Balance Training Reduces Falls Risk in Older Individuals With Type 2 Diabetes

Sixteen type 2 diabetic individuals (62.3 ± 5.5 years; average diabetes duration 15.2 ± 2.4 years) and 21 age-matched control subjects (64.7 ± 7.1 year) participated. Exclusion criteria included cardiovascular disease, unstable proliferative retinopathy, end-stage renal disease, uncontrolled hypertension, and/or participation in balance/resistance training during the previous year. All procedures complied with institutional review board guidelines.

Initial assessment included a complete history, physical examination, and full neurologic evaluation that included assessment for somatic/autonomic neuropathy (3). Warm-cold thermal perception, 128-Hz vibration perception, touch, pressure, and prickling pain perception were evaluated. An overall total neuropathy score was also calculated. The average A1C for the type 2 diabetic group was 7.5 ± 0.3%. Following screening, a record of previous falls, balance, reaction time, and falls risk assessments were completed. Individuals then completed a 6-week, thrice-weekly exercise program followed by posttraining evaluations.

Each exercise session consisted of a balance/posture component (e.g., lower-limb stretches and leg, abdominal, and lower-back exercises) and a resistance-/strength-training component (e.g., lower-/upper-limb exercises performed using strength-training machines). Participants performed 1–2 sets of 10–12 repetitions, with rests between exercises.

Risk of falling was determined using the long-form physiological profile assessment (PPA). This validated tool (4) assesses vision, sensation, proprioception, lower-limb strength, postural sway/coordination, and cognitive function.

Participants completed a simple reaction time (SRT) task where upper-limb (finger) and lower-limb (foot) responses were assessed. Individuals responded to a visual cue by depressing a timing switch. Fifteen trials were completed with each segment.

A repeated-measures, generalized linear model was used to assess for group and training effects. Significant effects were further examined using planned contrasts (one-way ANOVAs). Analyses were performed using SAS statistical software (SAS Institute) with P < 0.05.

Diabetic individuals exhibited significant differences in total neuropathy scores (left foot F_1,35 = 9.87; right foot F_1,35 = 8.86; all P < 0.05), BMI (F_1,35 = 22.50; P < 0.05), and percent body fat (F_1,35 = 7.11; P < 0.05). There were no significant group/exercise differences in blood pressure measurements for either sitting (pretraining: control subjects 128.2 ± 3.4/73.4 ± 2.3 mmHg, type 2 diabetes 133.0 ± 4.2/69.1 ± 2.7 mmHg; posttraining: control subjects 123.6 ± 2.8/70.7 ± 2.2 mmHg, type 2 diabetes 130.0 ± 4.6/67.4 ± 3.2 mmHg) or standing (pretraining control subects 125.8 ± 2.8/73.8 ± 2.2 mmHg, type 2 diabetes 133.4 ± 4.1/72.14 ± 3.3 mmHg; posttraining control subjects 119.8 ± 2.9/74.6 ± 2.5 mmHg, type 2 diabetes 127.7 ± 6.0/68.8 ± 3.2 mmHg). No significant group/exercise differences were found for the following measures of autonomic function: expiration-to-inspiration (E:I) ratio (control subjects 1.13 ± 0.02; type 2 diabetes 1.15 ± 0.02), Valsalva maneuver (control subjects 1.25 ± 0.03; type 2 diabetes 1.25 ± 0.07), and stand ratio (control subjects 1.36 ± 0.16; type 2 diabetes 1.48 ± 0.29).

A significant group difference was found for the average number of falls (F_1,35 = 4.44; P < 0.05), with type 2 diabetic subjects experiencing more falls over the past year. No group difference in falls was observed posttraining.

As shown in Fig. 1, type 2 diabetic subjects had a significantly higher falls risk score compared with control subjects (F_1,35 = 20.24; P < 0.05). Following training, both groups exhibited reduced falls risk, but this was only significant for type 2 diabetic individuals (F_1,35 = 33.03; P < 0.05). While no age effects were observed, correlation analysis revealed a significant falls risk–age relationship for the type 2 diabetic group (r = 0.519; P < 0.05).

Analysis of the individual PPA measures showed that type 2 diabetic individuals exhibited reduced proprioception (F_1,35 = 5.89; P < 0.05), sensation (F_1,35 = 5.78; P < 0.05), and ankle strength (F_1,35 = 4.17; P < 0.05) compared with control subjects. Following training, a significant group-by-exercise effect was seen for proprioception (F_1,35 = 4.54; P < 0.05), quadriceps (F_1,35 = 9.11; P < 0.05), and hamstring strength (F_1,35 = 5.07; P < 0.05). Planned contrasts revealed that both groups showed improvements in strength and proprioception postexercise.

There was a significant group difference for hand (F_1,35 = 7.22; P < 0.05) and foot SRT (F_1,35 = 9.64; P < 0.05), with the diabetic group being significantly slower (Fig. 1). Posttraining, a significant improvement in both SRT measures (hand F_1,35 = 11.87; foot F_1,35 = 14.52; all P < 0.05) was found. Planned contrasts revealed that both groups recorded faster hand SRT following exercise. However, only the type 2 diabetic group exhibited significantly faster foot SRT posttraining.

Normal aging is associated with slower cognitive processing (5), slower postural reactions (6), and decreased muscle strength (7), all of which are essential for optimal balance (8). The current study demonstrated that all older individuals showed a decline in SRT and strength, although the decrement was more pronounced for those with diabetes. The decline in function for older diabetic individuals was further compounded since they had a higher previous history of falls and all exhibited mild-to-moderate neuropathy, the latter being associated with increased falls risk (9). Consequently, the type 2 diabetic group was at greater falls risk, confirming the view that increasing age, previous falls history, increased postural sway, and presence of diabetes are major risk factors for falling (1,2,8,10,–12).

Following training, the diabetic group exhibited a significant decline in falls risk, dropping from a mild-to-moderate to a low-to-mild risk of falling. This decline was reflected by improved proprioception and increased hamstring/quadriceps strength. While increasing physical activity can lead to enhanced joint proprioception, a learning effect cannot be ruled out as a contributing factor for the improved lower-limb proprioception. Both groups also demonstrated significant improvements in SRT. The ability to respond quickly to any external perturbation is essential for correcting oneself to avoid possible falls (4,6,11). Unfortunately, many older individuals exhibit slower reaction times (5,11) and are at increased risk of falling since they respond slower under postural situations (6). The improved reaction time with exercise has obvious implications for individuals at high falls risk to correct themselves during balance-threatening situations. While hypoglycemia could be one reason for slower reaction times for the diabetic group (13), any decreased glucose levels would not explain the significantly improved SRTs seen postexercise. While increased strength correlates highly with improved balance and decreased falls risk (14,15), our results show that the benefits of exercise are not limited to muscle function. Rather, training resulted in improvements in a range of falls risk factors, impacting positively on sensory, motor, and cognitive processes.

Overall, this study demonstrated that older type 2 diabetic individuals are at increased falls risk. Following training, the type 2 diabetic group demonstrated improvements in balance, proprioception, lower-limb strength, reaction time, and, consequently, decreased risk of falling. The results support the practice of prescribing mild-to-moderate exercise to individuals with type 2 diabetes to alleviate falls risk.

No potential conflicts of interest relevant to this article were reported.