ORIGINAL INVESTIGATION. Alveolar and Postcranial Bone Density in Postmenopausal Women Receiving Hormone/Estrogen Replacement Therapy

ORIGINAL INVESTIGATION Alveolar and Postcranial Bone Density in Postmenopausal Women Receiving Hormone/Estrogen Replacement Therapy A Randomized, Double-blind, Placebo-Controlled Trial Roberto Civitelli, MD; Thomas K. Pilgram, PhD; Mary Dotson, RDH, MPH; Jane Muckerman, RN, BSN; Nancy Lewandowski, RN; Reina Armamento-Villareal, MD; Naoko Yokoyama-Crothers, BA; E. Eugenia Kardaris, DDS; Jay Hauser, DDS; Sheldon Cohen, DDS; Charles F. Hildebolt, DDS, PhD Background: We conducted a -year, double-blind, randomized, placebo-controlled study to determine whether the positive effects of hormone/estrogen replacement therapy (H/ERT) on postcranial bone density are accompanied by similar positive effects on oral bone mass. Methods: A total of 5 postmenopausal women (aged 4-7 years) with no evidence of moderate or severe periodontal disease were randomized to receive daily oral conjugated estrogen (Premarin;.65 mg) alone or in combination with medroxyprogesterone acetate (Prempro;.65 and.5 mg, respectively) or placebo. All subjects received calcium carbonate ( mg/d) and cholecalciferol (8 IU/d) supplements. The primary efficacy end points were the changes in alveolar crest height and alveolar bone density. Alveolar crest height was measured on bite-wing radiographs, and changes in alveolar bone mass were assessed by means of digital-subtraction radiography. Postcranial bone density was measured in the lumbar spine and left proximal femur by means of dualenergy x-ray absorptiometry. Results: Hormone/estrogen replacement therapy significantly increased alveolar bone mass compared with placebo (+.84% vs +.95% [P=.4]), and tended to improve alveolar crest height (+4.8% vs +.46% [P=.4]). Bone mineral density of the proximal femur significantly increased in the H/ERT compared with the placebo group (total proximal femur, +.59% vs +.% [P=.]; neck, +.5% vs.4% [P=.]; trochanter, +.49% vs +.8% [P.]), but not the lumbar spine (+.% vs +.7% [P=.9]). Changes in alveolar bone mass correlated with bone density changes in the total femur (r=.8 [P=.]) and femoral trochanter (r=.5 [P=.4]) in the H/ERT but not in the placebo group. Conclusions: Postcranial and oral bone mass were increased in postmenopausal women receiving H/ERT. Improvement in oral bone health constitutes an additional benefit of H/ERT. Arch Intern Med. ;6:49-45 From the Division of Bone and Mineral Diseases, Department of Internal Medicine (Drs Civitelli and Armamento-Villareal and Mss Muckerman and Lewandowski), the Department of Radiology (Drs Pilgram and Hildebolt and Ms Yokoyama-Crothers), and the Department of Dentistry (Ms Dotson and Drs Kardaris, Hauser, and Cohen), Washington University School of Medicine and Barnes-Jewish Hospital, St Louis, Mo. Dr Civitelli has stock ownership with Wyeth, Madison, NJ. OSTEOPOROSIS IS regarded as a risk factor for tooth loss and edentulism in the elderly. Accordingly, it is commonly believed that estrogen deficiency, use of corticosteroids, and other conditions leading to bone demineralization also cause oral bone loss. Although cross-sectional studies designed to examine the relationship between oral bone loss and osteoporosis have not been universally concordant, -6 a recent longitudinal survey demonstrated a higher degree of loss of alveolar crest height and alveolar bone density after years in osteoporotic and osteopenic postmenopausal women relative to women with normal bone mineral density (BMD) of the spine. 7 The notion that postcranial skeletal loss is accompanied by deterioration of oral bone health is supported by the evidence that subjects with osteoporosis were more likely to use dentures than were those without osteoporosis, 8 and by the demonstration of an increasing likelihood of losing teeth with decreasing BMD at the spine and the radius in partially edentulous women. 9 The same investigators later reported that women who lost teeth during a 7-year follow-up experienced less favorable changes in BMD at the spine, femoral neck, and total body. Thus, abundant epidemiological evidence supports the hypothesis that systemic bone loss leads to tooth loss in postmenopausal women, presumably as a consequence of oral bone loss. One of the major risk factors for osteoporosis is estrogen deficiency. Observational studies indicate that estrogen de- 49

