Abstract

Objectives: to compare the incidence of recurrent falls in older people with and without diabetes, and to examine diabetes- and fall-related risk factors explaining the increased risk of recurrent falls associated with diabetes.

Methods: population-based cohort study of 1,145 (85 with diabetes) community-dwelling participants, aged ≥65 years, from The Longitudinal Aging Study Amsterdam (LASA). Falls were assessed prospectively (every 3 months) during a 3-year follow-up period. Incidence of recurrent falls was estimated with Poisson regression analyses. The associations between diabetes and time to recurrent falls, defined as at least two falls occurring within a 6-month period, and the potential explanatory role of several risk factors herein, were analysed with the use of Cox-regression models.

Results: during a mean follow-up of 139 weeks, 30.6% of the individuals with and 19.4% of the individuals without diabetes fell recurrently [incidence rate of 129.7 versus 77.4 per 1,000 persons-years, respectively, HR = 1.67 (95% CI: 1.11–2.51)]. Adjustments for potential confounders did not change the increased risk associated with diabetes [HR = 1.63 (1.06–2.52)]. Factors that partly explained this increased risk were: greater number of medication, higher levels of pain, poorer self-perceived health, lower physical activity and grip strength, more limitations in ADLs, lower-extremity physical performance and cognitive impairment. Altogether, these variables accounted for ∼47% of the increased risk of recurrent falls associated with diabetes [adjusted HR = 1.30 (0.79–2.11)].

Conclusion: fall prevention efforts targeting the factors identified above may need to be incorporated into the care and treatment of older individuals with diabetes.

Introduction

About one-third of community-dwelling adults aged 65 and older fall each year, with approximately 1 in 10 falls resulting in a serious soft tissue injury, traumatic brain injury or fracture [1]. Even if non-injurious, falls carry serious consequences including decreases in social and physical activities, physical decline, disability, loss of independence and institutionalisation [2, 3]. Several studies have shown that older adults with diabetes fall more often than their peers without diabetes [4–9]. The prevalence of diabetes increases with age, and projections for the coming decades emphasise an increase in diabetes prevalence in particular in people >65 years of age [10, 11]. The expected increase in diabetes-related falls among older adults will likely impose a significant burden on health-care systems. Therefore, investigating the extent to which older individuals with diabetes have higher incidence of falls, and identification of diabetes- and falls-related factors amenable for intervention is of utmost importance.

So far, only two prospective studies, confined to women, have compared the incidence of falls among older individuals with and without diabetes, and in both the incidence of falls, particularly recurrent falls, was significantly higher particularly among women who were treated with insulin [6, 7]. A recent large population-based study of individuals with diabetes showed, however, that polypharmacy was associated with greater risk of incident falls [12]. In addition, and apart from the well-known cardiovascular complications, diabetes is associated with several ‘geriatric syndromes’ such as cognitive decline and dementia, depression, loss of muscle strength, functional limitations and physical disabilities, visual and hearing impairments, urinary incontinence and chronic pain [13, 14]. All of these are determinants of falls among community-dwelling older individuals [1, 15] and are thus likely to explain the excess risk of falls in individuals with diabetes [16]. Other potential factors that may be involved include inflammatory and hormonal changes [13], but their role in diabetes-related falls has never been investigated before.

We have therefore investigated, with a prospective-study design: first, the extent to which the risk of recurrent falls is higher in older community-dwelling subjects with diabetes relative to those without; and second, which diabetes- and fall-related risk factors could explain such an increased risk.

Methods

Study sample and design

The Longitudinal Aging Study Amsterdam (LASA) is an ongoing interdisciplinary cohort study on predictors and consequences of changes in physical, cognitive, emotional and social functioning in the ageing population of The Netherlands [17]. For the present study, the baseline study sample consisted of participants who completed both the main and medical interviews during LASA's second cycle of examinations and who were ≥65 years old as per 1 January 1996 (n = 1,509 from 1,720 eligible). These individuals were invited to participate in a 3-year fall and fracture follow-up study, but 82 did not participate (57 refused, 12 died and 13 became severely ill and thus ineligible before the start of follow-up). We further excluded 364 participants from the analyses: those who were not living independently in the community (n = 88) and/or who reported having had ≥2 falls during the previous year (n = 230). Therefore, the current study population consisted of 1,145 individuals (see further details and Flow-chart illustrating the study design and sample selection in the Supplementary data available in Age and Ageing online, Appendix 1).

Informed consent was obtained from all participants and the study was approved by the Medical Ethics Committee of the VU University Medical Center, Amsterdam.

Ascertainment of diabetes

Presence of diabetes mellitus (type 1 or 2, not specified) at baseline was defined with an algorithm combining three sources of information. First, the participants were assigned the diagnosis of diabetes (n = 64) when such diagnosis was present in the general practitioners' (GP) records. Second, a diagnosis of diabetes was assigned to participants whose GPs' records could not be retrieved (27% of the study population), if they self-reported diabetes and use of glucose-lowering medication was confirmed by a trained interviewer during the baseline medical interview (n = 21).

