Calcium Intake and Fracture Risk: Results from the Study of Osteoporotic Fractures

Transcription

1 American Journal of Epidemiology Copyright 1997 by The Johns Hopkins University School of Hygiene and Public Health All rights reserved Vol. 145, No. 10 Printed in U.S.A. Calcium Intake and Fracture Risk: Results from the Study of Osteoporotic Fractures Robert G. Cumming, 1 Steven R. Cummings, 2 Michael C. Nevitt, 2 Jean Scott, 3 Kristine E. Ensrud, 4 Thomas M. Vogt, 5 and Kathleen Fox 3 The relation between dietary calcium, calcium, and vitamin D supplements and the risk of of the hip (n = 332), ankle (n = 210), proximal humerus (n = 241), wrist (n = 467), and vertebrae (n = 389) was investigated in a cohort study involving 9,704 US white women aged 65 years or older. Baseline assessments took place in in four US metropolitan areas. Dietary calcium intake was assessed at baseline with a validated food frequency questionnaire. Data on new nonvertebral were collected every 4 months during a mean of 6.6 years of follow-up; identification of new vertebral was based on comparison of baseline and follow-up radiographs of the spine done_a mean of 3.7 years apart. Results were adjusted for numerous potential confounders, including weight, physical activity, estrogen use, protein intake, and history of falls, osteoporosis, and. There were no important associations between dietary calcium intake and. the risk of any of the studied. use of calcium supplements was associated with increased risk of hip (relative risk = 1.5, 95% confidence interval ) and vertebral (relative risk = 1.4, 95% confidence interval ) ; current use of Turns antacid tablets was associated with increased risk of of the proximal humerus (relative risk = 1.7, 95% confidence interval ). There was no evidence of a protective effect of vitamin D supplements. Although a true adverse effect of calcium supplements on fracture risk cannot be ruled out, it is more likely that our findings are due to inadequately controlled confounding by indications for use of supplements. In conclusion, this study did not find a substantial beneficial effect of calcium on fracture risk. Am J Epidemiol 1997;145: calcium; ; hip ; osteoporosis Although randomized trials have clearly demonstrated that calcium supplements reduce the rate of bone loss in postmenopausal women (1-6), the efficacy of calcium for fracture prevention remains unclear. Observational epidemiologic studies of dietary calcium and are inconsistent. There have been at least 14 studies of hip fracture and dietary calcium (7-20), and only three of these found a clearly protective effect. On the other hand, two small randomized trials have found a reduced rate of radiographic vertebral among subjects given calcium supplements (21, 22), and another small study Received for publication June 11, 1996, and accepted for publication January 14, Department of Public Health and Community Medicine, University of Sydney, Sydney, Australia. 2 Department of Epidemiology and Biostatistics, University of California, San Francisco, CA. 3 School of Medicine, University of Maryland, Baltimore, MD. 4 Division of Epidemiology, School of Public Health, University of Minnesota, Minneapolis, MN. 5 Kaiser Permanente Center for Health Research, Portland, OR. Reprint requests to Dr. Robert G. Cumming, Department of Public Health and Community Medicine, Building A27, University of Sydney, Sydney, NSW 2006, Australia. found a nonsignificant reduction in risk of symptomatic vertebral and nonvertebral (23). A large French trial found that a combination of calcium and vitamin D supplements halved the hip fracture rate among women living in nursing homes (24). To assess the relation between calcium intake and the risk of, we analyzed data from the Study of Osteoporotic Fractures, a cohort study of 9,704 white women aged 65 years and over who have been followed up for more than 6 years. We investigated dietary calcium, calcium supplements, vitamin D supplements, and the risk of at a number of sites: hip, ankle, proximal humerus, wrist, and vertebrae. MATERIALS AND METHODS Subjects From 1986 to 1988, women aged 65 years or more were recruited from Portland, Oregon; Minneapolis, Minnesota; Baltimore, Maryland; and the Monongahela Valley, Pennsylvania. Recruitment involved mailings to women on population-based lists (voter registration, driver's license, and health maintenance organization lists). We excluded black women, women 926

