Flexibility Training and Functional Ability in Older Adults: A Systematic Review

Background. As indicated in a recent systematic review relating to Canada's Physical Activity Guidelines for Older Adults, exercise interventions in older adults can maintain or improve functional abilities. Less is known about the role of flexibility in the maintenance or improvement of functional abilities, and there currently does not exist a synthesis of the literature supporting a consensus on flexibility training prescription. Purpose. To systematically review the effects of flexibility-specific training interventions on measures of functional outcomes in healthy older adults over the age of 65 years. Methods. Five electronic databases were searched for intervention studies involving concepts related to aging, flexibility, functional outcomes, and training interventions. After evaluating the articles for relevance, 22 studies were considered. Results. The results suggested that while flexibility-specific interventions may have effects on range of motion (ROM) outcomes, there is conflicting information regarding both the relationship between flexibility interventions and functional outcomes or daily functioning. Conclusions. Due to the wide range of intervention protocols, body parts studied, and functional measurements, conclusive recommendations regarding flexibility training for older adults or the validity of flexibility training interventions as supplements to other forms of exercise, or as significant positive influences on functional ability, require further investigation.


Introduction
As indicated in a recent systematic review relating to Canada's Physical Activity Guidelines for Older Adults, exercise interventions (comprised of aerobic and strength training) in older adults can maintain or improve functional abilities [1]. Less is known about the role of flexibility in the maintenance or improvement of functional abilities. While joint flexibility may decrease with age, with the potential to affect normal daily function, older adults do maintain the ability to improve flexibility through stretching exercises [1]. The 2009 American College of Sports Medicine (ACSM) position statement "Exercise and Physical Activity for Older Adults" [2] noted there is a lack of studies of the effects of range of motion exercises on flexibility outcomes in older populations and a lack of consensus regarding the prescription of stretching exercises for older adults. Despite the lack of a synthesis of the literature to support the recommendation of the inclusion of a flexibility component to older adult exercise programs, many older adult activity programs place a considerable emphasis on flexibility. Stretching exercises are used extensively in the rehabilitation context wherein injury or disease may have resulted in a restricted range of motion specific to given joints, and the goal is to regain "normal" range of motion [3]. However, the present paper is focused not on stretching exercise for rehabilitation purposes but for the role of flexibility in general exercise prescription for older adults.
In light of the significant benefits of an exercise program for an aging population, it is important to provide evidencebased prescription for older adult exercise programs and highlight areas of research requiring further investigation in order to maximize these benefits. The goal of a flexibility program is to improve range of motion in the major muscletendon groups in accordance with individualized goals [4]. For the majority of the aging population, the goals may not 2 Journal of Aging Research be related to athletic performance, but rather performance of functional abilities in activities of daily living. Nevertheless, there is relatively little research on the potential benefits of flexibility-specific training interventions for this population in that context. Despite the lack of research and no "known health benefits" [5], again, there is still a tendency in the literature to mention flexibility training as a presumed "component of fitness" and beneficial adjunct to other forms of exercise. Therefore, the purpose of this systematic review is to investigate the functional outcomes of flexibility specific training in older adults.

Methods
an intervention study, (4) flexibility training was an independent intervention or was used as a control, (5) aged population (mean age ≥ 65 years), and (6) the population was healthy but allowing for arthritis, osteoarthritis, and those residing in assisted living (based on age and risk, not diseases or other medical conditions). For this paper, healthy was operationally defined as community-dwelling and assisted living with the health and function and cognitive ability to participate in light physical activity interventions and complete physical function measures. Interventions targeting specific chronic conditions (aside from arthritis and osteoarthritis) were excluded from review. Despite their use in flexibility training, tai chi-and yoga-based studies were excluded from this paper because by nature they include strength components. The electronic search yielded 4037 citations. The citations and applicable electronic versions of the article (where available) were downloaded to an online research management system (RefWorks, Bethesda, MD, USA).

