The study was performed at a Danish university hospital (Herlev Hospital) from January 2004 to December 2005. The Local Ethics Committee accepted the study protocol (KA 03097, November 2003), and the study was performed in agreement with the Helsinki declaration. In 2005, the ED at Herlev Hospital had a local catchment area of approximately 200000 citizens and treated 100 surgical or medical patients daily. On average, 60% of treated patients were 50+ years old, and one third of contacts involved fractures, sprains, bruises, lacerations, and so forth.
Patients were eligible for inclusion in the study if they were aged 50–80 years and if they had attended the ED due to trauma sustained during a low-energy fall. A fall was defined as a sudden, unintentional change in position causing an individual to land at a lower level, on an object or the ground, rather than as a consequence of sudden onset of paralysis, epileptic seizure, or overwhelming external force [20
]. We considered a fall as low energy if the maximum displacement was to ground level from a standing position; thus, we excluded falls downstairs, falls from heights, and so forth. Patients were excluded from the study if they had acquired a bone fracture during the actual fall, if they became admitted to hospital after the trauma, or if they already received medical treatment for osteoporosis. Only community-dwelling patients able to communicate in Danish or English were included. Patients with dementia or with diseases making them unable to manage transportation to the hospital were excluded. We aimed to include a predefined number of 200 patients. The inclusion of patients was made on randomly selected days equally distributed over a year in order to facilitate sufficient time and capacity for scanning and to avoid seasonal changes of BMD.
ED files from randomly selected dates were used for prospective inclusion of patients. All patients were contacted by phone by one of the study investigators within 1–3 months after the visit to the ED if they had been treated for lacerations, sprains, or other trauma secondary to a fall. The phone interview was performed using a semistructured interview technique. The patients were questioned in order to secure that they fulfilled the inclusion criteria and that no exclusion criteria were present. Eligible patients received a letter with written information about the study. All participants gave written and verbal informed consent before inclusion.
All included patients filled in a questionnaire concerning body height reduction since their mid-twenties, lifestyle (exercising habits, dietary habits, and vitamin intake), smoking habits, medication use (patient self-report), alcohol abuse (more than 14 drinks a week for women and more than 21 drinks a week for men (one drink is approximately 12 grams of alcohol)) [21
], predisposition to osteoporosis, history of bone fractures within the previous 20 years (proximal humerus, distal antebrachium, femoral neck, thoracic and lumbar spine, or pubic bone), vision, chronic disease, mobility, and employment status. Women additionally answered questions concerning menopausal status and previous childbirths. One of two of the study investigators (BG or BZ) went over the questionnaire with all participants in order to clarify uncertainties and potential misunderstandings.
All patients had their body weight and body height measured by a trained nurse prior to densitometry. DXA scans (Lunar Prodigy, GE Medical Systems, Madison, WI, USA) were performed on both femoral necks and the lumbar spine (L2–L4) in all included patients. The data presented in the following paper are based on computerized numerical BMD values of the femoral neck and the lumbar spine for each patient. A specialist in densitometry evaluated all scans. The T- and Z-scores were defined according to the World Health Organization's recommendations [7
]. A T-score below −2.5 SD was considered diagnostic of osteoporosis, and a T-score below −1 SD but higher than or equal to −2.5 SD was diagnostic of osteopenia.
Patients referred from general practice represent by far the largest group among those who are DXA scanned at our department, and the frequency of osteoporosis in this group is thus a suitable reference point. Therefore, we selected 201 subjects routinely referred to the department of clinical physiology, Herlev Hospital from local general practices on the suspicion of osteoporosis. The patients were selected at random in a retrospective manner after the recruitment of fallers was completed. The patients were selected by a study staff member with no knowledge of the purpose of the study. Inclusion criteria were referral during the time period January 2004 till December 2005, age between 50 and 80 years, patients able to self-transport to hospital, and no previous DXA scan. Selection was made as a simple random sample so that all patients in the time period had equal probability of selection. Referral diagnoses, age, height, weight, and T-scores were recorded. As patients routinely referred from general practice do not answer questionnaires or report previous fractures or falls, these data were not available.