Ascertainment of (recurrent) falls

Falls were prospectively assessed during 3 years using a fall calendar [18, 19]. A fall was defined as ‘an unintentional change in position resulting in coming to rest at a lower level or on the ground’ [20]. Participants were asked to tick every week whether or not they had fallen. Once every 3 months, the calendar page was mailed to the Research Institute. If the calendar page was not received (even after a reminder), or if the page was completed incorrectly, the participants (or proxies if participants were unable to respond) were contacted by telephone.

Participants were categorised as recurrent fallers if they fell at least twice within a 6-month period [18, 19] and as an occasional faller if they fell at least once during follow-up, but did not meet the criteria for recurrent falls [19]. Distinction between occasional and recurrent falls was made deliberately to discern falls likely to have been coincidental from those more likely related to intrinsic disease-related causes [3, 21]. Time from baseline (date of medical interview) to the date of the first (occasional) fall and to the date of the first recurrent fall were recorded in week. Subjects who did not report (recurrent) falls during follow-up were censored. Time to censor was the study maximal duration (156 weeks) among participants who completed all falls calendar sheets; or time to first missed calendar sheet (due to reasons such as death, too ill or refusal to continue follow-up) among those who did not complete and were thus coded as lost to follow-up (13% of study participants).

Potential confounders and mediators

Listed in Table 1 (see detailed description in Supplementary data available in Age and Ageing online, Appendix 1).

Table 1.

Baseline characteristics of study's participants stratified by diabetes or recurrent falls status (n = 1,145)

Characteristics Missing observations Diabetes status
 
Recurrent falling during follow-up
 
  No (n = 1,060) Yes (n = 85) No (n = 913) Yes (n = 232) 
Potential confounders 
 Socio-demographic and home characteristics 
  Sex, % women — 49.3 56.5 49.4 51.7 
  Age, years — 75.4 ± 6.5 77.2 ± 6.6* 75.2 ± 6.4 76.9 ± 6.9*** 
   Education level [low (≤6 years)/medium (7–10 years)/high (≥11 years)], % — 41.0/30.3/28.7 45.9/30.6/23.5 41.7/31.1/27.2 40.1/27.2/32.7 
  Level of urbanisation      
   <500/500 to 1,000/1,000 to1,500/1,500 to 2,500/>2,500 no. addresses/km2, % 22.3/20.9/13.9/18.9/24.1 21.2/25.9/20.0/12.9/20.0 23.1/21.3/13.5/18.6/23.5 18.5/21.1/17.7/17.7/25.0 
  Pets (cats and/or dogs) in household, % — 16.6 16.5 15.9 19.4 
  Special adjustments in house, % 27.2 32.9 26.1 33.6* 
 Lifestyle risk factors 
  Alcohol drinkers, % 76.2 68.2 75.0 77.9 
  Current smoker, % 18.5 18.8 18.8 17.3 
  Body mass index, kg/m2 13 26.9 ± 4.3 27.9 ± 4.2* 27.1 ± 4.4 26.4 ± 3.9* 
Potential mediators 
 Co-morbidities and medication 
  Cardiovascular diseases, % 34.3 36.5 33.8 36.8 
  Cancer, % 12.0 11.8 12.4 10.4 
  Chronic obstructive pulmonary disease/asthma, % 14.8 11.8 15.1 12.6 
  Arthritis (rheumatoid/osteoarthritis), % 43.5 49.4 42.3 50.6* 
