Abstract

Background : Older people not only consume more medication but they also represent a group at high risk for adverse effects such as injurious falls. This study examines the association between the medications most commonly prescribed to older people in Sweden and fall injuries. Methods: This is a population-based, matched, case-control study of 64 399 persons aged ≥ 65 years in Sweden admitted to hospital because of a fall injury between March 2006 and December 2009, and four controls per case matched by gender, date of birth and place of residence. The prevalence of the 20 most commonly prescribed medications was compiled for the 30-day period before the index date. The association between those medications and injurious falls was estimated with odds ratios and corresponding 95% confidence intervals using conditional logistic regression. Results: Ten of the top 20 most commonly prescribed medications, and in particular the three medications affecting the central nervous system (CNS), significantly increased the risk of fall injuries (highest for opioids and antidepressants) but not the seven cardiovascular ones, who had a protective effect (lowest for angiotensin converting enzyme inhibitors and selective calcium channel blockers). Conclusions: The adverse effect of several commonly prescribed medications may seriously threaten their positive effects on the well-being and quality of life of older people. Their association with injurious falls is of particular concern as falls are prevalent and often leading to severe consequences. This needs to be acknowledged so physicians and patients can make informed decisions when prescribing and using them.

Introduction

The world population is aging both in absolute and relative terms. This demographic transition is paralleled by a change in the burden of disease, whereby non-communicable diseases are becoming more prevalent than communicable ones, and premature death is progressively giving way to a higher life expectancy and increased morbidity. 1 Among the main causes of disability, fall injuries rank high. In older people, they often lead to hospitalization and a long period of medical treatment, 2,3 and they can cause considerable morbidity with significant physical, psychological and social consequences. 2 In many instances, fall-related injuries can be fatal. 2,4

Individual factors that have a well-documented association with falls are age, 2 sex 2 and health status, 5 though these factors can be difficult to modify. Prescribed medications, the focus of this study, may also significantly affect the risk of injurious falls among older people, 6–13 and they are somewhat easier to act upon. Medications can impact on the risk of older people falling in different ways, including the expected and adverse effects of either single medications or drug–drug interactions 14 and dosage and change of medications. 15 Age-related factors can also come into play here by influencing either the potential side effects of a medication or the effect of the dosage compared with younger patients. 9

A recent study identified the 20 medications most commonly prescribed in the general population of people aged ≥65 years in Sweden 16 and revealed that such medications were used by up to 35% of community-dwelling older adults. A closer look at the list provided showed that several medications were known fall-inducing drugs, including hypnotics and sedatives, 9,11,17 antidepressants, 6 opioids, 12 other analgesics, 7 antipsychotics, 6 thiazide diuretics, 15 angiotensin converting enzyme (ACE) inhibitors, 13 thyroid hormones, 8 constipation drugs 8 and non-steroidal anti-inflammatory drugs (NSAID). 11 The knowledge at hand came from different study populations varying in age range and in health condition. In light of their potential adverse effects specifically on older people, in particular injurious falls and the risk of serious complications thereof, commonly prescribed drugs can have large individual and societal impacts, not least economic ones. 18

In this nationwide study, we assess the risk of fall injury leading to hospitalization associated with each of those most commonly prescribed medications in the general population of Swedish older people. 16 The medication-specific risk is assessed in the same population and separately for men and women.

Methods

Study design

This is a population-based, matched, case-control study, nested in a national cohort of 6 981 010 individuals born in 1958 or before and domiciled in Sweden at some point from 1973 onwards. The cohort was identified using the Swedish Total Population Register. The case-control study includes individuals ≥65 years who sustained a fall injury in the period between 1 March 2006 and 31 December 2009 and four matched controls per case.

Definitions of cases and controls

A case was defined as a person who sustained a fall injury leading to hospitalization (at least one night) registered in the National Inpatient Register based on the International Classification of Diseases, 10th Revision (Codes W00–W19). We considered only the first fall injury during the study period and excluded planned medical inpatient care for fall injury. Controls were randomly selected from the cohort among persons not having been hospitalized because of a fall injury during the study period. They were matched to the cases by sex, date of birth and area of residence. Date of admission, and corresponding date for the matched controls, was used as the index date.

In total, 64 399 cases (22 190 males and 42 209 females) were identified—including 257 596 controls—adding up to 321 995 persons. Among these, 29.4% were 65–79 years of age, 30.4% were 80–85, 21.2% were 86–90 and 19.1% were ≥90 years of age. There were more males in the younger groups and more females in the older groups ( table 1 ). The differences between cases and controls as regards educational level and civil status were only marginal, but there were noticeable differences between males and females with more males with higher education, more females widowed and males married ( table 1 ).

