgen levels are reduced in these women.(8,51)
On the other hand, highly detailed studies of a small number
of women suggest that the androgens in postmenopausal
women are not ovarian but rather adrenal in origin and that
prophylactic oophorectomy need not be abandoned.(9)Re-
gardless of the underlying pathophysiology involved, we
have documented a significant increase in fracture risk
among these women, which may be reduced somewhat by
estrogen replacement therapy. Although recent results from
the Women’s Health Initiative cast doubt on the wisdom of
using combined estrogen and progestin therapy for this
purpose,(40)the estrogen-only arm of that trial continues,
and it is the latter results that are relevant to these women,
almost all of whom are without a uterus. Further research
will be needed to define the impact of postmenopausal
oophorectomy on androgen and estrogen metabolism and its
effects on bone pathophysiology and to devise the most
suitable strategies for the care of women who have under-
gone bilateral oophorectomy at an older age.
The authors thank Leona Bellrichard, RN, Barbara Nolte,
RN, and Kristine Otto-Higgins, RN for their help with data
collection, and Mary Roberts for preparing the manuscript.
This work was supported in part by Grants AG-04875 and
AR-30582 from the National Institutes of Health, U.S. Pub-
lic Health Service.
1. Albright F, Smith PH, Richardson AM 1941 Postmenopausal os-
teoporosis. Its clinical features. JAMA 116:2465–2474.
2. Riggs BL, Melton LJ III 1986 Medical Progress: Involutional
osteoporosis. N Engl J Med 314:1676–1686.
3. Riggs BL, Khosla S, Melton LJ III 2002 Sex steroids and the
construction and conservation of the adult skeleton. Endocrine Rev
4. Judd HL, Lucas WE, Yen SSC 1974 Effect of oophorectomy on
circulating testosterone and androstenedione levels in patients with
endometrial cancer. Am J Obstet Gynecol 118:793–798.
5. Vermeulen A 1976 The hormonal activity of the postmenopausal
ovary. J Clin Endocrinol Metab 42:247–253.
6. Hughes CL Jr, Wall LL, Creasman WT 1991 Reproductive hor-
mone levels in gynecologic oncology patients undergoing surgical
castration after spontaneous menopause. Gynecol Oncol 40:42–
7. Sluijmer AV, Heineman MJ, De Jong FH, Evers JLH 1995 Endo-
crine activity of the postmenopausal ovary: The effects of pituitary
down-regulation and oophorectomy. J Clin Endocrinol Metab 80:
8. Laughlin GA, Barrett-Connor E, Kritz-Silverstein D, von Mu ¨hlen
D 2000 Hysterectomy, oophorectomy, and endogenous sex hor-
mone levels in older women: The Rancho Bernardo Study. J Clin
Endocrinol Metab 85:645–651.
904 MELTON ET AL.
9. Couzinet B, Meduri G, Lecce MG, Young J, Brailly S, Loosfelt H, Download full-text
Milgrom E, Schaison G 2001 The postmenopausal ovary is not a
major androgen-producing gland. J Clin Endocrinol Metab 86:
10. Slemenda C, Longcope C, Peacock M, Hui S, Johnston CC 1996
Sex steroids, bone mass, and bone loss: A prospective study of
pre-, peri, and postmenopausal women. J Clin Invest 97:14–21.
11. Greendale GA, Edelstein S, Barrett-Connor E 1997 Endogenous
sex steroids and bone mineral density in older women and men:
The Rancho Bernardo Study. J Bone Miner Res 12:1833–1843.
12. Cummings SR, Browner WS, Bauer D, Stone K, Ensrud K, Jamal
S, Ettinger B 1998 Endogenous hormones and the risk of hip and
vertebral fractures among older women. The Study of Osteoporotic
Fractures Research Group. N Engl J Med 339:733–738.
13. Stone K, Bauer DC, Black DM, Sklarin P, Ensrud KE, Cummings
SR 1998 Hormonal predictors of bone loss in elderly women: A
prospective study. The Study of Osteoporotic Fractures Research
Group. J Bone Miner Res 13:1167–1174.