  Chronic diseases (excluding DM), no. 1 (0–2) 1 (0–2) 1 (0–2) 1 (1–2) 
  Psychotropic drugs, % 14.4 20.0 12.7 23.4*** 
  Blood pressure-lowering drugs, % 31.8 48.2** 33.6 30.7 
  Lipid-lowering drugs, % 4.6 4.7 5.0 3.0 
  Glucose-lowering drugs, % — 88.2 5.7 10.0* 
   Insulin,/Oral/Both, % — 40.0/67.1/14.1 2.6/4.5/1.1 4.3/6.9/0.9 
  Medicines, no. 2 (0–3) 3 (2–5)*** 2 (0–3) 2 (1–4)* 
  Polypharmacy (≥4 drugs), % 22.7 48.2*** 23.2 29.9* 
 Physical impairment and general health      
  Orthostatic hypotension, % 46 34.4 41.0 33.9 38.5 
  Regular dizziness, % 13.6 15.5 12.2 20.9*** 
  Foot problems, % 27.3 22.4 26.6 28.1 
  Incontinence, % 21.7 28.2 20.7 28.1* 
  Visual impairment (poor/insufficient/good), % 15.9/25.7/88.5 21.2/24.7/54.1 14.8/25.1/60.1 22.1/27.7/50.2** 
  Hearing impairment (poor/insufficient/good), % 10.3/7.7/82.0 20.0/8.2/71.8** 10.3/7.3/82.4 13.9/9.1/77.0 
  Subjective body pain score (range 0–6) 94 0 (0–2) 1 (0–2)** 0 (0–1) 1 (0–1)*** 
   Any pain, %  31.9 46.2** 31.4 38.8* 
  Self-rated health (range 1–5)a 2 (2–3) 3 (2–3)*** 2 (2–3) 3 (2–3)* 
  Hospital admission past 6 months, % – 8.9 12.9 8.0 13.8** 
  Occasional fall in previous year, % – 19.8 17.6 16.9 30.6*** 
 Activity, muscle strength and mobility 
  Physical activity, min/d 41 130 (76–201) 122 (30–180)* 134 (75–203) 118 (70–181)* 
  Grip strength, kg 48 57.5 ± 20.4 50.3 ± 19.7** 57.9 ± 19.9 53.1 ± 22.1** 
  Functional limitations in activities of daily living (ADLs) (range 0–6) 10 1 (0–2) 2 (1–4)*** 1 (0–2) 1 (0–3)** 
  Physical performance (range 0–12)b 43 8 (6–10) 6 (2–9)*** 8 (6–10) 7 (4–9)*** 
Psychosocial functioning      
  Cognitive impairment score (Mini-Mental Scale Examination, MMSE) (range 0–30) 28 (26–29) 27 (24–28)** 28 (26–29) 27 (25–29) 
   MMSE ≤23, %  10.7 21.4** 10.6 14.7 
  Depression score (Center for Epidemiological Studies Depression Scale—CES-D) (range 0–60) 30 6 (2–11) 6 (4–11) 5 (2–10) 8 (3–13)*** 
   CES-D≥16, %  14.5 8.4 13.5 16.5 
  Fear of falling (range 0–30) 1 (0–3) 1 (0–4) 0 (0–3) 1 (0–3)*** 
  Loneliness (range 0–11) 2 (0–4) 2 (0–4) 1 (0–3) 2 (0–4)** 
  Living alone, % – 37.5 49.4* 37.3 42.2 
 Biomarkers 
  25-Hydroxyvitamin D, nmol/l 111 55.4 ± 24.0 48.4.5 ± 20.2* 54.7 ± 23.3 55.7 ± 25.6 
  Sex hormone-binding globulin, nmol/l 110 43.8 (32.5–58.7) 41.3 (27.0–55.7) 42.5 (31.4–57.9) 46.3 (34.4–60.6)* 
  Insulin-like growth factor-1, nmol/l 112 14.0 ± 5.2 13.5 ± 5.9 14.0 ± 5.3 13.5 ± 5.2 
  Osteocalcin, nmol/l 112 2.2 ± 1.1 2.1 ± 1.1 2.1 ± 1.1 2.2 ± 1.1 
  Parathyroid hormone, pmol/l 111 3.1 (2.4–4.2) 3.3 (2.4–4.4) 3.1 (2.4–4.2) 3.1 (2.4–4.1) 
  Interleukin-6, pg/ml 135 1.9 (1.1–3.1) 1.9 (1.4–3.3) 1.9 (1.1–3.1) 1.8 (0.9–3.1) 
  C-reactive protein, mg/l 134 2.9 (1.4–6.4) 4.0 (2.5–8.2)** 3.2 (1.5–6.8) 2.8 (1.2–5.9) 
  Albumin, g/l 109 44.5 ± 2.7 44.3 ± 2.8 44.6 ± 2.8 44.3 ± 2.6 
  Creatinine, µmol/l 108 90 (79–103) 89 (81–108) 90 (79–103) 93 (80–106) 
  Glomerular filtration rate, ml/min per 1.73 m2 108 58.0 ± 11.9 56.2 ± 12.5 58.2 ± 12.0 56.6 ± 11.9 
Characteristics Missing observations Diabetes status
 
