Vestibular Function Predicts Balance and Fall Risk in Patients with Alzheimer’s Disease

Abstract

Background:

Patients with Alzheimer’s disease (AD) are at high risk for falls. Vestibular dysfunction predicts balance impairment in healthy adults; however, its contribution to falls in patients with AD is not well known.

Objective:

The objective of this study was to assess whether vestibular function contributes to balance and fall risk in patients with AD.

Methods:

In this prospective observational study, we assessed vestibular function using measures of semicircular canal (vestibulo-ocular reflex (VOR) gain) and saccular function (cervical vestibular-evoked myogenic (cVEMP) response), and we assessed balance function using the Berg Balance Scale and quantitative posturography. We evaluated falls incidence for a mean 1-year follow-up period (range 3–21 months) in 48 patients with mild-moderate AD.

Results:

Relative to matched controls, AD patients exhibited increased medio-lateral (ML) sway in eyes-open (0.89 cm versus 0.69 cm; p = 0.033) and eyes-closed (0.86 cm versus 0.65 cm; p = 0.042) conditions. Among AD patients, better semicircular canal function was associated with lower ML sway and antero-posterior (AP) sway in the eyes-closed condition (β= –2.42, 95% CI (–3.89, –0.95), p = 0.002; β= –2.38, 95% CI (–4.43, –0.32), p = 0.025, respectively). Additionally, better saccular function was associated with lower sway velocity (β= –0.18, 95% CI (–0.28, –0.08); p = 0.001). Finally, we observed that better semicircular canal function was significantly associated with lower likelihood of falls when adjusted for age, sex, and MMSE score (HR = 0.65; p = 0.009).

Conclusion:

These results support the vestibular system as an important contributor to balance and fall risk in AD patients and suggest a role for vestibular therapy.

Keywords

Alzheimer’s disease cognitive aging falls postural balance vestibular function tests

INTRODUCTION

Patients with Alzheimer’s disease (AD) have frequent falls. The annual incidence of falls among AD patients is 60–70%, which is twice the incidence in cognitively intact older adults [1 –3]. Falls are also among the most devastating outcomes experienced by AD patients. AD patients who fall are five times more likely to be institutionalized compared to AD patients who do not fall, even after accounting for severity of dementia [4]. As the population ages and the number of individuals with AD in the US grows to a projected 14 million in 2050, managing falls in AD will consume an ever-increasing proportion of societal resources.

The pathogenesis of falls in AD is not fully understood. Falls in AD are likely caused by fall risk factors that occur in all older adults, but that are amplified by neurodegeneration (Fig. 1). Loss of balance and mobility is one of the most important fall risk factors in AD patients [1, 5]. Although impairments in balance and gait are typically associated with other causes of dementia including Parkinson’s disease dementia, dementia with Lewy bodies, or vascular dementia [6], emerging evidence using more sensitive balance and gait tests demonstrates significant balance and gait impairment even in the earliest stages of AD [7, 8]. Patients with mild-moderate AD have been shown to have greater postural instability on an unstable support surface or with a reduced base of support compared to age-matched controls [9 –11]. Additionally, AD patients have significantly slower gait speed and higher gait variability relative to controls [2, 11]. AD patients with balance and gait impairment have been shown to have an increased risk of falls compared to AD patients without balance and gait impairment, independent of dementia severity [12 –14].

Fig. 1

Conceptual model of pathogenesis of falls in Alzheimer’s disease (AD). Falls in patients with AD are likely caused by fall risk factors that occur in all older adults, including gait impairment, spatial cognition, global cognition, and behavioral patterns, all of which are impacted by aging. In addition, age-related vestibular loss may interact with the neurodegenerative processes of AD to accelerate known fall risk factors.

Whether vestibular (inner ear balance) loss contributes to balance and mobility impairment and falls risk in AD patients is unknown. The vestibular system consists of three semicircular canals (superior, posterior, and horizontal), which detect rotational head movement, and two otolith organs (saccule and utricle), which sense linear head movements and the orientation of the head with respect to gravity. The vestibular sensory system is known to play a critical role in the control of posture, gait, and falls risk in cognitively unimpaired adults [15 –18]. In recent work, we found that vestibular loss is two-fold more common in AD patients (∼50% prevalence) relative to cognitively unimpaired age-matched controls (∼25%) [19]. The reason for the higher prevalence of vestibular loss in AD patients is unclear. It is possible that central neurodegeneration may impair vestibular reflexes by degrading central vestibular pathways in the brainstem. Alternatively, peripheral vestibular loss may contribute to neurodegeneration and specifically AD [20 –22]. Regardless of whether vestibular loss is a cause or consequence of AD, AD patients with vestibular impairment may have higher levels of balance and gait impairment resulting in an increased risk of falls. We observed that AD patients may have bilateral partial vestibular losses [19], which typically manifest as postural and gait instability (rather than vertigo which is characteristic of acute unilateral vestibular loss) [23]. Due to the unique pathogenesis of cognition and vestibular dysfunction in AD, it is possible that balance impairment differs between patients with AD and cognitively unimpaired older adults, as has been shown previously [24].