SUBJECTS AND METHODS SUBJECTS From March, 994, through April, 995, 55 postmenopausal women were recruited from patients attending the Dental Clinic of Barnes-Jewish Hospital, St Louis, Mo. To be included in the study, a subject had to be postmenopausal for at least year, in good medical health, and ambulatory; have at least teeth and no moderate or advanced periodontal disease (defined as periodontal pockets of more than 5 mm); and have no contraindications for ERT. Excluded were women who had more than documented vertebral fractures or any previous treatment with bisphosphonates, corticosteroids, or antiepileptics within months before entry into the study; had been treated with estrogen or calcitonin within years before entering the study; had any major medical condition that would interfere with compliance to the protocol; and had conditions associated with abnormalities of bone metabolism such as chronic liver disease, chronic renal failure, hyperparathyroidism, hypoparathyroidism, hyperthyroidism, hypercortisolism, multiple myeloma, or osteomalacia. DEMOGRAPHIC DATA A body mass index (BMI) was calculated by dividing the weight in kilograms by the square of the height in meters. The number of teeth was recorded at baseline and yearly until the end of the study. Cigarette smoking was estimated by multiplying the number of packs smoked per day by the number of years of active smoking (pack-years). Data were collected on the following factors that affect estrogen exposure: age at menarche, history of menstrual irregularities, use of birth control pills, number of pregnancies, number of pregnancies to term, and months of lactation. STUDY DESIGN This study was a double-blind, placebo-controlled, randomized -year trial followed by a -year, open-label extension (Figure ). We report herein the results of the -year controlled phase of the study. The study consisted of the following treatment arms: the hormone replacement therapy/ert (H/ERT) arm and the placebo arm. Women with an intact uterus (n=86) were randomized to receive a combination tablet consisting of.65 mg of conjugated equine estrogen and.5 mg of medroxyprogesterone acetate (Prempro) once a day, or a placebo look-alike tablet. Women without a uterus (n=49) were randomized to receive either.65 mg of conjugated equine estrogen (Premarin) once a day or a placebo look-alike tablet. All women enrolled received daily supplements of mg of elemental calcium as calcium carbonate and 8 IU of cholecalciferol. All women also received dental care and cleanings every months. Dental services outside of the study were permitted as desired by the patient. This study protocol was approved by the Human Studies Committee of Washington University Medical Center, St Louis, and all subjects gave an informed written consent to participate in the study. ASSESSMENT OF POSTCRANIAL BMD Bone mineral density of the lumbar spine and proximal left femur was measured by means of dual-energy x-ray absorptiometry, with a bone densitometer (Hologic QDR- ; Hologic Inc, Waltham, Mass). At our center, the precision of this technique is.% at the lumbar spine and.7% at the total proximal femur. 7 Bone mineral density was expressed as grams per square centimeters. Bone mineral density measurements were performed at baseline and every months thereafter for years. ASSESSMENT OF ALVEOLAR BONE Depending on the number of remaining teeth, up to 7 vertical bite-wing radiographs per patient were obtained, using customized film-positioning, beam-alignment devices with occlusal registration material. Dental radiographs were obtained at baseline and repeated every months. A single bite-wing radiograph (chosen by lot) was obtained a second time in patients at baseline to determine measurement reliability. To obtain quantitative measurements of the alveolar bone, bite-wing radiographs were digitized at 5 µm spatial resolution and -bit gray scale resolution. We made types of measurements. First, we calculated a linear measurement of the alveolar crest height as the distance from the cementoenamel junction to the alveolar crest (CEJ-AC); second, we performed a radiodensity measurement (we herein use the terms alveolar bone radiodensity and alveolar bone mass interchangeably). In brief, after the digitized radiographic images were enhanced, CEJ-AC measurements were made by one of us (N.Y.-C.) at the mesial, distal, and midtooth areas (at the points of maximum bone loss) of each tooth for up to 8 teeth per patient (the third molars were not measured). We used method error (ME; SD of the differences between baseline and repeated ficiency is also associated with a greater loss of the dental attachment apparatus (cementum, periodontal ligament, and alveolar bone) than in conditions of estrogen sufficiency. 7, Three large observational studies -4 in different cohorts of postmenopausal women confirmed the potential beneficial effect of estrogen on dental health. The number of teeth was higher and the odds of being edentulous or using dentures were reduced in estrogen users compared with nonusers. -4 Although the evidence of the beneficial effect of estrogen on BMD at postcranial skeletal sites has been well established by longitudinal trials, 5 such evidence does not exist at present for alveolar bone. One noncontrolled study of estrogentreated women suggested a positive influence of the hormonal treatment on jaw bone mass, 6 giving further support to the hypothesis that estrogen replacement therapy (ERT) may protect from oral bone loss in postmenopausal women. The objective of the present study was to determine whether the positive effects on the postcranial skeleton that can be obtained with ERT are associated with similar positive effects on oral bone health. By extension, protection from alveolar bone loss or increases in oral bone mass may result in reduced tooth loss. 4