Recurrent falling during follow-up
 
  No (n = 1,060) Yes (n = 85) No (n = 913) Yes (n = 232) 
Potential confounders 
 Socio-demographic and home characteristics 
  Sex, % women — 49.3 56.5 49.4 51.7 
  Age, years — 75.4 ± 6.5 77.2 ± 6.6* 75.2 ± 6.4 76.9 ± 6.9*** 
   Education level [low (≤6 years)/medium (7–10 years)/high (≥11 years)], % — 41.0/30.3/28.7 45.9/30.6/23.5 41.7/31.1/27.2 40.1/27.2/32.7 
  Level of urbanisation      
   <500/500 to 1,000/1,000 to1,500/1,500 to 2,500/>2,500 no. addresses/km2, % 22.3/20.9/13.9/18.9/24.1 21.2/25.9/20.0/12.9/20.0 23.1/21.3/13.5/18.6/23.5 18.5/21.1/17.7/17.7/25.0 
  Pets (cats and/or dogs) in household, % — 16.6 16.5 15.9 19.4 
  Special adjustments in house, % 27.2 32.9 26.1 33.6* 
 Lifestyle risk factors 
  Alcohol drinkers, % 76.2 68.2 75.0 77.9 
  Current smoker, % 18.5 18.8 18.8 17.3 
  Body mass index, kg/m2 13 26.9 ± 4.3 27.9 ± 4.2* 27.1 ± 4.4 26.4 ± 3.9* 
Potential mediators 
 Co-morbidities and medication 
  Cardiovascular diseases, % 34.3 36.5 33.8 36.8 
  Cancer, % 12.0 11.8 12.4 10.4 
  Chronic obstructive pulmonary disease/asthma, % 14.8 11.8 15.1 12.6 
  Arthritis (rheumatoid/osteoarthritis), % 43.5 49.4 42.3 50.6* 
  Chronic diseases (excluding DM), no. 1 (0–2) 1 (0–2) 1 (0–2) 1 (1–2) 
  Psychotropic drugs, % 14.4 20.0 12.7 23.4*** 
  Blood pressure-lowering drugs, % 31.8 48.2** 33.6 30.7 
  Lipid-lowering drugs, % 4.6 4.7 5.0 3.0 
  Glucose-lowering drugs, % — 88.2 5.7 10.0* 
   Insulin,/Oral/Both, % — 40.0/67.1/14.1 2.6/4.5/1.1 4.3/6.9/0.9 
  Medicines, no. 2 (0–3) 3 (2–5)*** 2 (0–3) 2 (1–4)* 
  Polypharmacy (≥4 drugs), % 22.7 48.2*** 23.2 29.9* 
 Physical impairment and general health      
  Orthostatic hypotension, % 46 34.4 41.0 33.9 38.5 
  Regular dizziness, % 13.6 15.5 12.2 20.9*** 
  Foot problems, % 27.3 22.4 26.6 28.1 
  Incontinence, % 21.7 28.2 20.7 28.1* 
  Visual impairment (poor/insufficient/good), % 15.9/25.7/88.5 21.2/24.7/54.1 14.8/25.1/60.1 22.1/27.7/50.2** 
  Hearing impairment (poor/insufficient/good), % 10.3/7.7/82.0 20.0/8.2/71.8** 10.3/7.3/82.4 13.9/9.1/77.0 
  Subjective body pain score (range 0–6) 94 0 (0–2) 1 (0–2)** 0 (0–1) 1 (0–1)*** 
   Any pain, %  31.9 46.2** 31.4 38.8* 
  Self-rated health (range 1–5)a 2 (2–3) 3 (2–3)*** 2 (2–3) 3 (2–3)* 
  Hospital admission past 6 months, % – 8.9 12.9 8.0 13.8** 
  Occasional fall in previous year, % – 19.8 17.6 16.9 30.6*** 
 Activity, muscle strength and mobility 
  Physical activity, min/d 41 130 (76–201) 122 (30–180)* 134 (75–203) 118 (70–181)* 
  Grip strength, kg 48 57.5 ± 20.4 50.3 ± 19.7** 57.9 ± 19.9 53.1 ± 22.1** 
  Functional limitations in activities of daily living (ADLs) (range 0–6) 10 1 (0–2) 2 (1–4)*** 1 (0–2) 1 (0–3)** 
  Physical performance (range 0–12)b 43 8 (6–10) 6 (2–9)*** 8 (6–10) 7 (4–9)*** 
Psychosocial functioning      
  Cognitive impairment score (Mini-Mental Scale Examination, MMSE) (range 0–30) 28 (26–29) 27 (24–28)** 28 (26–29) 27 (25–29) 
   MMSE ≤23, %  10.7 21.4** 10.6 14.7 
  Depression score (Center for Epidemiological Studies Depression Scale—CES-D) (range 0–60) 30 6 (2–11) 6 (4–11) 5 (2–10) 8 (3–13)*** 
   CES-D≥16, %  14.5 8.4 13.5 16.5 
  Fear of falling (range 0–30) 1 (0–3) 1 (0–4) 0 (0–3) 1 (0–3)*** 
  Loneliness (range 0–11) 2 (0–4) 2 (0–4) 1 (0–3) 2 (0–4)** 
  Living alone, % – 37.5 49.4* 37.3 42.2 
 Biomarkers 
  25-Hydroxyvitamin D, nmol/l 111 55.4 ± 24.0 48.4.5 ± 20.2* 54.7 ± 23.3 55.7 ± 25.6 
  Sex hormone-binding globulin, nmol/l 110 43.8 (32.5–58.7) 41.3 (27.0–55.7) 42.5 (31.4–57.9) 46.3 (34.4–60.6)* 
  Insulin-like growth factor-1, nmol/l 112 14.0 ± 5.2 13.5 ± 5.9 14.0 ± 5.3 13.5 ± 5.2 
  Osteocalcin, nmol/l 112 2.2 ± 1.1 2.1 ± 1.1 2.1 ± 1.1 2.2 ± 1.1 
  Parathyroid hormone, pmol/l 111 3.1 (2.4–4.2) 3.3 (2.4–4.4) 3.1 (2.4–4.2) 3.1 (2.4–4.1) 
  Interleukin-6, pg/ml 135 1.9 (1.1–3.1) 1.9 (1.4–3.3) 1.9 (1.1–3.1) 1.8 (0.9–3.1) 
  C-reactive protein, mg/l 134 2.9 (1.4–6.4) 4.0 (2.5–8.2)** 3.2 (1.5–6.8) 2.8 (1.2–5.9) 
  Albumin, g/l 109 44.5 ± 2.7 44.3 ± 2.8 44.6 ± 2.8 44.3 ± 2.6 
  Creatinine, µmol/l 108 90 (79–103) 89 (81–108) 90 (79–103) 93 (80–106) 
  Glomerular filtration rate, ml/min per 1.73 m2 108 58.0 ± 11.9 56.2 ± 12.5 58.2 ± 12.0 56.6 ± 11.9 

Data are mean ± standard deviation, median (inter-quartile range) or frequency (%).

Higher scores indicate: apoorer self-rated health; bbetter performance; cmore fear *< 0.05; **< 0.01; ***< 0.001.

