The average American woman has nearly a one in three chance of fracturing a hip at some time during her life.1 Most often the fracture occurs late in life and is a result of a minor injury to an osteoporotic femur. The personal costs of the fracture are pain, disability, dependency, prolonged need for medical care, a marked increase in the risk of secondary illness, and a substantial increase in the probability of premature death. The economic costs to the Nation are staggering. The approximately 200,000 fractures that occur every year represent a total cost in the billions of dollars.2 The total number and cost of such fractures may well double by the turn of the century as a result of escalating numbers of elderly. These considerations underline the importance of research aimed at the prevention of hip fractures.
Although prevention of accidents is an essential part of the total prevention strategy, the greatest hope for reducing hip fractures among the elderly is related to delaying the onset or retarding the rate of progression of the osteoporotic process associated with aging. Therapies currently being evaluated and/or recommended include the use of estrogena, an increased intake of dietary calcium, vitamin D, and fluorides.3 Most of the research now underway is focused on ameliorating osteoporosis that has already been diagnosed. It will be many years before prospective studies are able to address the efficacy of these approaches in delaying the onset or slowing the loss of bone during normal aging.
An important epidemiologie research approach involves comparing the characteristics (personal, demographic) of persons who experience fractures with those of people who escape them. This is because personal attributes associated with a low or high risk of fracture may reflect important prior exposures and life experiences pathogemcally. The approach involves identifying population subgroups of unusually high or unusually low risk, then asking what common factors in their backgrounds might have influenced the risk of osteoporosis or fracture.
We have begun to examine these questions by comparing age-specific incidence curves for subpopulations of specific race and sex. Although the figures for the general population are quite large, the actual annual number of fractures in certain subpopulation age strata (for example, among black females age 75 to 79) is small. This problem is due to the lack of good descriptive information on the age-specific incidence in such groups. In order to accomplish this objective, we examined hospital discharge data from a 6-year time period drawn from the National Hospital Discharge Survey.4'10 The use of this very large data set has allowed us to estimate with unprecedented precision the curves of increasing risk with advancing age in white females, black females, white males, and black males in the United States.
The results of these epidemiologie investigations quantify the clinical observation that hip fractures occur more commonly in females, and especially in white females. However, the difference between white and black females, a two-fold increase in risk associated with being white, is substantially less than many orthopedic surgeons might have guessed. The practicing surgeon is likely to see many more fractures among white women than among black women. This is largely because the likelihood of a black woman surviving to the age of very high risk for hip fracture is substantially lower than the same likelihood for a white woman.
There appears to be something associated with being a black woman which is protective. At every age, black women are at approximately half the risk of white women for experiencing a hip fracture. The question we must ask is, "Does the protection against hip fracture conferred by being a black woman (rather than a white woman) reflect something different in the environment and/or life-style of the black woman, or is it merely an immutable expression of racial difference?" If the difference in risk is mostly due to environmental or life-style differences, perhaps careful study will lead us to define useful prevention strategies which may be applicable to all.
Although the answer to the question of possible differences in environment or lifestyle is not yet known, our analysis of the hospital discharge survey data suggests that the difference is not attributable to race alone. There seems to be no difference in the rate at which the risk of fracture increases with age among black males as compared with white males.
The other interesting observation is that the curve of increasing risk with advancing age seems to be the same for white males, black males, and black females. The other side of the coin, then, is that the increased risk associated with being female does not hold up if one looks at black individuals only. The common (and apparently incorrect) clinical impression that females are at greater risk than males at any age and in both races could easily be due to differential survival to the age at which fractures become common. This is because white females are the most likely to survive, followed by white males, then black females, and finally black males. The resulte we observed took this into accunt and examined the risk of a fracture without being influenced by differential survival.
The rate at which the risk of a fracture increases with advancing age doubles about every five to six years, as shown in the accompanying figure. It is clear that all groups show this increasing risk with advancing age. The two-fold increased risk at any specific age associated with being a white female can be viewed as a result of the incidence curve beginning to rise above baseline levels five to six years earlier in white females as compared with the other three groups. The curves are smooth throughout, showing no change in the rate of accelerating risk around the time of menopause. Since the smooth acceleration of the incidence curve apparently begins well before menopause, the term "postmenopausal osteoporosis" may be misleading. The factors that define its onset and rate probably have their origins well before menopause. In addition, the influence of age seem to be equally expressed in males.
