“Koshimagari Exercise” for Adult Spinal Deformity in Older Adults: Assessment of Home-Based Exercise Outcomes in a Prospective Multicenter Study

Abstract

Introduction

Adult spinal deformity (ASD) is prevalent among older adults, considerably affecting their quality of life. Although surgical interventions are effective, they have high complication rates and medical costs. Furthermore, there is a lack of evidence supporting the effectiveness of nonsurgical treatments (e.g., physical therapy) in patients with ASD. This study aimed to investigate the impact of “Koshimagari exercise,” a specific home-based exercise regimen designed for patients with ASD, and to evaluate its effects on clinical outcomes in older adults.

Methods

A total of 144 participants aged 50-80 years with chronic low back pain (LBP) due to spinal deformities were included in this multicenter prospective study. Qualified physiotherapists conducted intervention sessions at the hospital once a week, and self-exercise was performed at home three times a week. After 3 months, the frequency of self-exercise at home increased to four times a week. Clinical evaluations were conducted using the Oswestry Disability Index (ODI), five-level classification system of EuroQol-5 Dimensions (EQ-5D), Japanese edition of Scoliosis Research Society-22r (SRS-22r), and visual analog scale (VAS) for LBP at baseline and 3, 6, and 12 months. Radiographic evaluations were performed in standing and supine positions.

Results

Of 130 participants who provided written informed consent, 98 completed the 6-month follow-up and were included in the analysis. Significant improvements observed in ODI, EQ-5D, and VAS scores were observed at 3 months, with SRS-22r scores improving throughout the study period. Radiographically, there were significant differences in the sagittal vertical axis and pelvic tilt at 12 months. Sufficient compliance with the self-exercise program was reported by 96%, 86%, and 73% of participants at 3, 6, and 12 months, respectively.

Conclusions

The “Koshimagari Exercise” program led to significant short-term improvements in health-related quality of life and pain among elderly patients with ASD. This home-based self-exercise program is an excellent nonsurgical treatment option for patients with ASD.

Introduction

Spinal malalignment is a three-dimensional condition that presents with abnormalities in the coronal and/or sagittal planes^1,2). Sagittal plane malalignment is commonly associated with aging, intervertebral disc degeneration, weakened spinal extensor muscles, and reduced joint mobility of the hips and knees^1,3). Adult spinal deformity (ASD) with sagittal malalignment considerably reduces the quality of life (QoL) of older individuals, leading to chronic back pain, gastroesophageal reflux, respiratory dysfunction, falls, fractures, depression, and social isolation^4-8).

Surgical intervention is the most established evidence-based treatment for ASD. Schwab et al.'s classification, which incorporates pelvic incidence (PI), lumbar lordosis (LL), and pelvic tilt (PT), is pivotal in defining the ideal sagittal alignment⁹⁾. Subsequently, the development of various formulas for predicting the ideal sagittal alignment while considering factors such as age, along with the use of anterior interbody fusion and navigation systems, has led to less invasive surgeries^10-14). This progress has enabled corrective surgeries, even for patients with severe kyphotic deformities and older patients, which were previously challenging^15,16). Despite the demonstrated efficacy of surgery in improving activities of daily living, perioperative and postoperative complications remain common, with high rates of adjacent segment disease, including fractures, often requiring reoperations^17-20). Additionally, the increasing costs associated with extensive implant use further highlight the need for effective nonsurgical treatment options^1,21,22).

There is limited evidence supporting brace treatment and physical therapy. Brace treatment is effective in managing adult scoliosis with coronal malalignment but has shown limited efficacy in addressing sagittal plane issues^8,23,24). Physical therapy for sagittal malalignment primarily targets thoracic kyphosis (TK). Pawlowsky et al. reported that a 12-week twice-weekly exercise program improved kyphosis, knee/hip range of motion, and walking speed in older patients aged 72 years on average²⁵⁾. Similarly, Jang et al. observed improvements in spinal kyphosis and SF-36 scores after an 8-week exercise program in older women aged approximately 75 years on average²⁶⁾. Nevertheless, these studies primarily targeted TK, which is considered different from the pathology of sagittal malalignment accompanied by LL loss²⁷⁾. Hongo et al. showed that back muscle exercises improved the muscle strength and QoL in patients with osteoporosis²⁸⁾. However, the effects of exercise on sagittal malalignment in patients with ASD remain unclear. Additionally, prolonged physical therapy can be expensive, making it less sustainable in the long term. Therefore, we propose home-based exercise as a cost-effective and sustainable treatment option for managing this chronic condition.

