Efficacy of treatment of osteoporotic fractures

Efficacy of Treatment of Osteoporotic Fractures JOHNA. EISMAN, M.B., B.S., Ph.D., Geneva, Switzerlandand Sidney, NSW, Australsa

The efficacy o f o s t e o p o r o s i s t h e r a p y can be c o n s i d e r e d at s e v e r a l d i f f e r e n t levels, ranging from t h e overall c o m m u n i t y to the individual patient. Efficacy has d i f f e r e n t c o n n o t a t i o n s for each group within t h e c o m m u n i t y and for the individual, in w h o m life e x p e c t a n c y - among o t h e r p a r a m e t e r s h i s an i m p o r t a n t consideration. Some o f the q u e s t i o n s add r e s s e d in this article are the scope o f the o s t e o p o r o s i s problem, how many people m u s t be t r e a t e d f o r p r e v e n t i o n o f one f u r t h e r fracture, w h e t h e r t h e r e is a role for targeting likely high risk/high r e s p o n s e groups, and what e x a c t l y is a successful outcome. Clearly, t h e r e are n o i d e a l t h e r a p i e s as y e t for osteoporosis, in the c o n t e x t o f either p r e v e n t i o n or t r e a t m e n t . Moreover, m o s t studies have b e e n p e r f o r m e d in s e l e c t e d groups within the community, and the e x t r a p o l a t i o n to w i d e r clinical s u b s e t s is n o t straightforward. F o r example, d a t a are almost n o n e x i s t e n t for men or p r e m e n o p a u s a l women; also it is n o t c l e a r w h e t h e r s t u d i e s on carefully s e l e c t e d healthy elderly w o m e n can be simply e x t r a p o l a t e d to t h e i r less healthy cohorts. Similarly, s t u d i e s using v e r t e b r a l d e f o r m i t i e s as the primary e n d p o i n t may be difficult to r e l a t e to symptomatic v e r t e b r a l fractures, and o t h e r peripheral f r a c t u r e s including proximal f e m u r . Lastly, t h e efficacy of some t r e a t m e n t s must be r e l a t e d to the risk o f f u r t h e r f r a c t u r e s . Thus, many y e a r s o f t h e r a p y may r e l a t e to p r e v e n t i o n o f p e r h a p s only one symptomatic event, with equivocal c o s t - b e n e f i t for the i n d i v i d u a l As n e w t h e r a p i e s are u n d e r development, t h e s e q u e s t i o n s n e e d to be a d d r e s s e d f o r t h e e f f i c a c y of o s t e o p o r o s i s t h e r a p y to be considered.

From the Division of Clinical Pathophys~ology, Cantonal Hospital of the University of Geneva, Geneva, Switzerland. Requests for reprints should be addressed to Professor John A. Eisman, Head, Bone and Mneral Research D~wsion,Garvan Institute of Med,cal Research, St. Vlncent's Hospital, Sydney, NSW 2010, Australia.

to consideration of the efficacy of treatp rior ment of osteoporotic fractures, and indeed the optimum intervention in osteoporosis, it is appropriate to consider the setting in which this question is posed. Should this be from the perspective of a treating doctor, the perspective of a heaith care provider/insurer, or from that of the individual patient? Each of these will have a valid viewpoint. However, the final answer must and should be that from the viewpoint of each individual patient. At times through this article I will return to this viewpoint to underline the difficulty of answering this question in a general but meaningful way. These considerations depend on the answers to a number of key questions: (a) What is the scope of the osteoporosis problem to be considered (i.e., what types of fractures are considered)? (b) How effective is therapy in preventing further osteoporotic fractures (i.e., how many people must be treated for prevention of one further fracture)? (c) Is it possible to target likely high risk/high response groups (e.g., subjects with low bone mass and fractures)? (d) What is a successful outcome (stopping further fractures, increasing bone mass, stopping bone loss)? (e) Do we have any alternatives? Before addressing these questions, it is important to state that the first response to an osteoporotic fracture is treatment of the acute event with adequate acute pain relief, surgical intervention as required, and active rehabilitation to encourage return to the individual's previous level of independence. Some studies suggest that educational programs may limit risk of further disability or even fracture, but such approaches are beyond our consideration due to the dearth of adequately controlled studies to evaluate. If we could step back in time, there can be little doubt that prevention of the first fracture would be optimal. Although requiring more people to be treated, prevention is the optimum from the individual's viewpoint. Yet this may not be the most cost-beneficial approach from the point of view of a healthcare provider nor do preventive approaches typically reach a large proportion of their target population unless aimed shotgun fashion at the entire population, irrespective of risk. Although some studies suggest that advice for particular therapy and compliance may be improved by knowledge of risk, it is clear that for osteoporosis prevention