The overall pattern was for alveolar bone mass to increase at a statistically significant level in both arms of the study, and for the increase in bone mass to be greater in the H/ERT group than in the placebo group. The difbaseline radiographic measurements divided by the square root of ) 8 to determine CEJ-AC measurement reliability, which was. mm. Additional details on CEJ-AC measurements and their reliabilities are given elsewhere. 9, Since the CEJ-AC is an inverse measurement of crest height, a decrease over time in the CEJ-AC distance indicates an increase in alveolar crest height (ie, a gain in alveolar bone). However, for consistency with all the other measurements, changes indicating a gain in alveolar bone height are indicated as positive. For alveolar bone radiodensity measurements, baseline and follow-up digitized radiographs of alveolar bone were registered using ANALYZE software (Mayo Foundation, Rochester, Minn), cropped automatically to ensure image homology, and histogram matched to correct for exposure and processing errors. Radiodensities were determined for regions of interest corresponding to the alveolar bone between the roots of the posterior teeth. The coefficient of variation of the ME 8 (CV ME =[ME/(xfl +xfl )] ) was used to determine the reliability of this method, which was.9%. As for alveolar crest height, an increase in alveolar bone mass is expressed as a positive change. A mean CEJ-AC distance and a mean alveolar bone radiodensity were calculated for each patient for data analysis. ASSESSMENT OF ADVERSE EVENTS Since the active medications included estrogen, mammograms were obtained at baseline and repeated at years and 4. Subjects complaining of breakthrough or abundant vaginal bleeding were referred to a gynecologist for further evaluation and for endometrial biopsy, if necessary. Subjects were questioned at each visit about adverse events. Telephone calls from patients reporting adverse events were also recorded. Adverse events were counted and classified as serious or minor, regardless of whether they were believed to be related to the active medications. Those events that resulted in death, hospitalization, cancer, or permanent disability or that were considered life-threatening were categorized as serious adverse events. Excessive loss of BMD was also considered a serious adverse event and was defined as a 5% loss at year or a % loss at years at any postcranial site. In such cases, the participant was offered the choice to discontinue the study. All categories and frequencies of adverse events are reported herein. ASSESSMENT OF COMPLIANCE Compliance with the regimen of study medications was assessed by a pharmacist every 6 months by counting the remaining tablets in the bottles returned by patients at each scheduled refill. Adherence to treatment was similar in both treatment groups; 59 (88%) of 67 women receiving placebo and 5 (78%) of 68 women receiving estrogen took more than 8% of study medication. STATISTICAL ANALYSIS The study was powered to detect a.5-mm change in alveolar crest height, a clinically significant change, with and (type I and II error rates) set at.5 and., respectively. On the basis of a previous pilot study, we determined that 4 patients per treatment arm would be required to unambiguously detect such a change. Expecting approximately a 4% dropout rate, we targeted enrollment to 67 subjects per treatment arm. Data were analyzed on an intention-to-treat basis. All missing data from subjects who had at least the baseline measurements were replaced with the last value obtained (last value carried forward). Of the 5 subjects entered into the study (68 in the estrogen arm and 67 in the placebo arm), 7 ( in the H/ERT arm and 6 in the placebo arm) dropped out after the first (baseline) visit; 4 ( in the estrogen arm and in the placebo arm), after the second visit; 8 (5 in the estrogen arm and in the placebo arm), after the third visit; and 86 subjects (49 in the estrogen arm and 7 in the placebo arm) completed the entire study (Figure ). Measured values were analyzed as raw data and with the final value expressed as a percentage of the baseline value; only the latter results are presented herein. Alveolar radiodensity values are expressed and analyzed only as percentages of baseline values. Data collected at a nominal or ordinal scale of measurement were analyzed with contingency tables. When 8% or more of the table s cells had expected values of greater than 5, tests were used; in all other cases, -tailed Fisher exact tests were used. Data collected at an interval scale of measurement were analyzed by t tests when simple changes were examined, or by multivariate analysis of variance when repeated measures were included. The P values reported for the multivariate analysis of variance are for the interaction between time and drug arm. The role of demographic and historical variables was examined by including them as an additional variable in the model. We analyzed the association between changes in bone measurements in different locations by calculating the Pearson correlation coefficient of the changes of each variable from baseline to the end of the study. We analyzed the data using JMP software (SAS Institute Inc, Cary, NC). RESULTS One hundred thirty-five women aged 4 to 7 years underwent randomization and follow-up in the study. Almost all were white (n=7 [94.