Screening.
Two reviewers independently (RL, LS) evaluated the articles for relevance using standard systematic review methodology leading to further consideration of 22 articles.
Two reviewers independently completed standardized data extraction forms for each level of screening. Three levels of screening were utilized. Level 1 screening was based on article titles, Level 2 was based on the title and abstract, and Level 3 was a full text screening. The articles that progressed through to Level 3 were retrieved electronically or manually via the Canadian interlibrary system and were printed from electronic copy. Any cross-referenced articles from the reference section of Level 3 articles were hand-screened.
Disagreements regarding inclusion were resolved through discussion with a third reviewer (DP).

Data Extraction.
Data from the included studies were extracted (Table 1) and organized by the target muscle groups of the flexibility interventions. Two reviewers completed standardized data extraction forms. One reviewer performed the data extraction for each paper assigned to them and the extraction was verified by another reviewer. The reviewers were not blinded to the journal or the author names when extracting information from the articles.

Level of Evidence.
The approach used to establish the level and grade of evidence was consistent with Lau et al. [6] which provide predefined and objective criteria. Thus, the strength of the evidence was assessed for flexibility interventions and functional outcomes in older adults with respect to general recommendations and appropriate dose.

Quality
Assessment. Quality assessment of the included studies was also performed ( Table 1). The Downs and Black [7] scale was selected to assess the quality of each study as it is appropriate to evaluate nonrandomized investigations, and it contained the highest number of relevant items for the needs of this paper. However, as not all items were relevant to the various study types included in this paper, a modified version of the checklist was employed for each of RCTs (randomized control trials), and non-RCTs study types. Thus, the quality of each study was also established similar to the method of Gorber et al. [8] to include the most relevant components of the scoring tool. Therefore, a modified version of the Downs and Black checklist was used with the final checklist consisting of 22 items with a maximum score of 24 points for the studies of a RCT design; 22 items for non-RCT designs with a maximum score of 23; experimental single group interventions had a maximum score of 18 from 18 items; experimental single-group and single-session studies were based on 14 items for a maximum score of 14. Higher scores reflected a superior quality of investigation.

Integration of Findings.
Due to the heterogeneity across study populations, methods used, and outcomes assessed, we conducted a narrative synthesis of the results.

Description of Studies.
The initial search yielded 4037 articles. Twenty-two articles were ultimately included after meeting Level 3 inclusion criteria ( Figure 1). Of the final 22 articles, 18 were from electronic database searching, and 4 were found by hand searching. Quality assessment indicated that the RCT studies (n = 13) were of good quality with an average score of 18 out of 24. The non-RCT studies (n = 6) had an average score of 14 out 23. An average score of 12 out of 18 was assessed for the experimental single-group studies (n = 3). Fourteen articles were conducted in the United States [10-12, 14, 15, 17, 18, 20, 22, 25-29], while the remainder of the studies were from Japan [19], Brazil   Difference between groups was "almost significant."

Response time
Unchanged in both groups. Sensation No apparent differences.
The control group lost a small amount of strength and balance in just 3 months, even though flexibility improved.
These results suggest that the more comprehensive the exercise intervention, the greater the likely scope of improvement in frailty.
14/24 6 Journal of Aging Research   with no between-groups differences.
The exercise with no weights groups had nearly 2.5x greater increase in ROM than exercise with weights for shoulder abduction. No other flexibility comparisons in the exercise groups were significant.
No functional outcomes.
Exercise can increase flexibility in healthy, older women by improving shoulder flexion and abduction, ankle plantar flexion, and cervical rotation.
For shoulder abduction, a nonweighted exercise program can produce greater flexibility gains in older adults than a weighted exercise program, and should be considered if flexibility is the primary goal.  Found that the addition of weights enhanced effectiveness of stretching exercise. It is likely that a positive effect was noted for 2 reasons: increased resistance during exercise movement and exercises were performed through full ROM.

13/23
It seems plausible to hypothesize that greater effects would be shown by addition of weights to stretching routines for nursing home clients or free-living, otherwise sedentary elderly.
14/23    Authors stated that participating in daily flexibility group exercise increases ROM and causes some improvement in balance.
When compared to a randomly selected hospital applied group, the rest home group has better balance, trunk flexibility and functional ability. Evidence is provided from the results that joint motion is a modifiable impairment that can be effectively targeted for older people with simple, static stretching home-based intervention. Supervised stretching program is effective to alter a number of gait variables. Aged participants displayed gait parameters which were similar to those of young adults.

