Clinical trials of treatment agents impose strict and often necessary inclusion and exclusion criteria, while patients presenting to physicians for treatment frequently exhibit complicating features that would have excluded them from entry into study. To quantify the degree of discordance between ordinary patients and study subjects, a retrospective chart review was carried out of all new patients with osteoporosis seen in an academic medical center within a consecutive 40-month period, meeting clinical treatment criteria. Each patient chart was reviewed for the inclusion and exclusion criteria of four large, multicenter study protocols. There were 120 consecutive female patients seeking health care, with bone density T-scores below -2.0 and/or with one or more low-trauma fractures. The four trials would have accepted 4, 5, 25 and 8 of our 120 patients. The trial with the most liberal inclusion criteria would have taken only 21% of the total. Principal reasons for ineligibility were comorbidity, prior treatment with bone-active agents, and current therapy with glucocorticoids, anticoagulants and anticonvulsants. Some of these exclusions inevitably reflect the patient mix of a referral center; nevertheless, comorbidity and its therapy are common in the age range in which osteoporosis is prevalent and would, therefore, be expected to be present in patients in general medical practice as well. Thus a large fraction, perhaps the majority, of patients with diagnoses of osteoporosis who are candidates for treatment by their physicians, are not eligible for entry into typical treatment trials. The results of such trials may, therefore, have uncertain applicability to types of patients excluded, both for safety and for efficacy.