%]), reflecting the ethnic group distribution of patients referred to the Barnes- Jewish Hospital Dental Clinic. Women in the H/ERT arm were approximately years older than the women in the placebo arm, but the number of years since menopause was not different between the groups (Table ). Both arms of the study were well matched for BMI, smoking history, total number of pregnancies, number of pregnancies to term, alveolar crest height, and number of teeth at baseline. Bone mineral density was not a criterion for entry into the study. Average densities of the lumbar spine and all regions of interest on the proximal femur were similar in both arms of the study and in the osteopenia range. ALVEOLAR BONE 4

Screened Patients (N = 55) Not Randomized (n = ) Randomized (n = 5) Cumulative Dropouts: Months (n = 6) 4 Months (n = 7) 6 Months (n = ) Placebo Group (n = 67) All Received Calcium and Cholecalciferol Supplements and Yearly Dental Cleaning H/ERT Group (n = 68) All Received Calcium and Cholecalciferol Supplements and Yearly Dental Cleaning Cumulative Dropouts: Months (n = ) 4 Months (n = 4) 6 Months (n = 9) Follow-up BMD and Dental Measurements Months (n = 4) 4 Months (n = 4) 6 Months (n = 7) Follow-up BMD and Dental Measurements Months (n = 57) 4 Months (n = 54) 6 Months (n = 49) Figure. Patient recruitment and follow-up. H/ERT indicates hormone/estrogen replacement therapy; BMD, bone mineral density. Treatment arms are described in detail in the Study Design subsection of the Subjects and Methods section. Table. Baseline Data* Placebo Arm H/ERT Arm P Value Age, y 58. ± 6.8 6. ± 5.5.7 Years since menopause.7 ± 7.6.5 ± 7.6.59 BMI 7. ± 4.8 5.5 ± 4.4. Current smokers, No. (%) 7 () 9 ().79 Parity.6 ±.65. ±.4.7 Gravidity.4 ±.5.7 ±.95. Alveolar crest height, mm. ±.. ±..8 No. of teeth present 4. ± 4. 4.7 ±.7.4 Bone mineral density Lumbar spine g/cm.94 ±.5.95 ±.5.59 T-score. ±.. ±.4.99 Femoral neck g/cm.7 ±..7 ±..9 T-score.94 ±..99 ±.7.8 Total proximal femur g/cm.8 ±..8 ±..6 T-score.48 ±.9.44 ±.9.8 Compliance 8%, No. 59/67 5/68.6 *Unless otherwise indicated, data are given as mean ± SD. Treatment arms are described in detail in the Study Design subsection of the Subjects and Methods section. H/ERT indicates hormone/estrogen replacement therapy; BMI, body mass index (calculated as weight in kilograms divided by the square of height in meters). ference between groups was statistically significant for alveolar bone mass, but not for the crest height (Figure ). There was a 4.8% increase in alveolar crest height in the H/ERT group (P=.) and a.46% increase in the placebo group (P=.). The.7% greater increase in the H/ERT group compared with the control group was not statistically significant (P=.4). There was a.84% increase in alveolar bone mass in the H/ERT group (P.), and a.95% increase in the placebo group (P.). The.9% greater increase in the H/ERT group was statistically significant (P=.4). POSTCRANIAL BMD The overall pattern in postcranial bone was for the H/ERT group to gain BMD and for the placebo group to maintain BMD (Figure ). The amount of BMD gain for the H/ERT group and the difference in BMD change between the H/ERT and placebo groups were statistically Changes From Baseline, % 7 6 5 4 Alveolar Crest Height H/ERT Arm Placebo Arm significant for the femur, but not for the spine. Women in the H/ERT group experienced a.5% increase in BMD in the femoral neck (P=.), compared with a.4% decrease in density for the women in the placebo group (P=.74), for a difference of.9% between the groups (P=.). In the total femur, the gain in the H/ERT group was.59% (P.) compared with.% (P=.7), for a difference of.7% (P=.). In the trochanter, the gain in the H/ERT group was.49% (P.) compared with.8% (P=.86), for a difference of.4% (P.). In the lumbar spine, the changes at the end of years were a gain of.% for women in the H/ERT group (P=.) and a gain of.7% in the placebo group (P=.8), with a difference of.84% between the groups (P=.9). Although none of the demographic or historical variables differed at a statistically significant level between the arms of the study, those showing a trend for a difference were examined for influence on the study outcome. None of them altered the findings of the study. CORRELATIONS BETWEEN ORAL AND POSTCRANIAL BMD CHANGES We found positive correlations between changes in the alveolar crest height and the BMD of the lumbar spine and femoral neck among H/ERT-treated patients, indicating that an increase in postcranial BMD is accompa- Alveolar Bone Mass P =.4 P =.4 Figure. Changes in oral bone variables in women treated with hormone/estrogen replacement therapy (H/ERT) or placebo. Data represent mean±se obtained from 67 women in the placebo arm and 68 in the H/ERT arm and reflect an intention-to-treat analysis. P values represent the significance level of treatment effect estimated by multivariate analysis of variance. Treatment arms are described in detail in the Study Design subsection of the Subjects and Methods section. 4

