Risk factors for falling in community-dwelling older adults: Which of them are associated with the recurrence of falls?

Abstract

Background

The aim of the study was to determine which risk factors were associated with the recurrence of falls among community-dwelling older adults.

Methods

Based on a cross-sectional design, 1066 community-dwelling volunteers aged 65 and older were recruited. The use of psychoactive drugs, the number of drugs taken per day, the basic mobility assessed with the Timed Up & Go test (TUG), the maximal isometric voluntary contraction strength of hand, the lower limb proprioception, the distance binocular vision, the fear of falling and the history of falls during the past year were recorded. Subjects were separated into 4 groups based on the number of falls: 0, 1, 2 and > 2 falls.

Results

Among the 395 (37.1%) fallers, 291 (27.3%) were single fallers and 104 (9.8%) were recurrent fallers (i.e., > 2 falls). The numbers of falls increased significantly with age (Incident Rate Ratio (IRR)=1.03, p<0.001), female gender (IRR=1.95, p<0.001), institutionalization (IRR=1.66, p=0.002), number of drugs taken per day (IRR=1.05, p<0.001), use of psychoactive drugs (IRR=1.29, p=0.009), increased time of TUG Test (IRR=1.02, p<0.001), use of a walking aid (IRR=1.59, p=0.002), and fear of falling (IRR=3.08, p<0.001). In addition, a high score at the handgrip test (IRR=0.97, p<0.001) and distance binocular vision (IRR=0.92, p<0.001) were associated with a decreased number of falls. After adjustment for potential confounders, only female gender (IRR=1.44, p<0.001), vision (IRR=0.95, p=0.006) and lower limb proprioception (IRR=0.95, p=0.046), and fear of falling (IRR=2.68, p<0.001) were still significantly associated with the number of falls.

Conclusion

The current study shows that female gender, poor vision and lower limb proprioception, and fear of falling were associated with the recurrence of falls.

Introduction

Recurrent falls were usually defined as two or more falls in a 12-month period (1). Around 25% of adults aged over 80 years suffers at least two falls per year (2, 3, 4). Although recurrent falls cause greater morbid events such as injuries, fear of falling or a loss of independence, their mechanism has been few studied compared to singe fall (1, 2, 3, 4). For instance, Nevitt et al. showed in a cohort of 325 elderly community-dwelling persons that the risk of major injury per fall was higher in persons having a previous fall (5). In addition, it has been shown that subjects who reported fear of falling after a first fall experienced a higher in balance and gait impairment over time, resulting in a decrease in mobility level compared to non-recurrent fallers (6). Because of these fall-related adverse outcomes which impose high costs to public health and social services (4,7), the prevention of recurrent falls is mandatory.

The efficiency and cost-effectiveness of fall prevention strategies require identifying older adults at high risk of falls (3, 4, 7). The fall is seldom due to a single cause (2, 3). The association between risk factors for falls is complex and probably contributes to falls in different ways in single and recurrent fallers. More than 400 risk factors for falls have been identified (3, 4, 7, 8, 9). Most of them are intrinsic (i.e., subject-related) and usually include polymedication, psychoactive drugs, muscle weakness, sensory deficits, the use of a walking aid, a cognitive decline and the fear of falling (1, 2, 3, 4, 5, 6, 7, 8, 9). To date, few studies have specifically examined the risk factors for recurrent falls (1, 4, 10, 11). The recent guideline of management of recurrent falls in older adults by the French Geriatrics and Gerontology Society (FGGS) highlighted that only around 5% of published manuscripts in the field of falls focused on their recurrence (12). These prior studies showed that sensory deficits such as vision or proprioception impairment seem to be particularly related to the recurrence of falls (1, 3, 4, 11).

Despite the great merit in identifying the risk of the first fall, the recurrence of falls appears to be a greater primary outcome in clinical routine as shown in the last Guideline of the FGGS because of fall-related adverse outcomes (12). Complex interaction among risk factors may lead to a vicious cycle, with an increased risk of falls and probably more recurrent falls. The majority of previous models of fall risk did not provide any information about the risk factors associated with the recurrence of falls (1, 2, 3, 4, 5, 6, 7, 8, 9). Finding out more about the factors associated to recurrent falls could improve the development of specific models of the risk of recurrent falls and related adverse outcomes. The aim of the study was to determine which risk factors were associated with the recurrence of falls among community-dwelling older adults.

