The pace and prognosis of peripheral sensory loss in advanced age: association with gait speed and falls

Falls are a leading cause of morbidity and mortality among older adults, occurring in approximately one-third of community-dwelling persons over age 65 and costing over $31 billion in related injuries annually [1‐3]. The presence of peripheral sensory loss is considered to be one of many different risk factors for falls [3‐8]. A few previous studies demonstrated that older adults with peripheral sensory loss at the outset had an increased risk of subsequent falls [5‐9]. However, these studies relied on single baseline measures of sensory function and did not examine the effect of changes over time [7, 8]. The longitudinal relationship between changes in peripheral sensory function and the development of falls is not known.

Peripheral neuropathy is common among older adults, [10] especially in those with diabetes [11, 12]. The prevalence of peripheral neuropathy has been reported to be approximately 7% in older adult populations worldwide [11, 13]. While there are many causes, including alcohol ingestion, vitamin B₁₂ deficiency, cancer, chemotherapy, chronic kidney disease, and paraproteinemias, diabetes is the most frequent, accounting for 32–44% of patients with polyneuropathy [14, 15].

Many factors associated with peripheral neuropathy may increase the risk of falls, including impaired balance, muscle weakness, nutritional deficiencies, and medications [2, 16]. One mechanism by which the progression of sensory neuropathy may predispose people to falls is through alterations in gait. While slow gait speed has been reported to be cross-sectionally associated with distal sensory neuropathy [17], this could be either a cause of falls or protective mechanism to prevent falls. The present study takes advantage of longitudinal data from the Maintenance Of Balance, Independent Living, Intellect, and Zest in the Elderly (MOBILIZE) Study in Boston, MA, to assess the prevalence and progression of peripheral sensory deficits in the feet, and whether they are associated with slowing of gait and falls over a 5-year period.

The characteristics of the full study cohort and those with (mild, moderate, or severe) and without (intact) baseline sensory impairment are shown in Table 3. At baseline, participants with any sensory impairment were four years older on average than their intact counterparts, were more likely to be male, and had diminished executive function (as indicated by 20s longer Trails B time on average). They also exhibited more multimorbidity and diminished physical functioning than their intact counterparts, reporting a greater prevalence of diabetes and peripheral neuropathy, and 0.3 greater mean comorbidity count on average. They were twice as likely to report use of a walking aid and exhibited diminished physical function, scoring an average of 1.4 points lower on the SPPB and 0.1 m/s slower gait speed, both differences well above established thresholds of clinical significance. Despite these differences, however, a history of one or more falls was similar in the two groups.

The relationship between changes in peripheral sensory function and change in gait speed over 18 and 60 months is illustrated in Fig. 1. After model adjustment, participants in the three groups had comparable mean baseline gait speed. The group of subjects with consistently intact sensory function over this time period had the smallest declines in gait speed over 60 months (− 0.12 m/s; 95% CI: -0.15 to − 0.08). This decline in gait speed was of similar magnitude to the decline observed in subjects whose sensory function progressed to impairment (− 0.13 m/s; − 0.16 to − 0.10).Those whose sensory function was consistently impaired had a steeper decline in gait speed (− 0.23 m/s; 95% CI: -0.28 to − 0.18). The difference between the consistently impaired group and the others was statistically significant and consistent with a ‘substantially’ meaningful difference for gait speed metrics in older populations [36].

Declines in peripheral sensory function were also associated with fall risk. Table 4 shows the absolute and relative risk of falls by temporal pattern of impairment. Those whose sensory function progressed to impairment during followup had a greater risk of falls than those whose sensory function was consistently intact (adjusted risk ratio = 1.57 (95% confidence interval = 1.12 to 2.22). The group that remained consistently impaired over the 5 years had an elevated fall rate, but it was not statistically significantly different from that of the consistently intact group (adjusted risk ratio = 1.47, 95% confidence interval = 0.89 to 2.45).