These ideas have been corroborated by objective studies of bone density that suggest demineralization usually begins in the 20s or 30s for both men and women.11 The greater age-specific incidence of hip fractures in white women is thought to reflect a lesser bone strength, in turn related to a lesser mineralization of the bony elements associated with resistance to fracture. If the osteoporotic process progressed at the same rate and began at the same time in all four race/sex groups, but if white females simply had a lesser reservoir of calcified bone to begin with , the process would be expressed as an increased risk of fracture at an earlier age in white females as compared with the other three groups. This reasoning leads to an interesting approach to fracture prevention. It suggests that effective methods for the prevention of hip fractures in later life might well include therapies, diets, or life-style changes which increase bone mineralization and strength in the teens and 20s, followed by steps to sustain a good level of bone strength into middle and later life.
The test of whether or not a prevention strategy could then be expected to reduce the age-specific risks of hip fracture in late life might be the extent to which the strategy increased and/or maintained bone mineralization in early adult life. This suggests that prevention approaches should be focused much earlier than usually considered, and the efficacy of such factors as diet and exercise should be tested by monitoring effects on bone mineralization during that early period of life. This would represent true primary prevention. The use of estrogene and perhaps other medications in patients with documented osteoporosis would then become secondary prevention or therapy, depending upon the point of view.
A great deal of research remains to be done. For example, we need to compare the way in which nutrition and lifestyle factors are associated with the risk of a fracture across defined sex/race groups. We need to know if black women who have lived their lives in middle and upper socioeconomic strata have risks for fracture closer to those of black women of lower socioeconomic strata or to those of middle and upper socioeconomic white women. The corollary of this question is whether or not white women from lower socioeconomic strata have risks for fracture more similar to those of black women (considered as a total group) or to middle and upper socioeconomic white women.
Another possibility is that the lower risk among black women simply refleets a common habitus difference between black and white women, since many black women are heavier than their white counterparts. This consideration reflects both the observation that increased weight seems to protect against the risk of a hip fracture, and the physiologic knowledge that the strength of a bone reflects the stresses which have been placed upon it. In the case of the femur, this may well be represented by the time spent in weight-bearing, taken together with the actual weight that the bones must bear. Another consideration is that there may be estrogen level differences associated with differential body fat composition, or with the nutritional behaviors of different race, sex, or socioeconomic groups.
These research avenues are currently being explored in clinics, laboratories, and research institutes around the world. We are optimistic that the emerging knowledge will ultimately be expressed in a reduction in the risk of hip fracture in late life, in our own generations as well as those that follow us.
- 1. Gallagher JC, Melton LJ, Riggs BL, Bergstrath BA: Epidemiology of fractures of the proximal femur in Rochester, Minnesota, Clin Onhop 1980; 150:163-171.
- 2. Owen RA, Melton LJ, Gallagher JC, et al: The national cost of acute care of hip fractures associated with osteoporosis, Clin Orthop 1980; 150:172-176.
- 3. Whedon GD: Osteoporosis, New England J M ed 1981; 305:397-399.
- 4. National Center for Health Statistics, lnpatient utilization of short-stay hospitals by diagnosis. United States 1975. Hyattsville, MD: National Centerfor Health Statistics, 1978. (Vital and health statistics, Series 13: Data from the National Health Survey, no. 35) (DHEW publication no. (PHS) 78-1786).
- 5. National Center for Health Statistics. Utilization of short-stay hospitals: annual summary for the United States, 1976. Hyattsville, MD: National Center for Health Statistics, 1978. (Vital and health statistics, Series 13: Data from the National Health Survey, no. 37) (DHEW publication no. (PHS) 78-1788).
- 6. National Center for Health Statistics. Utilization of short-stay hospital annual summary of the United States, 1977. Hyattsville, MD: National Centerfor Health Statistics, 1979. (Vital and health statistics. Series 13: Data from the National Health Survey, no. 41) (DHEW publication no. (PHS) 79-1557).
- 7. National Center for Health Statistics. Utilization of short-stay hospitals: annual summary for the United States, 1978. Hyattsville, MD: National Center for Health Statistics, 1980. (Vital and health statistics, Series 13: Data from the National Health Survey, no. 46) (DHEW publication no. (PHS) 80-1797).
- 8. National Center for Health Statistics. Utilization of short-stay hospitals; annual summary for the United States, 1979. Hyattsville. MD: National Center for Health Statistics, 1981 . (Vital and health statistics, Series 13: Data from the National Health Survey, no. 60) (DHHS publication no. (PHS) 82-1721).
- 9. National Center for Health Statistics. Eighth revision. International classification of diseases: adapted for use in the United States. Hyattsville, MD: National Center for Health Statistics, 1967. (PHS publication no. 1693).
- 10. Farmer ME, White LR, Brody JA: Race and sex differences in hip fracture incidence. Am J Pub Health (In press).
- 11. Riggs BL, Wahner HW, Seeman E, et al: Changes in bone mineral density of the proximal femur and spine with aging, J Clin invest 1982; 70:716-723.