In this study, we investigated the cross-sectional relationship between vestibular physiologic function and balance function in a cohort of patients with mild-moderate AD. The primary objective of the study was to assess the impact of baseline vestibular impairment on balance and fall events in patients with AD over a longitudinal follow-up period of up to two years. In addition to this aim, we compared functional balance measures between patients with AD and matched cognitively unimpaired controls from the Baltimore Longitudinal Study of Aging (BLSA). This comparison with controls allowed for greater interpretation and context for the balance findings in patients with AD. These data provide key insights into the role of the vestibular system in balance control and falls risk in individuals with AD and suggest that vestibular interventions may play an important role in reducing falls and their highly morbid sequelae among individuals with AD.

MATERIALS AND METHODS

Study participants

This was a prospective observational study. Forty-eight participants were recruited between March 2018 and January 2020 from the Johns Hopkins Memory and Alzheimer’s Treatment Center and the Johns Hopkins Alzheimer’s Disease Research Center. Inclusion criteria for the study were 1) Age≥60 years; 2) Diagnosis of mild or moderate AD based on National Institute on Aging-Alzheimer’s Association 2011 criteria and Clinical Dementia Rating score of 1 or 2 [4, 25]; and 3) Mini-Mental State Exam (MMSE) score≥18 [26]. We have shown in prior work that an MMSE score of 13 is the threshold above which patients could reliably complete vestibular testing procedures [27]. As MMSE score may decline 2-3 points per year, an MMSE score cutoff of≥18 was used to maximize the chance that participants could continue participation during two-year follow-up [27]. The study was approved by the Johns Hopkins Institutional Review Board. All methods involving human participants were carried out in accordance with the ethical principles of the Declaration of Helsinki.

Vestibular and postural stability results were obtained from a cohort of cognitively unimpaired older participants, matched by age (within 5 years) and sex, from the BLSA. The BLSA is a longitudinal study of aging in healthy older adults that was established in 1958 [28]. For 4 AD patients, we were unable to find a unique control by sex and age; as such these patients were matched 2 : 1 with a BLSA control, yielding a control sample size of 44 participants (i.e., 40 controls matched 1 control: 1 AD patient; and 4 controls matched 2 controls:1 AD patient). As indicated below, there was no significant difference between the mean age or sex between the AD patients and BLSA controls.

Vestibular testing

For the cohort of patients with AD and BLSA controls, vestibular physiologic testing at baseline included assessment of semicircular canal and saccular function, per published procedures [27]. The EyeSeeCam video head impulse testing system (Interacoustics, Eden Prairie, MN) was used to assess horizontal semicircular canal function [29]. Semicircular canal function was measured as the ratio of eye velocity to head velocity, a quantity referred to as vestibulo-ocular reflex (VOR) gain. The mean VOR gain between right and left eyes was recorded. Saccular function was assessed using the cervical vestibular-evoked myogenic potential (cVEMP) [30 –32]. A 500 Hz, 125 dB sound pressure level acoustic tone burst was delivered monaurally to each ear. Surface electrodes were used to record an electromyogram of the myogenic response from the ipsilateral sternocleidomastoid (SCM) muscle, a response mediated by the saccule. cVEMP response amplitudes (in μV) were recorded and were corrected for background SCM contraction to account for differences in contraction strength. cVEMP measurements were taken from the better ear. Certain vestibular function tests in this cohort were found to be unsatisfactory due to the presence of background artifacts or recording errors on the response traces in the regions of interest and were thus removed (cVEMP data for 2 patients, VOR data for 8 patients excluded).

Functional balance testing

We used both qualitative and quantitative measures to assess balance function. In the cohort of patients with AD, we used the Berg Balance Scale, which has been well-validated in individuals with AD, as a qualitative balance function measure [33]. The Berg Balance Scale assesses the ability to maintain balance during 14 simple tasks such as moving from sitting to standing, transferring from one chair to another, and turning 360°. Each item is graded on a five-point scale, ranging from zero (patient unable to perform task) to four (patient can independently complete the task for the full duration) for a total possible score of 56.

Balance function was measured quantitatively using the BalanSens (BioSensics LLC, Brookline, MA) waist-worn accelerometer for both patients with AD and matched BLSA controls [34, 35]. Participants stood with feet together and arms at their sides, and center of mass sway area, sway velocity, and sway range were calculated in the medio-lateral (ML) and antero-posterior (AP) directions. Participants stood first with eyes open and then with eyes closed. In each condition, participants were provided up to three attempts to successfully complete one trial lasting 40 s. Both the eyes open and eyes closed conditions were used, given that vestibular deficits are more likely to manifest in eyes closed conditions, where vision is unavailable to help maintain standing balance. Among the group of patients with AD, 44 patients were able to safely provide postural stability measurements in both eyes open and eyes closed states. All matched BLSA controls were able to safely provide measurements.