Table 1

Characteristics (%) of the study population regarding age, education, civil status and area of residence ( N = 321 995)

Characteristic  Male n = 110 950
 
Female n = 211 045
 
All  Cases n = 22 190   Controls n = 88 760  All  Cases n = 42 209   Controls n = 168 836  
Age group (years)       
    65–79 36.7 Matching variable 25.6 Matching variable 
    80–85 30.7   30.2   
    86–90 18.9   22.4   
    >90 13.8   21.9   
Educational level       
    Primary or lower secondary school 38.9 39.4 38.8 45.2 47.6 44.6 
    Upper secondary 23.60 22.2 24.0 20.2 20.1 20.2 
    Post-secondary 10.3 9.1 10.7 7.1 6.7 7.2 
    Postgraduate 0.7 0.6 0.7 0.1 0.1 0.1 
    Unknown/missing 26.5 28.7 25.9 27.4 25.4 27.9 
Civil status       
    Unmarried 9.4 11.2 8.9 6.6 7.3 6.4 
    Married 56.2 48.3 58.2 23.6 20.2 24.5 
    Divorced 10.3 12.9 9.6 10.4 11.4 10.2 
    Widowed 24.0 27.6 23.1 59.4 61.2 58.9 
    Missing 0.1 <0.1 0.1 0.1 <0.1 0.1 
Area of residence       
    Stockholm 17.7 Matching variable 18.0 Matching variable 
    Central eastern Sweden 12.6   12.4   
    West Sweden 22.0   21.4   
    South Sweden 12.5   13.5   
    Småland and Islands 8.0   7.9   
    Central northern Sweden 13.1   14.0   
    Central Norrland 6.4   6.1   
    Northern Norrland 7.5   6.8   
Characteristic  Male n = 110 950
 
Female n = 211 045
 
All  Cases n = 22 190   Controls n = 88 760  All  Cases n = 42 209   Controls n = 168 836  
Age group (years)       
    65–79 36.7 Matching variable 25.6 Matching variable 
    80–85 30.7   30.2   
    86–90 18.9   22.4   
    >90 13.8   21.9   
Educational level       
    Primary or lower secondary school 38.9 39.4 38.8 45.2 47.6 44.6 
    Upper secondary 23.60 22.2 24.0 20.2 20.1 20.2 
    Post-secondary 10.3 9.1 10.7 7.1 6.7 7.2 
    Postgraduate 0.7 0.6 0.7 0.1 0.1 0.1 
    Unknown/missing 26.5 28.7 25.9 27.4 25.4 27.9 
Civil status       
    Unmarried 9.4 11.2 8.9 6.6 7.3 6.4 
    Married 56.2 48.3 58.2 23.6 20.2 24.5 
    Divorced 10.3 12.9 9.6 10.4 11.4 10.2 
    Widowed 24.0 27.6 23.1 59.4 61.2 58.9 
    Missing 0.1 <0.1 0.1 0.1 <0.1 0.1 
Area of residence       
    Stockholm 17.7 Matching variable 18.0 Matching variable 
    Central eastern Sweden 12.6   12.4   
    West Sweden 22.0   21.4   
    South Sweden 12.5   13.5   
    Småland and Islands 8.0   7.9   
    Central northern Sweden 13.1   14.0   
    Central Norrland 6.4   6.1   
    Northern Norrland 7.5   6.8   

Exposure

We considered as ‘common medications’ among persons aged ≥65 years living in Sweden those 20 highlighted previously 16 and focuses on their recent use during the 30 days before the index date. The Swedish Prescribed Drug Register (SPDR) contains information on all prescribed and dispensed drugs at all pharmacies in Sweden and uses the five-level Anatomical Therapeutic and Chemical (ATC) classification system. 19 The information, including dispensation date, for all expenditures is recorded in the computerized system of the pharmaceutical services. Medications based on the third level, the therapeutic and pharmacological subgroup, were examined. Dispensations on the index day were excluded.

Potential confounders

Besides the matching variables, we considered two additional demographic attributes (level of education and civil status) as potential confounders due to their unequal distribution between males and females; see table 1 . Information on level of education and civil status, as well as on sex and area of residence, was retrieved from the Total Population Register and Sweden’s Total Enumeration Income Survey and Longitudinal Database for Sickness Insurance and Labour Market Studies. We also considered two measurements of comorbidity that aimed to capture the difference in health status: the Charlson Comorbidity Index and the total number of medications. These were the Charlson Comorbidity Index derived from diagnoses recorded in the National Patient Register during the 1-year period before the index date. 20 The total number of medications was measured considering the 120 days before the index date and determined using the SPDR. The following categories were created based on the third-level ATC code: no medication, 1–5, 6–10, 11–15, 16–20 and >20 medications.

Statistical analyses

First, descriptive statistics were used to describe and estimate the prevalence of the medications with 95% confidence intervals (95% CIs) among the cases and controls considering different demographic characteristics. Further, the association between each medication and fall injuries was assessed using conditional logistic regression to estimate crude and adjusted odds ratios (ORs) with 95% CIs stratified by gender. Potential confounding factors were decided upon by adjustment with a statistically significant effect; hence, only the total number of medications was adjusted for in the final analyses.

Missing data only occurred for education and civil status and were treated as a separate category in the analyses. Stata/IC 12.0 and Microsoft Excel 2010 were used to perform the statistical analyses.

The study was approved by the Regional Ethical Review Board in Stockholm, Sweden.

Results

The prevalence of the 20 most commonly prescribed medications among the older people studied ranged between 0.5 and 15.3% and, in most instances, the prevalence was significantly higher among cases than controls, both among males and females ( table 2 ). The prevalence of specific medications differed significantly between the genders.