Methods

Study population

Between September and October 2005, 1066 community-dwelling volunteers aged 65 and older were recruited in 8 medical health institutes localized in Eastern France during a free medical examination. The exclusion criteria were as follows: acute medical illness in the past 3 months; neurological disease such as Parkinson's disease, cerebellar disease, myelopathy, peripheral neuropathy; major orthopaedic diagnoses involving the lumber vertebra, pelvis or lower extremities.

Assessment

The use of psychoactive drugs including benzodiazepines, antidepressants or neuroleptics, and the number of drugs taken per day were recorded. Basic mobility was assessed with the Timed Up & Go test (TUG) (13). The maximal isometric voluntary contraction (MVC) strength of hand was measured with computerized hydraulic dynamometers (Martin Vigorimeter, Medizin Tecnik, Tutlingen, Germany). The test was performed one time on each side. The highest MVC value recorded was used in the present data analysis. Cognitive function was assessed using the patient section score of the General Practitioner Assessment of Cognition (GPcog) (14). This test was composed of 9 spatio-temporal questions, which were equally weighted and summed to calculate a global score ranged from 0 to 9, with 0 as the worst cognitive situation. Lower limb proprioception was evaluated with a graduated diapason placed on the tibial tuberosity. The mean value obtained for the left and right sides was used in the present data analysis. Distance binocular vision was measured at 5 m with a Snellen letter test chart (15). Vision was assessed with corrective lenses on if used by the subject. The subjects were interviewed using a standardized questionnaire, gathering information on the history of falls over the past year. This face-to-face interview was based on 22 standardized questions exploring the number, delay and place of falls (i.e., inside or outside the subject's house), the evoked causes and circumstances of falls (i.e., syncope or other acute medical event, body transfer from sit position or walking or other physical activities such as cycling), and all physical traumatisms and inability to get up from ground after a fall. A fall was defined as an event resulting in a person coming to rest unintentionally on the ground or at other lower level, not as the result of a major intrinsic event or an overwhelming hazard. Thus, falls resulting from acute medical events and/or external force were excluded from the analysis. The fear of falling was also recorded using the question “Are you afraid of falling?” (16). The study was conducted in accordance with the ethical standards set forth in the Helsinki Declaration (1983). The local ethics committee approved the project.

Statistical analysis

The subjects' baseline characteristics were summarized using means and standard deviations or frequencies and percentages, as appropriate. The normality of the parameters' distribution was verified with skewness and kurtosis tests before and after applying the usual transformations to normalize non-Gaussian variables. Subjects were separated into 4 groups based on the number of falls: 0, 1, 2 and > 3 falls. Firstly, comparisons between groups were performed using independent samples t-test and Chi-square test for trend, as appropriate. Secondly, unadjusted ordered logistic regression was performed to characterize the effect (expressed as ordered odds ratio [ordered OR]) of each baseline characteristic on experiencing an additional fall. Thirdly, unadjusted and full adjusted Poisson regression analyses were performed to specify the association (expressed as incidence rate ratio, IRR) between baseline characteristics and the number of falls, taking into account the diminishing prevalence as the number of falls increases. P-values lower than 0.05 were considered as statistically significant. All statistics were performed using the Stata Statistical Software release 10.0 (17).

Results

As shown in Table 1, 395 (37.1%) participants were fallers. Two hundred ninety one (27.3%) were single fallers and 104 (9.8%) were recurrent fallers (i.e., > 2 falls). The number of falls per subject increased with age (P-trend=0.001, ordered OR=1.05 [1.02-1.08]), female gender (P-trend<0.001, ordered OR=2.50 [1.92-3.25]), institutionalization (P-trend=0.004, ordered OR= 2.38 [1.41-4.00]), the number of drugs taken per day (from 3.5±3.2 among non-fallers to 4.8±3.1 among subjects who reported >2 falls with P-trend<0.001, ordered OR=1.08 [1.04-1.12]), the use of psychoactive drugs (from 17.4% among non-fallers to 18.6% among subjects who reported >2 falls with P-trend=0.001, ordered OR=1.59 [1.20-2.12]), a high score at TUG (P-trend=0.003, ordered OR=1.04 [1.01-1.06]), the use of a walking aid (P-trend=0.003, ordered OR=2.06 [1.27-3.32]) and the fear of falling (P-trend<0.001, ordered OR=5.06 [3.89-6.58]). In addition, a low score at the handgrip strength (P-trend<0.001, ordered OR=0.96 [0.95-0.97]) and distance binocular vision (P-trend<0.001, ordered OR=0.90 [0.86-0.95]) were associated with an increased number of falls. There were no significant differences between groups for the other characteristics.

