The aging process modifies normal postural alignment, and flexed posture commonly increases with age. Thoracic kyphosis and protrusion of the head, and in more severe cases, knee flexion, characterize flexed posture 12345.

Although the precise etiology of flexed posture (FP) is unknown, its pathophysiology in the elderly is most likely multifactorial and can be associated with low bone mineral density and consequently vertebral fractures, and also degenerative alterations of intervertebral disks 26789101112. It is also related to musculoskeletal and neuromuscular impairments 131415: deficit in spinal extensor muscle strength and shoulder and hip range of motion decreasing have been correlated with flexed posture 11415161718. Increased flexed posture has been associated with less independence when performing activities of daily living and reduced physical performance, such as impaired balance, reduced postural control and slower walking, as well as risk of falling 113141920212223242526. Moreover kyphosis and compensative cervical and lumbar spine hyperlordosis can cause pain due to ligaments and muscles impairment 1122.

There is no standard approach to measure flexed posture and classification methods are very complex 27. Clinical assessments present the advantages of simplicity, low-cost and wide application, such as the occiput-to-wall distance 1 that classifies the severity of flexed posture. On the other hand, instrumental evaluation enables quantitative analysis of the global body alignment in flexed posture patients and especially the compensatory axial alterations at the head and lower limbs. In the literature, several studies have reported the use of goniometers, inclinometers 2829, electrogoniometers 30, stadiometers 313233 and photometric techniques 29343536. The more recent stereophotogrammetric systems 373839 seem to realize more reliable and valid evaluations of postural alignment analysis.

A sedentary life style is supposed to play a fundamental role in developing a flexed posture and there is evidence in literature that appropriate physical activity programs can correct this attitude 1.

There are few studies that investigate methods to improve flexed posture, and especially the relationship with the multiple impairments associated to flexed posture 174041. Sinaki 1416424344454647 focused most research on the correlation among osteoporosis, kyphosis and back extensors strength and defined an exercise program based on isometric back-extensor strengthening and proprioceptive postural retraining to contain flexed posture 17484950. In different studies, the same author analyzed the role of the physical activity program in improving balance, gait and quality of life and the consequences in reducing back pain, risk of falls and vertebral fractures 1746495051. The efficacy of these protocols seems to correlate to exercise specificity, but we failed to find sufficient randomized controlled trials in the literature that compare different programs.

Based on Sinaki's experience, we hypothesized that an Adapted Physical Activity (APA) program with specific exercises for flexed posture would improve postural alignment and physical performance in a more effective way than a non-specific physical activity protocol for the elderly. The purpose of this clinical study was to systematically compare the effects of a physical activity program that specifically address the flexed posture that marks a certain percentage of elderly individuals with a non specific exercise program for 3 months. A clinical multidimensional assessment 1, including evaluation of musculoskeletal impairment, motor function, and disability, and an instrumental assessment of global postural alignment were used as measures of outcome.

Considering the multifactorial pathophysiology of FP, we enrolled patients according to strict inclusion criteria to avoid confounding factors.

The multidimensional clinical assessment 1 was an excellent mean to characterize the elderly population because, as documented in literature 52535460616364, specific for studying the complex clinical profiles and impairments in the geriatric field.

Based on the well known concept that back extensors weakness is definitely recognized as a key element in the pathophysiology of FP 11467, exercised for strengthening of back extensor muscles were included in the APA protocol. Other exercises in the APA protocol were addressed to the stretching of muscles of hip and shoulder flexors. Besides a hip flexor static contracture, common in the elderly 68, we have to consider in fact that retraction of hip flexors and of anterior muscles of shoulders and neck is an expected consequence of flexed posture.

The results support our primary hypothesis that a Physical Activity program adapted to the specific impairment is more effective in improving flexed posture in the elderly than a non-specific protocol. This effect is evident when considering the spine extensor muscle strength, significantly increased only in the APA group. Moreover, even if both exercise programs had positive effects on knee ROM and hamstrings flexibility, the specific APA protocol was more effective in those districts mainly involved in flexed posture 1224044, modifying ROM of hip joint and increasing the flexibility of hip flexor muscles and of the pectoralis major. The improvement of these measures only in the Group APA confirmed the specificity of the APA program in modifying soft tissue retraction associated to flexed posture.

As a global measure of outcome in quantifying flexed posture, it is relevant that the occiput-to-wall distance, considered a specific indicator of severity of FP 1, showed a statistically significant decrease only in the APA Group.