14. Ettinger B, Pressman A, Sklarin P, Bauer DC, Cauley JA, Cummings
SR 1998 Associations between low levels of serum estradiol, bone
density, and fractures among elderly women: The Study of Osteopo-
rotic Fractures. J Clin Endocrinol Metab 83:2239–2243.
15. Garnero P, Sornay-Rendu E, Claustrat B, Delmas PD 2000 Bio-
chemical markers of bone turnover, endogenous hormones and the
risk of fractures in postmenopausal women: The OFELY Study.
J Bone Miner Res 15:1526–1636.
16. Melton LJ III 1996 History of the Rochester Epidemiology Project.
Mayo Clin Proc 71:266–274.
17. Melton LJ III, Bergstralh EJ, Malkasian GD, O’Fallon WM 1991
Bilateral oophorectomy trends in Olmsted County, Minnesota,
1950–1987. Epidemiology 2:149–152.
18. Melton LJ III, Crowson CS, Malkasian GD, O’Fallon WM 1996
Fracture risk following bilateral oophorectomy. J Clin Epidemiol
19. Melton LJ III 1997 The threat to medical-records research. N Engl
J Med 337:1466–1470.
20. Cooper C, Atkinson EJ, O’Fallon WM, Melton LJ III 1992 Inci-
dence of clinically diagnosed vertebral fractures: A population-
based study in Rochester, Minnesota, 1985–1989. J Bone Miner
21. Rose SH, Melton LJ III, Morrey BF, Ilstrup DM, Riggs BL 1982
Epidemiologic features of humeral fractures. Clin Orthop 168:24–30.
22. Melton LJ III, Atkinson EJ, Madhok R 1996 Downturn in hip
fracture incidence. Public Health Rep 111:146–150.
23. Melton LJ III, Amadio PC, Crowson CS, O’Fallon WM 1998
Long-term trends in the incidence of distal forearm fractures.
Osteoporos Int 8:341–348.
24. Melton LJ III, Crowson CS, O’Fallon WM 1999 Fracture inci-
dence in Olmsted County, Minnesota: Comparison of urban with
rural rates and changes in urban rates over time. Osteoporos Int
25. Cox DR 1953 Some simple approximate tests for Poisson variates.
26. Kaplan EL, Meier P 1958 Non-parametric estimation from incom-
plete observations. J Am Stat Assoc 53:457–481.
27. Kalbfleisch JD, Prentice RL 1980 The Statistical Analysis of
Failure Time Data. John Wiley & Sons, New York, NY, USA.
28. Cox DR 1972 Regression models and life-tables. J Royal Stat Soc
29. Simpson E, Rubin G, Clyne C, Robertson K, O’Donnell L, Jones
M, Davis S 2000 The role of local estrogen biosynthesis in males
and females. Trends Endocrinol Metab 11:184–188.
30. Labrie F, Luu-The V, Lin S-X, Simard J, Labrie C 2000 Role of
17?-hydroxysteroid dehydrogenases in sex steroid formation in
peripheral intracrine tissue. Trends Endocrinol Metab 11:421–427.
31. Adashi EY 1994 The climacteric ovary as a functional
gonadotropin-driven androgen-producing gland. Fertil Steril 62:
32. Heshmati HM, Khosla S, Robins SP, O’Fallon WM, Melton LJ III,
Riggs BL 2002 Role of low levels of endogenous estrogen in
regulation of bone resorption in late postmenopausal women.
J Bone Miner Res 17:172–178.
33. The ATAC Trialists’ Group 2002 Anastrozole alone or in combi-
nation with tamoxifen versus tamoxifen alone for adjuvant treat-
ment of postmenopausal women with early breast cancer: First
results of the ATAC randomised trial. Lancet 359:2131–2139.
34. Cauley JA, Gutai JP, Kuller LH, LeDonne D, Powell JG 1989 The
epidemiology of serum sex hormones in postmenopausal women.
Am J Epidemiol 129:1120–1131.
35. Barrett-Connor E, Mueller JE, von Mu ¨hlen DG, Laughlin GA,
Schneider DL, Sartoris DJ 2000 Low levels of estradiol are asso-
ciated with vertebral fractures in older men, but not women: The
Rancho Bernardo Study. J Clin Endocrinol Metab 85:219–223.