February 27, 1995 The AmericanJournal of Medtclne Volume98 (suppl 2A)

2A-17S

SYMPOSIUMON OSTEOPOROSIS/ EISMAN there is little evidence of widespread attention at the community level [1]. In other health areas, there have been some successes with public health education with respect to diet and life-style--for example, in relation to diet and cholesterol, smoking, and heart disease--but it remains to be shown conclusively that comparable strategies are appropriate for osteoporosis prevention. However, irrespective of the seductiveness of the preventive approach and irrespective of these limitations, we cannot opt for (primary) prevention when someone presents with a fracture, as they have already had an event. Thus, any advantages of preventive approaches aside, an individual presenting with a fracture requires primarily consideration about interventions designed to reduce risk of subsequent fractures. What is the scope of the osteoporosis problem to be considered? Rationally, all types of osteoporotic fractures must be considered, yet most studies have not taken this point of view. To the contrary, virtually all double-blind, randomized controlled studies have focused on vertebral fracture subjects who have been studied with respect to subsequent vertebral fractures, with and without intervention. Some studies have included so-called peripheral fractures, but these have generally been of small scale with limited statistical power and clinical value. Although studies on hip fracture are underway, outcome data are not yet available. Several large-scale studies are now starting to address various fractures in osteoporosis in community and volunteer populations [2-6]. Osteoporotic fractures are sometimes considered as being synonymous with hip fractures, based on the argument that proximal femoral fractures are responsible for the majority of osteoporotic fracture morbidity, mortality, and costs. However, our recent studies tracking all symptomatic fractures indicate that hip fractures are relatively unimportant until after the age of approximately 80 years in both men and women [2; Randell et al, submitted]. Prior to that age a variety of other fractures are quantitatively important, causing morbidity and generating medical costs. These comments should not be seen to underrate hip fracture costs but more to stress that other fractures are important as well. In this context it is important to note that this wide range of other fractures are largely ignored in current studies of osteoporosis, at least in terms of primary endpoints. An important question is, How does this community-level problem relate to the numerous studies of osteoporosis treatment? This relates in part to the 2A-18S

careful screening and selection of subjects for clinical studies. In fact virtually all high-quality doubleblind randomized controlled studies use a range of exclusion criteria. Some are entirely rational, including ability to give informed consent and likelihood to complete the study on the basis of life expectancy. Others are more difficult, since they exclude a large number of common diseases and commonly used therapies in the community at large. It is not atypical to have to exclude 75-80% of potential study candidates based on such criteria. It is quite clear that such exclusion criteria are used to avoid confounding of the study results; however, some obvious problems with the selection process should be considered. First, the normal population may well have a different overall risk of fracture from that in our atypically healthy study subjects and may respond differently, better or worse, to therapy. An example of this might be the vitamin D and calcium hip fracture prevention study [7]. If that study had carefully excluded any individual with low serum vitamin D levels, there might have not been any benefit--yet that is presumably the precise mechanism of at least a major part of the benefit. Thus, can we legitimately extrapolate the results of selective studies in atypically healthy individuals to the community at large? In the absence of clear evidence to the contrary, we generally, if cautiously, conclude in the positive. Another aspect of the problem is that of extrapolating the studies on women to men with osteoporosis. This is certainly no less uncertain than extrapolating the results in studies in highly selected subjects to the general population. Clearly, if men are considered, then some of the most common interventions either have not been tested at all in men or in some cases are patently inappropriate. Given our lack of information on this aspect, we can only say what works in women may work in men but, until appropriate studies are done, this can only be at best a rather shaky projection. How effective is therapy in preventing further osteoporotic fractures? In order to consider the efficacy of treatment, consideration must be given to the number of subject-years of treatment for every fracture avoided. Only double-blind randomized controlled studies have been considered from this viewpoint and only those with fracture number as their endpoint. For example, a hypothetical treatment may absolutely reliably prevent one fracture per 100 subject-years of therapy. Such a treatment would hardly be likely to be used. Another treatment may be highly variable in its effects but overall may prevent, e.g., 50 fractures per 100 subject-years of treatment. The

