First published August 2006, updated July 2021

Key points:

  • With falls in the elderly being an important public health concern, it is important to be aware of the modifiable risk factors.
  • Though evidence from published reports may be conflicting, it is clear that multiple medication load is a predisposing factor for falls.
  • The evidence to implicate benzodiazepines as a risk factor is overwhelming, especially in the initial stages of starting the medication.
  • There is a small positive association between cardioactive medications and falls.

Injuries due to falls represent an important public health concern in terms of morbidity, mortality and cost to health and social services1. About 30 per cent of people over 65 living in the community have a fall at least once each year and this figure is higher in institutional care. One-fifth of these incidents require medical attention. This, in turn, can result in significant morbidity and healthcare costs, restrictions on the patient’s activity level, a fear of subsequent falls and – in the worst case scenario – it can lead to institutionalisation.

It is estimated that about 40–60 per cent of falls lead to injuries, with 30–50 per cent being minor and five per cent being fractures2,3. Up to one per cent of falls in older people are said to result in a hip fracture2,4. In one study of 204 community dwelling people who had experienced more than one fall in the previous year with consequences, the results were5:

  • physical injury (68.1 per cent);
  • health service use (23.5 per cent);
  • treatment (17.2 per cent);
  • decline in functional status (35.3 per cent); > decline in social activity (16.7 per cent); and
  • decline in physical activity (15.2 per cent).

The main contributory factors associated with functional decline were being female, increased medication use and depressive symptoms.

There are a number of risk factors that have been identified. The Effective Health Care Bulletin placed the causes into five major categories:

  • environmental (eg, loose carpets, poor lighting, unsafe stairs);
  • medical conditions and changes associated with ageing (poor vision, cognitive impairment);
  • nutritional (eg, calcium and vitamin D deficiency);
  • lack of exercise; and
  • medication related.

The majority of non-syncopal falls do not have well-defined aetiologies and are often felt to be the result of multiple interacting factors. The contributing factors have also been described as ‘intrinsic’ (i.e., age-related changes and health status-related factors) and extrinsic (eg, the physical environment surrounding the patient).

The use of drugs in this population is often unavoidable and may be considered as either an intrinsic or an extrinsic factor. The role of medications in contributing to falls has received much attention over the past two decades, as medication exposure is potentially modifiable as a strategy to their prevention. Two landmark case control studies by Ray et al6,7 suggested that use of long-elimination half-life hypnotic anxiolytic agents in general6 – and use of long-elimination half-life benzodiazepines7 in particular – increase the risk of hip fractures in elderly patients.

Current National Institute for Clinical Health and Excellence (NICE) guidelines8 recommend a medication review be an integral part of the initial multifactorial assessment in those who fall. Indeed modifications to be considered include the following:

  • primary care physicians should review the need for medications on a regular basis;
  • rescheduling the taking of medications to minimise any cumulative effects;
  • close monitoring of a patient when new medications are commenced.

The management of medicines is a fundamental component of each of the National Services Framework (NSF) standards9. The NSF standard on falls points out that polypharmacy is a risk factor. In patients taking medications known to contribute to falls, medication review can play an important part in falls prevention. The majority of older people take medicines, making this a significant issue for the NSF.

Evidence implicating medications as a risk factor

The evidence for drug-related falls comes from numerous small-scale studies, meta-analyses, systematic reviews and community-based case control/cohort studies. There are no randomised control trials to produce evidence, as this would be unethical to carry out.

A meta-analysis by Leipzig et al10 pooling the results of 40 studies carried out between 1966–1996 indexed under ‘accidents/accidental falls’ and ‘aged/age factors’ revealed a small – but consistent – association between the most classes of psychotropic drugs and falls.

The pooled odds ratio for one or more falls was odds ratio (OR)1.73 for any psychotropic use, OR1.50 for neuroleptic use, OR1.54 for sedative/ hypnotic use, OR1.66 for any antidepressant use, OR1.51 for only tricyclic anti-depressant (TCA) use and OR1.48 for benzodiazepine use, with no difference between short and long-acting formulas. The evidence used in this meta-analysis was based solely on observational data and most of the studies evaluated did not take into account confounding factors at the time of fall. Subsequent studies have found confounders to provide a significant contribution to falls, thus this factor is a significant limitation in Leipzig’s meta-analysis.