Table 1.

Comparison of subject baseline characteristics according to history of falls (n=1066)

		Fallers
Characteristics	Non-fallers (n=671)	(n=291)	2 (n= 61)	>2 (n=43)	P-trend	Ordered OR	[95%CI]	P-Value
Age (years), mean ± SD	74.7±4.1	75.4±4.3	76.1±3.8	75.9±4.6	<0.001	1.05	[1.02-1.08]	0.001
Female, n (%)	336 (50.1)	209 (71.8)	38 (62.3)	37 (86.1)	<0.001	2.50	[1.92-3.25]	<0.001
Institutionalized†, n (%)	20 (3.0)	22 (7.6)	5 (8.2)	3 (7.0)	0.004	2.38	[1.41-4.00]	0.001
Body mass index (kg/m2), mean ± SD	26.3±3.8	26.6±4.4	26 8±4.00	27.3±6.3	0.476	1.03	[0.99-1.06]	0.100
Number of drugs/day, mean ± SD	3.5±3.2	4.3±3.2	4.4±3.4	4.8±3.1	<0.001	1.08	[1.04-1.12]	<0.001
Use of psychoactive drugs‡, n(%)	117 (17.4)	79 (27.2)	17 (27.9)	8 (18.6)	0.001	1.59	[1.20-2.12]	0.001
GPcog§ Score (/9), mean ±SD	7.0±2.0	6.8±2.1	7.0±1.7	6 2±2 3	0.084	0.96	[0.90-1.102]	0.175
Timed “Up & Go” (seconds), mean ±SD	11.1±6.3	12.0±5.9	12.6±5.4	14.0±6.0	<0.001	1.04	[1.01-1.06]	0.003
Handgrip strength||(N.m-2), mean±SD	29.5±12.1	24.6±104	26.3±10.6	21.5±7.3	<0.001	0.96	[0.95-0.97]	<0.001
Vision¶ (/10), mean ±SD	6.9±2.3	6.6±2.5	64±2.2	5.5±2.3	<0.001	0.90	[0.86-0.95]	<0.001
Lower limb proprioception# (/8), mean ± SD	5.8±1.5	5.9±1.5	5.2±2.0	5.3±1.8	0.105	0.95	[0.88-1.02]	0.160
Use of walking aid∗∗, n (%)	30 (4.5)	22 (7.6)	8 (13.1)	4 (9.3)	0.003	2.06	[1.27-3.32]	0.003
Fearoffalling††, n(%)	148 (22.1)	153 (52.6)	41 (67.2)	35 (81.4)	<0.001	5.06	[3.89-6.58]	<0.001

SD: standard deviation; OR: odds ratio; CI: confidence interval; †: Living in senior housing facilities or home care; ‡: Use of benzodiazepines or antidepressants or neuroleptics; §: Patient section of General practitioner assessment of cognition; ||: Mean value of the highest value of maximal isometric voluntary contraction strength measured with computerized dynamometers expressed in Newton per square meter; ¶: Binocular vision acuity at distance of 5 m with a Snellen letter test chart; #: Mean value of left and right side and based on graduated diapason placed on the tibial tuberosity; ∗∗: Use a walking stick; ††: Self-reported with the question “Are you afraid of falling?”; P significant (i.e., <0.05) indicated in bold

Table 2 displays the results of unadjusted and full adjusted Poisson regressions. It shows that age, female gender, institutionalization, a high number of drugs taken per day, the use of psychoactive drugs, increased time of TUG, muscle weakness, poor vision, poor lower limb proprioception, the use of a walking aid, and the fear of falling were separately associated with incident falls (P<0.020 for all variables). After adjustment for age and gender, a high BMI (IRR=1.02 with P=0.043), the number of drugs taken per day (IRR=1.03 with P=0.006), increased time of TUG (IRR=1.02 with P<0.001), the use of a walking aid (IRR=1.40 with P=0.028) and the fear of falling (IRR=2.73 with P<0.001) were associated significantly with the recurrence of falls, whereas a high muscle strength (IRR=0.99 with P=0.034), a high vision (IRR=0.94 with P=0.001) and lower limb proprioception (IRR=0.92 with P=0.004) scores were associated with a low number of falls. In addition, the stepwise backward model showed that female gender (IRR=1.44 with P<0.001), vision (IRR=0.95 with P=0.006) and lower limb proprioception (IRR=0.95 with P=0.046) scores, and the fear of falling (IRR=2.68 with P<0.001) were still significantly associated with the number of falls.