Statistical analyses

Incidence rates of recurrent falls were estimated with Poisson regression analyses. Kaplan–Meier curves were created to illustrate and compare, with log-rank tests, the cumulative incidence of recurrent falls between participants with versus without diabetes. Baseline characteristics were compared between participants with and without diabetes and between recurrent and non-recurrent fallers with Student's t-tests, Mann–Whitney U tests and χ2 tests.

We used multiple imputation by chained equations to impute missing data on study covariates (2.5% of all required values), to boost power and enable more efficient analyses [22] (for details see Supplementary data available in Age and Ageing online, Appendix 1).

Cox-proportional hazard regression models were used to account for the effects of confounders and potential mediators in the associations between diabetes and incident recurrent falls. After a first unadjusted analysis (model 1), successive models were created to examine: first, the extent to which this association was confounded by as age, sex, other socio-demographic factors, BMI, smoking, alcohol consumption (model 2); and second, explore potential pathways that may explain the association (subsequent models). Potential mediators were retained in the models if the strength of the regression coefficient (Ln-HR) reflecting the association between diabetes and recurrent falls was attenuated by ≥5%.

All analyses were carried out with Predictive Analytics SoftWare (PASW) Statistics software version 18.0 (SPSS, Inc., Chicago, IL, USA).

Results

Comparisons of participants' characteristics at baseline according to their diabetes or recurrent falls status are shown in Table 1. Eighty-five participants (7.4%) had an adjudicated diagnosis of diabetes (6.4% in men and 8.4% in women, P = 0.201). During a mean follow-up duration of 139 weeks, 232 participants (20.3%) fell recurrently. Non-recurrent fallers (n = 913) included 366 individuals who were classified as ‘occasional fallers’, but the incidence of occasional falls did not significantly differ by diabetes status (30.5 versus 40.7%, P = 0.132). In contrast, the incidence of recurrent falls was greater among individuals with than without diabetes (30.6 versus 19.4%, P = 0.017). The incidence rate of recurrent falls (per 1,000 persons-years) was 129.7 and 77.4 for individuals with and without diabetes, respectively: incidence rate ratio (IRR) = 1.67 (95% CI: 1.11–2.51) (Figure 1). These estimates did not differ between individuals with diabetes who were or not treated with insulin [122.2 versus 134.8, respectively, IRR = 0.91 (0.41–2.00)].

Figure 1.

Kaplan–Meier curves for recurrent falls during 3-years of follow-up by diabetes status.

Figure 1.

Kaplan–Meier curves for recurrent falls during 3-years of follow-up by diabetes status.

The risk for recurrent falls remained higher among individuals with diabetes even after adjustments for potential confounders: HR = 1.63 (1.06–2.52) (Table 2, model 2). These estimates were slightly higher in women [HR = 1.82 (1.06–3.13)] than in men [HR = 1.41 (0.73–2.72)], albeit not significantly so (P-interaction = 0.577). Also, age did not modify the association between diabetes and incident falls (P-interaction = 0.662). Factors that explained, in part, the increased incidence of recurrent falls associated with diabetes were the greater number of medicines used; higher levels of body pain; poorer self-perceived health; lower levels of physical activity and grip strength; more limitations in ADLs; lower physical performance (including gait and balance), and cognitive impairment (models 3–6). Altogether, these variables accounted for ∼47% of the increased risk associated with diabetes (model 7).

Table 2.

Associations between diabetes and incident risk of recurrent falls and the role of selected mediators herein

Model Adjustments HR 95% CI P-value Coefficient (Ln-HR) Change in coefficient versus model 2 (%) 
None 1.67 1.11–2.51 0.014   
Model 1 + potential confounders    — — 
 Age, sex, education level, level of urbanisation, dogs and cats in household, special adjustments in house, body mass index, alcohol and smoking status 1.63 1.06–2.52 0.020 0.491 — 
Model 1 + medication      
3a + Number of medicines 1.50 1.00–2.27 0.058 0.408 17 
Model 1 + physical impairments and general health      
4a + Pain score 1.54 1.01–2.34 0.044 0.431 12 
4b + Self-perceived health 1.55 1.05–2.31 0.039 0.441 10 
 + all variables included in models 4a–b 1.51 0.99–2.29 0.056 0.410 17 
Model 1 + activity and mobility      
5a + Physical activity 1.59 1.05–2.41 0.030 0.461 
5b + Grip strength 1.58 1.05–2.40 0.031 0.459 
5c + Functional limitations in ADLs 1.51 0.99–2.30 0.054 0.412 16 
5d + Physical performance 1.45 0.95–2.21 0.084 0.371 24 
 + All variables included in models 5a–d 1.41 0.93–2.16 0.107 0.344 30 
Model 1 + Psychosocial functioning      
 + MMSE score≤23 1.59 1.05–2.42 0.029 0.465 
Fully adjusted modela 1.30 0.79–2.11 0.300 0.259 47 
Model Adjustments HR 95% CI P-value Coefficient (Ln-HR) Change in coefficient versus model 2 (%) 
None 1.67 1.11–2.51 0.014   
Model 1 + potential confounders    — — 
 Age, sex, education level, level of urbanisation, dogs and cats in household, special adjustments in house, body mass index, alcohol and smoking status 1.63 1.06–2.52 0.020 0.491 — 
Model 1 + medication      
3a + Number of medicines 1.50 1.00–2.27 0.058 0.408 17 
Model 1 + physical impairments and general health      
4a + Pain score 1.54 1.01–2.34 0.044 0.431 12 
4b + Self-perceived health 1.55 1.05–2.31 0.039 0.441 10 
 + all variables included in models 4a–b 1.51 0.99–2.29 0.056 0.410 17 
Model 1 + activity and mobility      
5a + Physical activity 1.59 1.05–2.41 0.030 0.461 
5b + Grip strength 1.58 1.05–2.40 0.031 0.459 
5c + Functional limitations in ADLs 1.51 0.99–2.30 0.054 0.412 16 
5d + Physical performance 1.45 0.95–2.21 0.084 0.371 24 
 + All variables included in models 5a–d 1.41 0.93–2.16 0.107 0.344 30 
Model 1 + Psychosocial functioning      
 + MMSE score≤23 1.59 1.05–2.42 0.029 0.465 
Fully adjusted modela 1.30 0.79–2.11 0.300 0.259 47 