Table 2

Prevalence of the most commonly prescribed and dispensed medication among cases and controls, stratified by gender

Medication ATC code  Prevalence in % (95% CI)
 
Male n = 110 950
 
Female n = 211 045
 
Cases n = 22 190   Controls n = 88 760   Cases n = 422 09   Controls n = 168 836  
Antithrombotic agents B01A 17.0 (16.5–17.5) 11.2 (11.0–11.4) 14.4 (14.1–14.8) 9.8 (9.7–9.9) 
Drugs for peptic ulcer and gastro-oesophageal reflux disease A02B 8.6 (8.2–9.0) 4.4 (4.3–4.5) 9.6 (9.3–9.8) 5.7 (5.6–5.8) 
Beta–blocking agents C07A 2.9 (2.6–3.1) 2.9 (2.8–3.0) 3.3 (3.1–3.4) 3.1 (3.0–3.2) 
High-ceiling diuretics C03C 4.0 (3.7–4.2) 2.1 (2.0–2.2) 4.3 (4.1–4.5) 2.9 (2.8–3.0) 
Vitamin B12 and folic acid B03B 4.1 (3.9–4.4) 2.1 (2.0–2.2) 4.3 (4.1–4.5) 2.7 (2.6–2.7) 
Constipation drugs A06A 3.8 (3.6–4.1) 2.1 (2.0–2.2) 3.6 (3.4–3.7) 2.4 (2.3–2.5) 
Calcium A12A 1.2 (1.1–1.4) 0.6 (0.6–0.7) 4.7 (4.5–4.9) 2.8 (2.8–2.9) 
Glucose-lowering drugs A10B 3.2 (2.9–3.4) 2.3 (2.2–2.4) 2.9 (2.7–3.0) 2.0 (1.9–2.0) 
Hypnotics and sedatives N05C 3.7 (3.4–3.9) 1.6 (1.5–1.7) 3.3 (3.1–3.5) 2.3 (2.2–2.3) 
Other analgesics and antipyretics N02B 2.7 (2.5–2.9) 1.1 (1.0–1.2) 3.1 (2.9–3.2) 2.0 (1.9–2.1) 
Opioids N02A 3.9 (3.6–4.1) 1.1 (1.1–1.2) 4.1 (3.9–4.3) 1.5 (1.5–1.6) 
NSAIDS M01A 1.5 (1.3–1.6) 1.2 (1.1–1.3) 1.6 (1.5–1.8) 1.2 (1.2–1.3) 
ACE inhibitors C09A 1.3 (1.1–1.4) 1.3 (1.2–1.4) 1.0 (0.9–1.1) 1.0 (0.9–1.0) 
Selective calcium channel blockers with mainly vascular effects C08C 0.9 (0.8–1.0) 1.1 (1.0–1.2) 1.0 (0.9–1.1) 1.2 (1.2–1.3) 
Thiazide/low-ceiling diuretics C03A 0.7 (0.6–0.8) 0.7 (0.7–0.8) 1.1 (1.0–1.2) 1.2 (1.1–1.2) 
Antidepressants N06A 1.4 (1.3–1.6) 0.5 (0.5–0.6) 1.8 (1.7–1.9) 0.9 (0.8–0.9) 
Oestrogens G03C 1.1 (1.0–1.2) 1.2 (1.2–1.3) 
Lipid-modifying agents C10A 0.8 (0.7–0.9) 1.0 (0.9–1.1) 0.7 (0.6–0.7) 0.8 (0.8–0.9) 
Thyroid preparations H03A 0.3 (0.3–0.4) 0.3 (0.2–0.3) 1.1 (1.0–1.2) 0.9 (0.9–1.0) 
AT II antagonist C09C 0.4 (0.4–0.5) 0.5 (0.5–0.6) 0.5 (0.5–0.6) 0.6 (0.6–0.6) 
Medication ATC code  Prevalence in % (95% CI)
 