22
Journal of Aging Research 2x per day (approximately 5 min).

Treatment group
Hip stretching exercise.

Control group Deltoid stretching exercise. Assessment
Static hip extension range, goniometer.
Pelvic and bilateral lower-extremity joint motion and joint kinematics.

Analysis
Unpaired and paired t tests, paired Student t tests, and Bonferroni adjustment. Only peak ankle dorsiflexion and plantar flexion during swing increased significantly with treatment.
No major changes with control group. Both groups had similar minor increases in comfortable walking speed.
Tendency toward a reduction in anterior pelvic tilt implies that modest improvement in hip extension range allowed for a slight decrease in anterior pelvic tilt. These findings support the hypothesis that an increase in pelvic tilt in the elderly is a compensation for hip contracture rather than a compensation or a direct result of another impairment. Improvement in peak ankle plantar flexion with hip stretching and a trend toward improved ankle plantar flexor power at comfortable walking speed were found. These results suggest that these outcomes in the elderly may be secondary to proximal impairment rather than to impairment at the ankle or ankle musculature per se. There is a meaningful positive association between regular flexibility training and spinal mobility in older population studies and may reduce the potential for age-related spinal deconditioning.

17/24
14/24 24 Journal of Aging Research increased significantly in all measurements for groups II and III.
All groups increased significantly more than control.
Stretching therapy is recommended as an adjuvant treatment to isokinetic exercise for patients with knee OA. PNF is more effective than static stretching exercise.