This study aimed to investigate whether a home-based exercise program for elderly individuals could improve their sagittal malalignment, a condition expected to become more prevalent with population aging, to evaluate the extent of improvement in health-related QoL (HRQoL) and pain, and to determine the duration of these benefits after the intervention.

Materials and Methods

Ethical statements

The study protocol was approved by the Institutional Review Board of our institution. Written informed consent was obtained from all participants.

Study design and population

A multicenter prospective study was conducted in 13 facilities. A total of 144 participants were recruited from December 2020 to June 2022. The inclusion criteria were as follows: (1) age of 50-80 years; (2) chronic low back pain (LBP), lasting for more than 3 months, without severe neurological symptoms in the lower extremities; (3) LBP associated with spinal malalignment during standing and walking but not specific at rest or during postural changes; and (4) standing parameters of sagittal vertical axis (SVA) >50 mm, PT >25°, or TK >60° on whole lateral radiographs²⁹⁾. The exclusion criteria were as follows: severe lower-limb pain due to neurological or arthritic conditions, acute or nonunion osteoporotic vertebral fractures, ankylosing spondylitis, history of spinal fusion surgery, collagen disease, medical conditions preventing exercise (such as cardiopulmonary diseases), neurological disorders (such as dementia and Parkinson's disease), and psychiatric disorders.

Intervention

Spine surgeons and exercise specialists developed a new self-exercise program specifically for patients with ASD, named the “Koshimagari Exercise,” based on existing exercise therapies. Qualified physiotherapists conducted intervention sessions once a week at each hospital, and self-exercise was performed at home three times a week. The exercise instruction aimed at improving self-efficacy and self-management through education and encouraging active and continuous participation in exercise programs. After 3 months, they performed self-exercise four times a week at home. The program comprised three key components: (1) “self-exercise in the supine position” to mobilize the spine, thorax, and pelvis; (2) “lumbar back muscle exercise” to stimulate and strengthen the lumbar muscles; and (3) “aerobic exercise” performed in the standing position (Fig. 1, 2). In addition to the three main exercise programs, patients were encouraged to incorporate brief exercises into their daily routines. These exercises, such as trunk extension, PT, and leg raises, were performed for short durations and aimed at enhancing motor learning and pain management while promoting an upright posture during walking (Fig. 2). Detailed instructions for these exercises are provided in Supplement 1.

Figure 1.

(A–F) Self-exercise in the supine position: (A, B) trunk extension (prone position); (C, D) psoas major, lumbar back, and psoas muscle movement; (E, F) thoracic expansion exercise. (G–K) Lumbar back muscle exercises: (G, H, I) trunk extension exercise (sitting position); (J) trunk extension exercise (four-crawling position); (K) trunk extension exercise (prone position).

Figure 2.

(A, B) Aerobic exercise in the standing position: (A) walking with a backpack; (B) walking with poles. (C–F) Exercises in daily life (brief exercises): (C, D) pelvic forward/backward tilt movement (sitting position); (E) movement for trunk extension (standing); (F) lower-limb raising exercise with anterior pelvic tilt (sitting position).

Clinical evaluation

Three patient-oriented questionnaires, the Oswestry Disability Index (ODI) and the five-level classification system of the EuroQol-5 Dimensions (EQ-5D), the Japanese edition of the Scoliosis Research Society-22r (SRS-22r), and the visual analog scale (VAS) for LBP were used to evaluate HRQoL at baseline and after 3, 6, and 12 months of intervention. Compliance with the exercise program was categorized as excellent, good, and fair and evaluated at 3, 6, and 12 months. Excellent indicated that the patient fully adhered to the protocol, good indicated that the patient performed the exercises more than twice a week, and fair indicated that the patient performed the exercises once a week or less.