Male n = 110 950
 
Female n = 211 045
 
Cases n = 22 190   Controls n = 88 760   Cases n = 422 09   Controls n = 168 836  
Antithrombotic agents B01A 17.0 (16.5–17.5) 11.2 (11.0–11.4) 14.4 (14.1–14.8) 9.8 (9.7–9.9) 
Drugs for peptic ulcer and gastro-oesophageal reflux disease A02B 8.6 (8.2–9.0) 4.4 (4.3–4.5) 9.6 (9.3–9.8) 5.7 (5.6–5.8) 
Beta–blocking agents C07A 2.9 (2.6–3.1) 2.9 (2.8–3.0) 3.3 (3.1–3.4) 3.1 (3.0–3.2) 
High-ceiling diuretics C03C 4.0 (3.7–4.2) 2.1 (2.0–2.2) 4.3 (4.1–4.5) 2.9 (2.8–3.0) 
Vitamin B12 and folic acid B03B 4.1 (3.9–4.4) 2.1 (2.0–2.2) 4.3 (4.1–4.5) 2.7 (2.6–2.7) 
Constipation drugs A06A 3.8 (3.6–4.1) 2.1 (2.0–2.2) 3.6 (3.4–3.7) 2.4 (2.3–2.5) 
Calcium A12A 1.2 (1.1–1.4) 0.6 (0.6–0.7) 4.7 (4.5–4.9) 2.8 (2.8–2.9) 
Glucose-lowering drugs A10B 3.2 (2.9–3.4) 2.3 (2.2–2.4) 2.9 (2.7–3.0) 2.0 (1.9–2.0) 
Hypnotics and sedatives N05C 3.7 (3.4–3.9) 1.6 (1.5–1.7) 3.3 (3.1–3.5) 2.3 (2.2–2.3) 
Other analgesics and antipyretics N02B 2.7 (2.5–2.9) 1.1 (1.0–1.2) 3.1 (2.9–3.2) 2.0 (1.9–2.1) 
Opioids N02A 3.9 (3.6–4.1) 1.1 (1.1–1.2) 4.1 (3.9–4.3) 1.5 (1.5–1.6) 
NSAIDS M01A 1.5 (1.3–1.6) 1.2 (1.1–1.3) 1.6 (1.5–1.8) 1.2 (1.2–1.3) 
ACE inhibitors C09A 1.3 (1.1–1.4) 1.3 (1.2–1.4) 1.0 (0.9–1.1) 1.0 (0.9–1.0) 
Selective calcium channel blockers with mainly vascular effects C08C 0.9 (0.8–1.0) 1.1 (1.0–1.2) 1.0 (0.9–1.1) 1.2 (1.2–1.3) 
Thiazide/low-ceiling diuretics C03A 0.7 (0.6–0.8) 0.7 (0.7–0.8) 1.1 (1.0–1.2) 1.2 (1.1–1.2) 
Antidepressants N06A 1.4 (1.3–1.6) 0.5 (0.5–0.6) 1.8 (1.7–1.9) 0.9 (0.8–0.9) 
Oestrogens G03C 1.1 (1.0–1.2) 1.2 (1.2–1.3) 
Lipid-modifying agents C10A 0.8 (0.7–0.9) 1.0 (0.9–1.1) 0.7 (0.6–0.7) 0.8 (0.8–0.9) 
Thyroid preparations H03A 0.3 (0.3–0.4) 0.3 (0.2–0.3) 1.1 (1.0–1.2) 0.9 (0.9–1.0) 
AT II antagonist C09C 0.4 (0.4–0.5) 0.5 (0.5–0.6) 0.5 (0.5–0.6) 0.6 (0.6–0.6) 

Most medications showed a positive association with fall injuries ( table 3 ), with the exception of the majority of the cardiovascular medications and oestrogens. Medications prescribed for cardiovascular diseases showed either no significant or a slightly protective effect. The only cardiovascular medication associated with a higher risk of fall injury was high-ceiling diuretics. Central nervous system (CNS) drugs such as opioids, hypnotics and sedatives, other analgesics and antipyretics and antidepressants, all showed an increased risk of fall injuries as did constipation drugs, antithrombotic agents, glucose-lowering drugs, thyroid preparations, NSAIDs, drugs for peptic ulcer, calcium and vitamin B12 and folic acid. The associations were similar for males and females.

Table 3

OR of fall injury for the most commonly prescribed and dispensed medication among older people

Medication ATC code  Expected effect a  Male n = 110 950
 
Female n = 211 045
 
OR (95% CI)
 
OR (95% CI)
 
Crude  Adjusted b Crude  Adjusted b 
Antithrombotic agents B01A − 1.63 (1.56–1.70) 1.17 (1.12–1.22) 1.56 (1.51–1.61) 1.17 (1.13–1.21) 
Drugs for peptic ulcer and gastro-oesophageal reflux disease A02B − 2.05 (1.94–2.17) 1.21 (1.14–1.29) 1.74 (1.67–1.81) 1.13 (1.09–1.18) 
Beta-blocking agents C07A − 0.98 (0.90–1.07) 0.77 (0.70–0.84) 1.05 (0.99–1.12) 0.89 (0.84–0.95) 
High-ceiling diuretics C03C ↑ 1.94 (1.79–2.11) 1.32 (1.22–1.44) 1.52 (1.44–1.61) 1.14 (1.08–1.20) 
Vitamin B12 and folic acid B03B − 2.03 (1.88–2.20) 1.54 (1.42–1.68) 1.65 (1.56–1.75) 1.30 (1.22–1.37) 
Constipation drugs A06A ↑ 1.88 (1.73–2.05) 1.23 (1.13–1.34) 1.50 (1.41–1.59) 1.07 (1.00–1.13) 
Calcium A12A ↓ 2.04 (1.77–1.36) 1.27 (1.09–1.47) 1.71 (1.62–1.80) 1.24 (1.18–1.31) 
Glucose-lowering drugs A10B − 1.40 (1.28–1.53) 0.93 (0.85–1.01) 1.46 (1.37–1.57) 1.05 (0.98–1.13) 
Hypnotics and sedatives N05C ↑ 2.32 (2.13–2.54) 1.76 (1.61–1.93) 1.47 (1.38–1.57) 1.21 (1.14–1.29) 
Other analgesics and antipyretics N02B ↑ 2.45 (2.21–2.72) 1.74 (1.57–1.94) 1.56 (1.46–1.66) 1.22 (1.14–1.30) 
Opioids N02A ↑ 3.54 (3.22–3.88) 2.30 (2.09–2.53) 2.73 (2.57–2.91) 2.00 (1.87–2.12) 
NSAIDS M01A ↑ 1.22 (1.08–1.39) 0.99 (0.87–1.13) 1.37 (1.25–1.49) 1.14 (1.04–1.24) 
ACE inhibitors C09A ↑ 0.98 (0.86–1.12) 0.77 (0.67–0.88) 1.04 (0.93–1.16) 0.87 (0.78–0.97) 
Selective calcium channel blockers with mainly vascular effects C08C − 0.83 (0.72–0.97) 0.67 (0.57–0.78) 0.83 (0.75–0.92) 0.72 (0.65–0.80) 
Thiazide/low-ceiling diuretics C03A ↑ 0.99 (0.83–1.18) 0.85 (0.71–1.02) 0.93 (0.84–1.03) 0.83 (0.75–0.91) 
Antidepressants N06A ↑ 2.82 (2.44–3.25) 2.26 (1.95–2.62) 2.04 (1.87–2.23) 1.76 (1.61–1.93) 
Oestrogens G03C − n.a. n.a. 0.85 (0.77–0.95) 0.70 (0.63–0.78) 
Lipid-modifying agents C10A − 0.83 (0.71–0.98) 0.63 (0.54–0.75) 0.79 (0.77–0.95) 0.65 (0.57–0.74) 
Thyroid preparations H03A ↑ 1.26 (0.98–1.63) 1.07 (0.83–1.40) 1.18 (1.06–1.31) 1.04 (0.94–1.16) 
AT II antagonist C09C − 0.86 (0.69–1.07) 0.66 (0.53–0.83) 0.93 (0.80–1.07) 0.76 (0.65–0.87) 
Medication ATC code  Expected effect a  Male n = 110 950
 