2 Calcium Intake and Fractures 927 who were unable to walk without assistance, and women who had had bilateral hip replacements. Calcium intake Dietary calcium was estimated with a validated food frequency questionnaire, developed from the Second National Health and Nutrition Examination Survey (25). Subjects were asked how often in the last year they usually ate each of 20 foods. Food models were used to assist subjects to rate their usual portion size for each food as small, medium, or large. Inclusion of foods on the questionnaire was based on their contribution to total calcium and protein intake in US adults. The foods accounted for about 80 percent of calcium intake, and the correlation between calcium intake in elderly women assessed with the questionnaire and 7-day food records was 0.76 (26). The questionnaire also inquired about the frequency of drinking glasses of milk during teenage years, between the ages of 18 and years, since the age of years, and during pregnancy and breastfeeding. Women were asked to bring all medications to the baseline clinic visit. For women who had ever used a calcium supplement, we collected information on the name(s) and dose(s) of currently used calcium supplements and the calcium supplement they had used for the longest time. We also asked about the age at which calcium supplements had first been used and whether there was any period when they had not been used. Similar data were collected for Turns antacid tablets, a brand name medication containing calcium which is sold as an antacid (SmithKline Beecham, Philadelphia, Pennsylvania). Multivitamin supplements containing vitamin D were included in the vitamin D supplement category. We obtained data on the duration of use of vitamin D supplements but not on dose. Other measurements Baseline data on a wide range of possible risk factors for osteoporotic were collected by questionnaire and examination in outpatient clinics (27, 28). Among the factors assessed by questionnaire were alcohol use, smoking, caffeine intake, parental history of, falls during the past year, self-rated health, walking for exercise, and difficulties in activities of daily living. A modified Paffenbarger questionnaire was used to assess recreational physical activity in the past year, expressed in kilocalories per week (29). We asked about physician-diagnosed since the age of, osteoporosis, spine fracture, hyperthyroidism, Parkinson's disease, and stroke. Women were also asked if they had ever taken estrogen, thiazide diuretics, thyroid hormones, or psychotropic medications. The food frequency questionnaire described above was also used to obtain an approximate estimate of dietary protein intake. Among the factors assessed by examination were height, weight, the ability to stand up from a chair without using arms, the time to complete a measured course walking normally and with feet in tandem, depth perception using a Howard-Dohlman device (Lafayette Instruments, Lafayette, Indiana), and contrast sensitivity using a Vistech VCTS 60 wall chart (Vistech Company, Dayton, Ohio) (28, 30, 31). Ascertainment of incident We contacted women by postcard or telephone every 4 months and asked whether they had had a fracture. Over 99 percent of these contacts were completed. Radiology reports were reviewed (32), and only radiologically confirmed are included in this paper. In addition, hip were confirmed and classified by type (trochanteric or femoral neck) after review of preoperative radiographs by the study radiologist (H. K. Genant, University of California, San Francisco, California). Wrist include all of the distal radius or ulna, and humeral include only of the proximal humerus. Fractures that resulted from major trauma, such as motor vehicle accidents, were excluded (5 percent of all ). Vertebral were assessed by morphometry (quantitative assessment of vertebral heights by digitization), with comparison of baseline and follow-up lateral thoracic and lumbar spine films obtained a mean of 3.7 years apart. An incident vertebral fracture was defined as at least a 20 percent and 4-mm decrease in the anterior, middle, or posterior height of the vertebral body (33). The study radiologist reviewed discrepancies between morphometry and a technician's qualitative assessment of radiographs; the radiologist also visually confirmed all incident vertebral (34, 35). Follow-up radiographs were not done for 2,466 women. The baseline prevalence of vertebral deformity was very similar in women with (20 percent) and without (21 percent) follow-up radiographs. Statistical analysis We used proportional hazards models to assess the relation between calcium intake and the various types of nonvertebral. Logistic regression was used to analyze the vertebral fracture data, as these were identified by a planned radiographic survey. For simplicity, we use the term relative risk to describe