Table 2.

Risk estimates of recurrent falls based on simple and multiple Poisson regression models (n=1066)

	IRR	Model 1 95%CI	P-value	IRR	Model 2 95%CI	P-value	IRR	Model 3 95%CI	P-value
Age (years), mean ±SD	1.03	[1.02-1.06]	< 0.001	-	-	-
Female, n (%)	1.95	[1.61-2.36]	< 0.001	-	-	-	1.44	[1.18-1.77]	< 0.001
Institutionalized†, n (%)	1.66	[1.21-2.29]	0.002	1.24	[0.88-1.75]	0.218	1.29	[0.93-1.79]	0.125
Body mass index (kg/m2), mean ± SD	1.02	[1.00-1.04]	0.051	1.02	[1.00-1.04]	0.043
Number of drugs/day, mean ±SD	1.05	[1.03-1.07]	< 0.001	1.03	[1.01-1.06]	0.006
Use of psychoactive drugs‡, n (%)	1.29	[1.07-1.57]	0.009	1.11	[0.91-1.35]	0.320
GPcog§, mean ± SD	0.96	[0.93-1.00]	0.064	0.98	[0.94-1.02]	0.282
Timed “Up & Go”, mean ±SD	1.02	[1.01-1.02]	< 0.001	1.02	[1.01-1.03]	< 0.001
Handgrip strength||, mean ±SD	0.97	[0.96-0.98]	< 0.001	0.99	[0.97-1.00]	0.034
Vision¶, mean±SD	0.92	[0.89-0.95]	< 0.001	0.94	[0.91-0.98]	0.001	0.95	[0.92-0.99]	0.006
Lower limb proprioception#, mean ± SD	0.94	[0.89-0.99]	0.015	0.92	[0.88-0.98]	0.004	0.95	[0.90-0.99]	0.046
Use of walking aid∗∗, n (%)	1.59	[1.19-2.13]	0.002	1.40	[1.04-1.90]	0.028
Fearoffalling††, n(%)	3.08	[2.58-3.66]	< 0.001	2.73	[2.28-3.28]	< 0.001	2.68	[2.23-3.21]	< 0.001

Model 1: Crude IRR; Model 2: IRR adjusted for age and gender; Model 3: Stepwise backward model; SD: Standard deviation; IRR: incident rate ratio; CI: Confidence interval; †: Living in senior housing facilities or home care; ‡: Use of benzodiazepines or antidepressants or neuroleptics; §: Patient section of General practitioner assessment of cognition; ||: Mean value of the highest value of maximal isometric voluntary contraction strength measured with computerized dynamometers expressed in Newton per square meter; ¶: Binocular vision acuity at distance of 5 m with a Snellen letter test chart; #: Mean value of left and right side and based on graduated diapason placed on the tibial tuberosity; ∗∗: Use a walking stick; ††: Self-reported with the question “Are you afraid of falling?”; P significant (i.e., <0.05) indicated in bold

Discussion

Our findings showed that age, institutionalization, a high number of drugs taken per day, psychoactive drugs, abnormal basic mobility, and muscle weakness were separately related to fall but not to its recurrence, while female gender, poor vision and lower limb proprioception, and fear of falling were specific markers of the recurrence of falls.

Both aging and comorbitidies affect postural stability and increase the risk of falling (3, 8, 15, 16). Our results showed that while they are associated with falls, aging and comorbitidies have no relationship with the recurrence of falls. Firstly, the age-related physiological decline is only likely to lead to limited stability, which might explain the absence of link with the recurrence of falls (15). Secondly, institutionalization, the number of drugs taken per day and the use of psychoactive drugs were surrogate measures of comorbidities, not specifically related to postural control (18, 19). Regardless of underlying comorbid conditions, it has been suggested that only those responsible for a reduced functioning of one or several sensory systems may lead to greater instability and, as consequence, to the recurrence of falls (18). As shown in several studies (2, 3, 4), the current study also highlights the fact that muscle weakness was related to falls but not to their recurrence. The main explanation could be that the handgrip strength is a surrogate measure of lower limb muscle weakness, which is more related to postural instability than handgrip (20).