Radiographic and muscle mass assessment

The radiographic parameters assessed on standing posterior-anterior and lateral radiographs were as follows: Cobb angle of the main TL/L curve (TL/L), the distance between the C7-plumb line and the center of the sacral vertical line (C7-CSVL), the C7 SVA, the degree of TK (T5-12), thoracolumbar kyphosis (TLK; T10-L2), LL (T12-S1), distal LL (L4/S1; L4-S1), PT, PI, sacral slope (SS), and PI-LL. To evaluate sagittal flexibility, we measured the parameters in the supine position using lateral radiograph supine-TLK and LL lateral radiographs. Trunk mass and appendicular skeletal muscle mass (ASM) were measured using bioelectrical impedance analysis with a body composition analyzer. Assessments were conducted at 3, 6, and 12 months.

Statistical analysis

The Friedman test was used to compare all clinical scores and spinal parameters recorded at baseline and 3, 6, and 12 months after the intervention to analyze the differences among the groups. When there was a significant difference, subsequent post hoc analyses were conducted between baseline and 3, 6, and 12 months after the intervention. For univariate analysis between the two groups, Fisher's exact test and the Mann-Whitney U test was used for categorical and continuous variables, respectively. Missing HRQoL scores and radiographic data were imputed using the last observation carried forward. Statistical significance was set at P<0.05 with a two-tailed t-test. All analyses were performed using R software (version 3.5.1, Patched; R Foundation, Vienna, Austria).

Results

Of 144 recruited participants, 130 were enrolled in this study. Fourteen participants dropped out of the study before the exercise intervention. Moreover, 17, 15, and 10 participants were lost to follow-up at 3, 6, and 12 months, respectively. Fig. 3 shows detailed reasons for the loss to follow-up. The mean age of the participants was 73.0 years; 83 were female (85%). The average BMI was 23.3, and the Charlson Comorbidity Index (CCI) was 0.5. Furthermore, 27% of participants had old vertebral fractures. Thirteen (13%) and 31 (32%) participants occasionally and daily used painkillers, respectively. Table 1 shows smoking status, working status, painkiller usage, and baseline values of the ODI, EQ-5D, VAS, and SRS-22r scores.

Figure 3.

Flowchart of study enrollment and follow-up.

Table 1.

Baseline Characteristics.

Variable	Overall
Characteristics	n=98
Age (years)	73.0 (5.9)
Sex (female)	83 (85%)
BMI	23.3 (3.4)
CCI	0.5 (0.8)
Prevalent vertebral fractures	26 (27%)
Smoking
Current smoker	3 (3%)
Ex-smoker	14 (14%)
Working
Desk work	8 (8%)
Standing work	10 (10%)
Heavy work	9 (9%)
Pain killer use
Sometimes	13 (13%)
Everyday	31 (32%)
Acetaminophen	5 (5%)
NSAIDs	15 (15%)
Tramadol	4 (4%)
Others (not opioid)	20 (20%)
HRQoL
EQ-5D	0.63 (0.11)
ODI	35.3 (14.4)
VAS	53.1 (28.0)
SRS-22r subtotal	3.04 (0.57)
SRS-22r function	3.22 (0.72)
SRS-22r mental	3.13 (0.88)
SRS-22r pain	3.36 (0.70)
SRS-22r self-image	2.40 (0.64)

Data are presented as mean±standard deviation (SD).

BMI, body mass index; CCI, Charlson Comorbidity Index; EQ-5D, EuroQol-5 Dimensions; HRQoL, health-related quality of life; NSAIDs, nonsteroidal anti-inflammatory drugs; ODI, Oswestry Disability Index; SRS-22r, The Scoliosis Research Society-22r (SRS-22r); VAS, visual analog scale

Changes in the ODI, EQ-5D, and VAS

The ODI decreased from 35.3 before the intervention to 32.3 and 33.0 at 3- and 6-month follow-up, respectively; however, there was no significant difference at 12 months. EQ-5D increased from 0.63 before the intervention to 0.65 at 3-month follow-up, showing a significant difference, whereas there was no significant difference at 6 and 12 months. With respect to VAS scores, improvement was observed at all observation points, from 53.1 before the intervention to 42.8 at 3 months, 46.0 at 6 months, and 46.2 at 12 months (Fig. 4). The SRS-22r showed improvement in the pain and self-image domains at 3 and 12 months and at 3 months, respectively. The functional, mental, and subtotal domains of the SRS-22r improved significantly at all observation points (Fig. 5).

Figure 4.