3 928 dimming et al. both hazard ratios (from proportional hazard models) and odds ratios (from logistic regression models). Follow-up duration in proportional hazards models was the time from baseline interview/examination in to first relevant fracture, death, or end of follow-up (March 1, 1995), whichever came first. For hip fracture analyses, we excluded women with a history of hip fracture prior to enrolment in the study. This approach was not used for other types of fracture, because their self-report is less accurate (32). Unless otherwise stated, analyses involving new vertebral included women with and without vertebral on baseline radiographs. Selection of potential confounders for inclusion in multiple factor models was based on 1) previously published research on risk factors for (mostly research on hip ) and 2) the distribution of these risk factors across categories of dietary calcium and supplement use among study participants. Dietary calcium intake is a continuous variable; however, we also categorized dietary calcium on the basis of current beliefs about optimal calcium intake and the availability of sufficient numbers of subjects and in each category: less than 400 mg/day, mg/day, 800-1,199 mg/day, and 1,200 or more mg/day. We tested for trends in relative risks across categories by modeling the median calcium intakes for each category as a single continuous variable. Analyses for calcium supplements, Turns antacid tablets, and vitamin D supplements compared current users with never users and exusers with never users. total daily calcium intake was estimated by summing the milligrams of calcium consumed in the diet, in calcium supplements, and in Turns antacid tablets (each tablet contains about 200 mg of calcium). We assessed the interaction between calcium intake and the history of or osteoporosis using the Breslow-Day test for heterogeneity in stratified analyses and by including interaction terms in multiple factor models. We used similar techniques to assess the interaction between calcium intake and physical activity, use of estrogen replacement therapy, and family history of. All p values are two sided; a p value < 0.05 was used to define statistically significant associations. All analyses were done with SAS version 6.08 software (SAS Institute, Inc., Cary, North Carolina) using an IBM-compatible personal computer. Literature review Studies on calcium intake and were identified by reviewing all publications known to the authors that might contain relevant data. Reference lists of these articles were also reviewed, and possibly relevant articles were retrieved. A few additional articles were identified by conducting a computerized Medline search for the period from January 1966 to December 1995 using the text words "fracture" and "calcium." To be eligible for our review, studies had to have (or a particular type of fracture) as the outcome. RESULTS During a mean of 6.6 years of follow-up of 9,704 subjects, there were 332 hip, 210 ankle, 241 proximal humeral, and 467 wrist. The total number of nonvertebral was 1,9. There were 389 new vertebral among the 7,238 women who had both baseline and follow-up vertebral radiographs. The baseline distribution of dietary calcium intake and the use of calcium supplements, Turns antacid tablets, and vitamin D supplements among women in our study are shown in table 1. The mean daily dietary calcium intake was 714 mg (standard deviation, 425 mg). The mean dietary calcium intake varied by clinic from a low of 627 mg in Pittsburgh to a high of 772 mg in Minneapolis. TABLE 1. Distribution of intakes of dietary calcium, milk, calcium supplements, Turns antacid tablets, and vitamin D supplements, Study of Osteoporotic Fractures, from to 1995 dietary calcium (mg/day) < ,199 1,200 milk (glasses/day) Rarely/never < Past Turns antacid tablets Past Vitamin D supplements Past No. of subjects 2,418 3,992 2,073 1,221 3,361 2,302 3, , ,132 7, ,571 4, ,273 O/ vo

4 Calcium Intake and Fractures 929 Table 2 shows risk factors for that were associated with calcium intake in our data. Compared with women having lower dietary calcium intakes, those having higher intakes were more likely to have histories of osteoporosis and, to have fallen in the past year, and to use estrogen replacement therapy; they were less likely to use thiazide diuretics. These women also tended to have better contrast sensitivity scores, to eat more protein, to consume less caffeine, and to be more physically active. Dietary calcium intake was not associated with the use of calcium or Turns antacid tablets, but vitamin D use was more common among women with higher dietary calcium intakes. The major differences between users and nonusers of calcium supplements were that users were more likely to have histories of osteoporosis and and to have used estrogen replacement therapy (see table 2). Users also tended to be more physically active. Turns antacid tablets and vitamin D users were more likely than were nonusers to have histories of, osteoporosis, and estrogen use. We found a positive relation between the use of calcium and vitamin D supplements and a negative relation between the use of calcium supplements and Turns antacid tablets. Table 3 shows adjusted relative risks and 95 percent confidence