• polypharmacy;
  • falls;
  • elderly;
  • pharmaco-epidemiology;
  • risk drugs;
  • geriatric medicine


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References


Falls in the elderly are common and often serious. We studied the association between multiple drug use (polypharmacy) and falls in the elderly.


This was a population-based cross-sectional study, part of the Rotterdam Study. The participants were 6928 individuals aged ≥ 55 years. The prevalence of falls in the previous year was assessed. Medication use was determined with an interviewer-administered questionnaire with verification of use. Polypharmacy was defined as the use of four or more drugs per day.


The prevalence of falls strongly increased with age. Falls were more common in women than in men. Fall risk increased with increasing disability, presence of joint complaints, use of a walking aid and fracture history. The risk of falling increased significantly with the number of drugs used per day (P for trend < 0.0001). After adjustment for a large number of comorbid conditions and disability, polypharmacy remained a significant risk factor for falling. Stratification for polypharmacy with or without at least one drug which is known to increase fall risk (notably CNS drugs and diuretics) disclosed that only polypharmacy with at least one risk drug was associated with an increased risk of falling.


Fall risk is associated with the use of polypharmacy, but only when at least one established fall risk-increasing drug was part of the daily regimen.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References

Falls are a common phenomenon in the elderly and are associated with considerable morbidity and mortality [1]. They often lead to reduced functioning and to nursing home admissions [2]. The risk of falling increases dramatically with the number of risk factors, such as musculoskeletal problems, neurological diseases, psychosocial characteristics, functional dependency and drug use.

Polypharmacy, usually defined as the use of more than three or four medications, is regarded as an important risk factor for falling in the elderly [3–7]. A meta-analysis [6, 7] showed an increased fall risk in users of diuretics, antiarrhythmics and psychotropics. However, in a large population-based study it was concluded that comorbidity, being a relevant recognized risk factor for falling in the elderly, fully explains the increased risk associated with drug use [8].

Our hypothesis was that drugs can be an independent risk factor for falling but that polypharmacy itself is not a risk factor. In our hypothesis the association between polypharmacy and falling is explained by a higher probability of receiving a risk-increasing drug with the number of drugs taken. To investigate this issue, we assessed the association between polypharmacy and falling.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References

This cross-sectional analysis was part of the Rotterdam Study, a population-based prospective cohort study of 7983 people aged 55 years (mean age 70.6, range 55–106.2) [9]. Baseline examination was performed between 1990 and 1993.

We excluded people with dementia (n = 482) [10] or unknown mental state (n = 455) and those who could not give an adequate fall history at baseline (n = 118). During baseline interviews and subsequent physical and laboratory examinations, information was gathered on several relevant parameters such as age, gender, functional performance [11, 12] and blood pressure. A full assessment of medical and psychiatric comorbidity was also performed. Systolic and diastolic blood pressures were measured in a recumbent position, followed by subsequent measurements in an upright position after 1–5 min of standing. Orthostatic hypotension was defined as a systolic drop of ≥ 20 mmHg and a diastolic drop of ≥ 10 mmHg [13]. As exposure of interest we examined the use of drugs. Medication use was determined at baseline by interview and verified by a physician. Drugs were coded according to the Anatomical Therapeutic Chemical classification (ATC) system [14]. Although there is no uniform definition of polypharmacy, we defined it, in accordance with the literature, as the use of four or more medications [3, 5–7]. Drugs associated with falling in the fully adjusted model were classified as risk drugs. As the primary outcome we studied falling. A faller was defined as an individual with a history of one or more falls, without precipitating trauma (e.g. car accident or sports injury), in the 12 months preceding the baseline interview. Falling was assessed by structured personal interviews by trained research nurses.

The Medical Ethics Committee of the Erasmus MC, Rotterdam, the Netherlands, approved the study.


We analysed the association between risk factors and falling by means of multivariate logistic regression analysis. We performed an adjusted multivariate analysis adding all known risk factors for falling: age, gender, alcohol use, history of diabetes mellitus, myocardial infarction, hypertension, Parkinson's disease, stroke, thyroid diseases, depressive episodes, functional performance (described as disability index), dizziness, gait disturbance, home-bound life style, joint complaints, memory complaints, orthostatic hypotension, systolic and diastolic blood pressure after 5 min, postural disturbance and visual acuity. All analyses were performed using SPSS version 11.0.1 (SPSS Inc., Chicago, IL, USA; 2001).


