Subjective Motoric Complaints and New Onset Slow Gait

Abstract

Background

While reports of mobility problems are common with aging, their relationship to new onset of slow gait is unknown. Our objective was to examine the validity of subjective motoric complaints for predicting the incidence of slow gait.

Methods

Ambulatory community-residing participants (mean age 76.6, 55% women) with gait speeds in the normal range enrolled in an aging cohort. Five subjective motoric complaints were assessed. Incident slow gait (walking speed 1 SD below age and sex means) was the primary outcome.

Results

Of the 548 participants at baseline, 90 had prevalent slow gait and 253 participants (73.7%) reported one or more subjective motoric complaints. Subjective motoric complaints were more common in women than men (1.78 vs 1.23). Over a median follow-up of 3.34 years, 68 participants developed new onset slow gait. All 5 questions predicted incident slow gait (adjusted hazard ratios varying from 2.26 to 4.44). More subjective motoric complaints were associated with increased risk of developing incident slow gait (hazard ratio per complaint 1.81). Predictive validity of subjective motoric complaints for incident slow gait was unchanged when using alternate outcome definitions, accounting for diagnostic misclassification, recall bias, or adjusting for multiple confounders.

Conclusions

Subjective motoric complaints are a harbinger of mobility disability, and can help improve clinical risk assessments and identify high-risk individuals for interventions to prevent onset of slow gait.

Slow gait has emerged as an important and robust marker of health in aging. Presence of slow gait is included in definitions of dysmobility, sarcopenia, frailty, and disability (1–4). Slow gait has well-documented predictive value for falls, strokes, dementia, and death (5–8). A number of potentially modifiable risk factors for new onset slow gait in aging such as physical inactivity, cognitive impairment, muscle weakness, pain, obesity, visual loss, and falls have been identified (9). Intervening in early stages of disablement has been shown to improve functional outcomes in older individuals (10). Hence, early identification of older adults at risk of developing slow gait will be of assistance to clinicians in risk assessment and instituting preventive measures early.

Individuals may perceive and report difficulty in performing mobility-related tasks before developing objective mobility limitations (11). Subjective motoric complaints related to lower extremity function are common but not ubiquitous with aging with prevalence ranging from 20% to 40% for different motoric complaints at cross-section in an older community sample (12), suggesting that these motoric symptoms may be age related but are not age associated. Self-reported walking speed is correlated with actual walking speed, though significant variability exists. For instance, only about a third of community-dwelling older individuals who self-reported difficulty walking met objective slow gait criteria at the same visit (13). Other motoric complaints that assess impression of change in walking patterns over time have been shown to discriminate individuals with and without slow gait at cross-section (12). However, whether subjective motoric complaints related to lower extremity function and walking in ambulatory individuals precede objective signs of gait slowing in aging, and predict future development of slow gait is not established. Furthermore, the number or type of subjective motoric complaints related to lower extremity function that may impact risk of slow gait is unknown. In order to address these knowledge gaps, we examined the association of subjective motoric complaints in older individuals with gait speeds within the normal range (14) at baseline with incidence of slow gait who were identified from a prospective community-based cohort. We hypothesized that individual perceptions of walking difficulties or change in walking patterns will presage the development of slow gait in older adults.

Method

Participants

We studied community-residing adults aged 65 years and older enrolled in the “Central Control of Mobility in Aging” (CCMA) study (15). The goal of CCMA is to determine cognitive control of mobility. CCMA procedures are reported (15–18). Potential participants were identified by a convenience sample from local voter registration lists. Eligibility was assessed using a structured telephone interview including cognitive screeners to rule out dementia (16). CCMA exclusion criteria included inability to speak English, unable to ambulate independently, previously diagnosed dementia, significant visual or hearing loss, or active neurological or psychiatric disorders. Eligible individuals were scheduled for in-person visits at our center at baseline and annual follow-up visits. Study protocols were approved by the local institutional review board and conform to the provisions of the Declaration of Helsinki. Written informed consents were obtained from all participants prior to enrollment.