Falls incidence in patients with AD

An exploratory outcome in this study was incident fall, defined as a person “coming to rest inadvertently on the ground or other lower level [36].” To assess fall events, patients used monthly calendars.

Statistical analyses

Descriptive statistics were conducted to evaluate participant demographics, vestibular assessment, and BalanSens posturography results. Student’s t test and chi-square test were used to compare results between the cohort of patients with AD and the matched cohort of cognitively unimpaired participants from the BLSA. Linear regression was used to evaluate the relationship between vestibular function and balance assessments, including Berg Balance Scale score, eyes-open BalanSens assessment, and eyes-closed BalanSens assessment. Multiple linear regression models adjusting for demographic characteristics (model 1: age and sex; model 2: age, sex, and MMSE) were developed to explore the association between vestibular function and balance assessments. MMSE score was included as a potential confounder of cognitive impairment and disease severity, as the patients included in this study had varying degrees of mild to moderate AD. Furthermore, to evaluate whether baseline vestibular function predicted incident falls in patients with AD, an extension to the Cox proportional-hazards model was used to model falls incidence as a recurrent event [37]. Specifically, the Prentice, Williams, and Peterson gap time (PWP-GT) model was used. This model was selected as it assumes that fall order is stratified. In other words, each patient’s first fall event (n) was assumed to be distinct from each patient’s second fall event (n + 1), and the same distinction applies to subsequent fall events (n + 2, n + 3, etc.) [38]. Mean VOR gain was divided by the standard deviation of VOR gain for the sample to obtain a fall risk hazard ratio for each standard deviation unit of change in VOR gain. Hazards of falls analyses were adjusted by two models of demographic characteristics (model 1: age and sex; model 2: age, sex, and MMSE). Sensitivity analyses were conducted to adjust model 2 with additional factors that may impact fall risk, including vision, hearing, and cardiovascular disease. Vision was measured by Snellen visual acuity test (left and right; logMAR units) and Pelli-Robson contrast sensitivity test. Hearing was measured by average pure tone threshold in decibels across 4 different frequencies bilaterally (500, 1000, 2000, 4000 Hz). Cardiovascular risk factors were measured by number of medications used for cardiovascular disease. All analyses were performed using STATA version 15 (College Station, TX, USA).

RESULTS

The study sample consisted of 48 patients with mild-moderate AD (Table 1). Among AD patients, the mean (SD) age was 74.9 (7.4) years, 55% were male, and the mean (SD) MMSE score was 22.9 (3.8). The mean (SD) VOR gain was 0.91 (0.14), and the mean (SD) corrected cVEMP response amplitude was 0.80 (0.50). The mean (range) follow-up time for the cohort of patients was 11.4 months (3 to 21 months), with 12 patients (25%) having at least one fall event during follow-up. A total of 23 unique fall events occurred during the study period. Within the cohort of patients with AD, MMSE score was not significantly correlated with vestibular function or postural stability measurements (Supplementary Table 1).

Table 1

Demographic and Clinical Characteristics of AD Patients and BLSA Matched Controls

Characteristic	BLSA (n = 44)	AD (n = 48)	p
Demographic Characteristics
Age, Mean (SD)	75.6 (1.3)	74.9 (7.4)	0.651
Male Sex, N (%)	27 (56.25)	24 (54.6)	0.869
MMSE Total Score, Mean (SD)	28.4 (1.5)	22.9 (3.8)	< 0.001
Vestibular Function
Semicircular canal function (VOR), Mean (SD)	0.98 (0.14)	0.91 (0.14)	0.020
Saccular function (cVEMP), Mean (SD)	1.81 (1.68)	0.80 (0.50)	< 0.001
Falls
Follow-Up Time in Months, Mean (Range)	11.4 (3–21)
AD Patients with at least 1 Fall Event, N (%)	12 (25)
Total Number of Falls, N	23
Balance
Berg Balance Scale Score, Mean	50.4 (5.3)
Posturography
Eyes Open
Medio-Lateral Range	0.69 (0.38)	0.89 (0.45)	0.033
Antero-Posterior Range	1.05 (0.63)	1.05 (0.49)	0.994
Sway Area	0.52 (0.50)	0.72 (0.84)	0.176
Sway Velocity	0.63 (0.09)	0.66 (0.12)	0.097
Eyes Closed
Medio-Lateral Range	0.65 (0.24)	0.86 (0.61)	0.042
Antero-Posterior Range	0.98 (0.41)	1.13 (0.76)	0.252
Sway Area	0.50 (0.44)	0.92 (1.98)	0.168
Sway Velocity	0.68 (0.13)	0.70 (0.18)	0.526

BLSA, Baltimore Longitudinal Study of Aging; VOR, vestibulo-ocular reflex gain, which is the ratio of eye velocity to head velocity, unitless; cVEMP, cervical vestibular-evoked myogenic potential, which is corrected for background SCM contraction.