Female n = 211 045
 
OR (95% CI)
 
OR (95% CI)
 
Crude  Adjusted b Crude  Adjusted b 
Antithrombotic agents B01A − 1.63 (1.56–1.70) 1.17 (1.12–1.22) 1.56 (1.51–1.61) 1.17 (1.13–1.21) 
Drugs for peptic ulcer and gastro-oesophageal reflux disease A02B − 2.05 (1.94–2.17) 1.21 (1.14–1.29) 1.74 (1.67–1.81) 1.13 (1.09–1.18) 
Beta-blocking agents C07A − 0.98 (0.90–1.07) 0.77 (0.70–0.84) 1.05 (0.99–1.12) 0.89 (0.84–0.95) 
High-ceiling diuretics C03C ↑ 1.94 (1.79–2.11) 1.32 (1.22–1.44) 1.52 (1.44–1.61) 1.14 (1.08–1.20) 
Vitamin B12 and folic acid B03B − 2.03 (1.88–2.20) 1.54 (1.42–1.68) 1.65 (1.56–1.75) 1.30 (1.22–1.37) 
Constipation drugs A06A ↑ 1.88 (1.73–2.05) 1.23 (1.13–1.34) 1.50 (1.41–1.59) 1.07 (1.00–1.13) 
Calcium A12A ↓ 2.04 (1.77–1.36) 1.27 (1.09–1.47) 1.71 (1.62–1.80) 1.24 (1.18–1.31) 
Glucose-lowering drugs A10B − 1.40 (1.28–1.53) 0.93 (0.85–1.01) 1.46 (1.37–1.57) 1.05 (0.98–1.13) 
Hypnotics and sedatives N05C ↑ 2.32 (2.13–2.54) 1.76 (1.61–1.93) 1.47 (1.38–1.57) 1.21 (1.14–1.29) 
Other analgesics and antipyretics N02B ↑ 2.45 (2.21–2.72) 1.74 (1.57–1.94) 1.56 (1.46–1.66) 1.22 (1.14–1.30) 
Opioids N02A ↑ 3.54 (3.22–3.88) 2.30 (2.09–2.53) 2.73 (2.57–2.91) 2.00 (1.87–2.12) 
NSAIDS M01A ↑ 1.22 (1.08–1.39) 0.99 (0.87–1.13) 1.37 (1.25–1.49) 1.14 (1.04–1.24) 
ACE inhibitors C09A ↑ 0.98 (0.86–1.12) 0.77 (0.67–0.88) 1.04 (0.93–1.16) 0.87 (0.78–0.97) 
Selective calcium channel blockers with mainly vascular effects C08C − 0.83 (0.72–0.97) 0.67 (0.57–0.78) 0.83 (0.75–0.92) 0.72 (0.65–0.80) 
Thiazide/low-ceiling diuretics C03A ↑ 0.99 (0.83–1.18) 0.85 (0.71–1.02) 0.93 (0.84–1.03) 0.83 (0.75–0.91) 
Antidepressants N06A ↑ 2.82 (2.44–3.25) 2.26 (1.95–2.62) 2.04 (1.87–2.23) 1.76 (1.61–1.93) 
Oestrogens G03C − n.a. n.a. 0.85 (0.77–0.95) 0.70 (0.63–0.78) 
Lipid-modifying agents C10A − 0.83 (0.71–0.98) 0.63 (0.54–0.75) 0.79 (0.77–0.95) 0.65 (0.57–0.74) 
Thyroid preparations H03A ↑ 1.26 (0.98–1.63) 1.07 (0.83–1.40) 1.18 (1.06–1.31) 1.04 (0.94–1.16) 
AT II antagonist C09C − 0.86 (0.69–1.07) 0.66 (0.53–0.83) 0.93 (0.80–1.07) 0.76 (0.65–0.87) 

a: The effect classification was performed to put the results in a context. Previous research was scanned using the following search terms in PubMed: name of the medications, fall injuries, fall risk and/or elderly.

b: Adjusted by number of medications, n.a.: not applicable.

