Changes in Flexed Posture, Musculoskeletal Impairments, and Physical Performance After Group Exercise in Community-Dwelling Older Women

Methods 

Primary Outcome Measures 

Flexed posture 

Three primary outcome measures of flexed posture were identified for this study: thoracic kyphosis, forward head posture, and height. Each measure was repeated during “usual” and “best” posture. “Usual” posture was standardized at full exhalation, whereas “best” posture measurements were performed at full inhalation and participants were cued to “stand as upright and tall as you can.” Participants’ feet were traced to ensure the same positioning on future measurement. Thoracic kyphosis defined as the Cobb angle between the T2 and T3 vertebrae and the T11 and T12 vertebra was measured with a Debrunner kyphometerb by using a standard protocol.4 The researcher placed the kyphometer over appropriate landmarks, and an assistant read and recorded the measurement. Height measurements were performed with a standard protocol by using the Harbinger taximeter.18 Forward-head posture measurements were calculated as the distance from the tip of the tragus to the wall. Mantex Precision calipers,c adapted with a measurement level, were used as participants stood with their heels against a 5.1-cm (2-in) block to prevent trunk leaning. All measurements had good test-retest reliability (table 1), and t2 to t1 test-retest differences were nonsignificant (table 2).

Table 1. Descriptive Characteristics, Measures of Flexed Posture, Strength, ROM, and Physical Performances at Baseline (N=21)

	Characteristics	Mean ± SD or Median	Range (min–max)	ICC3,1 (95% CI)
	Age (y)	72.0±4.2	66–80	NA
	Vertebral compression fractures (n)	2	0–3	NA
	BMD (T score)	−1.5±1.3	−3.6 to 0.7	NA
	Flexed posture
	Kyphosis: usual (deg)	57±5.0	47–65	.93(.87–.99)
	Kyphosis: best (deg)	50±6.0	37–60	.94(.88–.99)
	Forward head: usual (cm)	22.4±1.8	18.8–25.7	.76(.57–.94)
	Forward head: best (cm)	21.3±2.8	16.0–28.5	.86(.75–.97)
	Height: usual (cm)	160.9±7.7	147.5–176.3	1.0(1.0–1.0)
	Height: best (cm)	161.8±7.9	148.4–176.9	1.0(1.0–1.0)
	Strength and ROM
	Biodex extension strength (%BW)	35.2±14.8	9.8–57.4	.84(.71–.96)
	Prone extension (kg)	7.9±1.7	5.1–10.9	.47(.14–.80)
	Shoulder flexion right (deg)	162±12.5	140–180	.82(.69–.96)
	Shoulder flexion left (deg)	159±11.7	135–176	.74(.54–.93)
	Popliteal angle right (deg)	141±9.2	123–155	.85(.74–.99)
	Popliteal angle left (deg)	142±11.0	124–162	.66(.42–.90)
	Modified Thomas test right (deg)	27±6.4	15–37	.87(.76–.97)
	Modified Thomas test left (deg)	27±6.7	13–38	.86(.75–.97)
	Physical performance
	Balance Master SOT (100 points)	73±8	58–86	.65(.40–.90)
	COP velocity (cm/s)	1.5±0.7	0.7–3.6	.86(.74–.97)
	COP excursion (% stance length)	53.6±7.0	38.7–67.3	.79(.62–.95)
	Modified PPT (36 points)	30±2.6	25–34	.73(.53–.93)
	Jug test (s)	13.5±3.3	8.9–22.6	.92(.85–.98)
	Gait speed (m/s)	1.35±0.18	1.03–1.67	.92(.85–.98)

Abbreviations: % BW, percentage body weight; CI, confidence interval; COP, center of pressure; ICC3,1, intraclass correlation coefficient model 3,1; Max, maximum; Min, minimum; NA, not applicable; PPT, Physical Performance Test; SD, standard deviation; SOT, Sensory Organization Test.