Oswestry Disability Index (ODI), EuroQol-5 Dimensions (EQ-5D), and visual analog scale (VAS) for low back pain (LBP) scores throughout the study period. Significant differences are marked: *p<0.05, **p<0.01, ***p<0.001.

Figure 5.

Scoliosis Research Society-22r (SRS-22r) scores throughout the study period. Significant differences are marked: *p<0.05, **p<0.01, ***p<0.001.

Radiographic and physical evaluations

Table 2 shows the physical and radiographic evaluations performed throughout the study period. Physical evaluation of trunk mass and ASM showed no significant differences in trunk mass at any of the measurement points. The only significant differences between pre- and post-intervention were the SVA (88.3-78.4) and PT (31.8-32.7) at 12 months. There were no changes in the parameters representing sagittal alignment such as TK and LL after the intervention.

Table 2.

Physical and Radiographic Parameters throughout the Study Period.

Variable	Pre	3M	6M	12M	P for all	Pre vs.
						3M	6M	12M
Physical
Trunk mass	20.5 (3.7)	20.5 (3.5)	20.4 (3.4)	20.4 (3.6)	0.561
ASM	15.7 (3.7)	15.8 (3.7)	15.8 (3.6)	15.9 (3.9)	0.084
Radiograph
SVA	88.3 (62.2)	84.1 (63.0)	84.0 (68.4)	78.4 (69.6)	0.002	0.246	0.116	0.023
SS	19.2 (7.4)	19.5 (7.4)	19.0 (8.0)	19.0 (7.2)	0.501
TK	25.1 (16.2)	25.4 (16.1)	25.0 (16.2)	24.7 (15.0)	0.667
TLK	25.2 (17.3)	26.1 (18.7)	25.7 (17.9)	23.8 (18.9)	0.191
L4/S1	30.7 (15.7)	31.5 (15.7)	31.0 (16.7)	29.3 (15.1)	0.214
LL	18.6 (16.8)	19.1 (16.7)	18.1 (17.5)	19.3 (17.0)	0.406
PI	51.0 (11.3)	50.9 (11.1)	51.1 (10.9)	51.5 (10.5)	0.090
PI-LL	32.4 (20.4)	31.6 (19.7)	33.0 (20.5)	31.9 (19.9)	0.359
PT	31.8 (10.8)	31.3 (9.9)	32.1 (10.5)	32.7 (9.5)	0.014	0.460	0.316	0.016
Coronal
T/L Cobb	13.4 (15.5)	NA	13.6 (16.1)	13.9 (16.0)	0.179
L Cobb	18.5 (15.5)	NA	18.3 (15.7)	18.4 (15.6)	0.270
C7-CSVL	22.7 (22.5)	NA	21.8 (20.1)	21.5 (20.0)	0.888
Dynamics
Supine LL (n=92)	27.9 (15.0)	NA	28.0 (15.1)	28.9 (14.9)	<0.001	NA	0.019	0.005
Supine TLK (n=87)	12.5 (14.3)	NA	12.3 (13.7)	12.3 (13.9)	0.974

Data are presented as mean±standard deviation (SD). The bold type indicates statistical significance.

ASM, appendicular skeletal muscle mass; C7-CSVL, distance between C7-plumb line and center for sacral vertical line; LL, lumbar lordosis; PI, pelvic incidence; PI-LL, pelvic incidence minus lumbar lordosis; PT, pelvic tilt; SVA, sagittal vertical axis; SS, sacral slope; TK, thoracic kyphosis; TLK, thoracolumbar kyphosis; T/L Cobb, Cobb angle of the main thoracolumbar curve

Compliance

Compliance with the self-exercise program was categorized as excellent at 65%, 43%, and 33%; good at 31%, 43%, and 40%; and fair at 4%, 13%, and 27% at 3-, 6-, and 12-month follow-up, respectively. We compared patients based on exercise compliance at 6 months (excellent, good, or fair). A total of 42 patients were classified into the excellent group, whereas 55 were in the good or fair groups. One patient was excluded due to missing compliance data. Table 3 describes the characteristics of these two groups and the clinical changes over 6 months of intervention. There were no significant differences in age or sex between the compliance groups. However, the EQ-5D and all SRS-22r scores showed significant differences between the two groups, except for the pain domain. Among the radiographic parameters, only SVA was significantly different between the groups.