AD patients and matched controls

Demographic information of the cohort of matched participants from the BLSA are also provided in Table 1. The mean (SD) age of the control group was 75.6 (1.3) years and 56% were male. Compared to cognitively unimpaired controls, patients with AD had a significantly lower mean MMSE score (22.9 versus 28.4; p < 0.001). With respect to vestibular measures, patients with AD exhibited significantly lower mean VOR gain (0.91 versus 0.98; p = 0.020) and lower corrected cVEMP amplitude (0.80 versus 1.81; p < 0.001).

With respect to balance function, patients with AD had mean (SD) Berg Balance Scale score of 50.4 (5.3). Patients with AD had higher ML sway range in both the eyes-open (0.89 cm versus 0.69 cm; p = 0.033) and eyes-closed (0.86 cm versus 0.65 cm; p = 0.042) conditions compared to cognitively unimpaired controls, although these were borderline significant when adjusted for multiple comparisons (Table 1). Additional postural stability measurements, including AP sway range, sway area, and sway velocity, were observed to be higher in patients with AD, but these differences were not statistically significant.

Vestibular function and balance, postural stability in patients with AD

Multiple linear regression was conducted to analyze the cross-sectional relationship between vestibular physiologic function and balance function measures among patients with AD (Table 2). No vestibular measures were found to be significantly associated with Berg Balance Scale score in either Model 1 (adjusted for age and sex) or Model 2 (adjusted for age, sex, and MMSE score).

Table 2

Multiple Linear Regression of Vestibular Function and Balance Outcomes in Patients with AD

	Model 1^a		Model 2^b
Variable	β (95% CI)	p	β (95% CI)	p
Berg Balance Scale
Semicircular canal function (VOR)	5.17 (–5.28, 15.62)	0.322	5.52 (–4.55, 15.58)	0.273
Saccular function (cVEMP)	1.78 (–0.98, 4.55)	0.200	1.86 (–0.91, 4.63)	0.181
Posturography
Eyes Open
	Medio-Lateral Range
Semicircular canal function (VOR)	–0.61 (–1.46, 0.24)	0.155	–0.46 (–1.52, 0.59)	0.376
Saccular function (cVEMP)	–0.05 (–0.30, 0.21)	0.715	–0.05 (–0.30, 0.21)	0.722
	Antero-Posterior Range
Semicircular canal function (VOR)	–0.50 (–1.59, 0.59)	0.360	–0.02 (–1.34, 1.29)	0.975
Saccular function (cVEMP)	0.01 (–0.29, 0.32)	0.924	0.03 (–0.28, 0.34)	0.845
	Sway Area
Semicircular canal function (VOR)	–0.54 (–2.49, 1.42)	0.579	–0.36 (–2.78, 2.07)	0.767
Saccular function (cVEMP)	–0.09 (–0.63, 0.46)	0.750	–0.08 (–0.64, 0.48)	0.772
	Sway Velocity
Semicircular canal function (VOR)	–0.11 (–0.34, 0.13)	0.369	–0.01 (–0.30, 0.28)	0.941
Saccular function (cVEMP)	–0.06 (–0.13, 0)	0.061	–0.06 (–0.12, 0.01)	0.078
Eyes Closed
	Medio-Lateral Range
Semicircular canal function (VOR)	–1.46 (–2.76, –0.17)	0.028	–2.42 (–3.89, –0.95)	0.002
Saccular function (cVEMP)	–0.12 (–0.52, 0.27)	0.522	–0.12 (–0.52, 0.28)	0.537
	Antero-Posterior Range
Semicircular canal function (VOR)	–1.60 (–3.30, 0.10)	0.064	–2.38 (–4.43, –0.32)	0.025
Saccular function (cVEMP)	–0.14 (–0.64, 0.36)	0.567	–0.15 (–0.66, 0.35)	0.542
	Sway Area
Semicircular canal function (VOR)	–2.70 (–7.19, 1.80)	0.231	–4.26 (–9.76, 1.25)	0.125
Saccular function (cVEMP)	–0.40 (–1.67, 0.88)	0.534	–0.42 (–1.72, 0.88)	0.518
	Sway Velocity
Semicircular canal function (VOR)	0.12 (–0.25, 0.49)	0.512	0.22 (–0.23, 0.67)	0.318
Saccular function (cVEMP)	–0.18 (–0.28, –0.08)	0.001	–0.18 (–0.27, –0.08)	0.001

^aModel adjusted for age and sex. ^bModel adjusted for age, sex, and MMSE score. VOR, vestibulo-ocular reflex gain, which is the ratio of eye velocity to head velocity, unitless; cVEMP, cervical vestibular-evoked myogenic potential, which is corrected for background SCM contraction.

Multiple linear regression was conducted in a similar manner to analyze the relationship between vestibular function and quantitative posturography measures in patients with AD (Table 2). No associations were found to be significant in the eyes-open state. However, in the eyes-closed state, better semicircular canal function was significantly associated with lower ML sway range (β= –2.42; p = 0.002) and AP sway range (β= –2.38; p = 0.025) when adjusted for age, sex, and MMSE score. Additionally, there was a significant inverse relationship between saccular function and sway velocity in the eyes-closed state (β= –0.18; p = 0.001) in this model.