Table 2. Differences (t2–t1) and Change (t3–t2) in Dependent Measures

	Dependent Measures	t2–t1	P	t3–t2	P
	Flexed Posture
	Kyphosis: usual (deg)	1(−1to2)	.67	−6±3†	<.001⁎
	Kyphosis: best (deg)	0(−2to2)	.99	−5±3†	<.001⁎
	Forward head: usual (cm)	0.0(−1.0to0.8)	.99	−0.3±1.3	.82
	Forward head: best (cm)	0.3(−0.8to01.3)	.86	−0.5±1.8	.49
	Height: usual (cm)	0.2(−0.2to0.5)	.50	0.2±0.6	.41
	Height: best (cm)	0.1(−0.1to0.4)	.26	0.1±0.4	.98
	Strength and ROM
	Biodex extensor strength (% BW)	−4(−11to4)	.44	21±13†	<.001⁎
	Prone extension (kg)	0.45(−0.45to1.75)	.60	0.45±1.80	.75
	Shoulder flex right (deg)	0(−4to4)	.99	−0.5±6.4	.93
	Shoulder flex left (deg)	−1(−6to5)	.93	2±10	.56
	Popliteal right (deg)	−1(−4to3)	.83	7±7†	<.001⁎
	Popliteal left (deg)	0(−5to5)	.99	9±10†	<.001⁎
	Modified Thomas test (deg)	0(−2to3)	.97	4±5	.003
	Modified Thomas test (deg)	0(−3to2)	.91	2±6	.38
	Physical performance
	Balance Master SOT (100 points)	3(0to6)	.05	2±7	.40
	COP velocity (cm/s)	0.1(−0.1to0.3)	.45	−1.0±5	.77
	COP excursion (% stance length)	0.9(−3to5)	.81	−0.07±0.4	.62
	Jug test (s)	−0.5(−1.3to0.2)	.21	−1.4±1.3†	.001⁎
	Modified PPT (36 points)	1(0to2)	.23	2±2†	.001⁎
	Gait speed (m/s)	.02(−.03to.07)	.49	.05±.10	.06

NOTE. t1 is time 1, t2 is time 2 pretest, and t3 is time 3 posttest measurements. Data for t2 to t1 are reported as mean change (95% CI). Data for t3 to t2 are reported as mean change ± SD.

⁎ Significant difference t3–t2 at P<.025 (by paired t test). † t3 to t2 change meets or exceeds the upper and lower limits of the 95% CI for t2 to t1 measurement variability.

Results 

Thirty-six participants were enrolled in the study after meeting all eligibility requirements. Eleven participants withdrew before the intervention phase: 5 participants changed their mind, 3 sustained nonstudy related injuries, 2 had balance problems and could not safely participate in the group intervention, and 1 admitted to modifying her “usual” posture at screening and no longer qualified. Of the 25 who began the intervention, 3 participants withdrew during the intervention phase because of nonstudy-related injuries and 1 withdrew because of a family emergency. There were no injuries associated with study participation. Twenty-one participants completed the study and all 24 exercise sessions (see table 1). Participants who withdrew during the intervention were no different in age, kyphosis, or BMD T score than those who completed the study. Participants completing the study had a mean age ± standard deviation (SD) of 72±4.2 years, with lumbar spine BMD T score –1.5±1.3 and a median number of 2 VCFs.

Discussion 

These results support our primary hypothesis that improvements in flexed posture and physical performance would be observed after a 12-week multidimensional group exercise in women 65 years and older with flexed posture. The 6° change in “usual” kyphosis represents an 11% improvement, and the 5° change in “best” kyphosis represents a 10% improvement from baseline. This exceeds that described by Itoi and Sinaki11 after a 2-year prone trunk extension intervention. It is consistent with that reported after a 6-month use of passive postural bracing.13 This improvement in kyphosis in 12 weeks exceeds the amount of progression of kyphosis typically observed over a decade in older people and may have significant longitudinal benefits.

The magnitude of change in trunk extension strength in this study exceeds that reported in prior research. To compare our results to others, we converted peak torque/body weight ratio to peak torque multiplying by body weight. Our results show a 53% increase in peak torque after 3 months in our participants with a mean age ± SD of 72.0±4.2 years old. One study15 reports a 24% to 45% increase at 6 months in women 80.2±4.8 years, whereas another27 reports a 27% increase at 1 year in women 35.9±2.9 years. Although we used the same prone trunk extension strengthening intervention as Gold et al,15 we developed a multidimensional intervention including integrated proprioception training of postural alignment that may have enhanced our participants’ strength. However, the method for testing trunk extension strength was not identical across all studies.

This intervention increased ROM in the lower extremities. The popliteal angle increased bilaterally to age-matched normative values for this measure.28 There was no change in shoulder mobility as measured by shoulder flexion ROM. There may be better methods to measure shoulder mobility that quantify pectoral muscle length that can be used in future study.