Table 3.

Comparison of Characteristics, Changes in HRQoL, and Radiographic Parameters between the Two Groups.

Variable	Exercise compliance excellent	Exercise compliance good or fair	p-value
Characteristics	n=42	n=55
Age (years)	73.0 (5.9)	72.6 (6.6)	0.786
Sex (female)	47 (84%)	18 (82%)	1.000
HRQoL (change in 6 months)
EQ-5D	0.03 (0.09)	−0.01 (0.08)	0.015
ODI	−3.5 (13.0)	−1.3 (10.2)	0.372
VAS	−6.9 (29.6)	−7.3 (29.4)	0.951
SRS-22r subtotal	0.33 (0.51)	−0.02 (0.44)	<0.001
SRS-22r function	0.31 (0.54)	0.04 (0.51)	0.015
SRS-22r mental	0.45 (0.80)	−0.02 (0.65)	0.002
SRS-22r pain	0.21 (0.63)	0.02 (0.70)	0.158
SRS-22r self-image	0.34 (0.65)	−0.15 (0.56)	<0.001
Radiographic parameter
Sagittal (change in 6 months)
SVA	−13.2 (32.2)	2.9 (40.9)	0.040
PI-LL	0.2 (6.9)	1.0 (7.6)	0.628
PT	0.4 (3.8)	0.3 (4.0)	0.978

Data are presented as mean±standard deviation (SD). The bold type indicates statistical significance.

EQ-5D, EuroQol-5 Dimensions; HRQoL, health-related quality of life; ODI, Oswestry Disability Index; SRS-22r, The Scoliosis Research Society-22r; PI-LL, pelvic incidence minus lumbar lordosis; PT, pelvic tilt; SVA, sagittal vertical axis; VAS, visual analog scale

Discussion

There is limited evidence supporting the use of physical therapy and exercise for ASD¹⁾. To our knowledge, this study is the first to evaluate the effects of a uniform exercise program on ASD. The “Koshimagari Exercise” regimen was designed to be sustainable and cost-effective, involving physiotherapist-supervised sessions for the first 3 months, followed by self-directed exercises. The results indicated significant improvements in HRQoL, as measured by the ODI, EQ-5D, and VAS, at 3-month follow-up, with improvements in the SRS-22r and VAS scores persisting throughout the 12-month study.

The ODI and EQ-5D scores did not show any sustained improvement at 12 months. One possible explanation for the lack of sustained HRQoL improvement may be the absence of significant changes in key radiographic parameters, such as TK and LL. ASD is a multifactorial condition involving de novo kyphoscoliosis, degenerative disc disease, and iatrogenic deformities. These structural deformities cannot be effectively addressed by exercise alone. The slight improvement in the SVA at 12 months, although encouraging, did not correlate with other radiographic outcomes. Similar studies, such as those by Katzman and Jang, reported positive changes in TK after exercise^26,30,31). However, their cohorts primarily involved Caucasians with TK, whereas the participants in this study were older adults Asian patients with LL loss and sagittal plane abnormalities²⁷⁾. In addition, factors such as hip extension range of motion (ROM), back extensor endurance, and hip and knee muscle strength are critical for maintaining sagittal alignment and improving QoL³²⁾. Back, hip flexor, and knee extensor muscle strengths are associated with sagittal spinal alignment due to improvements in compensatory mechanisms³³⁾. Although the “Koshimagari Exercise” included ROM exercises for adjacent joints such as the hip and thoracic cage, we did not evaluate joint ROM or hip/knee strength in this study.

A notable finding of this study is the role of compliance in determining outcomes. Patient compliance is a significant challenge in this study. Patients with higher compliance rates showed significantly better improvements in the EQ-5D and SRS-22r scores than those with lower compliance. The rate of excellent compliance fell from 65% at 3 months to 33% at 12 months, underscoring the importance of patient engagement and adherence to exercise programs to achieve meaningful clinical benefits. Compliance is crucial for achieving optimal clinical and radiological outcomes, as emphasized by O'Connell et al. who advocated continuous professional intervention to maintain adherence to exercise programs³⁴⁾. Bridwell et al. similarly demonstrated that nonoperative treatments had little effect on improving QoL in patients over a 2-year follow-up period³⁵⁾. To improve long-term adherence, patients may require re-instruction and additional support, particularly after the initial 6-month period. Strategies to promote compliance, such as regular follow-ups, digital health tools for tracking progress, and social support systems, should be considered in future interventions.