The results of this study demonstrate the course of tactile sensory loss in the feet of older community-dwelling people over 5 years, and relationships between peripheral sensory loss and the concurrent slowing of gait and development of falls. Over the 5-year course of the study, older adults with consistently impaired peripheral sensory function had a significantly greater decline in gait speed compared to those with consistently intact or progressive impairment over 5 years. Compared to those with consistently intact sensory function, older adults who developed sensory impairments had a greater risk of falls. These results indicate that the loss of peripheral sensory function is a significant contributor to slowing of gait and an increased risk of falls in a community-dwelling older adult population.

The absolute number (59) and percent (17%) of participants with sensory impairment were relatively low in the MOBILIZE Boston population, but higher than the 7% reported in other older populations [11, 13]. Most of our participants had intact sensation at baseline, but this enabled us to observe the development of impairments over time. Nearly half (49%) of those initially intact developed sensory impairment over 5 years of followup. This high incidence of sensory loss has not been fully appreciated in other studies nor in clinical practice.

To our knowledge previous studies have not examined changes in peripheral sensory function over time nor their relationship to mobility outcomes. However, several cross-sectional and longitudinal studies have examined relationships between baseline abnormalities in foot sensation (including sense of vibration and touch), functional impairments, and fall risk. Most of these studies demonstrated that the loss of vibratory sensation was associated with increased fall risk [5, 37, 38]. For example, Patel et al. [37] found among older women with diabetes that fallers had higher vibratory sensory thresholds than non-fallers, but both fallers and non-fallers had similar touch sensation. Using cross-sectional data from the 2003–2004 National Health and Nutrition Examination Survey (NHANES), Wilson et al. [38] found no association between sensory loss in the feet, as measured using 10 g SWMT, and a subjective report of “difficulty with falls during the past 12 months,” assessed by questionnaire. However, after adjustments, participants who exhibited peripheral neuropathy had an increased risk of balance impairment compared to those without peripheral neuropathy. In a cross-sectional study exploring the effect of different sensory systems on postural stability, Lord and colleagues [5] found that poor tactile sensitivity to Semmes-Weinstein monofilaments at the lateral malleolus of the dominant ankle was associated with increased body sway. They also reported that peripheral sensation is the most important sensory system in the maintenance of static postural stability. Richardson et al. [39] found a cross-sectional relationship between peripheral neuropathy diagnosed by electromyography in a small referral population and a self-reported history of falls during the past year. Using longitudinal data from the Health, Aging and Body Composition study, Strotmeyer et al. [16] found that insensitivity to the 10 g Semmes Weinstein monofilament at the great toe was associated with lower quadriceps muscle strength. Thus, a number of neuromuscular deficits are associated with peripheral neuropathy and may lead to mobility impairments and falls.

Our longitudinal results are most consistent with those of Lord et al. [7] and Luukinen et al. [8] who found relationships between various modalities of sensory loss at the knees and recurrent falls in community-dwelling populations followed prospectively with fall questionnaires or diaries over 1 or 2 years, respectively. In Lord et al’s study, touch sensation did not discriminate non-fallers from single or recurrent fallers. However, in another one-year prospective study of older adults living in an intermediate care facility, Lord and Clark [40] found that tactile sensitivity using a Semmes Weinstein Pressure Aesthesiometer did discriminate between fallers and non-fallers.

The strength of the current study lies in its 5-year longitudinal design and rigorous detection of falls using state-of-the-art falls calendars and follow-up phone calls. There are several weaknesses to the current study. Participants did not receive a full neuropathic evaluation and the monofilament assessment of sensory function is only a quick screen that is specific, but not very sensitive to neuropathic sensory loss [41]. However, this simple, widely available, bed-side test of sensation on the dorsum of the great toe was sensitive to changes over time and predictive of a decrease in gait speed and increase in fall rate, even in the absence of more rigorous assessments of vibratory, proprioceptive, and motor nerve function. Although observed relationships could be confounded by underlying diseases such as diabetes, which affect sensation, gait and falls, the findings persisted after multivariate adjustments.

Our longitudinal results provide validation of the belief that peripheral sensory loss is a risk factor for mobility impairment. They also suggest that a decline in touch sensation at the great toe over time may be a more important predictor of slow gait speed and falls than a baseline sensory deficit. Therefore, simple, repeated clinical assessments of sensory function in the feet of older adults may be helpful in identifying and treating those at risk of falls to prevent their morbid consequences.