Neutel et al11 produced figures in keeping with those by Leipzig while examining drug use and falls in an institutionalised population. The relationship between benzodiazepines and analgesics were statistically significant, but only when unadjusted for confounders. They found that residents starting a new (defined as a new prescription of the medication, or an increase in dose) benzodiazepine or antipsychotic were greatly at risk of having a fall and patients starting a ‘new’ drug of any class were at nearly twice as much at risk of having a fall within two days. Residents taking multiple drugs were also at particular risk of falling when compared with those taking less than five medications.

Narcotics, anticonvulsants and antidepressants were associated with a significant increased risk of an injurious fall, even after accounting for underlying medical conditions and confounders in one large community-based case-control study12. This used 2,278 individuals presenting to any of six regional emergency departments in Alberta, Canada, classified by ICD-9 criteria as having an ‘injurious fall’. The strengths of the study included the selection bias being minimal, as all cases from the database were included and all individuals over 65 years of age residing in that area have premium-free health insurance. The limitations were that only falls related to emergency and hospital treatment were included and that non-prescription medications, alcohol or illicit drug use could not be accounted for. However, the statistically significant, positive association with the aforementioned medication classes adds strongly to the existing evidence base.

Other studies have examined associations between certain medications and hip fractures. A French study13 failed to show an association between benzodiazepine use and hip fracture when reviewing 245 institutionalised and community-dwelling persons admitted to hospital with hip fracture resulting from a fall. Use of the medication was assessed by a patient questionnaire, medical records and plasma assay. Hip fracture was associated with the use of two or more benzodiazepines as determined by the questionnaire and medical records, but not from plasma samples, therefore showing long-life benzodiazepines or increased doses are not risk factors. They did, however, confirm an increased risk from TCAs but as this was not the main objective of their study, this relationship may be tenuous.

Wagner14 found an increased incidence of hip fracture in people over 65 years of age in 24 per cent of those exposed to benzodiazepines compared with those without exposure. This was even after adjustment for confounding factors. They also found the incidence rate to be significantly higher within the first two-week period of starting a benzodiazepine, with the risk declining as time went on. This data adds strength to studies previously reported15-17. Another finding in this large study controlling for multiple confounders was that an increased risk of hip fracture was associated with the use of short half-life high potency benzodiazepines but not low potency long half-life or mixed benzodiazepine exposure.

This is important as short half-life preparations are often considered to be the best choice of drug in this class for elderly patients. Evidence for this from previous studies is inconsistent18. The publication by Wagner should be interpreted with the following potential limitations in mind:

  • since exposure to benzodiazepines was inferred from pharmacy dispensing claims, it is therefore impossible to tell whether increased risk of fracture in the first two weeks was due to an increased compliance in this time period; and
  • whether patients were actually taking their medications when the hip fracture occurred.

However, other studies were unable to confirm an association between benzodiazepines or anticonvulsants and falls/hip fracture18,19.

Other classes of medications have also been investigated. Leipzig et al, published a second meta-analysis reviewing the role of cardiac and analgesic medications as an aetiological factor20. Methodology and data sources are the same as for their previously mentioned study. Very few of the sources had adjusted for confounders, which are a possible limitation of their findings. However, out of 14 medication groups analysed, only the use of diuretics, digoxin and type IA antiarrhythmics were associated with one or more falls. The pooled odds ratios for these classes were: diuretics OR1.08; digoxin OR1.22; type 1A antiarrhythmics OR1.59. They also found the use of more than three medications associated with an increased risk of recurrent falls. It is interesting to note that Pierfitte et al13 confirmed a protective effect from diuretics – as did Jones et al21 – although that was not the main objective of their study.

Drug-related orthostatic hypotension

Diuretics can induce falls by causing hypovolemia, dehydration and subsequent orthostatic hypotension. TCAs, by virtue of their anticholinergic and sedative properties, can potentially precipitate falls. The falls may be related to the underlying medical condition aggravated by these effects.

The role of orthostatic hypotension in this situation is debatable. Orthostatic hypotension does not appear to predict falls and this is due to a number of reasons. First, there is a large day-today variability in postural blood pressure readings22 and a frequent lack of associated symptoms23. It is also difficult to ascertain whether patients were actually taking the medication thought to be responsible for falls secondary to postural hypotension at or around the time of postural hypotension.