Outcome Measures.
Outcomes were measured using flexibility measurements, physical ability tests, and questionnaires. One study also used brain imaging for the purposes of identifying changes in hippocampal volume with training [20]. A common outcome measure was simply whether there was a change in range of motion usually assessed by goniometry [15,17,20,[24][25][26][27][28]30]. The inclusion of these studies in the review (although they reported no "functional outcome") was to provide the data for the purpose of determining whether older adults would, in fact, improve range of motion about different joints with various flexibility exercise programs. Functional outcomes were operationally defined as tests or measures designed to reflect abilities for various levels of daily activities and potentially related to maintenance of independence of older adults. In general, these tests assessed ability in a function that involved more than a single fitness component (e.g., not just a strength measure of a weight that could be lifted, but rather a performance that may involve strength and power as well as balance and agility). The most commonly used tests of functional outcome were gait and various walking speeds [11,13,14,19,21,22,26], the sit-and-reach test [10, 12-14, 19, 29], the sit-to-stand test [9,12,14,16,18,19], functional reach test [9,10,21,26], step test [9,16], timed up-and-go (TUG) [10,13,14,16,19,24,26], and Romberg test [11,21]. Other functional outcome measures were Berg balance scale [13], questionnaires [9,12], peg board [11], red-lightgreen-light [11], Lequesne's index of disability [30], the physical performance test (PPT) [11], and the gallon jug shelf test [14]. Gait and walking speed proved to be more positively affected by flexibility training than other outcome measures [14,19,21,22,26], although this was not entirely consistent [11,13]. Several studies showed increases in flexibility-related outcomes, but lacked significant changes in other more applicable and generic measures of functionality. Only one study followed up on outcome maintenance after the postintervention measurements. This study found that knee torque and timed up-and-go showed improvements compared to the control group, which persisted for four weeks after intervention [24].
In the sub-group containing eight studies of the very old (≥80 years), frail, and assisted-living populations [10,11,13,14,18,21,25,27], there were significant improvements seen in functional reach [21], sit-to-stand [10,14] and 30 m walk times [21], but no changes in the PPT [11] and mixed results for flexibility, strength, balance, and TUG tests. As compared to studies on less aged independently-living populations, the outcomes results of this sub-group were similar, except that the more aged/dependent group had less consistent flexibility outcomes; there were some neutral outcomes and some negative changes following flexibility training.
Six out of the 22 studies included several functional outcome measures and were considered most relevant to this paper [11-14, 16, 19], and are individually reviewed herein. Four of the six conducted randomized control trials (RCT) [11][12][13][14], and only two studies reported sample sizes less than 68 participants. Brown et al. [11] conducted a 12week RCT (n = 87), wherein the flexibility trained group demonstrated only flexibility measure improvements and no change in functional outcomes. The measures included were very practical and included the chair stand, picking up a penny, putting on and taking off a coat, and the Romberg balance test. However, it should be noted that this study also showed minor strength and balance losses. In a one year RCT, King et al. [12], similarly, showed no change in most functional outcomes in the flexibility trained group. Functional measures included lift-and-reach, sit-to-stand, sit-and-reach, self-rated physical performance, self-efficacy for physical performance scale, and perceived functioning and well-being. There were only significant increases for the sit-and-reach test in men only (10.4 to 11.9 inches) and decreases in self-rated daily bodily pain scores (by 7.3% for women and 9.4% for men). Lazowski et al. [13] conducted a 16-week RCT (n = 68). Timed up-and-go increased (worsened) from 27 to 33 seconds, and no changes were seen in any other functional measures or in the sit-andreach test. Additional functional measures included strength tests, Berg balance scale, self-paced and fast-paced walk tests, stair-climbing, and the functional independence measure for functional capacity. Stanziano et al. [14] employed an 8-week RCT (n = 17), where the experimental flexibility group significantly improved in every measure: chair stand repetitions (11 to 13), modified ramp power (69 W to 86 W), arm curl repetitions (12.9 to 18.8), gallon jug shelf test (13.4 to 11.5 seconds), 8 foot timed up-and-go (8.7 to 7.6 seconds), and 50 foot gait speed (13.9 to 12.3 seconds). The study by Bird et al. [16] was a 32-week randomized crossover design (n = 32) with 16 weeks spent in the flexibility group. Four functional outcomes improved significantly for the flexibility group: TUG (7.6 to 6.6 seconds), sit-to-stand (22.6 to 18.0 seconds), step test (13.5 to 17.6 steps), and mediolateral sway range (eyes open: 4.16 to 2.97 cm; eyes closed: 6.87 to 5.64 cm). Strength was also tested, with no improvement by the flexibility group. In a 12-week non-RCT (n = 117), Takeshima et al. [19] reported no improvements for the flexibility trained group in the 12-minute walk, arm curls, chair stand, TUG, functional reach, back scratch, and sit-and-reach. These six studies exemplify the mixed results of the functional outcome measures in this paper. They serve as a strong representation of the lack of consistency of functional outcomes in the literature and therefore, any specific recommendation regarding type or frequency of stretching exercises is premature.
Overall, seven of 22 studies demonstrated mostly positive functional outcomes [9,10,14,16,21,23,29], while six reported mostly negative functional results, that is, no improvement in variables [11-13, 18, 19, 22]. Ten studies showed no functional outcome measures [15,17,20,[24][25][26][27][28][29][30] and only reported changes in flexibility and ROM (although they purported to relate flexibility to function, and were included for their results related to the ability of older adults to improve flexibility). There were no obvious differences between the characteristics of the studies with positive, negative, or no functional outcomes.