The associations remained mainly similar after adjustment for total number of medications, although with lowered effect estimates. For glucose-lowering drugs, the adjusted analyses showed no significant association with fall injury, among either males or females. Additionally, among males, there was no association found for NSAIDs and among females for thyroid preparations. Among females, a slightly protective effect was observed for low-ceiling diuretics.

Discussion

In this large population-based study, we provide additional evidence that several of the 20 medications most commonly prescribed to older people (one in two) are positively associated with fall injuries leading to hospitalization. From the 10 most common ones, 8 have a positive association among males and 7 among females. The gender differences are otherwise marginal.

Unique to this study is that all assessments are made in the same population, concern community dwellings, by far the largest segment of the population of older people in Sweden, and focus on falls with potentially serious consequences, as they all led to hospitalization at least one night. We acknowledge that several associations were already known and that some studies provided far greater insight into the mechanisms coming into play. Those studies, however, describe the situation of a variety of target populations and often are medication specific.

Single or groups of medications known to be of concern are CNS drugs, 6,9,11,17 high-ceiling diuretics 8 and constipation drugs. 8 NSAIDs for their part have a positive (and expected) 11 association among women but not among men; a difference that we cannot explain from the data at hand. It is of note that the positive association between antithrombotic agents, drugs for peptic ulcer and vitamin B12 has not been shown in previous studies, but one could posit that the underlying diseases can contribute to explaining the association found. 21 An indicator for this confounding is the observation, that medications with similar indications such as cardiovascular medications have shown the same trend, though these medications also could have different effects on the human body. Antithrombotic agents are the most commonly prescribed medications among older people and are very likely to be prescribed together with other medications. History of stroke as an underlying disease can influence the risk of falling. 22 We observed an increased risk, and a higher rate of complications after a fall due to the longer bleeding time is possible, although by adjusting the analyses for comorbidity, this confounding was at least partially taken care of.

Medications for peptic ulcer and gastro-oesophageal reflux diseases showed only a slightly elevated risk when adjusted for total number of medications. These medications are often prescribed to reduce the risk of gastro-intestinal complications from other medications, especially NSAIDs 23 prescribed for pain, creating the potential for confounding by indication as well as drug interactions. Our measures of comorbidity do not cover the majority of gastro-oesophageal diseases, so residual confounding could explain the association found.

Vitamin B12 and folic acid prescribed for anaemia might, by leading to dizziness, confound the estimates. Vitamin B12 has also been described as protective for cardiovascular disease, osteoporosis and mental disorders. 24 One explanation for the observed effect could hence be that persons prescribed vitamin B12 were at a higher risk for these diseases and risk of falls before the deficiency was diagnosed. Further, previous research has pointed to various factors such as civil status, socio-economic position, age and the dosing regimen 25 that can influence compliance. Psychiatric diseases such as depression, 25 side effects of the medication 26 as well as the number of medications 26 are factors known to affect compliance in a negative way. The cases more prone to these factors are more likely to have a negative compliance, and therefore, the observed effect could be slightly overestimated.

Although we expected a protective effect 27 or no effect 28 for calcium, our study shows an increased risk. The reason for this observation can be due to calcium being prescribed to persons with osteoporosis leading to a higher risk of injuries due to falling. 21,22,28,29 In contrast, most medications prescribed for cardiovascular diseases as well as oestrogens were negatively associated with the risk of injurious fall, whereas no particular effect was expected for angiotensin (AT) II antagonists, selective calcium channel blockers and beta blocking agents. 13 This does not apply to ACE inhibitors 13 and glucose-lowering drugs, 8 which were both expected to increase the risk of falls, but such an association was not found in this study.

Finally, contrary to expectations, no association was found with thyroid preparations and low-ceiling diuretics. 8,13 Oestrogens, prescribed for post-menopausal symptoms, have not been related to fall injuries previously, but seem to lower the risk of injurious falls. This observation can be due to diminished severity of falls through strengthening of the bones, 30 leading to less severe injuries.

Although this study focused on the 20 most common medications among older people in Sweden, it is of note that the level of prescription of these medications differs considerably in range and to some extent between males and females. Nonetheless, most medications are more prevalent among the fall-injured than among their controls. Because of the fact that these medications are the most common ones, 16 their adverse effects on the likelihood of serious fall injuries are of great importance 14 for both attending physicians and public health agencies. Furthermore, older people are often frail because of morbidity and therefore more prone to polypharmacy, 31 and future studies therefore need to do in-depth analyses of the potential interaction effects.

Strengths and limitations

Noteworthy strengths are that this study is population-based, large in size and based on data of documented quality and coverage. 19,32 The nested case-control design used yields a slightly lower level of statistical efficiency compared with the cohort design. 33 As a result of its population-based design, this study yields results that can be generalized to the entire elderly population in Sweden. Further, the recall bias and selection bias are minimized.

This study captures only fall injuries leading to inpatient medical care. Because of this, we cannot assume that the results can be generalized to less severe falls not leading to hospitalization or treated elsewhere.