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References

A total of 6928 subjects (87%) were eligible for our study, of whom 1144 (16.5%) experienced one or more falls in the previous year. The prevalence of falls strongly increased with age. Falls were more common in women than in men. In addition, fall risk increased with increasing disability, staying indoors because of health, joint complaints, dizziness, gait or postural disturbance, orthostatic hypotension, history of diabetes mellitus, hypertension, Parkinson's disease, stroke, depressive episodes and presence of memory complaints (Table 1). Almost 72% (n = 4983) of the participants were taking at least one drug, and 20.3% (n = 1407) were taking four or more drugs. The risk of falling increased significantly with the number of drugs used per day (P for trend < 0.001) (Figure 1). In the univariate analysis, 28 drugs were associated with falling and were therefore considered as potential risk drugs (Table 2). After adjustment for age, gender, comorbid conditions and disability, falling remained associated with the use of central acting antiobesity products, calcium preparations, potassium sparing diuretics, oxicams, quinine and derivatives, anilides, anxiolytics-benzodiazepine derivatives, hypnotics-benzodiazepine derivatives (Table 2). These drugs were considered as risk drugs.

Table 1.  Patient characteristics and risk of falling (n = 6928)
Characteristic(n = 5784)Without fall(n = 1144)With fallOdds ratio*(95% CI)
n% (SD)n% (SD)
  • *

    Adjusted for age and gender.

  • Orthostatic hypotension was defined as a systolic drop of ≥20 mmHg, and a diastolic drop of ≥ 10 mmHg.

Age category (years)
 75–84108018.7%34330.0%2.52(2.11, 3.00)
 > 85261 4.5%15613.6%4.31(3.40, 5.46)
Mean age (SD)68.6(8.6) 73.2(9.8)  
Female gender326256.4%86875.9%2.43(2.10, 2.81)
Staying indoors242 4.2%16914.8%2.19(1.74, 2.76)
Disability index
 Not disabled457479.1%59552.0%1.00Ref
 Mildly disabled81514.1%27223.8%2.02(1.70, 2.40)
 Moderately disabled254 4.4%14512.7%2.93(2.29, 3.74)
 Severe disabled141 2.4%13211.5%4.53(3.38, 6.07)
Alcohol use234640.6%34430.1%0.98(0.84, 1.16)
Joint complaints278948.2%70161.3%1.51(1.32, 1.73)
Visual acuity
 Both eyes intact386066.7%60853.1%1.00Ref
 One eye impaired88415.3%20317.7%1.11(0.92, 1.34)
 Both eyes impaired63310.9%23620.6%1.23(0.99, 1.52)
Dizziness165728.6%55748.7%1.98(1.74, 2.27)
Gait disturbance318 5.5%18115.8%2.47(1.99, 3.07)
Postural disturbance149 2.6% 87 7.6%2.17(1.62, 2.91)
Orthostatic hypotension44 0.8% 22 1.9%2.10(1.23, 3.61)
History of diabetes mellitus328 5.7% 96 8.4%1.29(1.01, 1.65)
History of heart attack523 9.0%101 8.8%1.01(0.80, 1.28)
History of hypertension80413.9%21418.7%1.25(1.05, 1.50)
History of Parkinson's disease28 0.5% 24 2.1%3.27(1.84, 5.82)
History of stroke192 3.3% 83 7.3%1.89(1.43, 2.51)
History of thyroid diseases477 8.2%13712.0%1.17(0.95, 1.45)
History of depressive episodes176930.6%43538.0%1.30(1.13, 1.50)
Memory complaints100717.4%29926.1%1.49(1.28, 1.74)

Figure 1. Influence of the number of medications on falling adjusted for age & gender (p for trend <0.001)

Download figure to PowerPoint

Table 2.  Drugs associated with falling
ATC codeDescriptionCasesPercent caseswithin usersOR*(95% CI)OR(adj.)(95% CI)
  • *

    Corrected for age, gender.