Hazards of falls analysis in patients with AD

To analyze falls incidence as a consequence of baseline vestibular function and balance metrics in the prospective cohort of patients with AD, the PWP-GT extension to the Cox proportional-hazards model was utilized (Table 3). In unadjusted analysis, each standard deviation increase in mean VOR gain (each increase in VOR gain by 0.14) was associated with 22% reduced likelihood of recurrent falls (p = 0.042). This remained significant when adjusted for age, sex, and baseline MMSE score (HR = 0.65; p = 0.009), such that in the adjusted analysis, each standard deviation increase in VOR gain was associated with a 35% reduction in fall risk. The adjustment factors in Model 2 (age, sex, MMSE score) were not significantly associated with fall risk (Supplementary Table 2). Additional sensitivity analyses showed that semicircular canal function continued to significantly predict falls when adjusted for vision, hearing, and cardiovascular risk factors (Supplementary Table 3). Saccular function was not observed to be a significant predictor of incident falls. None of the sway measures in eyes open or eyes closed states were significant predictors of incident falls in adjusted analyses (Supplementary Table 4). We explored whether the sway measures may mediate the link between vestibular function and incident falls by adding them into the PWP-GT proportional hazards models. We specifically considered the eyes closed measures, given their significant association with vestibular function. We did not observe any significant change in the hazard ratios for semicircular canal function with the addition of eyes closed ML sway (HR = 0.64, p = 0.025) or eyes closed AP sway (HR = 0.66, p = 0.016).

Table 3

Cox Proportional Hazard Models of Falls in Patients with AD (PWP-GT)

	Unadjusted HR (95% CI)	p	Model 1 HR (95% CI)^b	p	Model 2 HR (95% CI)^c	p
Semicircular canal function (VOR SD units^a)	0.78 (0.61, 0.99)	0.042	0.67 (0.50, 0.90)	0.007	0.65 (0.47, 0.90)	0.009
Saccular function (cVEMP)	0.85 (0.38, 1.89)	0.685	0.87 (0.35, 2.16)	0.770	0.81 (0.32, 2.02)	0.646

^aMean VOR Gain divided by 1 standard deviation (0.14). ^bModel adjusted for age and sex. ^cModel adjusted for age, sex, and MMSE score. HR, hazard ratio; CI, confidence interval; VOR, vestibulo-ocular reflex gain, which is the ratio of eye velocity to head velocity, unitless; cVEMP, cervical vestibular-evoked myogenic potential, which is corrected for background SCM contraction.

DISCUSSION

There is an increasingly important need to characterize balance function and fall risk in patients with AD, as balance impairment and falls in AD are associated with high risk for institutionalization and early mortality [39, 40]. In this study, which to our knowledge represents the largest series examining balance and prospective falls in AD, we observed that patients with AD had increased ML instability when compared to cognitively unimpaired matched controls. Moreover, we found that better vestibular function (specifically semicircular canal and otolith function) was associated with greater postural stability in the eyes-closed state in patients with AD. Finally, in a recurrent-event hazards analysis of incident falls, we observed that better vestibular function (specifically semicircular canal function) was associated with a significantly lower likelihood of falls over a mean ∼1-year follow-up period.

Consistent with prior reports, we observed that patients with AD had poorer vestibular function (including both semicircular canal function and otolith function) relative to matched controls [27, 41]. Additionally, we found that patients with AD had greater ML sway in both eyes-open and eyes-closed conditions. ML sway, or lateral stability, is known to be biomechanically more challenging than AP or frontal stability, in part because of the reduced base of support provided by the elongated foot in the lateral versus frontal axes [42]. Lateral stability declines more quickly with age [43], and has specifically been identified as a fall risk factor in cognitively unimpaired older adults [44 –46]. Moreover, when considering lateral stability in both eyes-open and eyes-closed states, studies suggest that impaired eyes-closed ML sway (lateral stability) is the single greatest predictor of incident falls in community-dwelling older adults [47]. In the eyes-closed state compared to the eyes-open state, the main contributors to postural stability are vestibular and proprioceptive function. Indeed, in this study, we observed a significant relationship between vestibular function (specifically semicircular canal function) and eyes-closed postural stability, in both the lateral/ML and frontal/AP directions. Prior studies have reported that vestibular function preferentially influences ML over AP stability, possibly because the paired vestibular sensors on each side work in a coordinated manner in the AP dimension in contrast to a push-pull manner in the ML dimension, with the coordinated signaling allowing for more accurate encoding of redundant information [48].