aModel adjusted for all variables included in models 2–6; ADLs, activities of daily living; MMSE, Mini-Mental Scale Examination.

For additional information on the lack of a mediating role of the other variables examined, and the lack of impact of loss to follow-up on the estimates and inferences arising from our analyses please see Supplementary data available in Age and Ageing online, Appendix 2).

Discussion

We examined the association between diabetes and incident falls among community-dwelling older individuals and found that individuals with diabetes had a 67% higher risk of recurrent falls than their counterparts without diabetes.

In contrast to previous studies [6, 7], we did not find that the incidence of recurrent falls was particularly higher among the individuals with diabetes who were treated with insulin. Instead, and in line with previous observations [12] polypharmacy, not the use of any specific medication per se, explained in part the increased risk of recurrent falls in individuals with diabetes. [12]. These observations may have important clinical implications. Diabetes care places a great focus on the prevention of micro- and macrovascular complications, which are often targeted following multidrug regimens in order to control hyperglycaemia and commonly coexisting morbidities such as hypertension, dyslipidaemia and renal dysfunction. In order to maximise benefits to patients, standardisation among health-care providers is encouraged through disease-specific treatment guidelines [23]. However, the long-term benefits and harms associated with combination of drugs that are taken by individuals with comorbidities is less clear [24]. In addition, given that levels of clinical control are often sub-optimal, intensification of the overall medication prescribed to individuals with diabetes has been advocated. However, an increased risk of mortality associated with the more intensive use of glucose-lowering drugs [25] has raised questions about the wisdom of this approach. The potential to increase the incidence of recurrent falls illustrates the complexity of polypharmacy management of older individuals with diabetes.

High levels of chronic musculoskeletal pain are common in the older, particularly those with chronic diseases such as diabetes [26, 27]. Our findings suggest that pain may be an important factor to consider in the prevention of diabetes-related falls. The presence of chronic pain may limit the ability of individuals to adhere to self-management practices such as regular exercise [26], and is also associated with poorer mental health and physical functioning [27], and thus places an individual with diabetes at greater risk for recurrent falls [28]. Indeed, among all the potential mediators investigated, poorer levels of physical activity and performance (including gait and balance) seemed to contribute to the increased risk of recurrent falls attributable to diabetes. These findings are in line with several studies showing accelerated loss of muscle strength, functional limitations in ADLs and impairment in physical performance related to diabetes [29–33]. Importantly, they may be amenable to modification through exercise programmes such as resistance/balance training [34] thereby leading to decreases in falls [35]. In addition, although exercise may not be an appropriate pain management strategy for every individual, it may reduce pain and improve function, in contrast to the more widely and potentially counterproductive adopted strategies of rest and inactivity [36]. However, the benefits of exercise as a means to prevent falls among older individuals with diabetes may be hampered in the presence of (severe) cognitive impairment [37]. Indeed, diabetes is a major risk factor for cognitive impairment and dementia [38] and, in our study, signs of cognitive impairment explained a small part of the increased falls associated with diabetes as well.

Strengths and unique features of this study are its frequent assessment of incident falls over the course of a long-enough period to capture the specific phenotype of interest—recurrent falls; the representative population, including both men and women, for the community-dwelling older population of the Netherlands; and the assessment of a large set of potential confounders and mediators of the associations between diabetes and recurrent falls. Despite the comprehensive approach adopted, about half of the diabetes-related increased incidence of recurrent falls remained unexplained. This may have been due to imprecision in identifying relevant factors that mediate an increased risk of falling, such as the presence of peripheral neuropathy [16]. However, peripheral neuropathy may have been, at least in part, captured by individuals' self-reported levels of pain and/or foot problems, and the measured levels of participants' walking speed, gait and balance (included in the physical performance score), most of which were indeed significantly higher or impaired among the individuals with diabetes. Other potential limitations may include, the fact that we examined (recurrent) falls but not injurious falls specifically.