Although using the SPDR has major advantages, it also leads to some potential limitations. The register does not include information on medications dispensed over-the-counter or during hospitalization, 19 nor the actual intake, compliance to or indication for prescription. However, the drugs had been dispensed, which is a step closer to intake than mere prescription. Though over-the-counter medications might be common and also have an effect on the risk of fall injuries, their independent effect or potential interaction effect cannot be elucidated. Most of the 20 most common medications studied are not available over the counter in Sweden. Hence, most exposure was caught and potential misclassification would be non-differential between cases and controls.

By matching by sex, area of residence and date of birth, the influence of these factors was taken into account, and the effects presented cannot be explained by differences between cases and controls with regard to these variables. Other factors such as level of education, civil status and the Charlson Comorbidity Index were also explored as potential confounders but found to have only marginal influence on the effect estimates. More specific information about morbidities, and particular indication for prescription, is needed to fully address the potential confounding by indication, which was only partly adjusted for in this study.

Conclusions

Among the medications most commonly prescribed to older people, quite a few may pose a problem to their safety by increasing the risk of injurious falls. Fall injuries are major threat to the health and well-being of older people, far more than in populations on which the side effects of drugs are tested, and this is a serious concern. Weighing this in the clinical setting, however, may be challenging; although we can assume that the risk for individual patients to sustain injurious falls would be minimized by not prescribing these medications, they may still remain essential in other critical aspects of health and well-being. This is even more problematic in instances where there are no non-fall-inducing medications as substitute. Even drug–drug interactions are of course to be taken into account, although this was beyond the scope of the current study.

Funding

This work was supported by the Swedish Civil Contingencies Agency [grant number 2010-2808-13].

Conflicts of interest : None declared.

Keypoints

  • The consumption of medication is common among older people and drugs—or combinations of drugs—are prescribed for a variety of health conditions.

  • Fall injuries are a frequent adverse effect of the consumption of medication among older people not least some of those most commonly prescribed ones.

  • Eleven of the 20 most commonly prescribed medications are associated with severe injuries requiring hospitalization.

  • This study shows a positive association between antithrombotic agents, drugs for peptic ulcer and vitamin B12 and injurious falls, which has not been shown in previous studies. It is likely that the underlying disease contributes to explaining the association.

  • The side effects of pharmaceutical treatments important for older people’s health are sometimes poorly assessed. Falls, in particular, constitute a severe threat to their life and well-being, not least those leading to hospitalization. The fact that commonly prescribed medications also play a role for such injuries must be regarded as serious.