Subjective Motoric Complaints

See Table 1 for the 5 questions and responses related to lower extremity function and walking used in this study. We selected 4 questions from the Telephone Mobility Assessment questionnaire (TMAQ) that has moderate to excellent sensitivity (54%–78%) and specificity (64%–90%) in discriminating slow gait at cross-section in our catchment area (12). The TMAQ has been administered over the telephone and in person (13,16). We also included a question about difficulty climbing up or down stairs that is highly correlated with objective stair negotiation times (19,20). These 5 questions are included in national mobility and disability surveys (21,22). Test development for both questionnaires has been described (12,19). In brief, potential items for the TMAQ were selected from a large pool of mobility complaints identified by literature review, and a final set of 4 questions were selected by a multidisciplinary panel of experts. Based on sampling distributions, literature review and criteria used in national surveys abnormal responses (Table 1) were categorically defined as a walking distance of less than ¼ mile on questions 1 and 2. Self-reported difficulty in walking was considered abnormal on question 3. More or severe difficulty compared to 5 years ago was termed abnormal on question 4. Reporting difficulty on either climbing up or down stairs was abnormal on question 5, and was more strongly associated with activity limitations than either question alone in our local population (19). We also summed positive responses to individual questions to derive a subjective motoric complaint count (range 0–5).

Table 1.

Baseline Characteristics of Study Sample Overall as Well as by Sex

	Overall (n = 548)	Men (n = 244)	Women (n = 304)	p Value (Men vs Women)
Age	76.62 ± 6.46	76.32 ± 6.59	76.86 ± 6.36	.332
Education, years	14.59 ± 2.94	15.05 ± 3.14	14.23 ± 2.71	.001
Ethnicity
White, %	79.7	87.3	73.9	.001
Black, %	16.4	10.2	21.5
Other, %	3.7	2.5	4.6
Gait speed, cm/s	98.27 ± 22.46	100.33 ± 20.78	96.63 ± 23.62	.055
Subjective motoric complaints
1. How far can you walk without a break?a % < ¼ mile	18.4	10.2	25.0	<.001
2. How far can you walk in an hour?b % < ¼ mile	11.5	6.1	15.8	<.001
3. Do you have any difficulty in walking?c % Yes	23.9	18.9	28.0	.013
4. How is your walking compared to 5 years ago?d % with more difficulty	67.7	61.9	72.4	.009
5. Do you have trouble climbing up or down stairs?e % Yes	31.9	25.4	37.2	.003
Subjective motoric complaint count (range 0–5)	1.53 ± 1.44	1.23 ± 1.29	1.78 ± 1.51	<.001
Subjective motoric complaints ≥1, %	73.7	67.2	78.9	.002
Covariates
Clinical gait abnormality, %	44.9	40.6	48.4	.069
Fall in last 12 mo, %	19.3	17.6	20.7	.361
Pain, %	63.1	56.1	68.8	.002
MCI, %	13.9	14.3	13.5	.773
Multi-morbidity score (0–10)	1.6 ± 1.1	1.5 ± 1.1	1.6 ± 1.1	.274
Obesity (BMI ≥ 30), %	34.9	34.4	35.2	.851
Depressed, % GDS ≥10	11.3	8.6	13.5	.073
Physically inactive, %	34.1	32.0	35.9	.340
Subjective weak grip, %	23.2	15.2	29.6	<.001
Poor vision, %	27.4	25.8	28.6	.465
RBANS	91.39 ± 11.86	90.94 ± 10.56	91.74 ± 12.81	.425
Fair or poor self-rating of health, %	7.9	6.8	8.9	.368

Notes: All values are means ± SD unless otherwise specified. See Methods for definitions. BMI = Body mass index; GDS = Geriatric Depression Scale; MCI = Mild cognitive impairment syndrome; RBANS = Repeatable Battery for the Assessment of Neuropsychological Status.