In addition, although for most participants the diagnosis of diabetes was retrieved from GPs' records, in some (∼27%), these were missing. The diagnosis was therefore based on a combination of data obtained form self-reports and review of glucose-lowering medication. In a previous study, the self-reports of diabetes have shown a good agreement with GPs' records [39]. Still, the observed 7.4% prevalence of diabetes is likely an underestimate since the prevalence of reported diabetes is usually higher in those aged over 65 and diabetes is not infrequently undiagnosed in up to 50% of cases if screening using blood glucose tests is not performed [40]. However, if such an underestimate is present, then the strength of the association between diabetes and recurrent falls reported in our study is likely underestimated as well. The reverse is possible if non-diagnosed diabetic subjects fall less than those diagnosed with diabetes. We consider that, however, unlikely as in the diabetes group other risk factors for recurrent falling were more prevalent as well.

Duration of diabetes and measures of HbA1c were also missing and, therefore, we could not ascertain the extent of disease severity/glycaemic control among the individuals with diabetes. Finally, the study population consisted of relatively healthy community-dwelling white persons over 65 and our results may thus not apply to those who are institutionalised or belong to other ethnicities.

In conclusion, older community-dwelling individuals with diabetes have higher risk for recurrent falls than their peers without diabetes. Factors partly explaining this increased risk include diabetes-related polypharmacy, musculoskeletal pain and poorer self-perceived health, lower levels of physical activity and muscle strength, limitations in ADLs, impaired physical performance (including gait and balance) and cognitive impairment. In view of the current ageing of the population and the increase in the incidence of diabetes, particularly among older people, our findings support the view that fall prevention efforts targeting the factors identified above may need to be incorporated into the care and treatment of older individuals with diabetes.

Key points

  • Older community-dwelling individuals with diabetes have higher risk for recurrent falls than their peers without diabetes.

  • Factors explaining this increased risk need to be comprehensively identified.

  • Among others, polypharmacy, pain, impaired ADLs and physical activity, strength and mobility were critical factors identified.

  • Fall prevention efforts targeting the factors identified may need to be incorporated into the care of older diabetic individuals.

Conflict of interest

None declared.

Funding

This study is based on data collected in the context of the Longitudinal Aging Study Amsterdam, which has been largely funded by the Ministry of Health, Welfare and Sports of The Netherlands. The sponsors did not have any role in the design, execution, analysis and interpretation of data or writing of the study.

Supplementary data

Supplementary data mentioned in the text is available to subscribers in Age and Ageing online.

References

The long list of references supporting this study has meant that only the most important are listed here and are represented by bold type throughout the text. The full list of references is available on Supplementary data are available in Age and Ageing online, Appendix 3.

1
Tinetti
ME
Kumar
C
The patient who falls: it's always a trade-off
JAMA
 , 
2010
, vol. 
303
 (pg. 
258
-
66
)
3
Cumming
RG
Kelsey
JL
Nevitt
MC
Methodologic issues in the study of frequent and recurrent health problems. Falls in the elderly
Ann Epidemiol
 , 
1990
, vol. 
1
 (pg. 
49
-
56
)
4
Miller
DK
Lui
LY
Perry
HM
III
Kaiser
FE
Morley
JE
Reported and measured physical functioning in older inner-city diabetic African Americans
J Gerontol A Biol Sci Med Sci
 , 
1999
, vol. 
54
 (pg. 
M230
-
6
)
5
Hanlon
JT
Landerman
LR
Fillenbaum
GG
Studenski
S
Falls in African American and white community-dwelling elderly residents
J Gerontol A Biol Sci Med Sci
 , 
2002
, vol. 
57
 (pg. 
M473
-
8
)
6
Schwartz
AV
Hillier
TA
Sellmeyer
DE
, et al.  . 
Older women with diabetes have a higher risk of falls: a prospective study
Diabetes Care
 , 
2002
, vol. 
25
 (pg. 
1749
-
54
)
7
Volpato
S
Leveille
SG
Blaum
C
Fried
LP
Guralnik
JM
Risk factors for falls in older disabled women with diabetes: the women's health and aging study
J Gerontol A Biol Sci Med Sci
 , 
2005
, vol. 
60
 (pg. 
1539
-
45
)
8
Quandt
SA
Stafford
JM
Bell
RA
Smith
SL
Snively
BM
Arcury
TA
Predictors of falls in a multiethnic population of older rural adults with diabetes
J Gerontol A Biol Sci Med Sci
 , 
2006
, vol. 
61
 (pg. 
394
-
8
)
9
Gregg
EW
Beckles
GL
Williamson
DF
, et al.  . 
Diabetes and physical disability among older U.S. adults
Diabetes Care
 , 
2000
, vol. 
23
 (pg. 
1272
-
7
)
12
Huang
ES
Karter
AJ
Danielson
KK
Warton
EM
Ahmed
AT
The association between the number of prescription medications and incident falls in a multi-ethnic population of adult type-2 diabetes patients: the diabetes and aging study
J Gen Intern Med
 , 
2010
, vol. 
25
 (pg. 
141
-
6
)
13
Volpato
S
Maraldi
C
Fellin
R
Type 2 diabetes and risk for functional decline and disability in older persons
Curr Diabetes Rev
 , 
2010
, vol. 
6
 (pg. 
134
-
43
)
14
Lu
FP
Lin
KP
Kuo
HK
Diabetes and the risk of multi-system aging phenotypes: a systematic review and meta-analysis
PLoS One
 , 
2009
, vol. 
4
 pg. 
e4144
 