Additionally, we observed a significant association between better saccular function and lower sway velocity in the cohort of patients with AD. Sway velocity is another dimension of balance control that is distinct from sway amplitude: for example, individuals may have normal sway amplitudes though high sway velocities which bring them to their limit of stability and increase the risk of a fall [47]. The saccule is the vestibular sensory organ that detects tilt of the head with respect to vertical gravitation [49]. The saccular graviceptors may play a specific role in providing rapid signaling to the central nervous system that drives compensatory ankle and hip torques which limit sway velocity and increase postural stability. Altogether, these findings suggest that for AD patients who present with balance problems that are exacerbated with darkness or eyes-closed state, vestibular diagnostic testing should be considered. If vestibular impairment is identified, the patient may be a candidate for vestibular therapy. Vestibular therapy, which is a set of exercises that improves vestibular adaptation (i.e., strengthen residual vestibular function) and compensation (i.e., recruit other mechanisms including vision or proprioception to compensate for vestibular loss), is known to be effective in reducing falls among cognitively unimpaired adults. Further studies will be needed to potentially extend the benefits of this existing, low-risk therapy to AD patients who face disproportionately higher fall risk compared to cognitively unimpaired individuals.

In exploratory analyses considering vestibular function and balance control as predictors of incident falls in this small cohort, we observed that better vestibular function (specifically semicircular canal function) was significantly associated with a reduced hazard of falls. The semicircular canals in the vestibular organ detect angular accelerations of the head in each canal plane, and semicircular canal function has been shown to predict postural stability, gait speed, and fall risk in cognitively unimpaired populations [16, 50]. To our knowledge, this study is among the first to link vestibular loss to increased falls in patients with AD, which has substantial implications with respect to the potential effectiveness of vestibular therapies in reducing falls among AD patients. Interestingly, although prior reports noted the importance of sway measures as predictors of fall risk, we did not observe any significant relationship between postural measures and incident falls in this analysis. Moreover, the significant association between vestibular function and incident falls was not attenuated by the addition of balance measures to proportional hazards models. It is possible that in AD patients, in whom the variability of sway measures is much larger, a larger sample size will be needed to demonstrate significant relationships. Additionally, vestibular impairment may increase fall risk via mechanisms that do not involve postural stability, such as via the known contribution of the vestibular system to spatial abilities such as spatial memory and spatial navigation in both cognitively unimpaired and cognitively impaired older individuals [51, 52].

We note several limitations of this study. The sample size for the cohort of patients with AD was small, preventing subset analyses based on severity of AD or mild versus moderate cognitive impairment. The comparison between patients with AD and the matched cohort of cognitively unimpaired participants does not include all confounders, including visual, proprioceptive, and motor dysfunction, all of which may impact postural stability. The use of monthly calendars, while considered the current gold standard for falls tracking in healthy adults, may lead to recall bias and poor adherence. Future studies in falls tracking may utilize wearable technologies or text-messaging platforms, which hold promise in improving accuracy in falls assessment [53, 54]. Furthermore, due to the relatively short mean follow-up period of ∼1 year, there may not have been sufficient time to reveal effects of other variables, such as postural stability measures, on the likelihood of recurrent falls in this cohort. Indeed, further studies with larger sample sizes and increased patient follow-up time are warranted for better understanding the relationship between vestibular function, postural stability, and fall risk in patients with AD. Despite these limitations, our study identifies vestibular dysfunction as a key risk factor for falls in patients with AD and may inform patient and provider decisions regarding falls preventive strategies.

Conclusion

Patients with AD are at high risk of falls, which is a devastating outcome in this population linked to earlier institutionalization and mortality. This study provides preliminary evidence that vestibular function is a significant determinant of postural stability and falls risk in patients with AD. Further studies in larger cohorts with longer follow-up periods will be needed to better understand whether or not postural stability mediates the link between vestibular function and fall risk in AD patients. Additionally, these data support the need for future studies on the potential benefit of vestibular interventions, including vestibular physical therapy, as well as more novel vestibular therapeutics such as vestibular prosthetic devices, in reducing fall risk in patients with AD.

Footnotes

ACKNOWLEDGMENTS

Dr. Yuri Agrawal is receiving grant funding from NIA (#RO1 AG057667), NIH/NIA (#R01 AG061786), and NIH/NIA (#R01 AG065259). Dr. Oh is receiving grant funding from NIA/NIH R01AG057725. Kevin Biju was a participant in the Medical Student Training in Aging Research (MSTAR) Summer Program, which was funded by NIA (#T35AG026758).

Authors’ disclosures available online ().

The supplementary material is available in the electronic version of this article: .

References

van Dijk

, Meulenberg

, Van de Sande

, Habbema

JDF

(1993) Falls in dementia patients. Gerontologist 33, 200–204.

Visser

(1983) Gait and balance in senile dementia of Alzheimer’s type. Age Ageing 12, 296–301.

Muir

, Gopaul

, Odasso

MMM

(2012) The role of cognitive impairment in fall risk among older adults: A systematic review and meta-analysis. Age Ageing 41, 299–308.

Morris

(1997) Clinical dementia rating: A reliable and valid diagnostic and staging measure for dementia of the Alzheimer type. Int Psychogeriatr 9 Suppl 1, 173–176; discussion 177-178.

Kosse

, de Groot

, Vuillerme

, Hortobágyi

, Lamoth

(2015) Factors related to the high fall rate in long-term care residents with dementia. Int Psychogeriatr 27, 803–814.