Intervention Characteristics.
Twelve studies used flexibility training as the sole intervention [10,11,14,17,[21][22][23][24][25][26][27]29], four studies used flexibility training as a significant part of an intervention protocol [15,16,19,30], five studies used flexibility as a control group to compare with various other exercise interventions [9, 11-13, 18, 20], and four studies utilized flexibility exercises along with strength or aerobic exercise in the intervention protocol but in the control group, the entire exercise protocol was flexibility exercises [9,11,13,20]. The types of flexibility training methods varied from simple static stretches (19 studies) to different proprioceptive neuromuscular facilitation (PNF) techniques (1 study). Passive static stretching was the most common method used [9, 11-13, 16, 18-29], while passive static stretching with added weights was used once [17], activeassisted (AA) was used twice [14,30], active and passive were used in conjunction once [15], contract-relax (CR) PNF was used once [30], contract-relax-agonist contract (CRAC) PNF was used once [30], and hold-relax-agonist contract (HRAC) PNF was used once as well [30]. One study compared multiple methods with each other [30]. Activeassisted stretching had positive and sometimes significant improvements in several outcome measures as compared to the inactive control group, but less significant than the improvements seen with the PNF techniques [30]. Weighted flexibility exercises were similar to nonweighted exercises in one study [15], but significantly better than nonweighted exercises in another [17].
The mean length of the included studies was 14.2 weeks, ranging from 4 weeks to one year [12,20]. Results did not differ significantly throughout the range of intervention durations. In the three studies of at least 25 weeks in duration [12,15,20], no functional outcome measures were improved other than flexibility and ROM. In the four studies with durations of 6 weeks or less [23][24][25]27] TUG improved from 8.4 to 7.2 seconds [24], walking velocity increased 1.07 to 1.22 m/s [23], and flexibility and ROM were improved overall.
The mean frequency was 4-exercise sessions per week, ranging from the lowest frequency of twice per week [9,10,12,22,24] to 14x/week [27,28]. Two studies did not report exercise frequency [20,22]. Some results from these studies include a TUG time decrease from 8.4 to 7.2 seconds [24] and a no-change [10], sit-to-stand time decreased from 10.2 to 9.2 seconds [9] and 9.3 to 7.9 seconds [10] and a no-change [12], a step test increased in reps from 16.5 to 20.2 [9], and an increase in freely chosen gait speed from 1.23 to 1.30 m/s [22]. The 5x/week and 14x/week studies did not include functional outcomes similar to the twice weekly studies. The 3x/week studies included results such as an improved TUG time from 7.6 to 6.6 seconds [16] and a worsening time from 26.8 to 33.0 seconds [13], a functional reach improvement of 16.0 to 19.6 cm [21] and two no-changes [19,26], a sitto-stand improvement of 22.6 to 18.0 seconds [16], a step test improvement of 13.5 to 17.6 repetitions [16], a 10 m walk time improvement from 6.44 to 5.99 seconds [26], and a 30 m walk time improvement from 28.1 to 20.0 seconds [21]. Although limited in number, these results show that the frequency of the flexibility training interventions had no noticeable differences compare with 2 and 3 times per week.
The mean flexibility exercise session time was 32 minutes, ranging from 30 seconds [29] to 85 minutes [19].