The main finding of our study was that female gender, poor vision and lower limb proprioception, and fear of falling were specifically related to the recurrence of falls. The role of gender in the recurrence of falls is unclear. It has been demonstrated that falls are more common in female subjects than in male subjects (2, 4, 8). More specifically, we established that recurrent falls were also more frequent in female subjects. It has been suggested that the effects of chronic diseases and physiologic decline could be different between women and men (15, 19), notably the greater loss of fat-free mass with aging. This may therefore contribute to a vicious circle, with an increased risk of falls in women compared to men, particularly since a low functionality and a high level of motor impairment are considered as risk factors for recurrent fallers. Furthermore, it has been suggested that the circumstances and the mechanism of falls in women and men might differ, which could explain a higher recurrence of falls in women compared to men (2, 8, 18, 19).

A large proportion of falls in the elderly occurs while walking or standing up and sitting down (4, 8). An efficient dynamic stability, defined as the ability to control whole body position during motor activities in which the body's center of mass is displaced outside the base of support (21, 22, 23), is required during these daily motor activities. Human postural stability depends on the interaction of multiple sensory, motor and integrative systems (15, 22, 23). It has been shown that the sensory system and in particular vision and proprioception are important to keep control of body position (15, 21, 22, 23). Similarly to our findings, the rare previous studies showed that impaired performance in vision and proprioception were related to falls and their recurrence (1, 3, 4, 10). The fact that a worse performance was observed for vision and proprioception in multiple fallers compared to single- and non-fallers in the current study is in agreement with this statement. In addition, the fear of falling is a well-known fall-related adverse outcome (16). The main consequences of the fear of falling have been described as physical, functional, psychological and social changes (6, 16). It partially depends on the self-perception of the mobility problem, a recognized independent clinical predictor of recurrent falls (23). Our study showed that the fear of falling was the main risk factor for the recurrence of falls, affecting the postural behavior and influencing the walking process and the onset of new falls.

Whilst cognitive decline has been previously identified as a risk factor for falling, we did not find any association with the recurrence of falls in our study. Two main explanations are possible. First, this result may be related to the test used to access cognitive impairment. Differences with prior studies may be due to the poor capacity of GPcog to show an association between falls and cognitive impairment, all the more we used only the patient section score (14). Second, it could be hypothesized that cognitive impairment is related only to the first fall but not to its recurrence.

Our study has some limitations. Firstly, the study cohort was restricted to community-dwelling older adults who were probably healthier, more motivated and also showed greater interest in health issues than the general population of older adults. Secondly, the findings should take into account the limitation of the study design. The cross-sectional design of the present study may represent a limitation to the exploration of the association between risk factors for falls and their recurrence, compared to prospective cohort study design. Thirdly, although we were able to control for many characteristics likely to modify this relationship, residual potential confounders might still be present. Fourthly, the fear of falling, which was based on self-report, might have required more refined observations based on recognized ordinal measures. Fifthly, limitations of this study include a potential recall bias. Falls are usually underreported because of a cognitive decline of fallers who forget to report the falls, and depends on the occurrence of fall-related adverse health outcomes (24). Actually, it has been shown that fallers who had physical traumatisms reported more frequent falls compared to those who had not (1). Furthermore, in their systematic review, Hauer et al. also showed that the recall bias could be also related to the method used to report falls (1). Prospective registration systems, shorter recall period or the use of fall diaries have also proven their superiority on the other methods of data collection, and could lead to a substantial increase in reported falls. Lastly, another limitation could be related to the exclusion criteria used in our study. We chose to exclude subjects with a high risk of falls related to neurological diseases or to major orthopaedic diagnoses involving the lower back, pelvis or lower extremities, but these criteria might be too severe.

To conclude, we can say that several intrinsic factors are related to falls while few of them are associated with their recurrence. Although risk factors for falls are well-identified, those related to the recurrence of falls remain less known. The current study confirms that female gender, poor vision and lower limb proprioception, and fear of falling are specific markers of recurrent falls.