Article Text

Cohort study
People with Alzheimer's disease are at increased risk of hip fracture and of mortality after hip fracture
  1. Kilian Rapp
  1. Department of Clinical Gerontology, Stuttgart and Institute of Epidemiology, Ulm University, Stuttgart, Germany
  1. Correspondence to Kilian Rapp
    Department of Clinical Gerontology, Robert-Bosch-Hospital, Auerbachstr 110, 70376 Stuttgart, Germany; kilian.rapp{at}

Statistics from

Commentary on: OpenUrlAbstract/FREE Full Text.

Implications for practice and research

  • People with Alzheimer's disease (AD) are a high-risk population for hip fractures.

  • Mortality after a hip fracture is particularly high in such people. Therefore, prevention and management of potential causes like delirium and an early rehabilitation should be priorities in AD patients after a hip fracture.


AD as well as hip fractures increase exponentially with age. Both diseases are frequent and have a considerable effect on the person's function. Previous studies have demonstrated an association between AD and hip fractures. The objective of the study by Baker and colleagues was to evaluate whether AD is an independent risk factor for hip fractures. Furthermore, the study compared mortality after hip fracture in people with and without AD.


People with AD were identified by (1) diagnosis ‘AD’ or (2) AD-typical medications from an electronic database. The data were derived from general practitioner practices in the UK and were collected during routine medical care. To analyse the influence of AD on hip fracture incidence, each person with AD was matched by sex and age to a person without AD. A Cox regression analysis was applied and adjusted for potential confounders like comorbidities or medications (prospective analysis).


The risk of a hip fracture was more than three times higher in people with AD compared with those without. After hip fracture, people with AD also had higher mortality.


Falls and fall-related fractures in older people are usually the consequence of a combination of risk factors like gait and balance disorders, urinary incontinence, visual disorders, postural hypotension, depression, psychotropic drugs or cognitive impairment.1 2 Many of these risk factors can be symptoms of AD. Thus, it is expected that AD is a strong predictor for hip fractures. It is, therefore, not completely clear what Baker and colleagues meant when they aimed to evaluate AD as an ‘independent’ risk factor for hip fractures. They try to demonstrate that there is a causal relation between AD and hip fractures beyond the pathways of the well-known risk factors and speculate, for example, that people with AD may fall in a (different) way that increases the risk of a fracture.

In an observational study, the methodological way to analyse an independent effect of a factor on an outcome is to control for potential confounders. The authors adjusted for comorbidities, smoking status and many drug groups like benzodiazepines, antidepressants or antipsychotics. However, they did not adjust for physical function or other typical risk factors for falls. It is not wrong not to adjust for these variables as they may act as so-called ‘intermediate factors’ between AD and hip fractures. However, in this case, it is impossible to differentiate between an effect caused through known risk factors for falls and an additional AD-specific hypothetical effect. To summarise, the study shows impressively that AD is a strong predictor for hip fractures. Most of the effect is probably caused or mediated through well-known risk factors.

The comparison of mortalities after hip fracture in people with and without AD is the more interesting part of the study. The authors found an increased relative mortality risk of 1.5 in people with AD. The survival curves show that (1) the excess mortality was limited to the first half year after the hip fracture and (2) the relative mortality risk during this time was clearly higher than 1.5. Therefore, the relative mortality risk should have been analysed and stratified by different time periods (eg, 0–6 moths; >6 months). Unfortunately, the potential reasons for the observed excess mortality in people with AD were not discussed by the authors. One reason for the excess mortality could be the higher comorbidity in people with AD. From clinical experience, AD-specific problems may have also contributed to the observed excess mortality: after surgery, people with AD have an extremely high risk of delirium that is associated with an increased mortality3; only few AD patients get an inpatient rehabilitation, and therefore, such patients have an increased risk of immobilisation that is associated with complications like infections or pulmonary embolism.

Potential implications of these results could be (1) better prevention and management of delirium in AD patients with a hip fracture, such as rooming in of relatives, use of eyeglasses or hearing aids, non-pharmacologic approaches for agitation and so on3; (2) commencement of rehabilitation immediately after surgery and (3) continued rehabilitation in a familiar environment since an unknown environment is usually problematic for people with dementia.

Results from fall-prevention studies in cognitively impaired people are inconsistent. Some have found a beneficial effect,4 5 whereas others have not.6 7 The clinical experience is that fall prevention strategies like strength and balance training are feasible in ambulatory people with dementia and that they can be a way to increase quality of life.

View Abstract


  • Competing interests None.

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.