Femoral Neck Total Proximal Femur Lumbar Spine Changes From Baseline, % 4 H/ERT Arm Placebo Arm P =. 5 4 P =. 4 P =.9 Figure. Changes in postcranial bone density in women treated with hormone/estrogen replacement therapy (H/ERT) or placebo. Data represent mean±se obtained from 67 women in the placebo arm and 68 in the H/ERT arm and reflect an intention-to-treat analysis. P values represent the significance level of treatment effect estimated by multivariate analysis of variance. Treatment arms are described in detail in the Study Design subsection of the Subjects and Methods section. Table. Correlations Between Changes on Alveolar and Postcranial Bone Density* Alveolar Crest Height nied by an increase in alveolar crest height (Table ). By contrast, no significant correlations were found between any measures of postcranial BMD and alveolar crest height in the placebo-treated women. Likewise, BMD changes at most femoral regions were positively correlated with changes in alveolar BMD in women receiving H/ERT, whereas no similar correlations were observed among the placebo-treated women (Table ). NUMBER OF TEETH Alveolar Bone Mass r P Value r P Value H/ERT arm Lumbar spine.9..6.9 Femoral neck.6...8 Femoral trochanter..9.5.4 Total femur.4.5.8. Placebo arm Lumbar spine.6.6..95 Femoral neck.6...84 Femoral trochanter.6.64..8 Total femur.4.7.5. *Treatment arms are described in detail in the Study Design subsection of the Subjects and Methods section. H/ERT indicates hormone/estrogen replacement therapy. At the end of the -year study, there was no difference in the mean number of remaining teeth between the H/ERT (4.4±.69) and placebo (4.47±.47) groups, and there was no difference in the average number of teeth lost per patient between the H/ERT (.47±.9) and placebo (.5±.54) groups. Furthermore, no significant correlation was seen in the number of teeth lost and oral bone measures during the -year study in both groups. ADVERSE EVENTS AND PATIENT RETENTION Adverse events developed in 9 women in the placebo and 57 women in the H/ERT groups (Table ). Only 4 Table. Adverse Events During Trial* Events Placebo Arm (n = 67) H/ERT Arm (n = 68) Minor Headaches (5) 4 (6) Vaginal bleeding 7 () (8) Breast tenderness 6 (9) (8) Thrombophlebitis () Mood changes () () Hot flushes () (4) Leg pain () 4 (6) Gastrointestinal tract upset 4 (6) (4) Cholecystitis () Vaginitis () 5 (7) Vaginal dryness (5) Urinary urgency (5) () Arrhythmias () () Weight gain () () Anxiety () Serious Breast cancer () Endometrial cancer () Colon cancer () Transient ischemic attack () Ankle fracture () Excessive decrease in BMD () () *For all variables, P. for difference between treatment arms ( test). Treatment arms are described in detail in the Study Design subsection of the Subjects and Methods section. H/ERT indicates hormone/estrogen replacement therapy; BMD, bone mineral density. (6%) of 68 women in the H/ERT group and (4%) of 67 in the placebo group experienced adverse events that can be categorized as serious. The frequency of serious adverse events in the H/ERT group was lower than that reported during the larger Postmenopausal Estrogen/ Progestin Interventions trial, where the incidence of lifethreatening events was.7% in the estrogen-treated group and 4.6% in the placebo group. 5 Forty-nine (6.%) of the 5 women left the study prematurely, (45%) of 67 from the placebo group and 9 (8%) of 68 from the H/ERT group. Of these 49 subjects, (47%) withdrew because of an adverse event, (9%) of 67 from the placebo group and (5%) of 68 4

Table 4. Reasons for Discontinuation* Reason Placebo Arm H/ERT Arm Excluded after randomization () () Serious adverse events Breast cancer () Endometrial cancer () Ankle fracture () Excessive decrease in BMD () Minor adverse events Vaginal bleeding (5) () Breast tenderness 4 (6) () Thrombophlebitis () Other 4 (6) 4 (6) Protocol violation 5 (8) Lost to follow-up 5 (8) (4) Lost interest 5 (8) 4 (6) *Treatment arms are described in detail in the Study Design subsection of the Subjects and Methods section. H/ERT indicates hormone/estrogen replacement therapy; BMD, body mass index. Includes irritability, kidney stones, dyspepsia, hot flushes, leg pain, and headache. from the H/ERT group. Of the subjects who reported serious adverse events, a total of 4 withdrew from the study, in each group (Table 4). These included women in the H/ERT group (one with a new diagnosis of breast cancer and another with a new diagnosis of endometrial cancer) and women in the placebo group (one with a fractured ankle and another who experienced an excessive decrease in bone density). The other subjects with serious adverse events ( in the placebo group with colon cancer, in the H/ERT group with a transient ischemic attack requiring hospitalization, and in the H/ERT group with excessive bone density loss) elected to continue in the study. The higher withdrawal rate in the placebo group relative to the H/ERT group was unexpected. We believe this finding may relate to some patients having entered the study in hopes of receiving H/ERT. Some of the women assigned to placebo may have suspected they were not taking the active medication, and they requested active treatment from their physician, thus violating the protocol or they simply lost interest or failed to return for follow-up. COMMENT The preventive effect of H/ERT on postmenopausal bone loss and osteoporosis is established. 