February27, 1995 The Arner~canJournal of Medicine Volume98 (suppl 2A)

SYMPOSIUM ON OSTEOPOROSIS/ EISMAN latter, although less predictable, would certainly be worthy of consideration, especially if some way of more reliably predicting the outcome were available (vide infra). Another dimension of this approach relates to considerations of side-effects and/or costs of treatments. Another hypothetical therapy that reliably prevented 100 fractures per 100 subject-years of treatment but caused serious, permanent, and even life-threatening side-effects in even a small proportion of people would be unacceptable. Although monetary costs are not within the ambit of this article, human costs of discomfort or unacceptable routes of administration must be carefully considered.

Is it possible to target likely high risk/high response groups? This question could be paraphrased as, Can low bone mineral density (BMD) plus fractures identify a higher risk group where treatment could be expected to produce more effective outcomes ? The answer is yes and no. There is growing evidence that individuals with ->1 osteoporotic fractures may be at increased risk of further fractures, independently of their underlying BMD. This has been reported for peripheral and vertebral fractures with respect to further vertebral and nonvertebral fractures [8-10]. However, the nature of most treatment studies is such that it is not possible to answer whether someone at increased risk would have greater benefit. At one level it is clear that if a community were treated with some very safe easy treatment, if such were to exist, someone with a low risk is not likely to benefit even if the community as a whole did do so. However, we cannot say whether that benefit is greater in someone from a group at increased risk, even if it is true for the group as a whole. Let us consider treatment A, which can halve fracture risk in a low risk group (say 6 per 100 patient-years) and treatment B, which can reduce fracture risk by 30% but only in a high-risk group (say 40 per 100 patientyears). Considering 1,000 people-years in each group, treatment A in the low-risk group would reduce the fracture number from 60 to 30 per 1,000 patient-years. Treatment B in the high risk group would reduce the number of fractures from 400 to 300 per 1,000 patient-years. Thus, treatment B saves 100 fracture events versus the 30 fracture events of treatment A. Which is better? For the individual with a remaining life expectancy of 25 years, the saving in risk is 0.75 in one group and 2.5 in the other. The choice would seem obvious; however, none of these "fractures" may be symptomatic to any e x t e n t - - t h u s a difference of 1-2 possibly

asymptomatic vertebral deformities detectable over their remaining life expectancy. It is important to note that these figures were crudely drawn from two etidronate studies [11-13] and a subsequent epidemiologica] perspective [14], and they are based on radiologically defined vertebral deformities. Although these radiological changes are an invaluable guide to treatment effect, they are a surrogate for symptomatic fractures, and from other studies, such radiological changes, unless quite marked, are a poor indicator of symptomatic fractures [15-17]. Thus, these treatments with quite different effects, even in a high-risk individual, may prevent a small and perhaps, even to the individual, inconsequential number of symptomatic fractures. Due to the above difficulties, some analyses have concentrated on hip fractures--in which case the symptomatic nature, need for hospitalization, and subsequent morbidity, mortality, and costs are fairly clearly defined--even though they account for