^aResponses include less than quarter of a mile, half mile or less, less than a mile, 1–2 miles, and 2 or more miles. Less than quarter of a mile abnormal. ^bResponses include less than quarter of a mile, half mile or less, less than a mile, 1–2 miles, and 2 or more miles. Less than quarter of a mile abnormal. ^cResponses include “yes” or “no.” No abnormal. ^dResponses include not at all or very little, little difficulty, more difficult and severe difficulty. More difficult and severe difficulty abnormal. ^eResponses include “yes” or “no.” No abnormal.

Slow Gait

Quantitative gait assessments were performed by research assistants using an instrumented mat (GAITRite, CIR systems, USA) at study visits. Previous studies have shown strong validity and reliability for the GAITRite equipment (6,7,23). The system includes a 20-foot electronic mat imbedded with sensors connected to a computer. Subjects were asked to walk over the mat at their normal walking pace wearing comfortable footwear in a well-lit room. The walk was started and finished 4 feet from the recording field to allow for initial acceleration and terminal deceleration. The footfalls were recorded electronically, and data used to compute speed (centimeters/second).

Our primary outcome was incident slow gait defined as gait speeds 1 SD or more below age and sex means. Slow gait cutscores for women below age 75 were 84.70 cm/s and those aged 75 years and older were 66.10 cm/s. Cutscores in men were 86.20 cm/s and 76.40 cm/s, respectively. The same procedure was used to define slow gait in other cohorts (9,24); and is similar to how cutscores on neuropsychological tests is used to define cognitive impairment. Slow gait at baseline in our sample was associated at cross-section with difficulty performing daily activities such as climbing stairs (odds ratio [OR] adjusted for age, sex, and education [OR] 3.10, 95% confidence intervals [CI] 1.94–4.97) and bathing (OR 2.64, 95% CI 1.60–4.38); confirming the clinical relevance of this outcome.

Other Variables

Risk factors for new onset slow gait such as physical inactivity, presence of mild cognitive impairment syndrome (MCI), muscle weakness, pain, obesity (body mass index ≥30 kg/m²), poor vision (visual acuity > 70 in either eye), and falls in previous 12 months were assessed (9). Together, these 7 risk factors accounted for 77% of the Population Attributable Risk for incident slow gait in a nationally representative sample (9). Questionnaires were administered by research assistants at study visits using standardized forms to collect information on subjective measures included as covariates. Physical inactivity was defined as responding “yes” to the National Health Interview Survey question: “Have you been less physically active over the last year?” (25,26) MCI was diagnosed at consensus conferences using established criteria after reviewing all available clinical and neuropsychological data (27,28). Muscle weakness was defined as subjective report that the participant’s grip felt weak. Pain was assessed by the 7-item pain questionnaire from the Medical Outcome Study (29). Participants responding to the question: “How much bodily pain have you generally had during the past 4 weeks?” as “very mild” to “very severe” were considered to have pain. General mental status was measured by total score on the Repeatable Battery for the Assessment of Neuropsychological Status (RBANS); administered as part of the cognitive evaluation done by research assistants under the supervision of a licensed neuropsychologist at the in-person visits (30). A single-item subjective self-health rating measure (“In general, would you say your health is: excellent, very good, good, fair or poor?”) was used.

We also documented presence of clinical gait abnormalities, comorbid illnesses and depression. Clinical gait abnormalities were determined by study clinicians following visual inspection of gait patterns while participants walked up and down a well-lit hallway at their normal pace (31). Study clinicians determined whether gaits were normal or abnormal, and if abnormal, subtyped as neurological or non-neurological (31). A multi-morbidity score (range 0–10) was calculated by summing self-reported physician diagnosed diseases including, diabetes, angina, myocardial infarction, chronic heart failure, hypertension, major depression, stroke, Parkinson’s disease, chronic obstructive lung disease, and arthritis (16). Symptoms of depression were assessed with the 30-item Geriatric Depression Scale administered by research assistants (32). A validated cutscore of ≥10 was used to define depression (32).