intervals for the relation between dietary calcium intake and the risk of hip, ankle, proximal humeral, wrist, and vertebral. None of the individual relative risks or tests for trend was statistically significant, but there was some suggestion of TABLE 2. Selected baseline characteristics* of subjects according to dietary calcium intake and use of calcium supplements, Turns antacid tablets, and vitamin D supplements, Study of Osteoporotic Fractures, Dietary calcium (mg/day) <400 1,200 Turns antacid tablets vitamin D supplements Age (years) Physical characteristics Weight (kg)t Low frequency contrast sensitivity scoret Physical activity (teal/ week)f 1,466 1,874 1,449 1,797 1,586 1,687 1,517 1,697 Physical activity Take walks for exercise Caffeine Intake (lifetimeflt 8,329 7,020 8,497 7,567 8,247 7,380 8,601 7,441 nipt LJlOl Calcium Intake (mg/day)t Protein intake (9/day)t Health status Fall in pasi year osteoporosis history Fracture since age calcium supplements Tums antacid tablets vitamin D supplements estrogen thiazides Dietary calcium (mg/day) <400 1,200 Tums antacid tablets vitamin D supplements * Except for age, all listed characteristics were statistically significantly (p < 0.05) associated with at least one of trie four exposure variables (using a chi-squared test or analysis of variance), t Age-adjusted means. $ Number of mg/day of caffeine multiplied by number of years of use of caffeine

5 930 Cumming et al. TABLE 3. Adjusted relative risks* and 95% confidence intervals for dietary calcium and milk intake and risk of of various types, Study of Osteoporotic Fractures, from to 1995 Any nonvertebral fracture (n= 1,854) Hip (n = 306) Ankle (n = 195) Proximal humeral (n = 216) Wrist (n = 421) Vertebral. (n = 354) dietary calcium intake (mg/day) < ,199 1, ( )t 0.9( ) 0.9( ) 1.0( ) 0.8( ) 0.9( ) 1.2( ) 0.8( ) 0.6 ( ) 1.0( ) 0.9( ) 0.9( ) 1.0( ) 1.4( ) 0.9( ) 1.2( ) 1.2( ) 1.5( ) Trend (p) milk intake (glasses/day) Rarely/never < S3 Trend (p) 1.0( ) 1.0( ) 1.0( ) ( ) 0.9( ) 0.9 ( ) ( ) 0.7 ( ) 0.4 ( ) ( ) 1.0( ) 1.2( ) ( ) 1.1 ( ) 0.8( ) ( ) 1.3( ) 1.4( ) * Relative risks are adjusted for age, clinic, weight, history of osteoporosis, history of since age, fall in past 12 months, protein intake (g/week), caffeine intake (see table 2), recreational physical activity (teal/week), take walks for exercise (yes/no), impaired low frequency contrast sensitivity, estrogen replacement therapy (current/not current), thiazide use (current/not current), and use of calcium supplements, vitamin D supplements, and Turns antacid tablets. t Numbers in parentheses, 95% confidence interval reduced risk of ankle with increased dietary calcium intake (test for trend p = 0.08). Also shown in Table 3 are adjusted relative risks for the association between current milk drinking and. The findings for milk were essentially the same as for overall dietary calcium: no association with any type of fracture except ankle. We also calculated relative risks for milk consumption at the ages of 12-18,19-, and after the age of years and during pregnancy and lactation (data not shown). We used the same categories of milk consumption as for current age. The only statistically significant finding was a protective association between drinking milk during pregnancy and lactation and the risk of first vertebral fracture among women with no vertebral on baseline radiographs (test for trend p = 0.01). Results for the use of calcium supplements, Turns antacid tablets, and vitamin D supplements are shown in table 4. Compared with never users, current users of calcium supplements were at increased risk of hip (relative risk = 1.5) and vertebral (relative risk = 1.4), and current users of Turns antacid tablets were at increased risk of proximal humeral (relative risk = 1.7). Among current users, relative risks were similar for women who had used supplements for more than 10 years and women who had used them for shorter periods of time. Associations between the use of calcium supplements and fracture risk were the same in women with low and high dietary calcium intakes (data not shown). Since vitamin D and calcium are both needed to maintain calcium balance, we calculated relative risks comparing current users of both vitamin D and calcium supplements (n = 3,365) with people who had never used either (n = 2,705). There was no evidence that a combination of calcium and vitamin D supplements protected against ; adjusted relative risks were 1.2 (any fracture), 1.4 (hip ), 0.7 (ankle ), 1.4 (proximal humeral ), 1.0 (wrist ), and 1.3 (vertebral ). To minimize confounding by indications for use of supplements, we examined the relation between and the current use of calcium supplements, Turns antacid tablets, and vitamin D supplements in women who had neither a history of previous since age nor a history of osteoporosis (n = 5,230). were no longer associated with hip among these women, but the associations between calcium supplements and vertebral and between Turns antacid tablets and proximal humeral persisted (see table 5). The use of vitamin D supplements was associated with an increased risk of hip fracture among these women without previous or a history of osteoporosis. Higher current daily intake of calcium from all sources (dietary calcium, calcium supplements, and Turns antacid tablets) was associated with increased