Allan

, Ballard

, Rowan

, Kenny

(2009) Incidence and prediction of falls in dementia: A prospective study in older people. PLoS One 4, e5521.

Pettersson

, Engardt

, Wahlund

L-O

(2002) Activity level and balance in subjects with mild Alzheimer’s disease. Dement Geriatr Cogn Disord 13, 213–216.

O’keeffe

, Kazeem

, Philpott

, Playfer

, Gosney

, Lye

(1996) Gait disturbance in Alzheimer’s disease: A clinical study. Age Ageing 25, 313–316.

Taylor

, Delbaere

, Lord

, Mikolaizak

, Brodaty

, Close

(2014) Neuropsychological, physical, and functional mobility measures associated with falls in cognitively impaired older adults. J Gerontol A Biol Sci Med Sci 69, 987–995.

10.

Chong

, Horak

, Frank

, Kaye

(1999) Sensory organization for balance: Specific deficits in Alzheimer’s but not in Parkinson’s disease. J Gerontol A Biol Sci Med Sci 54, M122–M128.

11.

Alexander

, Mollo

, Giordani

, Ashton-Miller

, Schultz

, Grunawalt

, Foster

(1995) Maintenance of balance, gait patterns, and obstacle clearance in Alzheimer’s disease. Neurology 45, 908–914.

12.

Camicioli

, Licis

(2004) Motor impairment predicts falls in specialized Alzheimer care units. Alzheimer Dis Assoc Disord 18, 214–218.

13.

Buchner

, Larson

(1987) Falls and fractures in patients with Alzheimer-type dementia. JAMA 257, 1492–1495.

14.

Nakamura

, Meguro

, Sasaki

(1996) Relationship between falls and stride length variability in senile dementia of the Alzheimer type. Gerontology 42, 108–113.

15.

Herdman

, Blatt

, Schubert

, Tusa

(2000) Falls in patients with vestibular deficits. Am J Otol 21, 847–851.

16.

Agrawal

, Davalos-Bichara

, Zuniga

, Carey

(2013) Head impulse test abnormalities and influence on gait speed and falls in older individuals. Otol Neurotol 34, 1729–1735.

17.

Agrawal

, Carey

, Della Santina

, Schubert

, Minor

(2009) Disorders of balance and vestibular function in US adults: Data from the National Health and Nutrition Examination Survey, 2001-2004. Arch Intern Med 169, 938–944.

18.

Whitney

, Marchetti

, Pritcher

, Furman

(2009) Gaze stabilization and gait performance in vestibular dysfunction. Gait Posture 29, 194–198.

19.

Harun

, Oh

, Bigelow

, Studenski

, Agrawal

(2016) Vestibular impairment in dementia. Otol Neurotol 37, 1137–1142.

20.

Previc

(2013) Vestibular loss as a contributor to Alzheimer’s disease. Med Hypotheses 80, 360–367.

21.

Smith

(2017) The vestibular system and cognition. Curr Opin Neurol 30, 84–89.

22.

Smith

, Zheng

, Horii

, Darlington

(2005) Does vestibular damage cause cognitive dysfunction in humans? J Vestib Res 15, 1–9.

23.

Minor

(1998) Gentamicin-induced bilateral vestibular hypofunction. JAMA 279, 541–544.

24.

Yoon

, Choi

, Jeong

, Park

, Kim

, Hwang

, Jang

, Kim

, Hong

, Lee

, Kang

, Yoon

(2020) Balance and mobility performance along the Alzheimer’s disease spectrum. J Alzheimers Dis 73, 633–644.

25.

Jack

Jr. , Bennett

, Blennow

, Carrillo

, Dunn

, Haeberlein

, Holtzman

, Jagust

, Jessen

, Karlawish

, Liu

, Molinuevo

, Montine

, Phelps

, Rankin

, Rowe

, Scheltens

, Siemers

, Snyder

, Sperling

(2018) NIA-AA Research Framework: Toward a biological definition of Alzheimer’s disease. Alzheimers Dement 14, 535–562.

26.

Folstein

, Folstein

, McHugh

(1975) “Mini-mental state”. A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 12, 189–198.

27.

Harun

, Oh

, Bigelow

, Agrawal

(2017) Vestibular function testing in individuals with cognitive impairment: Our experience with sixty participants. Clin Otolaryngol 42, 772–776.

28.

Ferrucci

(2008) The Baltimore Longitudinal Study of Aging (BLSA): A 50-year-long journey and plans for the future. J Gerontol A Biol Sci Med Sci 63, 1416–1419.

29.

Bartl

, Lehnen

, Kohlbecher

, Schneider

(2009) Head impulse testing using video-oculography. Ann N Y Acad Sci 1164, 331–333.

30.

, Zuniga

, Nguyen

, Carey

, Agrawal

(2014) How to interpret latencies of cervical and ocular vestibular-evoked myogenic potentials: Our experience in fifty-three participants. Clin Otolaryngol 39, 297–301.

31.

, Layman

, Carey

, Agrawal

(2015) Epidemiology of vestibular evoked myogenic potentials: Data from the Baltimore Longitudinal Study of Aging. Clin Neurophysiol 126, 2207–2215.