The instrumental assessment of posture allowed us to measure the global postural alignment and especially to analyse the possible compensatory strategies to FP. The experimental set up was quite simply and easily reproducible. Considering the complexity of classification methods for posture in the literature 27, we designed a specific biomechanical model based on five different angles on the sagittal plane. These angles, clinically meaningful, (head protrusion, trunk flexion, hip flexion, knee flexion, ankle dorsiflexion) referred to precise compensatory axial deviations to kyphosis, that are important components of FP 121. The three different visual tests (static, extension, and flexion) were performed to study the postural adaptations during eye-level modifications; this analysis allowed us to define the impact that the axial deformity of FP has on daily living activities (ADL), simulating dynamic activities which could be performed during ADL. While any compensatory movement is expected in normal people when looking at the top or down due to normal postural alignment and adjustment ability, we hypothesized that people with flexed posture and reduced mobility at the level of upper spine would adopt possible compensation strategies, mainly bending the knees and increasing ankle dorsiflexion in order to control balance.

Even in the instrumental evaluation, the specific APA program provided more significant improvements in the control of postural alignment than the non-specific protocol. In detail, both in static and flexion positions we observed a reduction in the flexion of the head and dorsiflexion of the ankle. In the extension position, the two programs determined the same positive effects with the decrease in knee flexion.

The more aligned head position instrumentally measured in the static posture and in the "extension test" in the APA Group confirms the reduction of the "occiput to wall distance" clinically measured. Moreover instrumental postural assessment highlights the increased ankle dorsiflexion as a compensatory strategy adopted by patients in order to maintain balance before the physical activity period, which reduces after training. The recovery of a better alignment of head and spine after specific exercise reduces the need of a biomechanical compensation at the ankle. The same mechanism can be supposed for the reduction of knee flexion in the "flexion test". In this case the improvement was observed in both groups, even more evident in the APA Group.

General physical performance increased in both groups, as the SPPB scores demonstrated. This was probably due to the positive effects that exercises, in general, have on balance, gait and motor function in the elderly. Conversely, after analyzing the Barthel Index 64 and the Nottingham Extended Activities of Daily Living Index 65 we could not find any improvement in the disability of the elderly after exercise. The reason for this can be attributed to the high-level of independence of the elderly population before treatment with a possible ceiling effect of score systems used. The slight association with disability might be due to the use of effective compensatory strategies even in the presence of severe FP, as previously hypothesized 1. This finding confirms the results of other authors who found a weak correlation between severity of FP and disability 1421. The measure of cognitive status 52, depression 60, and fatigue 61 did not reveal any statistically significant changes; these results were consistent with the high values found before exercise. All patients complained of lumbar pain, improved after both the physical activity programs. As back pain is probably related to abnormal stress of muscles and ligaments 1122, it is reasonable that even a postural non-specific physical activity program can be effective in relieving this symptom.

The Adapted Physical Activity program was inspired by the well-documented Sinaki approach, based on the very important role of Spinal Proprioceptive Extension Exercise Dynamic (SPEED) program 515556. Sinaki amplified her own protocol with the use of a spinal weighted kypho-orthosis (WKO) to increase a patient's perception of spinal positioning 174350. We limited our program to the 10 selected exercises without using any orthosis.

The good results of the present trial are in agreement with findings of previous studies 141740414448 that investigated methods to improve flexed posture, all based on back-extension strengthening exercises. However, it was not possible to make a precise comparison of previous findings due to the different measurement systems adopted. The peculiarity of our research consisted of the randomized controlled clinical trial and the instrumental quantitative analysis of flexed posture. A limitation is however related to the small sample size of eligible subjects due to strict criteria of inclusion and different number of subjects which concluded the study in the two groups. The larger benefit achieved with the APA program compared to the non-specific activity program essentially refers to the investigation herein presented as a preliminary study. Further development of the present research is required particularly to calibrate the optimal amount of exercise to be administered taking into account even individual impairment. As Sinaki 46 affirmed, an exercise program aimed at maintaining muscular strength and flexibility is characterized by the principle of reversibility, so discontinuation reverses the improvement to pre-exercise levels. For this reason, the results of this research may be useful in developing a long-term Adapted Physical Activity program for the elderly aimed at containing and preventing FP.

The Physical Activity program adapted for people with flexed posture improved postural alignment and musculoskeletal impairment more effectively than a non-specific physical activity protocol. The increasing of back extensors strength, the increasing in the flexibility of pectoralis, hip flexors and hamstrings muscles correspond to the reduction of FP, as measured by means of the occiput-wall distance. The instrumental assessment, based on a clinically oriented, reliable biomechanical model, allowed to measure the global postural alignment in patients with FP before and after physical activity trials and especially to analyse the possible compensatory strategies at the head and lower limbs.

The authors declare that they have no competing interests.

MGB made substantial contributions to conception and design of the study, analysis and interpretation of data and she was involved in drafting the manuscript and revising it critically for relevant intellectual content. LB performed all the measurements and was involved in drafting the manuscript. CP and AF participated in data acquisition and analysis. SG gave final approval of the version to be published.