6 Calcium Intake and Fractures 931 TABLE 4. Adjusted relative risks* and 95% confidence intervals for current use of calcium supplements, Turns antacid tablets, and vitamin D supplements and risk of of various types in all study subjects, Study of Osteoporotic Fractures, from to 1995 Ex- <10 years 510 years Any nonvertebraj fracture (n= 1,854) 1.1 ( )f 1.1 ( ) 1.2( ) 1.1 ( ) HI) (n=306) 0.9( ) 1.5( ) 1.6( ) 1.2( ) Ankle (n=195) 1.1 ( ) 1.1 ( ) 1.1 ( ) 1.0( ) Proximal humeral tract uros II OwtUJ 5> (n = 216) 0.9( ) 1.2( ) 1.2( ) 1.2( ) Wrist (n = 421) 1.0( ) 1.2( ) 1.2( ) 1.1 ( ) Vertebral (n = 354) 1.3( ) 1.4( ) 1.5( ) 1.5( ) Turns antacid tablets Ex- <10 years 510 years 1.1 ( ) 1.0( ) 1.0( ) 1.0( ) 1.5( ) 0.8( ) 0.7( ) 1.0( ) 0.9( ) 0.8( ) 0.7( ) 1.0( ) 1.4( ) 1.7( ) 1.9( ) 0.9 ( ) 1.3( ) 1.0( ) 1.0( ) 1.0( ) 1.2( ) 1.1 ( ) 1.1 ( ) 1.0( ) Vitamin D supplements Ex- <10 years 510 years 1.1 ( ) 1.1 ( ) 1.0( ) 1.1 ( ) 0.7( ) 1.1 ( ) 1.0( ) 1.2( ) 0.8( ) 0.7( ) 0.7( ) 0.8 ( ) 0.9( ) 1.0( ) 1.1 ( ) 0.8( ) 1.1 ( ) 0.9( ) 0.8( ) 1.0( ) 0.8( ) 0.8( ) 0.9( ) 0.8( ) * Relative risks are adjusted for age, clinic, weight, history of osteoporosis, history of since age, fall in past 12 months, dietary calcium intake (mg/week), protein intake (g/week), caffeine intake (see table 2), recreational physical activity (kcal/week), take walks for exercise (yes/no), impaired low frequency contrast sensitivity, estrogen replacement therapy (current/not current), thiazide use (current/not current), and use of calcium supplements, vitamin D supplements, and Turns antacid tablets (where relevant). t Numbers in parentheses, 95% confidence interval. risks of hip (test for trend p = 0.04), proximal humeral (test for trend p = 0.05), and vertebral (test for trend p < 0.001) (data not shown). Adjusted relative risks comparing those in the top fifth (s 1,600 mg/day of calcium) with those in the bottom fifth (<400 mg/day of calcium) were 1.6 (hip ), 1.7 (proximal humeral ), and 2.5 (vertebral ). We examined the two subtypes of hip fracture (trochanteric and femoral neck ) separately (data not shown). There were no differences for dietary calcium and milk, but current use of calcium supplements appeared to be more strongly associated with trochanteric than with femoral neck. Among all subjects, adjusted relativeriskscomparing current with never users were 1.8 for trochanteric and 1.3 for femoral neck. There were no meaningful interactions between calcium intake and physical activity, estrogen replacement therapy, or family history of. DISCUSSION In this large cohort study, we found no associations between dietary calcium intake and the risk of several different types of. were associated with increased risk of some types of fracture. Interpretation of our findings for calcium supplements is difficult, because these are now widely recommended for prevention and treatment of osteoporosis. It is probable that people at high risk of are more likely to start taking calcium supplements than are people at lower risk. In our study, for example, supplement users were more likely than were nonusers to have a history of falls,, or osteoporosis. We believe that this confounding by indication (36) is the most likely explanation for our finding of an increased risk of hip and vertebral fracture in users of calcium supplements and of proximal humeral in Turns antacid tablets users. We cannot, however, exclude the possibility that supplements taken by our subjects had an adverse effect on fracture risk. Other observational studies of calcium supplements have found increased risks of (14), no association (13, 37), and reduced fracture risk (38). The one study to find a protective effect was conducted among southern European women (38), a population in which the type of women who use supplements, and