Discussion
There are currently scientific discussions regarding the utility of stretching exercises which are regularly recommended and conducted as a part of preexercise protocols to reduce injury and increase performance. Earlier reviews of stretching and flexibility have questioned their value in terms of injury prevention, delayed onset of muscle soreness, and improvement of performance [31,32]. In fact, what occurs physiologically with stretching remains unknown [32]. Due to equivocal evidence thus far, the current American College of Sports Medicine's guidelines for exercise testing and prescription [33] recommended the removal of static stretching as part of a warm-up routine for strength and power activities. Additionally, based on available evidence, the 2011 ACSM position statement for guidance on prescribing exercise suggests performing flexibility after cardiorespiratory endurance or resistance exercise for general fitness programs. This position stand highlighted the need for further research to ascertain the effects of various flexibility prescriptions for various activities and performance goals. From the present review, there is not enough consistent evidence to make recommendations for any specific prescriptions of type, frequency, duration, or length of program related to flexibility training; particularly no specific recommendations can be made regarding the program or dose response of flexibility training to the focus of the present study, the transfer of flexibility gains to functions of daily life, or ability to live independently. A recent systematic review of 106 articles relating the effects of pre-exercise acute-passive static stretching on maximal muscle performance provided 74 methodologically sound studies providing 104 findings [34]. This paper showed that 50% of the 104 findings reported significant reductions in task performances, and the authors concluded that static muscle stretches totaling less than 45 s can be used in pre-exercise protocols without significant decrement to strength, power or speed type tasks; thus a conclusion was to recommend stretches held for at maximum 45 s to avoid loss of strength. Shrier [35] had also previously reported the potential negative acute effects of stretching on performance, but additionally reviewed the literature regarding regular stretching on performance which indicates that regular stretching improves force, jump height, and speed performance. Both reviews recommend further synthesis of the literature with respect to the effects of other forms of stretching on various performance measures.
The difficulties of the ability of this paper to provide a consensus on flexibility training prescription for healthy older adults include the lack of well-conducted studies focused on flexibility in older adults and the lack of consistency in the flexibility protocols employed, functional outcomes measured, and functional results observed. As such, according to the criteria used to assess level of evidence, to recommend stretching/flexibility exercises as a routine component of an exercise program for older adults to enhance health or functional abilities is Level 4, Grade C. The more influential studies in this paper (based on focused flexibility protocols with clear functional outcomes and relatively large sample sizes) [11-14, 16, 19] showed very comparable effects to the overall outcomes of the 26 studies, namely, an ambivalence in the value of flexibility training on functional outcomes that may be related to maintenance of independence in daily activities of older adults. Of these six studies, 5 were RCTs with an average quality assessment score of 18 out of 24. Only two of the six showed improvement in flexibility and functional outcomes ( [14] 16/24; [16] 19/24).
In the sub-group (≥80 years), frail, and assisted-living populations there were significant improvements seen in functional reach, sit-to-stand, and 30 m walk times, but no changes in the PPT, and mixed results were observed for flexibility, strength, balance, and TUG tests. This subgroup was similarly ambivalent in the role of flexibility training with functional outcomes to the rest of the study populations, although this group had less consistency in the flexibility-related outcome measures. Frequency and duration differences between studies showed no noticeable differences in outcomes. When different muscle groups were targeted, the flexibility outcomes were expectedly fairly body-part specific. Regarding the different flexibility training methods, active-assisted (AA) stretching had positive and sometimes significant improvements in several outcome measures as compared to the inactive control group, but less significant than the improvements seen with the PNF techniques. Weighted flexibility exercises were similar to nonweighted exercises in one study, but significantly better than nonweighted exercises in another. One study showed ACR-PNF to be much more effective than CR-PNF and static stretching for both ROM and EMG activity. The overall results point to PNF stretching being more effective than non-PNF techniques for improving flexibility outcomes, but not necessarily functional outcome measures.
While flexibility training interventions synthesized in the present paper have been shown to increase flexibility and joint ROM, no consistent increases in functional outcomes have been observed. Therefore, future studies should consider the relationship that increased flexibility and joint ROM have with functional outcomes to determine if the increased flexibility is beneficial and worthwhile in terms of maintaining or increasing functional capacity for healthy older adults. More research is also needed regarding the relationships between outcome variables (i.e., how one variable such as functional reach would relate to another variable such as the timed up-and-go) and on the relationships between outcome measures and quality of life through self-reported functioning/quality of life questionnaires to best determine the applicability of the outcome measures.
Older adults are less concerned with high performance benefits from increased flexibility and more focused on being safely active and safely performing activities of daily living [36]. Injury and fall prevention are also common motives for recommending flexibility programs to older adults. The 2011 ACSM position statement notes that flexibility training may enhance postural stability and balance when combined with resistance training; however, no consistent link has been shown between regular flexibility exercise and a reduction of musculoskeletal injuries or delayed onset of muscle soreness [4]. However, despite the growing literature describing the relationship of flexibility to injury risk in 28 Journal of Aging Research younger populations [31], there is little research regarding older adults.

Conclusions
This paper found that flexibility training interventions in older adults are often effective at increasing joint range of motion in various joints, and various functional outcomes can be improved. However, due to the wide range of intervention protocols, body parts studied, and functional measurements, conclusive recommendations regarding flexibility training and functional outcomes for older adults remain ungrounded. As such, a specific prescription of how long to hold a stretch, how many repetitions of each stretch to conduct, and the type of stretches to do, is not determinable at this point. Because there is conflicting information regarding both the relationship between flexibility training interventions and functional outcomes, and the relationship between improved flexibility and daily functioning and health benefits have not been established, future research studies should attempt to address these issues.
While there is a lack of evidence to recommend stretching routines outside of a rehabilitative context, there is no additional health or functional risk of including flexibility exercises. As such, in light of increases in functional outcomes achieved by other exercise modes (balance, aerobic exercise, and strengthening exercises), stretching exercises can be included as an adjunct to the above, but the current literature would indicate it would add little to the functional benefits of the other exercise modes. Of note, the evidencebased and expert consensus statements of "Physical Activity Guidelines for Older Adults" of the US, the UK, Canada, and the World Health Organization (Global recommendations) have not included flexibility as a component in the recommendations.