Data Analysis

Baseline characteristics were examined using descriptive statistics. The prevalence of subjective motoric complaints is presented for the overall sample as well as by sex (men and women). At baseline, we examined the association of risk factors for slow gait with the prevalent subjective motoric complaints using logistic regression models adjusted for age, sex, and education. Cox proportional-hazards models were utilized to evaluate the association of subjective motoric complaints (individual items and count) with risk of incident slow gait, adjusted for age, sex, and education. Hazard ratios (HR) adjusted for age, sex, and education with 95% CI were reported. We reran the model including the 7 risk factors for slow gait (9) as well as clinical gait abnormalities, multi-morbidity score, depression, and subjective self-health measure. Time to event was from entry to incident slow gait or final contact. Proportional hazards assumptions were met graphically and analytically. All analyses were conducted using SPSS version 25.0 (Armonk, NY: IBM Corp).

While the slow gait definition was consistent with our previous studies (9,18), we redid our analyses using alternate definitions to confirm reliability of our findings. We redefined slow gait as 1.5 SD below age and sex means. An alternate approach to define slow gait is to use a single cutscore (1,33), but there is no consensus on the value (1,34). Single cutscores also discriminate against individuals in older age groups and show racial/ethnic differences (35,36). But, to facilitate comparisons, we examined incident slow gait defined by one of the recommended cutscores (≤80 cm/s) (33).

We conducted sensitivity analyses. We excluded incident slow gait cases (using the 1 SD cut) in the first year of follow-up to account for diagnostic misclassification of individuals who were on the cusp of the slow gait criteria at baseline. We repeated the survival analysis excluding 3 persons who reported ever using walking aids at baseline. Of note, all participants completed our in-house gait speed test without using any walking aids. While our participants in the incidence analysis had gait speeds within the normal range (14) (mean 104.42 ± 18.67 cm/s) at baseline, there was variability in gait speeds. Hence, we reran the model adjusting for baseline gait speed. To account for potential recall bias or underreporting of subjective motor complaints by cognitively impaired individuals, we reran survival models excluding participants with MCI at baseline. We then adjusted for baseline RBANS score to account for cognitive status in this initially dementia and MCI free subsample.

Results

Study Population

Data collection began in June 2011 and ended August 2018. Of the 588 potential CCMA participants at baseline, 9 individuals who met criteria for dementia at consensus case conferences after review of all clinical and neuropsychological data at baseline were excluded. In addition, 29 participants who did not complete the mobility questionnaire at baseline and 2 participants without gait speed were excluded. Hence, there were 548 participants eligible for this analysis (55% women, mean age 76.6 years). All participants were independently ambulatory in the community at entry. Baseline characteristics of the sample are given in Table 1. Men had higher education than women. Women reported more pain and weaker grip more than men. There were no group differences in other covariates examined. Of the 548 participants in the baseline sample, 90 had slow gait at baseline, and were excluded from the incident analysis.

Subjective Motoric Complaints

Table 1 presents the prevalence of the 5 subjective motoric complaints overall and by sex. Change in walking pattern compared to 5 years ago was the most common subjective motoric complaint endorsed at baseline (67.7%). Reporting inability to walking more than ¼ mile was the least common subjective motoric complaint (11.5%). Women (range 15.8%–72.4%) reported subjective motoric complaints more often than men (range 6.1%– 61.9%). On the average, at baseline, men reported 1.23 subjective motoric complaints, whereas women reported 1.78 subjective motoric complaints. Age differences were seen in subjective motoric complaints. On the average, those aged 80 years and older reported 1.97 motoric complaints and those younger than 80 reported 1.36 motoric complaints.

Table 2 shows that all 5 subjective motoric complaints at baseline were associated with most but not all risk factors. For instance, none of the subjective motoric complaints were associated with presence of MCI or poor vision. The strength of the associations varied.

Table 2.