References

1
Murray
CJ
Vos
T
Lozano
R
, et al.  . 
Disability-adjusted life years (DALYs) for 291 diseases and injuries in 21 regions, 1990-2010: a systematic analysis for the Global Burden of Disease Study 2010
Lancet
 , 
2012
, vol. 
380
 (pg. 
2197
-
223
)
2
World Health Organisation
WHO Global Report on Falls Prevention in Older Age
 , 
2007
France
WHO
3
Alexander
BH
Rivara
FP
Wolf
ME
The cost and frequency of hospitalization for fall-related injuries in older adults
Am J Public Health
 , 
1992
, vol. 
82
 (pg. 
1020
-
3
)
4
Korhonen
N
Kannus
P
Niemi
S
, et al.  . 
Fall-induced deaths among older adults: nationwide statistics in Finland between 1971 and 2009 and prediction for the future
Injury
 , 
2013
, vol. 
44
 (pg. 
867
-
71
)
5
Akyol
AD
Falls in the elderly: what can be done?
Int Nurs Rev
 , 
2007
, vol. 
54
 (pg. 
191
-
6
)
6
Hartikainen
S
Lönnroos
E
Louhivuori
K
Medication as a risk factor for falls: critical systematic review
J Gerontol A Biol Sci Med Sci
 , 
2007
, vol. 
62
 (pg. 
1172
-
81
)
7
Kelly
KD
Pickett
W
Yiannakoulias
N
, et al.  . 
Medication use and falls in community-dwelling older persons
Age Ageing
 , 
2003
, vol. 
32
 (pg. 
503
-
9
)
8
Cumming
RG
Miller
JP
Kelsey
JL
, et al.  . 
Medications and multiple falls in elderly people: the St Louis OASIS study
Age Ageing
 , 
1991
, vol. 
20
 (pg. 
455
-
61
)
9
Huang
AR
Mallet
L
Rochefort
CM
, et al.  . 
Medication-Related Falls in the Elderly: causative factors and preventive strategies
Drugs Aging
 , 
2012
, vol. 
29
 (pg. 
359
-
76
)
10
Ziere
G
Dieleman
JP
Hofman
A
, et al.  . 
Polypharmacy and falls in the middle age and elderly population
Br J Clin Pharmacol
 , 
2006
, vol. 
61
 (pg. 
218
-
23
)
11
Woolcott
JC
Richardson
KJ
Wiens
MO
, et al.  . 
Meta-analysis of the impact of 9 medication classes on falls in elderly persons
Arch Intern Med
 , 
2009
, vol. 
169
 (pg. 
1952
-
60
)
12
Söderberg
KC
Laflamme
L
Möller
J
Newly initiated opioid treatment and the risk of fall-related injuries. A nationwide, register-based, case-crossover study in sweden
CNS Drugs
 , 
2013
, vol. 
27
 (pg. 
155
-
61
)
13
Gribbin
J
Hubbard
R
Gladman
JR
, et al.  . 
Risk of falls associated with antihypertensive medication: population-based case-control study
Age Ageing
 , 
2010
, vol. 
39
 (pg. 
592
-
7
)
14
Seymour
RM
Routledge
PA
Important drug-drug interactions in the elderly
Drugs Aging
 , 
1998
, vol. 
12
 (pg. 
485
-
94
)
15
Payne
RA
Abel
GA
Simpson
CR
Maxwell
SR
Association between prescribing of cardiovascular and psychotropic medications and hospital admission for falls or fractures
Drugs Aging
 , 
2013
, vol. 
30
 (pg. 
247
-
54
)
16
Johnell
K
Fastbom
J
Comparison of prescription drug use between community-dwelling and institutionalized elderly in Sweden
Drugs Aging
 , 
2012
, vol. 
29
 (pg. 
751
-
8
)
17
Herings
RM
Stricker
BH
de Boer
A
, et al.  . 
Benzodiazepines and the risk of falling leading to femur fractures. Dosage more important than elimination half-life
Arch Intern Med
 , 
1995
, vol. 
155
 (pg. 
1801
-
7
)
18
Panneman
MJ
Goettsch
WG
Kramarz
P
Herings
RM
The costs of benzodiazepine-associated hospital-treated fall injuries in the EU: a pharmo study
Drugs Aging
 , 
2003
, vol. 
20
 (pg. 
833
-
9
)
19
Wettermark
B
Hammar
N
Fored
CM
, et al.  . 
The new Swedish Prescribed drug register—opportunities for pharmacoepidemiological research and experience from the first six months
Pharmacoepidemiol Drug Saf
 , 
2007
, vol. 
16
 (pg. 
726
-
35
)
20
Charlson
ME
Pompei
P
Ales
KL
MacKenzie
CR
A new method of classifying prognostic comorbidity in longitudinal studies: development and validation
J Chronic Dis
 , 
1987
, vol. 
40
 (pg. 
373
-
83
)
21
Signorello
LB
McLaughlin
JK
Lipworth
L
, et al.  . 
Confounding by indication in epidemiologic studies of commonly used analgesics
Am J Ther
 , 
2002
, vol. 
9
 (pg. 
199
-
205
)
22
Schmid
AA
Kapoor
JR
Dallas
M
Bravata
DM
Association between stroke severity and fall risk among stroke patients
Neuroepidemiology
 , 
2010
, vol. 
34
 (pg. 
158
-
62
)
23
Le
Ray
I
Barkun
AN
Vauzelle-Kervroëdan
F
Bardou
M
Failure to renew prescriptions for gastroprotective agents to patients on continuous nonsteroidal anti-inflammatory drugs increases rate of upper gastrointestinal injury
Clin Gastroenterol Hepatol
 , 
2013
, vol. 
11
 (pg. 
499
-
504
)
24
Hughes
CF
Ward
M
Hoey
L
McNulty
H
Vitamin B12 and ageing: current issues and interaction with folate
Ann Clin Biochem
 , 
2013
, vol. 
50
 (pg. 
315
-
29
)
25
Vlasnik
JJ
Aliotta
SL
Delor
B
Medication adherence: factors influencing compliance with prescribed medication plans
Case Manager
 , 
2005
, vol. 
16
 (pg. 
47
-
51
)
26
Hughes
CM
Medication non-adherence in the elderly: How big is the problem?
Drugs Aging
 , 
2004
, vol. 
21
 (pg. 
793
-
811
)
27
Bischoff-Ferrari
HA
Orav
EJ
Dawson-Hughes
B
Effect of cholecalciferol plus calcium on falling in ambulatory older men and women: a 3-year randomized controlled trial
Arch Intern Med
 , 
2006
, vol. 
166
 (pg. 
424
-
30
)
28
Kärkkäinen
MK
Tuppurainen
M
Salovaara
K
, et al.  . 
Does daily vitamin D 800 IU and calcium 1000 mg supplementation decrease the risk of falling in ambulatory women aged 65-71 years? A 3-year randomized population-based trial (OSTPRE-FPS)
Maturitas
 , 
2010
, vol. 
65
 (pg. 
359
-
65
)
29
Uusi-Rasi
K
Kärkkäinen
MUM
Lamberg-Allardt
CJE
Calcium intake in health maintenance-a systematic review
Food Nutr Res
 , 
2013
, vol. 
57
 
30
Wren
BG
The benefits of oestrogen following menopause: why hormone replacement therapy should be offered to postmenopausal women
Med J Aust
 , 
2009
, vol. 
190
 (pg. 
321
-
5
)
31
Fulton
MM
Allen
ER
Polypharmacy in the elderly: a literature review
J Am Acad Nurse Pract
 , 
2005
, vol. 
17
 (pg. 
123
-
32
)
32
Ludvigsson
JF
Andersson
E
Ekbom
A
, et al.  . 
External review and validation of the Swedish national inpatient register
BMC Public Health
 , 
2011
, vol. 
11
 pg. 
450
 
33
Ernster
VL
Nested Case-Control Studies
Prev Med
 , 
1994
, vol. 
23
 (pg. 
587
-
90
)

Comments

0 Comments