  • Corrected for age, gender, alcohol use, history of diabetes, history of heart attack, history of hypertension, history of Parkinson's disease, history of stroke, history of thyroid diseases, history of depressive episodes, disability, dizziness, gait disturbance, staying indoors because of poor health, joint complaints, memory complaints, orthostatic hypotension systolic and diastolic after 5 min, postural disturbance and visual acuity.

AAlimentary tract and metabolism      
A06ABContact laxatives2637.1% 1.8(1.1, 3.0)1.3(0.6, 2.9)
A06AGEnemas 583.3%23.4(2.6, 207.8)0,0(0.0, ∞)
A08AACentral acting antiobesity products 758.3% 7.7(2.4, 24.8)4.9(1.0, 24.7)
A10BBSulphonamides urea derivatives6026.2% 1.5(1.1, 2.0)1.4(0.7, 3.1)
A11BAMultivitamins1930.2% 2(1.1, 3.5)2.0(0.9, 4.2)
A11EAVitamin b complex8226.8% 1.5(1.1, 2.0)1.2(0.8, 1.7)
A12AACalcium preparations3633.6% 1.9(1.3, 2.9)1.9(1.0, 3.3)
BBlood and blood-forming organs      
B03AAOral ferro preparations2446.2% 2.8(1.6, 5.0)2.3(0.8, 6.7)
B04ADBile acid sequestrants 642.9% 3.8(1.2, 11.8)2.3(0.4, 13.5)
CCardiovascular system      
C03BADiuretics—sulphonamides2631.0% 2.1(1.3, 3.4)1.4(0.7, 2.9)
C03DBDiuretics—potassium-sparing agents1836.7% 1.9(1.0, 3.5)3.6(1.1, 11.8)
C04AEErgot alkaloids 750.0% 3.3(1.1, 9.7)2.0(0.4, 9.9)
C05CABioflavonoids 550.0% 5(1.4, 17.9)3.3(0.8, 14.6)
GGenitourinary system and sex hormones      
G02CBProlactine inhibitors 770.0%10.5(2.6, 43.4)NA(0.0, 0.0)
MMusculoskeletal system      
M01ACOxicams1840.9% 3.1(1.6, 5.8)3.2(1.3, 7.9)
M09AAQuinine and derivatives3535.0% 1.8(1.2, 2.8)2.2(1.2, 4.2)
NNervous sytem      
N02BEAnilides21420.7% 1.4(1.1, 1.6)1.3(1.0, 1.6)
N03ABHydantoin derivatives933.3% 3(1.3, 7.0)1.1(0.2, 5.8)
N04AAAnticholinergic agents—tertiary amines360.0%12(2.0, 73.6)4.2(0.2, 80.2)
N04BADopa and dopa derivatives1750.0% 3.5(1.7, 7.1)0.8(0.1, 6.7)
N04BBAdamantane derivatives743.8% 3.2(1.1, 9.2)0.3(0.0, 4.0)
N04BDMao-inhibitors type b850.0% 4.6(1.7, 12.9)1.9(0.2, 20.0)
N05BAAnxiolytics, benzodiazepine-derivatives13226.1% 1.5(1.2, 1.9)1.3(1.0, 1.9)
N05CDHypnotics benzodiazepine derivatives14727.1% 1.3(1.0, 1.6)1.6(1.1, 2.1)
N07CAAntivertiginous drugs5933.0% 1.7(1.2, 2.4)1.0(0.6, 1.7)
RRespiratory system      
R03BBParasympathicolytics2129.2% 2.6(1.5, 4.5)1.1(0.4, 2.7)
R05CBMucolytics2526.9% 1.8(1.1, 2.9)1.0(0.4, 2.3)
SSensory organs      
S01AAOcular antibiotics444.4% 4.7(1.2, 18.1)5.9(0.9, 37.1)