Prevalent SMC and Risk Factors at Baseline

Risk Factors	SMC 1	SMC 2	SMC 3	SMC 4	SMC 5
Clinical gait abnormality	4.34 (2.58–7.29), <0.001	4.60 (2.38–8.89), <0.001	6.29 (3.90–10.15), <0.001	2.04 (1.36–3.04), 0.001	3.87 (2.59–5.78), <0.001
Falls last 12 months	1.93 (1.14–3.25), 0.014	1.75 (0.94–3.24), 0.073	2.12 (1.32–3.40), 0.002	2.45 (1.43–4.22), 0.001	1.93 (1.24–3.01), 0.003
Pain	1.91 (1.14–3.19), 0.013	1.89 (1.01–3.54), 0.046	4.02 (2.40–6.72), <0.001	2.31 (1.56–3.41), <0.001	2.07 (1.38–3.10), <0.001
MCI	0.98 (0.52–1.85), 0.960	1.30 (0.64–2.66), 0.459	0.84 (0.47–1.51), 0.574	1.69 (0.90–3.20), 0.101	1.24 (0.73–2.08), 0.419
Multi-morbidity score	1.62 (1.30–2.01), <0.001	1.35 (1.05–1.73), 0.017	1.58 (1.30–1.91), <0.001	1.51 (1.26–1.8), <0.001	1.30 (1.10–1.54), 0.002
Obesity	2.95 (1.83–4.75), <0.001	2.31 (1.32–4.04), 0.003	2.81 (1.84–4.30), <0.001	2.53 (1.65–3.88), <0.001	2.38 (1.62–3.50), <0.001
Depressed	1.77 (0.95–3.31), 0.071	1.43 (0.68–3.00), 0.344	2.15 (1.22–3.79), 0.008	4.51 (1.88–10.86), 0.001	2.00 (1.16–3.45), 0.012
Physically inactive	2.74 (1.72–4.34), <0.001	2.63 (1.51–4.56), 0.001	2.94 (1.94–4.45), <0.001	3.60 (2.26–5.74), <0.001	2.66 (1.81–3.89), <0.001
Subjective weak grip	1.20 (0.72–2.00), 0.468	1.66 (0.93–2.97), 0.084	2.15 (1.37–3.37), 0.001	4.69 (2.52–8.72), <0.001	2.16 (1.42–3.29), <0.001
Poor vision	1.24 (0.76–2.02), 0.380	0.76 (0.41–1.41), 0.393	1.19 (0.76–1.85), 0.429	0.67 (0.44–1.03), 0.074	0.81 (0.53–1.23), 0.339
Fair or poor self-rated health	9.77 (4.69–20.33), <0.001	8.72 (4.07–18.69), <0.001	5.91 (2.99–11.68), <0.001	1.07 (1.04–1.11), <0.001	1.04 (1.01–1.07), 0.015
Gait speed	0.95 (0.94–0.97), <0.001	0.95 (0.94–0.97), <0.001	0.95 (0.94–0.97), <0.001	0.97 (0.96–0.98), <0.001	0.97 (0.96–0.98), <0.001

Notes: Odds ratios with 95% confidence intervals adjusted for age, sex and education. See Methods for definitions. SMC = Subjective motoric complaints.

Incident Slow Gait

Over a median follow-up of 3.34 years (IQR 2.01–4.98), there were 68 cases of new onset slow gait. Blinded clinician evaluations of gait using standardized procedures (13,31) (described above) were done in 42 of the 68 new onset slow gait cases, and diagnoses included normal gait in 10 participants, non-neurological in 11, combined neurological and non-neurological in 9, and neurological gait abnormalities in 12. For those endorsing subjective motoric complaints at entry, median time to slow gait ranged from 1.13 years for those reporting that they could walk less than ¼ mile in 1 hour to 2.22 years for those reporting more difficulty walking compared to 5 years ago.