15
Ganz
DA
Bao
Y
Shekelle
PG
Rubenstein
LZ
Will my patient fall?
JAMA
 , 
2007
, vol. 
297
 (pg. 
77
-
86
)
16
Mayne
D
Stout
NR
Aspray
TJ
Diabetes, falls and fractures
Age Ageing
 , 
2010
, vol. 
39
 (pg. 
522
-
5
)
17
Deeg
DJ
van Tilburg
T
Smit
JH
de Leeuw
ED
Attrition in the Longitudinal Aging Study Amsterdam. The effect of differential inclusion in side studies
J Clin Epidemiol
 , 
2002
, vol. 
55
 (pg. 
319
-
28
)
18
Pluijm
SM
Smit
JH
Tromp
EA
, et al.  . 
A risk profile for identifying community-dwelling elderly with a high risk of recurrent falling: results of a 3-year prospective study
Osteoporos Int
 , 
2006
, vol. 
17
 (pg. 
417
-
25
)
19
Peeters
GM
van Schoor
NM
Pluijm
SM
Deeg
DJ
Lips
P
Is there a U-shaped association between physical activity and falling in older persons?
Osteoporos Int
 , 
2010
, vol. 
21
 (pg. 
1189
-
95
)
20
Kellogg International Work Group on the prevention of falls by the elderly
The prevention of falls in later life
Dan Med Bull
 , 
1987
, vol. 
34
 (pg. 
1
-
24
)
24
Tinetti
ME
Bogardus
ST
Jr
Agostini
JV
Potential pitfalls of disease-specific guidelines for patients with multiple conditions
N Engl J Med
 , 
2004
, vol. 
351
 (pg. 
2870
-
4
)
26
Krein
SL
Heisler
M
Piette
JD
Makki
F
Kerr
EA
The effect of chronic pain on diabetes patients' self-management
Diabetes Care
 , 
2005
, vol. 
28
 (pg. 
65
-
70
)
27
Bair
MJ
Brizendine
EJ
Ackermann
RT
Shen
C
Kroenke
K
Marrero
DG
Prevalence of pain and association with quality of life, depression and glycaemic control in patients with diabetes
Diabet Med
 , 
2010
, vol. 
27
 (pg. 
578
-
84
)
28
Leveille
SG
Bean
J
Bandeen-Roche
K
Jones
R
Hochberg
M
Guralnik
JM
Musculoskeletal pain and risk for falls in older disabled women living in the community
J Am Geriatr Soc
 , 
2002
, vol. 
50
 (pg. 
671
-
8
)
29
Volpato
S
Ferrucci
L
Blaum
C
, et al.  . 
Progression of lower-extremity disability in older women with diabetes: the Women's Health and Aging Study
Diabetes Care
 , 
2003
, vol. 
26
 (pg. 
70
-
5
)
30
Strotmeyer
ES
de Rekeneire
N
Schwartz
AV
, et al.  . 
The relationship of reduced peripheral nerve function and diabetes with physical performance in older white and black adults: the Health, Aging, and Body Composition (Health ABC) study
Diabetes Care
 , 
2008
, vol. 
31
 (pg. 
1767
-
72
)
31
De Rekeneire
N
Resnick
HE
Schwartz
AV
, et al.  . 
Diabetes is associated with subclinical functional limitation in nondisabled older individuals: the Health, Aging, and Body Composition study
Diabetes Care
 , 
2003
, vol. 
26
 (pg. 
3257
-
63
)
32
Park
SW
Goodpaster
BH
Strotmeyer
ES
, et al.  . 
Accelerated loss of skeletal muscle strength in older adults with type 2 diabetes: the health, aging, and body composition study
Diabetes Care
 , 
2007
, vol. 
30
 (pg. 
1507
-
12
)
33
Gregg
EW
Mangione
CM
Cauley
JA
, et al.  . 
Diabetes and incidence of functional disability in older women
Diabetes Care
 , 
2002
, vol. 
25
 (pg. 
61
-
7
)
34
Allet
L
Armand
S
de Bie
RA
, et al.  . 
The gait and balance of patients with diabetes can be improved: a randomised controlled trial
Diabetologia
 , 
2010
, vol. 
53
 (pg. 
458
-
66
)
35
Morrison
S
Colberg
SR
Mariano
M
Parson
HK
Vinik
AI
Balance training reduces falls risk in older individuals with type 2 diabetes
Diabetes Care
 , 
2010
, vol. 
33
 (pg. 
748
-
50
)
36
Butchart
A
Kerr
EA
Heisler
M
Piette
JD
Krein
SL
Experience and management of chronic pain among patients with other complex chronic conditions
Clin J Pain
 , 
2009
, vol. 
25
 (pg. 
293
-
8
)
37
Hauer
K
Becker
C
Lindemann
U
Beyer
N
Effectiveness of physical training on motor performance and fall prevention in cognitively impaired older persons: a systematic review
Am J Phys Med Rehabil
 , 
2006
, vol. 
85
 (pg. 
847
-
57
)

Supplementary data

Comments

0 Comments