The 2-point increase in modified PPT score matches that previously reported in the initial 3-month phase of a group exercise intervention in sedentary adults 83±4 years.24 There was no significant change in measures of balance and gait speed. However, our participants were comparable to age-matched normative populations on the Balance Master SOT and had slower COP velocity than healthy adults aged 66 to 70 years at baseline.10, 29 Furthermore, our participants were extremely robust, with a baseline gait speed of 1.35m/s. The lack of significant improvement is likely because of the relatively high-functioning baseline status of our participants.

We did find that some characteristics of the women were associated with the amount of change in flexed posture. Younger participants show greater change in “best” kyphosis consistent with prior findings that younger women have more ability to improve kyphosis compared to older women.5 Our finding that baseline modified PPT scores correlate with change in “usual” and “best” kyphosis suggests that higher-functioning patients exhibit greater postural change. We did not find a relation between the number of vertebral compression fractures and change in flexed posture. In fact, the 5 women with the greatest number of VCF (3) improved “usual” kyphosis by 7.4° and “best” kyphosis by 5.5°, results that are comparable to the group as a whole.

These results did not find that changes in strength or ROM correlate with change in flexed posture or that change in flexed posture correlates with change in physical performance. It is possible the strength measurements quantify lumbar extension strength and do not measure the change in thoracic extension strength that may be associated with improved kyphosis. We expected to find a correlation between improved hamstring flexibility, measured by popliteal angle, and improved kyphosis because improving hamstring flexibility theoretically allows greater mobility in the pelvis and spine. However, we did not measure lumbar lordosis, and it is likely that increasing hamstring flexibility allows the pelvis to rotate into more anterior rotation and increases lumbar lordosis rather than decreases thoracic kyphosis. We used static measurements of kyphosis that may not capture the dynamic nature of the physical performance tasks. Although there were significant changes in physical performance, small effect sizes may limit the ability to detect significant correlations less than .05. Furthermore, other covariates not measured, such as trunk proprioception, vital capacity, pain, comorbidities, and mobility self-confidence, may influence change in kyphosis and physical performance.

The participation rate in the 12-week program was exceptional and shows the appealing and feasible nature of the exercise program. Other than those who withdrew because of unrelated injuries or family emergency, participants completed all 24 sessions and exceeded the required 3-times-per-day home postural alignment practice.

We recognize that our study has several limitations. This study is a single-group pretest-posttest design and is not a randomized controlled trial. However, we tried to reduce the confounding effects of measurement variability and learning by comparing our results to the 95% CI of t2 to t1 measurements. In all cases, we met or exceeded this CI, evidence that t3 to t2 change is attributable to the intervention (see table 2). Additionally, we applied the most conservative correction for multiple comparisons to reduce the probability of making a type I error. The primary investigator was not blinded; she performed all measurements and taught the exercise classes. We decreased this potential bias by having a research assistant read and record all measurements. We enrolled a group of highly motivated, robust women and cannot generalize our results to less-motivated or frail populations. We had a small sample size limiting power to detect significant correlations (r<0.5) and restricting analysis of covariates known to affect variation in function from spinal deformity.30 We did not find significant change in other measures of flexed posture previously reported.12, 13

These results provide intriguing preliminary data that this type of exercise intervention may not only improve posture but also improve physical performance. Further research is needed to develop better methods to quantify standing height and forward head posture that incorporate postural sway and multiple degrees of freedom in the spine, pelvis, and legs, as well as to develop dynamic measures of flexed posture during functional activities. During our study, participants expressed improved self-confidence as their posture improved, and future work should incorporate measures of body image and quality of life. Ultimately, a randomized controlled trial with a larger sample size and longer-term outcomes would allow optimal analysis of pathways of change between flexed posture, impairments of strength and ROM, and physical performance.

Conclusions 

These results show statistically significant gains can be made in measures of flexed posture, strength, ROM, and physical performance in older women with increased kyphosis. This 3-month multidimensional exercise intervention yields greater improvements in kyphosis and spinal extension strength than other techniques have previously reported. Considering kyphosis progresses with age and is associated with decreased physical performance and increased fracture risk, targeted exercise that improves kyphosis may have significant functional implications. Multidimensional group exercise should be considered when developing a comprehensive program to improve posture, musculoskeletal impairments, and physical performance.