Table 3 shows that all 5 subjective motoric complaints (SMC) predicted incident slow gait (HR varying from 2.26 to 4.43). Reporting inability to walking more than ¼ mile (SMC 2) had the strongest association with incident slow gait in men (HR 39.12), and change in walking pattern compared to 5 years ago (SMC 4) in women (HR 7.97).

Table 3.

SMC and Risk of New Onset Slow Gait

Subjective Motoric Complaints	Overall (n = 458)	Men (n = 203)	Women (n = 255)
SMC 1. How far can you walk without a break?a	4.43 (2.52–7.78), <0.001	5.33 (2.10–13.53), <0.001	4.20 (2.05–8.59), <0.001
SMC 2. How far can you walk in an hour?b	3.87 (1.77–8.44), 0.001	39.12 (4.43–345.10), 0.001	3.35 (1.45–7.75), 0.005
SMC 3. Do you have any difficulty in walking?c	4.26 (2.55–7.09), <0.001	3.62 (1.64–8.00), 0.001	5.06 (2.51–10.19), <0.001
SMC 4. Do you find it more difficult to walk now compared to 5 years ago?d	3.21 (1.61–6.39), 0.001	2.07 (0.91–4.68), 0.080	7.97 (1.86–34.09), 0.005
SMC 5. Do you have climbing up or down stairs?e	2.26 (1.37–3.73), 0.001	2.18 (1.02–4.66), 0.042	2.41 (1.21–4.78), 0.012

Notes: Cox models reported as hazard ratios with 95% confidence intervals adjusted for age, sex and education. See Methods for definitions of abnormal responses on individual questions. SMC = Subjective motoric complaints.

^aResponses include less than quarter of a mile, half mile or less, less than a mile, 1–2 miles, and 2 or more miles. Less than quarter of a mile abnormal. ^bResponses include less than quarter of a mile, half mile or less, less than a mile, 1–2 miles, and 2 or more miles. Less than quarter of a mile abnormal. ^cResponses include “yes” or “no.” No abnormal. ^dResponses include not at all or very little, little difficulty, more difficult and severe difficulty. More difficult and severe difficulty abnormal. ^eResponses include “yes” or “no.” No abnormal.

Increasing number of subjective motoric complaints was associated with increased risk of incident slow gait (HR per complaint adjusted for age, sex, and education 1.82, 95% CI 1.53–2.16, p = <.001). The subjective motoric complaint count still predicted slow gait after additional adjustments for eleven risk factors in the model (aHR 1.40, 95% CI 1.12–1.74, p = .003)—see Supplementary Table 1. Table 4 shows that higher number of subjective motoric complaints was associated with increased HR (ranging from 7.54 for 3 complaints endorsed to 15.26 for all 5 complaints endorsed) of developing new onset slow gait.

Table 4.

SMC Count and New Onset Slow Gait

SMC count	N (Incident slow gait)	HR, 95% CI, p-value
0	134 (10)	Reference
1 only	171 (24)	1.57, 0.75–3.33, .234
2 only	76 (7)	1.45, 0.53–3.95, .470
3 only	41 (13)	7.54, 3.21–17.75, <.001
4 only	16 (7)	8.64, 3.00–24.83, <.001
5 only	20 (7)	15.26, 5.39–43.17, <.001

Note: Cox models reported as hazard ratios with 95% confidence intervals adjusted for age, sex, and education. CI = Confidence interval; SMC = Subjective motoric complaints.

Compared to those with no subjective motoric complaints, those with 3 to 5 subjective motoric complaints (HR adjusted for age, sex, and education 8.96, 95% CI 4.12–19.46) but not those with 1 to 2 subjective motoric complaints (adjusted HR 1.54, 95% CI 0.75–3.19) were at higher risk of developing slow gait. Survival plot (Figure 1) shows probabilities for incidence of slow gait based on subjective motoric complaint status at baseline.

Figure 1.

Kaplan–Meier hazard function for categories of subjective motoric complaints.