Interestingly, the study found a small but significant improvement in SVA and increasing in PT at 12 months. This result suggests that although overall sagittal alignment may not have changed, there may have been some improvement in postural control and balance. The increase in PT and reduction in SVA may be attributed to enhancements in muscle strength, joint flexibility, and compensatory mechanisms facilitated by the exercise program, helping patients maintain a more upright posture during daily activities. However, the clinical significance of this finding remains unclear because the change in SVA did not correlate with improvements in other radiographic parameters or HRQoL scores. Dynamic sagittal imbalance may influence the relationship between HRQoL and sagittal parameters. Thus, sagittal parameters could be more accurately assessed using radiographs taken during or after ambulation to reveal hidden sagittal imbalance in compensated deformities^36,37). However, exercise therapy seems slightly effective in decreasing pain and improving function in adults with nonspecific chronic LBP³⁸⁾. It has been shown that exercise reduces inflammation through various mechanisms, such as modulation of cytokines, Toll-like receptors, and vagal tone³⁹⁾. In addition, many studies have highlighted the positive impact of exercise on mental health, which could partly explain the improvements observed in HRQoL despite minimal changes in sagittal alignment³⁹⁾. Although exercise therapy was shown to improve VAS scores, two patients required spinal surgery during follow-up. This highlights the challenge of using exercise therapy alone to manage severe cases of sagittal malalignment and suggests that surgical intervention is still necessary for some patients.

This study had several limitations that must be acknowledged. First, the lack of a control group prevented a direct comparison of the effects of exercise therapy with those of natural progression or surgical treatment. Future studies should incorporate control groups to better understand the comparative effectiveness of interventions. Second, the study highlighted the importance of compliance. However, there is a potential bias, as patients with favorable outcomes may more likely continue to exercise. Third, dynamic assessments were not performed, despite muscle fatigue being a significant factor influencing clinical symptoms. Finally, we did not account for changes in pain medication, which could have influenced the observed improvements in pain scores. Despite these limitations, this is the first large-scale multicenter prospective study focusing on sagittal malalignment with regular radiographic and physical evaluations. The use of radiographic assessments provides strengths compared to prior studies that relied on clinical tools, such as the scoliometer.

In conclusion, this study showed that home-based exercise programs led to short-term improvements in HRQoL and pain among elderly patients with sagittal malalignment. However, these benefits were not sustained over time, likely due to declining compliance and the complex nature of sagittal plane deformities. Future research should focus on strategies to improve long-term adherence to exercise programs and explore the potential benefits of combining conservative treatment with surgical intervention.

Disclaimer: Shinji Takahashi, Koji Tamai, Kei Watanabe, Naohisa Miyakoshi, and Hiroshi Hashizume are members of the Editorial Committee of Spine Surgery and Related Research. They were not involved in the editorial evaluation or decision to accept this article for publication at all.

Conflicts of Interest: The authors declare that there are no relevant conflicts of interest.

Sources of Funding: This study was supported by the Japanese Society for Spine Surgery and Related Research.

Author Contributions: H. Te. managed project administration, wrote the original draft, and contributed to writing - review and editing. S.T. was responsible for project administration, formal analysis, writing - review and editing, and supervision. M.H. contributed to conceptualization, project administration, and supervision. H.T. conducted data curation, formal analysis, and visualization. K.T., T.O., T.N., G.S., M. Te., M. Mi., T. Ki., M.K., T.T., Y.T., T. Mi., K.I., T.E., Y.M., T.S., H.M., Y.K., T.H., K.M., Y.O., K.Y., R.K., and H.S. were involved in the investigation. K.W., N.S., and K.I. contributed to conceptualization and investigation. K.N., Y.Y., K.M., N.M., Y.M., H. Haro., H. Hashizume., H.Y., and T.K. managed project administration. All authors reviewed the manuscript.

Ethical Approval: The study protocol was approved by the Institutional Review Board of Osaka Metropolitan University (approval no. 2020-277).

Informed Consent: Written informed consent was obtained from all participants.