7 932 Cumming et al. TABLE 5. Adjusted relative risks* and 95% confidence intervals for current use of calcium supplements, Tums antacid tablets, and vitamin D supplements and risk of of various types, according to history of osteoporosis or fracture since the age of, Study of Osteoporotic Fractures, from to 1995 No history of osteoporosis or since age Tums antacid tablets vitamin D supplements Any nonvertebral (n = 801) 1.0 ( )f 1.0( ) 1.2( ) Hip (n=h3) 1.0( ) 0.8( ) 1.6( ) Ankle (n=105) 1.0( ) 0.8( ) 0.8( ) Proximal humeral (n = 82) 1.0( ) 2.0( ) 1.0( ) Wrist (n=181) 0.9( ) 0.8( ) 0.9( ) Vertebral (n=131) 1.6( ) 1.4( ) 0.8( ) Any nonvertebral (77=1,053) 1.2( ) Tums antacid tablets 1.0( ) Vitamin D su pplements 1.0 ( ) Hip (n=193) 2.0( ) 0.9( ) 0.8( ) History of osteoporosis or since age Ankle (n = 90) 1.4( ) 0.8( ) 0.6( ) Proximal humeral (n=134) 1.5( ) 1.6( ) 1.0( ) Wrist (n = 240) 1.4( ) 1.2( ) 0.8( ) Vertebral (r>=223) 1.3( ) 0.9( ) 0.8( ) * Relative risks are adjusted for age, clinic, weight, fall in past 12 months, dietary calcium intake (mg/week), protein intake (g/week), caffeine intake (see table 2), recreational physical activity (kcal/week), take walks for exercise (yes/no), impaired low frequency contrast sensitivity, estrogen replacement therapy (current/not current), thiazide use (current/not current), and use of calcium supplements, vitamin D supplements, and Turns antacid tablets (where relevant). t Numbers in parentheses, 95% confidence interval. the indications for their use, may be very different from the participants in North American studies. There have now been three small randomized trials of calcium supplements and radiographically defined, mostly asymptomatic, vertebral (21-23). Recker et al. (21) found a statistically significant reduction in fracture rate among women taking calcium supplements. Chevalley et al. (22) also found a protective effect, but the statistical significance of this effect was dependent on the way that vertebral were defined. Finally, although they found no difference in vertebral fracture rates, Reid et al. (23) reported fewer symptomatic among women taking calcium (two ) than among those on placebo (nine ). We found no evidence of a protective effect of vitamin D. A combination of vitamin D and calcium supplements has been shown to reduce the rate of in nursing home residents (24). However, the role of vitamin D supplements alone remains controversial. Ooms et al. (39) have shown that a small dose of vitamin D increases bone mass and, although unproven at present, vitamin D might also improve muscle strength and reduce falls (40). In a randomized trial of people living in the community in the Netherlands, Lips et al. (41) found no effect of daily use of 400 International Units of vitamin D on the risk of hip fracture. We found limited evidence of a protective effect of dietary calcium for, consistent with most previous studies. Of 14 previous epidemiologic studies of dietary calcium and the risk of hip (7-20), three found a statistically significant protective effect (9, 10, 18). One of these studies, a case-control study from Hong Kong, was done among people with a very low mean calcium intake (less than 170 mg/day (10)). The one North American study to find a protective effect was the Rancho Bernardo cohort study, where there was a 40 percent reduction in hip fracture risk for each 300-mg increase in dietary calcium (9). This was the one study to adjust for total caloric intake, using the nutrient density technique. This technique has several limitations (42). The other study to find a protective effect was conducted among women living in southern European countries and involved only milk consumption (18). Two large US cohort studies have found no association between dietary calcium and hip fracture (13, 16). Paganini-Hill et al. (13) followed 8,600 women for a mean of 5 years and found a relative risk for hip fracture of 1.1 (95 percent confidence interval ) for women consuming more than 0 mg/day of calcium compared with those consuming less than 280 mg/day. Looker et al. (16) followed 2,226 women for a mean of 15 years and found a relative risk for hip fracture of 0.9 (95 percent confidence interval ) for those consuming more than 1,000 mg/day of calcium compared with those consuming less than 400 mg/day. Case-control studies from the United States (15), the United Kingdom (7, 8, 11), Canada (12, 14), Scandinavia (19, 20), and Australia (17) have also failed to find a protective effect of dietary calcium.

8 Calcium Intake and Fractures 933 Apart from an earlier report from the Study of Osteoporotic Fractures (43), there has been one other study of dietary calcium and wrist (14). In a Canadian hospital-based case-control study, Kreiger et al. (14) found a statistically significant reduction in the risk of wrist in women who consumed more than 1,000 mg/day of calcium. There have been no previous studies of dietary calcium and the risk of vertebral, ankle, or humeral. For assessment of dietary calcium and, the Study of Osteoporotic Fractures has several strengths: prospective design, large numbers of of many different types, good measurement of dietary calcium, and assessment of, and control for, a wide range of potential confounders. However, dietary calcium intake is difficult to assess accurately, and "noise" inherent in our measurement may have caused us to underestimate its true effect. Our study was done among women with a mean dietary