Alternate Slow Gait Definitions

Analyses to examine alternative definitions of slow gait including 1.5 SD below age and sex means and a global cut score of ≤80 cm/s had similar results to the main findings for each of the individual subjective motoric complaint questions (data not shown), except for difficulty climbing up or down stairs, which did not attain significance (p = .189). The subjective motoric complaint count predicted incident slow gait defined as 1.5 SD below age and sex means (n = 52) (aHR 2.01 95% CI 1.64–2.46, p < .001) as well as incident slow gait defined as ≤80 cm/s (n = 18) (aHR 1.93 95% CI 1.36–2.73, p < .001).

Sensitivity Analysis

After excluding 18 individuals who developed incident slow gait in the first year of follow-up to account for diagnostic misclassification, the subjective motoric complaint count still predicted slow gait (aHR 1.60, 95% CI 1.27–2.00, p < .001). The subjective motoric complaint count predicted slow gait even after additional adjustment for baseline gait speed (aHR 1.55 95% CI 1.30–1.85, p < .001) or after excluding 3 individuals who reported ever using walking aids (aHR 1.84, 95% CI 1.40–2.41, p < .001). Excluding participants with MCI at baseline (n = 76), the subjective motoric complaint count remained a predictor of incident slow gait (aHR 1.87, 95% CI 1.54–2.25, p < .001). Further adjustment for general mental status using the RBANS in this initially MCI and dementia-free subsample did not effect the association of the subjective motoric complaint count with incident slow gait (aHR 1.85, 95% CI 1.53–2.24, p > .001).

Discussion

Subjective motoric complaints predicted development of incident slow gait in a community-based sample of ambulatory adults aged 65–95 years with gait speeds in the normal range for age and sex at baseline. All 5 subjective motoric complaints individually predicted new onset slow gait. Furthermore, increasing count of subjective motoric complaints was associated with a higher risk of developing slow gait; indicating a dose response effect. Those with 3 or more subjective motoric complaints in our cohort were at the highest risk of developing incident slow gait with an over 8-fold increase in risk compared to those without complaints. Given the importance of slow gait as a marker for function and harbinger of multiple adverse geriatric outcomes, these findings have implications in identifying high-risk individuals for further assessment and treatment.

Our results shed light on the natural history of slowing of gait in aging; older individuals perceive changes in their walking patterns in preclinical stages before developing slow gait. Subjective motoric complaints might capture the earliest stages of disability, and before developing objective or clinical signs of slowing of gait in aging (37). The temporal sequence of motoric symptoms in preclinical stages followed by the development of motoric signs of slow gait in our cohort is in line with findings in other geriatric syndromes such as dementia and disability. Subjective cognitive complaints, while also common with aging, have been correlated with Alzheimer pathology and may predict cognitive decline measured by objective tests (38,39). Women ages 70–80 years who reported modifying the way they carried out activities of daily living were found to be at higher risk of developing mobility disability (defined as self-report of difficulty in walking ¼ mile or climbing 10 steps) (11).

The prevalence of subjective motoric complaints was variable with change in walking pattern being the most common complaint endorsed. Interestingly, despite having gait speeds within the normal range, 19% of ambulatory participants in this cohort reported difficulty walking and 28% of these participants were diagnosed to have normal gaits by experienced clinicians. The prevalence rates for various motoric complaints are in the range of other national and population surveys. National health statistics indicate that 14% of U.S. adults aged 65–74 years and 27% of adults aged 75 and older report being unable or very difficult to walk a quarter mile (26). Mobility was the most common disability with 27% of U.S. adults aged 65 years and older reporting serious difficulty walking or climbing stairs in a 2016 survey (40). Subjective motoric complaints were more frequent in the 80 years and older age group, which parallels the increasing prevalence of mobility disability with aging (14,25,26,37).

Subjective motoric complaints were endorsed by women more than men in our cohort, which corresponds to the higher rates of mobility disability and disability reported in older women compared to older men (41). Interestingly, the motoric complaints that were most strongly associated with slow gait were different for men and women, suggesting that there are sex differences in the perception or evolution of mobility loss (42). Both sexes were found to report their disability accurately in a previous study (42), although, similar to our results, more women than men report disability and functional limitations (42).