Methodological issues

Having discussed the various studies, let us concentrate on some of the methodological issues in many of these studies. There appears to be lack of clarity in the way all the co-morbid conditions were recorded. The use of non-prescription drugs is difficult to assess or ascertain. Many studies were small scale and had a small sample size. There is also less clarity whether the medications were actually taken around the time of falls. Some were community-based studies, some were institution-based cohort case-controlled studies but there were no randomised controlled trials. There were innumerable confounding factors to adjust as well as a lack of an exact unit of reference in many studies (i.e., did they take total falls or fallers as the outcome of interest). It is only fair to say the recall for falls in the older adult can be very misleading and a gross underestimate


  1. Rizzo JA, Friedkin R, Williams CS et al. Health care utilisation and costs in a Medicare population by fall status. Med Care 1998; 36: 1174-88
  2. Tinneti ME, Speechley M, Ginter SF. Risk factors for falls among elderly persons living in the community. N Eng J Med 1988;319: 1701-07
  3. Campell AJ, Reinken J, Allan BC, Martinez GS. Falls in old age: a study of frequency and related clinical factors. Age ageing 1981; 10: 264-70
  4. Nevitt MC, Cummings SR. The study of Osteoporosis Fractures Research Group. Type of hip and wrist fractures: the study of osteoporotic fractures. J Am Geriatr Soc 1993; 41: 1226- 34
  5. Stel VS, Smit JH, Pluijm SM, et al. Consequences of falling in older men and women and risk factors for health service use and functional decline. Age and Ageing. 2004 Jan; 33(1): 58-65
  6. Ray WA, et al. Psychotropic drug use and the risk of hip fracture. N Eng J Medicine. 1987; 316: 363-369
  7. Ray WA, Griffi n MR, Schaffner W, et al. Benzodiazepines of long and short half life and risk of hip fracture. JAMA. 1989; 262: 3303-3307
  8. The assessment and prevention of falls in older people. NICE Nov 2004; CG021
  9. National Service Framework for Older People. Department of Health, March 2001
  10. Leipzig RM, et al. Drugs and Falls in Older People: A Systematic Review and Metaanalysis: I. Psychotropic drugs. JAGS. 1999; 47:30-39
  11. Neutel CI, Perry S, Maxwell, CJ. Medication use and risk of falls. Pharmacoepidemiology and Drug Safety. 2002; 11:97-104
  12. Kelly K, Picket W, Yiannakoulias N, et al. Medication use and falls in community-dwelling older persons. Age and Ageing. 2003; 32:503-509
  13. Pierfitte C, Macouillard G, Thicoïpe M, et al. Benzodiazepines and hip fractures in elderly people: A case-control study. BMJ. 2001; 322:704-708
  14. Wagner AK, Zhang F, Sourmerai SB, et al Benzodiazepine Use and Hip Fractures in the Elderly; Who is at Greatest Risk? Arch Int Med. 2004; 164:1567-1572
  15. Neutel CI, Hirdes JP, Maxwell CJ, et al. New evidence on benzodiazepine use and falls: the time factor. Age and Ageing. 1996; 25:273-278.
  16. Ray WA, Thapa PB, Gideon P, et al. Benzodiazepines and the risk of falling in Nursing Home Residents. J Am Geriatr Soc. 2000; 48:682-685
  17. Wang PS, Bohn R, Glynn R, et al. Hazardous benzodiazepine regimens in the elderly: effects of half-life, dosage, and duration of hip fracture. Am J of Psychiatry. 2001; 158: 892-898
  18. Ensrud KE, Blackwell T, Mangione CM, et al. Central nervous system active medications and risk for fractures in older women. Arch Intern Med. 2003; 163(8): 949-57
  19. Gales BJ, Mernard SM. Relationship between the administration of selected medications and falls in hospitalised elderly patients. Annals of Pharmacotherapy. 1995; 29(4): 354-358
  20. Leipzig RM, Cumming RG, Tinnetti, et al. Drugs and Falls in Older People: A Systematic Review and Meta-analysis: II. Cardiac and Analgesic Drugs. JAGS. 1999; 47:40-50
  21. Jones G, Nguyen T, Sambrook PN, et al. Thiazide diuretics and fractures: can metaanalysis help? J Bone Miner Res 1995; 10:106-111
  22. Lipsitz LA, Storch HA, Minaker, et al. Intraindividual variability in postural blood pressure in the elderly. Cli Sci 1985; 69:337-341
  23. Tinnetti ME, Williams TF, Mayewski R. Fall risk index for elderly patients based on number of disabilities. Am J Med 1986;80: 429-434