The probability of using a risk drug increased proportionally with the total number of medications taken, from 25% with the use of only one prescription daily to more than 60% when six or more drugs were prescribed (Figure 2). Women were using significantly more risk drugs than men [odds ratio (OR) 2.2, 95% confidence interval (CI) 1.9, 2.4]. After adjustment for comorbid conditions and disability, polypharmacy (i.e. the number of drugs) remained a significant risk factor for falling. The ORs increased from 1.4 (95% CI 1.0, 2.0) using three medications to 1.6 (95% CI 1.1, 2.1), using four or more medications (P for trend = 0.008). Considering the influence of the number of risk drugs, the ORs increased by 42% per risk drug (P for trend < 0.001), from 1.3 (95% CI 1.0, 1.6) using one risk drug to 2.5 (95% CI 1.7, 3.6) using two risk drugs. The analysis was also done separately for persons under and over age 75 years, and by introducing an interaction-term in the nonstratified statistical analysis. We found no statistically significant interaction (P = 0.698). In the age group > 75 years old there remained an association of using more than four drugs and falling, although this was no longer significant because of lack of power. Respective ORs were 1.58 (95% CI 1.08, 2.29) for persons 55–74 years old and 1.42 (95% CI 76, 2.67) for persons above 75 years old.


Figure 2. Polypharmacy and use of drugs associated with falling. Users of medications ( bsl00001 ), using one or more drugs associated with falling ( bsl00077 ), proportion using one or more drugs associated with falling ( • )

Download figure to PowerPoint

Stratification for presence or absence of at least one risk drug disclosed that polypharmacy is a risk factor for falling only if it includes a risk drug (P for trend = 0.004; Figure 3). In other words, polypharmacy itself is not a risk factor for falling unless a risk drug is part of the drug regimen.


Figure 3. Influence of polypharmacy on falling, stratified on use of risk drugs

Download figure to PowerPoint


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References

In this population-based study, fall risk was associated with the use of multiple drugs, but only when at least one established fall risk-increasing drug was part of the daily regimen. Part of the increased risk could be explained by comorbidity as shown in the fully adjusted model, but some drugs appeared to have a risk-increasing effect, independent of comorbidity. This is in contrast to the findings of Lawlor et al.[8]. They did, however, study composite groups of medications only. Possible explanations for the mechanism of action are numerous, e.g. diuretics can cause dizziness as a consequence of orthostatic hypotension, with falling as a result. Benzodiazepine derivatives may play a role by effects on the central nervous system. However, after adjustment for comorbid conditions and disability, polypharmacy (i.e. the number of drugs) remained a significant risk factor for falling.

In the Guideline for the Prevention of Falls in Older Persons [2], the assessment of persons having experienced a fall focuses on modifiable risk factors. Our results support the recommendation to assess medication use, being a modifiable risk factor for falling. According to our findings, the falls assessment should focus on identifying risk-increasing drugs rather than polypharmacy per se[1].

Limitations of the study

Being a cross-sectional study, our study may have some limitations. First, 37% of our population was younger than 65 years. This possibly explains the relatively low prevalence of falling (16.5%) in comparison with other studies [2], but is consistent with the large study of Lawlor et al.[8].

Because of the cross-sectional nature of this study, we cannot be certain that drug use preceded falling. The magnitude of this problem varies between the different observed associations. Calcium preparations, for example, may be prescribed as a consequence of falling to prevent fractures. However, in chronic disease situations, medications are generally prescribed on a continuous basis. Therefore it is likely that most of the drugs were already used before the assessment of falls. Third, it was not possible to control for ‘confounding by indication’, which is likely to play a role in the association between, for example, calcium preparations or laxative use and falling. Presumably, there is a clinically relevant association between osteoporosis and falling, or between disabling conditions, causing constipation and falling.

The majority of relevant comorbid conditions were taken into account in the analysis. However, we were not, for example, able to assess the influence of chronic pulmonary diseases on falling. Hence, some residual confounding may play a role in our study. Finally, there may be misclassification of the outcome, which was based on structured interview. The results are dependent upon recall of events, which might introduce ‘recall bias’ as a consequence of the retrospective character of our study. In an earlier study, falls were recalled with a specificity of 91.4%, and were more likely to be remembered if an injury had occurred. The number of falls was not accurately recalled in more than a third of the persons [15]. In our analysis we therefore dichotomized on falls vs. no falls. We have no reason to believe that misclassification of the outcome, if any, was differential. Moreover, the effect of under reporting of falls was minimized by exclusion of persons with an established cognitive disorder, mostly dementia [16].