Subjective motoric complaints were related to previously identified modifiable risk factors for slow gait (9). The strength of the association of the complaints with individual risk factors varied suggesting the complaints might be differentially capturing pathways to slow gait. The median time to onset of slow gait varied by complaint though the duration of complaints was not ascertained at baseline. Longitudinal follow-up is required to elucidate temporal sequences in the origin of various subjective complaints. Subjective motoric complaints predicted slow gait even after adjusting for covariates that account for the majority of the population attributable risk for new onset slow gait (9); supporting its incremental value over current risk assessments for mobility disability. Furthermore, this finding raises the possibility that subjective complaints may also tap into other unmeasured risk factors or pathways for mobility disability (37). These findings provide insights into potential intervention strategies that may be investigated in future studies.

Subjective motoric complaints were examined in this study to examine the natural history of mobility disablement and assess its role as a risk predictor for incident slow gait (and not as a cross-sectional diagnostic screen). The lower sensitivity and higher specificity of subjective motoric complaints for dysmobility at cross-section has led to investigators to suggest that older adults are more likely to accurately report what they cannot do than what they can do (43,44). But, in the context of prognostic tests, experts have stated that commonly used discriminative statistics for evaluating diagnostic tests, such sensitivity and specificity, are not as informative (45,46). The high predictive validity for incident slow gait of subjective motoric complaints and across different risk/dose levels supports it as a prognostic indicator (45,46).

Strengths include the large cohort with standardized mobility questionnaires and assessments as well as utilization of a prospective design to test temporal associations between symptoms and signs of mobility disability. We accounted for diagnostic misclassification and recall bias, adjusted for a number of confounders, and conducted a number of secondary analyses that corroborate our main findings. We used an objective marker to define the mobility outcome in contrast to previous studies (11). Our analysis accounted for the overlap of subjective and objective features of slow gait in this ambulatory sample by excluding those who converted to slow gait within the first year as well as accounting for baseline gait speed. Potential limitations are noted. Subjective motoric complaints may show racial/ethnic differences in prevalence, and should be examined in diverse populations. The subjective motoric complaints were systematically derived from a large pool of mobility items (12), but other unexamined motoric complaints may also predict slow gait. Our focus in this analysis was on subjective complaints related to lower extremity function and walking, but in other contexts complaints related to the upper limb could be considered. Our sample is reflective of the demographics of the older population in our catchment area in lower Westchester county but was recruited as a convenience sample and not a representative population. Our sample was mostly Caucasian. Ethnic differences in objective gait speed have been reported (35,36), and should also be explored for subjective motoric complaints. Furthermore, the applicability of our findings in younger age groups should also be examined though slowing of gait occurs at higher rate in later decades (14,37). We examined alternate definitions of slow gait but acknowledge that there are other approaches to defining mobility such as distance covered in fixed time (33,43,44). These different mobility tests are highly correlated (5,33), but should also be examined in future studies.

Our results have potential for research and clinical applications. These 5 questions can be administered in person or remotely to identify individuals at risk for mobility disablement for research studies or clinical interventions. Remote mobility assessment has become important in the context of the ongoing Covid-19 pandemic. These 5 subjective motoric complaints can be easily and quickly elicited in less than 5 minutes in clinical settings to aid decision making. Endorsing subjective motoric complaints, especially more than one, can prompt clinicians to do further testing to ascertain and treat potential modifiable risk factors. Knowledge of their risk status may encourage patients to engage in healthy lifestyle and preventive activities.

Funding

This work was supported by the National Institute on Aging at the National Institutes of Health (grant number RO1 AGO57548).

Author Contributions

J.V. developed the idea, obtained funding, and prepared the initial draft. E.A. developed the database, conducted analyses, and edited the manuscript.

Conflict of Interest

None declared.