Medication status and gait mechanics in older adults: a multivariate analysis

Key points

1. 01Participants on 4 or more medications had lower gait velocity (mean 66.4 cm/s) and shorter stride length (mean 81.9-82.0 cm) compared to those on 3 or fewer medications (velocity 72.7 cm/s, stride 88.1-88.5 cm)
2. 02Double-support time was significantly longer in the high-medication group (mean 0.50 vs 0.43 of gait cycle, F=6.886, P=.009), indicating a more cautious, unstable gait pattern
3. 03Cadence did not significantly differ between groups (P=.076), suggesting clinicians should prioritize stride length and walking speed over step rate when assessing medication-related fall risk
4. 04Multivariate analysis of pace and rhythm gait factors both showed significant differences by medication group (pace: F=2.416, P=.048; rhythm: F=2.315, P=.075)
5. 05Findings support considering de-prescribing as a strategy to reduce fall risk in older adults

How it was conducted

Design

Cross-sectional observational study with multivariate analysis

Participants

384 adults older than 60 years (mean age 73.2 years, SD 4.2) with self-reported history of at least 1 fall in the past 3 years, recruited across the Southwest United States

Groups

4 or more medications (n=262) vs 3 or fewer medications (n=122)

Gait measurement

Walkway gait analysis system (Tekscan) capturing cadence, gait velocity, stride length, swing time, and double-support time at 100 Hz

Primary analysis

Factor analysis to identify pace and rhythm gait factors, followed by multivariate analysis of variance (MANOVA) with Bonferroni-corrected univariate follow-ups

Primary outcome

Differences in pace factor (cadence, gait velocity, stride length) and rhythm factor (swing time, double-support time) between medication groups

What they found

• Multivariate test for pace factor: F(4,379)=2.416, P=.048, effect size=0.025, power>0.80
• Multivariate test for rhythm factor: F(3,380)=2.315, P=.075, effect size=0.018, power=0.67
• Univariate: gait velocity differed significantly between groups (F=5.952, P=.015)
• Univariate: right stride length differed significantly (F=5.517, P=.019)
• Univariate: left stride length differed significantly (F=5.932, P=.015)
• Univariate: total double-support time differed significantly (F=6.886, P=.009)
• Univariate: cadence did not differ significantly (F=3.156, P=.076)
• Univariate: right swing time did not differ significantly (F=1.728, P=.189); left swing time did not differ (F=0.097, P=.755)
• Mean gait velocity: 66.4 cm/s (4+ medications) vs 72.7 cm/s (fewer than 4 medications)
• Mean double-support time: 0.50 of gait cycle (4+ medications) vs 0.43 (fewer than 4 medications)

Limitations

• Respondent bias: medication status, fall history, and medical history were self-reported, not verified from medical records
• Inclusion of participants with lower extremity impairments may have lowered pace factor values compared to prior studies, potentially influencing results
• Cross-sectional design prevents causal inference; it is unknown whether medications cause gait changes or whether sicker patients both take more medications and walk differently
• The study only classified participants as taking 4 or more vs fewer than 4 medications and did not examine specific medication types or classes

Why it matters

For patients

Older adults who have fallen and take many medications may have measurably worse walking patterns, and reducing unnecessary medications could be one way to lower fall risk.

For clinicians

Rehabilitation assessments for older adults on polypharmacy should emphasize stride length and walking speed rather than cadence, and double-support time is the most relevant rhythm variable to monitor.

For readers

This study adds multivariate evidence that polypharmacy is associated with specific gait impairments in fall-prone older adults, supporting medication review as part of fall prevention strategies.