5 Observational studies have demonstrated that estrogen treatment also prevents bone loss in the mandibular/alveolar bone.,6 A direct effect of estrogen on oral bone loss as measured by BMD changes over time, however, has not been explored. To our knowledge, our study is the first to demonstrate in a prospective, controlled fashion that H/ERT benefits the postcranial skeleton and produces a significant improvement in alveolar bone mass. In addition to the clear, statistically significant increase of alveolar bone mass after years of treatment, we also observed an increase in alveolar crest height in the H/ERT-treated patients, although this increase was not significantly greater than that for non H/ERTtreated subjects. The increase in alveolar crest height in all subjects may seem surprising, as it implies an expansion of the alveolar bone. Although periosteal bone expansion is theoretically possible in the alveolar bone, a simpler alternative explanation may be that a denser alveolar process may project as a higher crest, simply because of its increased peripheral density. Regardless of the causes for the observed changes in the linear measurement of alveolar bone, the effect of H/ERT on the densitometric variables of oral bone mass was unambiguous. Therefore, the overall findings imply that the dental attachment apparatus is strengthened by H/ERT in postmenopausal women. Furthermore, the positive effects of H/ERT on alveolar and postcranial bone mass were correlated at most sites, including the spine and the alveolar crest height, where treatment effects were not significantly different from those of the placebo group, indicating that the benefits of H/ERT on alveolar bone are most likely mediated by its systemic action on bone remodeling. Epidemiological studies have shown a relationship between systemic bone loss and oral bone health, suggesting that tooth loss may be a manifestation of osteoporosis., This relationship was more apparent in women with edentulism. 4,5 A substantial body of evidence supports the notion that edentulous women are more likely to have low BMD than dentate women. 8,9,4 Although no data directly link loss in alveolar crest height or alveolar bone to tooth loss, in a relatively large cohort of postmenopausal women, those with osteoporosis had lower mandibular bone mass and thinner cortex at the gonion and had lost a significantly higher number of teeth than did nonosteoporotic women. 4 Therefore, it is reasonable to say that alveolar bone loss is the most important contributor to tooth loss and edentulism in women with osteoporosis. By contrast, estrogen users have more teeth and a lower prevalence of edentulism than nonusers, as indicated by at least large retrospective cohort studies. -4 Specifically, in the Leisure World cohort, the proportion of women with edentulism decreased with duration of estrogen exposure, and in the Framingham Heart Study cohort, duration of ERT was an independent predictor of the number of remaining teeth. We did not find differences in the numbers of teeth lost or remaining at the end of the study between the groups, as tooth loss was minimal during the study. This result is not surprising, as the study was not powered to detect differences in tooth loss, which would be a relatively rare event during a -year period in women in good oral health. In any case, our positive results on alveolar bone mass support the conclusions of the existing observational studies, reinforcing the notion that H/ERT strengthens the tooth attachment apparatus, thus improving dental health. This conclusion is highly relevant for clinical practice, since progressive loss of bony support for teeth after menopause may result in costly, lengthy, uncomfortable treatments, with eventual tooth loss and the inconvenience of wearing prosthetic appliances. In the presence of periodontal disease and poor oral hygiene, systemic bone loss plays a lesser role than oral disease in disrupting the attachment apparatus. This was not the case for our patient population, who did not have clinically significant periodontal disease. The women in the placebo group who received calcium and cholecal- 44

ciferol supplements and regular dental care during the study actually experienced a significant increase in alveolar bone density and a slight improvement in alveolar crest height. Thus, dental care and dietary supplementation with calcium and cholecalciferol may be sufficient to prevent postmenopausal oral bone loss and produce positive effects on dental health. This conclusion is consistent with that of a previous report, which demonstrated a lower incidence of tooth loss in postmenopausal women taking calcium supplements with or without cholecalciferol relative to subjects who were not using dietary supplements. Therefore, our results support a public health recommendation for regular dental care and sufficient calcium and cholecalciferol intake to maintain oral health in postmenopausal women. This study has some limitations. Most important