32.

Nguyen

, Welgampola

, Carey

(2010) Test-retest reliability and age-related characteristics of the ocular and cervical vestibular evoked myogenic potential tests. Otol Neurotol 31, 793–802.

33.

Berg

, Wood-Dauphinee

, Williams

, Maki

(1992) Measuring balance in the elderly: Validation of an instrument. Can J Public Health 83 Suppl 2, S7–11.

34.

Najafi

, Horn

, Marclay

, Crews

, Wu

, Wrobel

(2010) Assessing postural control and postural control strategy in diabetes patients using innovative and wearable technology. J Diabetes Sci Technol 4, 780–791.

35.

Anson

, Bigelow

, Swenor

, Deshpande

, Studenski

, Jeka

, Agrawal

(2017) Loss of peripheral sensory function explains much of the increase in postural sway in healthy older adults. Front Aging Neurosci 9, 202.

36.

(1987) The prevention of falls in later life. A report of the Kellogg International Work Group on the Prevention of Falls by the Elderly. Dan Med Bull 34 Suppl 4, 1–24.

37.

Gill

, Zou

, Jones

, Speechley

(2009) Comparison of regression models for the analysis of fall risk factors in older veterans. Ann Epidemiol 19, 523–530.

38.

Prentice

, Williams

, Peterson

(1981) On the regression analysis of multivariate failure time data. Biometrika 68, 373–379.

39.

Nevitt

, Cummings

, Hudes

(1991) Risk factors for injurious falls: A prospective study. J Gerontol 46, M164–M170.

40.

Dolatabadi

, Van Ooteghem

, Taati

, Iaboni

(2018) Quantitative mobility assessment for fall risk prediction in dementia: A systematic review. Dement Geriatr Cogn Disord 45, 353–367.

41.

Birdane

, Incesulu

, Gurbuz

, Ozbabalik

(2012) Sacculocolic reflex in patients with dementia: Is it possible to use it for early diagnosis? Neurol Sci 33, 17–21.

42.

Grüneberg

, Duysens

, Honegger

, Allum

(2005) Spatio-temporal separation of roll and pitch balance-correcting commands in humans. J Neurophysiol 94, 3143–3158.

43.

Gill

, Allum

, Carpenter

, Held-Ziolkowska

, Adkin

, Honegger

, Pierchala

(2001) Trunk sway measures of postural stability during clinical balance tests: Effects of age. J Gerontol A Biol Sci Med Sci 56, M438–M447.

44.

Hilliard

, Martinez

, Janssen

, Edwards

, Mille

, Zhang

, Rogers

(2008) Lateral balance factors predict future falls in community-living older adults. Arch Phys Med Rehabil 89, 1708–1713.

45.

Lord

, Rogers

, Howland

, Fitzpatrick

(1999) Lateral stability, sensorimotor function and falls in older people. J Am Geriatr Soc 47, 1077–1081.

46.

Rogers

, Mille

(2003) Lateral stability and falls in older people. Exerc Sport Sci Rev 31, 182–187.

47.

Maki

, Holliday

, Topper

(1994) A prospective study of postural balance and risk of falling in an ambulatory and independent elderly population. J Gerontol 49, M72–M84.

48.

Mbongo

, Patko

, Vidal

, Vibert

, Tran Ba Huy

, de Waele

(2005) Postural control in patients with unilateral vestibular lesions is more impaired in the roll than in the pitch plane: A static and dynamic posturography study. Audiol Neurootol 10, 291–302.

49.

Goldberg

, Wilson

, Cullen

, Angelaki

, Broussard

, Buttner-Ennever

, Fukushima

, Minor

(2012), The Vestibular System: A Sixth Sense, Oxford University Press.

50.

Ekvall Hansson

, Magnusson

(2013) Vestibular asymmetry predicts falls among elderly patients with multi-sensory dizziness. BMC Geriatr 13, 77.

51.

Bigelow

, Semenov

, Trevino

, Ferrucci

, Resnick

, Simonsick

, Xue

, Agrawal

(2015) Association between visuospatial ability and vestibular function in the Baltimore Longitudinal Study of Aging. J Am Geriatr Soc 63, 1837–1844.

52.

Wei

, Oh

, Harun

, Ehrenburg

, Agrawal

(2018) Vestibular loss predicts poorer spatial cognition in patients with Alzheimer’s disease. J Alzheimers Dis 61, 995–1003.

53.

Kamil

, Bakar

, Ehrenburg

, Frankenthaler

, Wei

, Anson

, Oh

, Agrawal

(2018) “Text It” program to track falls in patients with Alzheimer’s disease and dementia. Alzheimers Dement (N Y) 4, 137–140.

54.

Rucco

, Sorriso

, Liparoti

, Ferraioli

, Sorrentino

, Ambrosanio

, Baselice

(2018) Type and location of wearable sensors for monitoring falls during static and dynamic tasks in healthy elderly: A review. Sensors (Basel) 18, 1613.