36. Chapurlat RD, Garnero P, Bre ´art G, Meunier PJ, Delmas PD 2000
Serum estradiol and sex hormone-binding globulin and the risk of
hip fracture in elderly women: The EPIDOS Study. J Bone Miner
37. Cauley JA, Cummings SR, Black DM, Mascioli SR, Seeley DG
1990 Prevalence and determinants of estrogen replacement therapy
in elderly women. Am J Obstet Gynecol 163:1438–1444.
38. Recker RR, Davies M, Dowd RM, Heaney RP 1999 The effect of
low-dose continuous estrogen and progesterone therapy with cal-
cium and vitamin D on bone in elderly women. A randomized,
controlled trial. Ann Intern Med 130:897–904.
39. Villareal DT, Binder EF, Williams DB, Schechtman KB, Yarash-
eski KE, Kohrt WM 2001 Bone mineral density response to
estrogen replacement in frail elderly women. A randomized con-
trolled trial. JAMA 286:815–820.
40. Writing Group for the Women’s Health Initiative Investigators 2002
Risks and benefits of estrogen plus progestin in healthy postmeno-
pausal women. Principal results from the Women’s Health Initiative
Randomized Controlled Trial. JAMA 288:321–333.
41. Cummings SR, Melton LJ 2002 Epidemiology and outcomes of
osteoporotic fractures. Lancet 359:1761–1767.
42. Klotzbuecher CM, Ross PD, Landsman PB, Abbott TA III, Berger
M 2000 Patients with prior fractures have an increased risk of
future fractures: A summary of the literature and statistical syn-
thesis. J Bone Miner Res 15:721–727.
43. Vestergaard P, Tigaran S, Rejnmark L, Tigaran C, Dam M,
Mosekilde L 1999 Fracture risk is increased in epilepsy. Acta
Neurol Scand 99:269–275.
44. Jamal SA, Browner WS, Bauer DC, Cummings SR 1998 Warfarin
use and the risk for osteoporosis in elderly women. Study of
Osteoporotic Fractures Research Group. Ann Intern Med 128:
45. Caraballo PJ, Heit JA, Atkinson EJ, Silverstein MD, O’Fallon
WM, Castro MR, Melton LJ III 1999 Long-term use of oral
anticoagulants and the risk of fracture. Arch Intern Med 159:
46. Seeman E 2001 Effects of tobacco and alcohol use on bone. In:
Marcus R, Feldman D, Kelsey J (eds.) Osteoporosis, 2nd ed., vol.
1. Academic Press, San Diego, CA, USA, pp. 771–794.
47. Cauley JA, Salamone LM 2001 Postmenopausal endogenous and
exogenous hormones, degree of obesity, thiazide diuretics, and risk
of osteoporosis. In: Marcus R, Feldman D, Kelsey J (eds.) Osteo-
porosis, 2nd ed., vol. 1. Academic Press, San Diego, CA, USA, pp.
48. Brett KM, Pokras R, Madans JH, Peterson HB 1994 National
trends in bilateral oophorectomy, 1965–1990. J Women Health
Gend Based Med 3:337–345.
49. Beard CM, Hartmann LC, Atkinson EJ, O’Brien PC, Malkasian
GD, Keeney GL, Melton LJ III 2000 The epidemiology of ovarian
cancer: A population-based study in Olmsted County, MN, 1935–
91. Ann Epidemiol 10:14–23.
50. Carlson KJ, Nichols DH, Schiff I 1993 Indications for hysterec-
tomy. N Engl J Med 328:856–860.
51. Plouffe L Jr 1998 Ovaries, androgens and the menopause: Practical
applications. Semin Reprod Endocrinol 16:117–120.
Address reprint requests to:
LJ Melton III, MD
Division of Epidemiology
200 First Street S.W.
Rochester, MN 55905, USA
Received in original form July 25, 2002; in revised form October
25, 2002; accepted November 7, 2002.
905LATE OOPHORECTOMY AND FRACTURES