is the higher dropout rate in the placebo group relative to the H/ERT arm. This may have reduced the significance of some of the differences between groups, although the intentionto-treat analysis overcomes the potential bias that may be introduced by excluding the noncompleters from the final analysis. Another limitation is the enrollment of women in good dental health, and therefore at low risk for alveolar bone loss and tooth loss. This was a deliberate choice, because we wanted to test the hypothesis that estrogen affects the alveolar bone through its action on bone remodeling as it does in the rest of the skeleton. Including subjects with significant periodontal disease may have resulted in oral bone loss, thus increasing the chances of detecting a treatment effect, but it would have precluded the distinction between a systemic and a local effect of estrogen, which can interfere with locally produced inflammatory cytokines that stimulate bone resorption. Finally, the response to H/ERT in the lumbar spine was less than in the proximal femur, whereas the opposite is usually seen after hormonal therapy. There are several possible reasons for this unexpected result, although all are speculative. Most of our subjects had been postmenopausal for more than 5 years, and thus past the stage of accelerated bone loss which mainly affects trabecular bone. Furthermore, several women in our cohort had high BMIs. A high level of body fat diminishes the accuracy and the precision of dual-energy x-ray absorptiometry of the spine. Finally, some degree of spondyloarthritis in the more elderly women may have contributed to a further increase in the variability of the dualenergy x-ray absorptiometry measurements of the spine. CONCLUSIONS Hormone/estrogen replacement therapy benefits postcranial bone density and produces significant improvement in oral bone mass. Thus, improved oral bone health constitutes an additional therapeutic benefit of ERT in postmenopausal women. Research should be encouraged to develop simple dental radiographic methods for identification of women with low alveolar bone mass for further evaluation and consideration of treatment strategies to prevent bone loss and/or increase bone mass. Routine dental visits could then represent a convenient setting for identification of subjects with or at risk for osteoporosis. Accepted for publication October 5,. This work was supported by grant DE986 from the National Institutes of Health, Bethesda, Md. Additional support was provided by Wyeth-Ayerst Laboratories, St Davids, Pa, and by Smith-Kline Beecham, Parsippany, NJ. This study was presented in part at the 78th General Session of the International Association for Dental Research, Washington, DC, April 68, ; the 4th World Congress on Osteoporosis, Chicago, Ill, June 7, ; and the th North American Menopause Society Meeting, Orlando, Fla, September 8,. Corresponding author and reprints: Roberto Civitelli, MD, Division of Bone and Mineral Diseases, Barnes-Jewish Hospital, North Campus, Mailstop 9--656, 6 S Kingshighway Blvd, St Louis, MO 6 (e-mail: rcivitel@im.wustl.edu). REFERENCES. Hildebolt CF. Osteoporosis and oral bone loss. Dentomaxillofac Radiol. 997; 6:-5.. Wactawski-Wende J, Grossi SG, Trevisan M, et al. The role of osteopenia in oral bone loss and periodontal disease. J Periodontol. 996;67:76-84.. Mohajery M, Brooks SL. Oral radiographs in the detection of early signs of osteoporosis. Oral Surg Oral Med Oral Pathol. 99;7:-7. 4. von Wowern N, Storm TL, Olgaard K. Bone mineral content by photon absorptiometry of the mandible compared with that of the forearm and the lumbar spine. Calcif Tissue Int. 988;4:57-6. 5. Elders PJ, Habets LL, Netelenbos JC, van der Linden LW, van der Stelt PF. The relation between periodontitis and systemic bone mass in women between 46 and 55 years of age. J Clin Periodontol. 99;9:49-496. 6. Hildebolt CF, Pilgram TK, Yokoyama-Crothers N, et al. Alveolar bone height and postcranial bone mineral density: negative effects of cigarette smoking and parity. J Periodontol. ;7:68-689. 7. Payne JB, Reinhardt RA, Nummikoski PV, Patil KD. Longitudinal alveolar bone loss in postmenopausal osteoporotic/osteopenic women. Osteoporos Int. 999; :4-4. 8. Daniell HW. Postmenopausal tooth loss: contributions to edentulism by osteoporosis and cigarette smoking. Arch Intern Med. 98;4:678-68. 9. Krall EA, Dawson-Hughes B, Papas A, Garcia RI. Tooth loss and skeletal bone density in healthy postmenopausal women. Osteoporos Int. 994;4:4-9.. Krall EA, Garcia RI, Dawson-Hughes B. Increased risk of tooth loss is related to bone loss at the whole body, hip, and spine. Calcif Tissue Int. 996;59:4-47.. Payne JB, Zachs NR, Reinhardt RA, Nummikoski PV, Patil K. The association between estrogen status and alveolar bone density changes in postmenopausal women with a history of periodontitis. J Periodontol. 997;68:4-.. Krall EA, Dawson-Hughes B, Hannan MT, Wilson PW, Kiel DP. Postmenopausal estrogen replacement and tooth retention. Am J Med. 997;:56-54.. Paganini-Hill A. The benefits of estrogen replacement therapy on oral health: the Leisure World cohort. Arch Intern Med. 995;55:5-9. 4. Grodstein F, Colditz GA, Stampfer MJ. Post-menopausal hormone use and tooth loss: a prospective study. J Am Dent Assoc. 996;7:7-77. 5. The Writing Group for the PEPI Trial. Effects of hormone therapy on bone mineral density: results from the Postmenopausal Estrogen/Progestin Interventions (PEPI) Trial. JAMA. 996;76:89-96. 6. Jacobs R, Ghyselen J, Koninckx P, van Steenberghe D. Long-term bone mass evaluation of mandible and lumbar spine in a group of women receiving hormone replacement therapy. Eur J Oral Sci. 996;4:-6. 7. Leelawattana R, Ziambaras K, Roodman-Weiss J, et al. The oxidative metabolism of estradiol conditions postmenopausal bone density and bone loss. J Bone Miner Res. ;5:5-5. 8. Portney LG, Watkins MP. Foundations of Clinical Research: Application to Practice. East Norwalk, Conn: Appleton & Lange; 99:55-56. 9. Hildebolt CF, Pilgram TK, Yokoyama-Crothers N, et al. Reliability of linear alveolar bone loss measurements of mandibular posterior teeth from digitized bitewing radiographs. J Clin Periodontol. 998;5:85-856.. Hildebolt CF, Brunsden B, Yokoyama-Crothers N, et al. Comparison of reliability of manual and computer-intensive methods for radiodensity measures of alveolar bone loss. Dentomaxillofac Radiol. 998;7:45-5.. HildeboltCF,BartlettTQ,ShroutMK,Yokoyama-CrothersN,RupichRC.Image-based quantification of alveolar bone. Yearbook Phys Anthropol. 994;7:55-76.. May H, Reader R, Murphy S, Khaw KT. Self-reported tooth loss and bone mineral density in older men and women. Age Ageing. 995;4:7-.. Taguchi A, Tanimoto K, Suei Y, Wada T. Tooth loss and mandibular osteopenia. Oral Surg Oral Med Oral Pathol Oral Radiol Endod. 995;79:7-. 4. Kribbs PJ. Comparison of mandibular bone in normal and osteoporotic women. J Prosthet Dent. 99;6:8-. 5. Hirai T, Ishijima T, Hashikawa Y, Yajima T. Osteoporosis and reduction of residual ridge in edentulous patients. J Prosthet Dent. 99;69:49-56. 45

4. Pinede L, Ninet J, Duhaut P, et al. Comparison of and 6 months of oral anticoagulant therapy after a first episode of proximal deep vein thrombosis or pulmonary embolism and comparison of 6 and weeks of therapy after isolated calf deep vein thrombosis. Circulation. ;:45-46. 5. Agnelli G, Prandoni P, Santamaria MG, et al. Three months versus one year of oral anticoagulant therapy for idiopathic deep venous thrombosis. N Engl J Med. ;45:65-69. 6. PIOPED Investigators. Value of the ventilation/perfusion scan in acute pulmonary embolism: results of the prospective investigation of pulmonary embolism diagnosis (PIOPED). JAMA. 99;6:75-759. 7. Bertina RM, Koeleman BP, Koster T, et al. Mutation in blood coagulation factor V associated with resistance to activated protein C. Nature. 994;69:64-67. 8. Poort SR, Rosendaal FR, Reitsma PH, Bertina RM. A common genetic variation in the -untranslated region of the prothrombin gene. Blood. 996;88:698-7. 9. Kyrle PA, Minar E, Hirschl M, et al. High plasma levels of factor VIII and the risk of recurrent venous thromboembolism. N Engl J Med. ;4:457-46.. Brandt JT, Triplett DA, Alving B, Scharrer I, for the Subcommittee on Lupus Anticoagulant/Antiphospholipid Antibody of the Scientific and Standardisation Committee of the ISTH. Criteria for the diagnosis of lupus anticoagulants: an update. Thromb Haemost. 995;74:85-9.. Kalbfleisch JD, Prentice RL. The Statistical Analysis of Failure Time Data. New York, NY: John Wiley & Sons; 98.. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc. 958;5:457-48.. Columbus Investigators. Low-molecular-weight heparin in the treatment of patients with venous thromboembolism. N Engl J Med. 997;7:657-66. 4. Douketis JD, Kearon C, Bates S, Duku EK, Ginsberg JS. Risk of fatal pulmonary embolism in patients with treated venous thromboembolism. JAMA. 998;79: 458-46. 5. Murin S, Romano PS, White RH. Comparison of outcomes after hospitalization for deep vein thrombosis or pulmonary embolism. Thromb Haemost. ;88: 47-44. 6. Eichinger S, Pabinger I, Stümpflen A, et al. The risk of recurrent venous thromboembolism in patients with and without factor V Leiden. Thromb Haemost. 997; 77:64-68. 7. Eichinger S, Minar E, Hirschl M, et al. The risk of early recurrent venous thromboembolism after oral anticoagulant therapy in patients with the GA transition in the prothrombin gene. Thromb Haemost. 999;8:4-7. 8. Moser KM, Fedullo PF, LitteJohn JK, Crawford R. Frequent asymptomatic pulmonary embolism in patients with deep vein thrombosis. JAMA. 994;7:- 5. 9. Prandoni P, Lensing AWA, Cogo A, et al. The long-term clinical course of acute deep venous thrombosis. Ann Intern Med. 996;5:-7.. Prandoni P, Lensing AW, Prins MR. Long-term outcomes after deep venous thrombosis of the lower extremities. Vasc Med. 998;:57-6. Correction Error in Text. In the Original Investigation by Civitelli et al titled Alveolar and Postcranial Bone Density in Postmenopausal Women Receiving Hormone/ Estrogen Replacement Therapy: A Randomized, Double-blind, Placebo- Controlled Trial, published in the June 4,, issue of the ARCHIVES (; 6:49-45), the dosage of cholecalciferol was incorrectly given as 8 IU/ d, instead of the correct dosage 4 IU/d. (REPRINTED) ARCH INTERN MED/ VOL 64, JAN, 4 96 4 American Medical Association. All rights reserved.