calcium intake of 700 mg/day. Hence, our results do not rule out a large calcium effect in populations with much lower mean intakes. Despite its large size, our study had limited power to detect small associations between dietary calcium intake and specific fracture types. For hip, the study had 70 percent power to detect a relative risk of at least 0.7 for those with a calcium intake greater than 400 mg/day compared with those with lower intakes. Power was high for all combined; the study had 90 percent power to detect a relative risk of at least 0.85 comparing these two groups of women. Although the women in the Study of Osteoporotic Fractures were volunteers, this should not affect the generalizability of our findings for relations between risk factors and. We also note that fracture incidence in our cohort is only slightly lower than in more representative populations. For example, in the age group in which hip occur most frequently (80-84 years), hip fracture incidence in our study is 11.8 per 1,000 women per year, compared with 15.1 per 1,000 women per year in a population-based study in New England (44). In conclusion, we were unable to find a substantial beneficial effect of dietary calcium on fracture risk. Our data also confirm the difficulty of studying the effects of calcium supplements in observational studies. Randomized trials of calcium supplements with symptomatic as the outcome are the best way to determine the efficacy of calcium for the prevention of. A large trial of supplements of calcium together with vitamin D is currently underway in the United States, as part of the Women's Health Initiative that is sponsored by the National Institutes of Health. ACKNOWLEDGMENTS Supported by grants (1-RO1-AG05407, 1-RO1- AR35582, 5-RO1-AG05394, 1-RO1-AM35584, and 1-RO1-AR35583) from the Public Health Service. Investigators in the Study of Osteoporotic Fractures research group: University of California, San Francisco (Coordinating Center): S. R. Cummings (principal investigator), M. C. Nevitt (coinvestigator), D. G. Seeley (project director), D. M. Black (study statistician), H. K. Genant (director, central radiology laboratory), C. Arnaud, D. Bauer, W. Browner, L. Christianson, M. Dockrell, C. Fox, C. Gluer, S. Harvey, M. Jergas, Mario Jaime-Chavez, R. Lipschutz, G. Milani, L. Palermo, A. Pressman, R. San Valentin, K. Stone, H. Tabor, D. Tanaka, and C. Yeung. University of Maryland: J. C. Scott (principal investigator), R. Sherwin (co-principal investigator), M. C. Hocberg (coinvestigator), J. Lewis (project director), Cheryl Bailey (clinic coordinator), A. Bauer, L. Finazzo, G. Greenberg, D. Harris, B. Hohman, S. Kallenberger, E. Oliner, T. Page, A. Pettit, S. Snyder, L. Stranovsky, and S. Trusty. University of Minnesota: K. Ensrud (principal investigator), R. Grimm, Jr. (coinvestigator), C. Bell (project director), E. Mitson (clinic coordinator), M. Baumhover, C. Berger, S. Estill, S. Fillhouer, J. Hansen, K. Jacobson, K. Kiel, C. Linville, N. Nelson, E. Penland-Miller, and Jayne Griffith. University of Pittsburgh: J. A. Cauley (principal investigator), L. H. Kuller (co-principal investigator), M. Vogt (coinvestigator), L. Harper (project director), L. Buck (clinic coordinator), C. Bashada, A. Githens, A. McCune, D. McDue, M. Nasim, C. Newman, S. Rudovsky, N. Watson, and J. Carothers. Kaiser Permanente Center for Health Research, Portland, Oregon: T. M. Vogt (principal investigator), W. M. Vollmer, E. Orwoll, H. Nelson (coinvestigators), J. Blank (project director), S. Craddick (clinic coordinator), R. Bright, J. Wallace, F. Heinith, K. Moore, K. Redden, C. Romero, and C. Souvanlasy. REFERENCES 1. dimming RG. Calcium intake and bone mass: a quantitative review of the evidence. Calcif Tissue Int 1990;47: Dawson-Hughes B, Dallal GE, Krai! EA, et al. A controlled trial of the effect of calcium supplementation on bone density in postmenopausal women. N Engl J Med 199O;323: Reid IR, Ames RW, Evans MC, et al. Effect of calcium supplementation on bone loss in postmenopausal women. N Engl J Med 1993;328: Aloia JF, Vaswani A, Yeh JH, et al. Calcium supplementation with and without hormone replacement therapy to prevent postmenopausal bone loss. Ann Intern Med 1994;120: Elders PJ, Netelenbos JC, Lips P, et al. Calcium supplementation reduces vertebral bone loss in perimenopausal women: a controlled trial in 248 women between 46 and 55 years of age. J Clin Endocrinol Metab 1991;73: Nelson ME, Fisher EC, Dilmanian FA, et al. A 1-year walking program and increased dietary calcium in postmenopausal women: effects on bone. Am J Clin Nutr 1991;53:1304-ll. 7. Wootton R, Brereton PJ, Clark MB, et al. Fractured neck of femur in the elderly: an attempt to identify patients at risk. Clin Sci 1979;57: Cooper C, Barker DJP, Wickham C. Physical activity, muscle