Potentially, drug interactions can play a role in falling, but the methodology of our analysis was not suitable to address that issue.


In accordance with the meta-analyses by Leipzig et al.[6, 7], we also found an association between diuretics, quinine and derivatives, and psychotropic drugs (especially anxiolytics-benzodiazepine derivatives and hypnotics-benzodiazepine derivatives) with falling.

The major finding of our study is that the contribution of identifiable risk drugs to polypharmacy is associated with an increased fall risk, rather than polypharmacy itself. As a consequence, there is an opportunity for risk-reducing interventions in a frail elderly population, in whom polypharmacy is inevitable in order to control the underlying comorbidity.

We are grateful to Professor Th. Stijnen for helpful statistical advice.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References
  • 1
    Tinetti ME. Clinical practice. Preventing falls in elderly persons. N Engl J Med 2003; 348: 429.
  • 2
    American Geriatrics Society Panel on Falls Prevention OSP. Guideline for the prevention of falls in older persons. J Am Geriatr Soc 2001; 49: 66472.
  • 3
    Robbins AS, Rubenstein LZ, Josephson KR, Schulman BL, Osterweil D, Fine G. Predictors of falls among elderly people. Results of two population-based studies. Arch Intern Med 1989; 149: 162833.
  • 4
    Evans JG. Drugs and falls in later life. Lancet 2003; 361 (9356): 448.
  • 5
    Cumming RG, Miller JP, Kelsey JL, Davis P, Arfken CL, Birge SJ, Peck WA. Medications and multiple falls in elderly people: the St Louis OASIS study. Age Ageing 1991; 20: 45561.
  • 6
    Leipzig RM, Cumming RG, Tinetti ME. Drugs and falls in older people: a systematic review and meta-analysis. I. Psychotropic drugs. J Am Geriatr Soc 1999; 47: 309.
  • 7
    Leipzig RM, Cumming RG, Tinetti ME. Drugs and falls in older people: a systematic review and meta-analysis. II. Cardiac and analgesic drugs. J Am Geriatr Soc 1999; 47: 4050.
  • 8
    Lawlor DA, Patel R, Ebrahim S. Association between falls in elderly women and chronic diseases and drug use: cross sectional study. BMJ 2003; 327 (7417): 7127.
  • 9
    Hofman A, Grobbee DE, De Jong PT, Van Den Ouweland FA. Determinants of disease and disability in the elderly: the Rotterdam Elderly Study. Eur J Epidemiol 1991; 7: 40322.
  • 10
    Ott A, Breteler MMB, Van Harskamp F, Claus JJ, Van Der Cammen TJM, Grobbee DE, Hofman A. Prevalence of Alzheimer's disease and vascular dementia: association with education. The Rotterdam study. BMJ 1995; 310 (6985): 9703.
  • 11
    Pincus T, Summey JA, Soraci SA Jr, Wallston KA, Hummon NP. Assessment of patient satisfaction in activities of daily living using a modified Stanford Health Assessment Questionnaire. Arthritis Rheum 1983; 26: 134653.
  • 12
    Fries JF, Spitz PW, Young DY. The dimensions of health outcomes: the health assessment questionnaire, disability and pain scales. J Rheumatol 1982; 9: 78993.
  • 13
    Ooi WL, Hossain M, Lipsitz LA. The association between orthostatic hypotension and recurrent falls in nursing home residents. Am J Med 2000; 108: 10611.
  • 14
    ATC Index with DDDs. Oslo, Norway, 2002. WHO Collaborating Centre for Drug Statistics Methodology: ‘Guidelines for ATC Classification and DDD Assignment’, .
  • 15
    Peel N. Validating recall of falls by older people. Accid Anal Prev 2000; 32: 3712.
  • 16
    Cummings SR, Nevitt MC, Kidd S. Forgetting falls. The limited accuracy of recall of falls in the elderly. J Am Geriatr Soc 1988; 36: 6136.