9 934 Cumming et al. strength, and calcium intake in fracture of the proximal femur in Britain. BMJ 1988;297: Holbrook TL, Barrett-Connor E, Wingard DL. Dietary calcium and risk of hip fracture: 14-year prospective population study. Lancet 1988;2: Lau E, Donnan S, Barker DJP, et al. Physical activity and calcium intake in fracture of the proximal femur in Hong Kong. BMJ 1988;297: Wickham CAC, Walsh K, Cooper C, et al. Dietary calcium, physical activity, and risk of hip fracture: a prospective study. BMJ 1989;299: Jaglal SB, Kreiger N, Darlington G. Past and recent physical activity and risk of hip fracture. Am J Epidemiol 1993; 138: Paganini-Hill A, Chao A, Ross RK, et al. Exercise and other factors in the prevention of hip fracture: the Leisure World Study. Epidemiology 1991 ;2: Kreiger N, Gross A, Hunter G. Dietary factors and fracture in postmenopausal women: a case-control study. Int J Epidemiol 1992:21: Nieves JW, Grisso JA, Kelsey JL. A case-control study of hip fracture: evaluation of selected dietary variables and teenage physical activity. Osteoporos Int 1992;2: Looker AC, Harris TB, Madans JH, et al. Dietary calcium and hip fracture risk: the NHANES I Epidemiologic Follow-up Study. Osteoporos Int 1993:3: Cumming RG, Klineberg RJ. Case-control study of risk factors for hip in the elderly. Am J Epidemiol 1994; 139: Johnell OF, Gullberg B, Kanis JA, et al. Risk factors for hip fracture in European women: the MEDOS Study. J Bone Miner Res 1995; 10: Michaelsson K, Holmberg L, Mallmin H, et al. Diet and hip fracture: a case-control study. Int J Epidemiol 1995;24: Meyer HE, Henriksen C, Falch JA, et al. Risk factors for hip fracture in a high incidence area: a case-control study from Oslo, Norway. Osteoporos Int 1995;5: Recker RR, Hinders S, Davies KM, et al. Correcting calcium nutritional deficiency prevents spine in elderly women. J Bone Miner Res 1996;11: Chevalley T, Rizzoli R, Nydegger V, et al. Effects of calcium supplements on femoral bone mineral density and vertebral fracture rate in vitamin D-replete elderly patients. Osteoporos Int 1994:4: Reid IR, Ames RW, Evans MC, et al. Long-term effects of calcium supplementation on bone loss and in postmenopausal women: a randomized controlled trial. Am J Med 1995:98: Chapuy MC, Arlot ME, Duboeuf F, et al. Vitamin D 3 and calcium to prevent hip in elderly women. N Engl J Med 1992:327: Block G, Hartman AM, Dresser CM, et al. A data-based approach to diet questionnaire design and testing. Am J Epidemiol 1986:124: Cummings SR, Block G, McHenry K, et al. Evaluation of two food frequency methods of measuring dietary calcium intake. Am J Epidemiol 1987;126: Cummings SR, Nevitt MC, Browner WS, et al. Risk factors for hip in white women. N Engl J Med 1995;332: Bauer DC, Browner WS, Cauley JA, et al. Factors associated with appendicular bone mass in older women. Ann Intern Med 1993:118: Kriska AM, Sandier RB, Cauley JA, et al. The assessment of historical physical activity and its relation to adult bone parameters. Am J Epidemiol 1988;127: Ginsberg AP. A new contrast sensitivity vision test chart. Am J Optom Physiol Opt 1984;64: Gibson JJ. The perception of the visual world. Boston: Houghton Mifflin, Nevitt MC, Cummings SR, Browner WS, et al. The accuracy of self-report of in elderly women: evidence from a prospective study. Am J Epidemiol 1992; 135: Black D, Palermo L, Genant HK, et al. Determining the best morphometric definition of incident vertebral fracture: a prospective comparison of three approaches. (Abstract). J Bone Miner Res 1995;10:S Black DM, Palermo L, Nevitt MC, et al. Comparison of methods for defining prevalent vertebral deformities: the Study of Osteoporotic Fractures. J Bone Miner Res 1995;10: Genant HK, Wu CY, van Kuijk C, et al. Vertebral fracture assessment using a semiquantitative technique. J Bone Miner Res 1993:9: Greenland S, Neutra R. Control of confounding in the assessment of medical technology. Int J Epidemiol 1980;9: Paganini-Hill A, Ross RK, Gerkins VR, et al. Menopausal estrogen therapy and hip. Ann Intern Med 1981;95: Kanis JA, Johnell O, Gullberg B, et al. Evidence for efficacy of drugs affecting bone metabolism in preventing hip fracture. BMJ 1992;305:l Ooms ME, Roos JC, Bezemer PD, et al. Prevention of bone loss by vitamin D supplementation in elderly women: a randomized double-blind trial. J Clin Endocrinol Metab 1995;80: Gloth FM, Tobin JD. Vitamin D deficiency in older people. J Am Geriatr Soc 1995;43: Lips P, Graafmans WC, Ooms ME, et al. Vitamin D supplementation and fracture incidence in elderly persons. Ann Intern Med 1996;124: Willett W, Stampfer MJ. Total energy intake: implications for epidemiologic analyses. Am J Epidemiol 1986;124: Kelsey JL, Browner WS, Seeley DG, et al. Risk factors for of the distal forearm and proximal humerus. Am J Epidemiol 1992;135: Fisher ES, Baron JA, Malenka DJ, et al. Hip fracture incidence and mortality in New England. Epidemiology 1991;2: