Journal of Rehabilitation Research & Development (JRRD)

Quick Links

  • Health Programs
  • Protect your health
  • Learn more: A-Z Health
Veterans Crisis Line Badge
 

Volume 52 Number 1, 2015
   Pages 1 — 20

Spinal, pelvic, and hip movement asymmetries in people with lower-limb amputation: Systematic review

Hemakumar Devan, MPhty;1* Allan Carman, PhD;2 Paul Hendrick, PhD;3 Leigh Hale, PhD;1 Daniel Cury Ribeiro, PhD1

1Centre for Health, Activity, and Rehabilitation Research, School of Physiotherapy, University of Otago, Dunedin, New Zealand; 2Sports Performance Research Institute New Zealand, Auckland University of Technology, Auckland, New -Zealand; 3Division of Physiotherapy and Rehabilitation Sciences, University of Nottingham, Nottingham, United Kingdom

Abstract — Following amputation, people with transfemoral amputation (TFA) and transtibial amputation (TTA) adapt with asymmetrical movements in the spinal and lower-limb joints. The aim of this review is to describe the trunk, lumbopelvic, and hip joint movement asymmetries of the amputated limb of people with TFA and TTA during functional tasks as compared with the intact leg and/or referent leg of nondisabled controls. Electronic databases were searched from inception to February 2014. Studies with kinematic data comparing (1) amputated and intact leg and (2) amputated and referent leg of nondisabled controls were included (26 articles). Considerable heterogeneity in the studies precluded data pooling. During stance phase of walking in participants with TFA, there is moderate evidence for increased trunk lateral flexion toward the amputated limb as compared with the intact leg and increased anterior pelvic tilt as compared with nondisabled controls. None of the studies investigated spinal kinematics during other functional tasks such as running, ramp walking, stair climbing, or obstacle crossing in participants with TFA or TTA. Overall, persons with TFA adapt with trunk and pelvic movement asymmetries at the amputated limb to facilitate weight transfer during walking. Among participants with TTA, there is limited evidence of spinal and pelvic asymmetries during walking.

Key words: activity of daily living, adaptation, amputation, biomechanics, compensation, functional task, kinematics, lower limb, transfemoral, transtibial.

Abbreviations: LBP = low back pain, TFA = transfemoral amputation, TTA = transtibial amputation.
*Address all correspondence to Hemakumar Devan, MPhty; Centre for Health, Activity, and Rehabilitation Research, School of Physiotherapy, University of Otago, Dunedin, New Zealand; +64-3-479-9619.
Email: hemakumar.devan@gmail.com
http://dx.doi.org/10.1682/JRRD.2014.05.0135
INTRODUCTION

Lower-limb amputation, including transfemoral amputation (TFA) and transtibial amputation (TTA), is increasingly common secondary to vascular and nonvascular etiology [1–2]. Owing to the lack of intact knee and/or ankle joints, persons with TFA and TTA compensate by increased loading on the intact leg as compared with the amputated leg during walking [3–5]. Further, the mechanical limitations of the prosthesis to fully weight bear on the amputated limb and the loss of lower-limb musculature leads to compensatory movements in the hip [6], pelvis [7], and trunk [8–9] segments during walking. Such compensatory movements are asymmetric in nature, with either increased or decreased motion occurring in the joints of the amputated limb as compared with the intact limb and/or referent limb of nondisabled controls [3,5,9]. The terms "movement adaptations" [3,6], "compensatory movements" [10–11], and "asymmetrical movements" [5,9] are often used interchangeably in the amputation literature. While movement asymmetries are a form of adaptation following lower-limb amputation, the potential exists that some of the lumbopelvic and lower-limb movement asymmetries could be "maladaptive," potentially predisposing this population to musculoskeletal disorders such as low back pain (LBP) and osteoarthritis [12–13].

LBP is a common musculoskeletal condition following lower-limb amputation [14–15]. Evidence from prevalence studies confirm that LBP is reported as "more bothersome" than other comorbid conditions such as phantom limb pain and osteoarthritis [16–17]. Further, a majority of respondents with LBP consistently report the presence of LBP for more than 3 yr, which suggests the chronic ongoing nature of LBP in this population [15].1 Potential biomechanical contributing factors for ongoing LBP in this population include proximal movement asymmetries at the trunk and lumbopelvic segments secondary to walking with a prosthesis [12]. Increased lumbar transverse rotation has been reported during walking in persons with TFA and LBP as compared with persons with TFA without LBP (p < 0.05, effect size 1.03) [18]. Despite the cross-sectional study design, the result provides initial evidence for proximal movement asymmetries associated with LBP in people with lower-limb amputation. Such movement and muscle asymmetries in the trunk and lumbopelvic segments could lead to fatigue of spinal musculature and/or cumulative stress of osteoligamentous structures, potentially resulting in spinal instability and LBP [12,19].

Previous systematic reviews have focused mainly on the spatiotemporal parameters and kinetic variables of lower-limb joints during walking [4,20]. Soares et al. reviewed the biomechanical parameters in persons with TTA; however, firm conclusions could not be drawn from this literature review due to the lack of quality assessment of the included studies [20]. Recently, Sagawa et al. reviewed the interlimb movement asymmetries of persons with TFA during stance phase of walking and reported decreased hip motion at the amputated limb in the sagittal plane as compared with the intact limb (p < 0.05); nevertheless, the aim of the review was not specific to kinematics because it investigated various biomechanical parameters such as spatiotemporal parameters, kinetics, and electromyography [4]. Further, the review included studies conducted on participants with both unilateral and bilateral amputation [4]. A recent review reported the muscle compensatory strategies of persons with TFA and TTA during walking [3]; however, it solely investigated kinetic variables such as joint moment, power, and work of lower-limb joints.

Proximal movement asymmetries at the trunk and lumbopelvic segments during walking have received less attention in the lower-limb amputation literature. In addition to walking, it is equally important to investigate other daily tasks such as climbing stairs, walking uphill and downhill, and running. Understanding the proximal movement asymmetries at the trunk and lumbopelvic segments will inform future prospective studies specifically investigating the potential causal relationship between those movement asymmetries and musculoskeletal disorders such as LBP in this population. The aim of this review is to describe the trunk, lumbopelvic, and hip joint movement asymmetries of the amputated limb during functional tasks as compared with the intact and/or referent limbs of people with TFA and TTA.

METHODS
Eligibility Criteria

Our review was limited to observational studies, including cross-sectional, case series, and case studies, because they provide background information for future case-control and prospective studies [21]. Studies involving adults with unilateral TFA and TTA due to all causes of amputation were included. Participants had to be independent while performing functional tasks, which included every day activities such as, but not limited to, walking, stair climbing, lifting or bending, sit-to-stand, and running. The main outcome variable included kinematics of trunk, lumbopelvic, and/or hip joint during functional tasks. For the purpose of the review, asymmetry is defined as a statistically significant difference between the amputated limb and the intact limb and/or the joint segments of nondisabled controls [3–4]. For trunk and lumbopelvic segments, studies comparing amputated and intact sides and persons with amputation and nondisabled controls were included. For hip joint, studies comparing amputated and intact legs and amputated and referent legs of nondisabled controls were included [3]. Studies solely investigating postural control and physiological parameters of participants and comparing different prosthetic foot components during functional tasks were excluded. Peer-reviewed articles published in languages other than English and conference proceedings without full text were excluded.

Literature Search

A comprehensive search strategy was devised (Appendix, available online only) in consultation with a liaison librarian, including the key words "amputation," "adaptation," "asymmetry," "compensation," and "kinematics." The following databases were searched: MEDLINE (via Ovid), EMBASE, AMED (via Ovid), PsycINFO (via Ovid), Cochrane Library (via Ovid), PubMed, CINAHL, Academic Search Complete, SPORTDiscus (via EBSCO), Scopus, Science Direct, Web of Science, Google Scholar, and ProQuest (conference papers and proceedings) from inception to week 3 of February 2014. The primary investigator (H.D.) conducted a hand search of references from the included studies and previous systematic reviews [3–4,20]. The primary investigator also created electronic alerts for the search strategy in major databases such as PubMed, CINAHL, Scopus, and Web of Science to identify potential articles published until March 2014.

Study Selection

All the references from electronic databases were exported to Endnote X5 (Thomson Reuters; Philadelphia, Pennsylvania). Two reviewers (H.D. and P.S.) independently searched the electronic databases. Following duplicates exclusion, both reviewers independently screened the titles and abstracts for relevancy. Next, full-text articles were screened for potential inclusion. Throughout the process, a third reviewer (A.C.) was available to settle any disagreement between the reviewers.

Risk of Bias in Individual Studies

The included articles were assessed for methodological quality based on the modified Down and Black quality assessment tool (Table 1) [22]. This assessment tool was chosen due to its high interrater (r = 0.75) and test-retest (r = 0.88) reliability [22]. Because our review primarily investigated laboratory-based biomechanical studies, items 8, 9, 14, 17, 19, 21, 24, and 26 in the scale were removed because they are specific to randomized controlled trials. The modified tool had 19 items. Items 13 and 23 were modified, and the term "interventions" was replaced with "functional tasks." Item 4 was modified into "Are the methods clearly described?" Based on previous research, the percentage of total quality scores was classified as high (>75%), moderate (50%–74%), and low (<50%) quality [23]. Two reviewers (H.D. and P.S.) independently assessed the quality of included articles; any disagreement was resolved by mutual discussion and a third reviewer (A.C.) was available to resolve any disagreements.


Table 1. 
2. Are the main outcomes to be measured clearly described in the "Introduction" or "Methods" sections?
If the main outcomes are first mentioned in the "Results" section, the question should be answered "no."
Clear description of inclusion and exclusion criteria. (Studies should specify whether the participants were healthy with no pathology, illness, or injury affecting their functional task.)
4A + 4B = 0.5 mark: Adequate (≥3/5) = 0.5, No or Inadequate = 0; 4C + 4D = 0.5 mark: Adequate (≥3/5) = 0.5, No or Inadequate = 0.
5. Are the distributions of principal confounders in each group of subjects to be compared clearly described?
Adequate (≥4/8) = 1, No or Inadequate = 0.
Tables and graphs should be clearly presented so that the reader can check the major analyses and conclusions.
7. Does the study provide estimates of the random variability in the data for the main outcomes?
Normal distribution: SD, SE, or 95% CI. Non-normal distribution: interquartile range. If the distribution of the data was not described, it must be assumed that the estimates used were appropriate and the question should be answered "yes."
10. Have actual probability values been reported for the main outcomes except where the probability value is less than 0.001?
Actual p-value (e.g., p = 0.035 rather than p < 0.05).
11. Were the subjects asked to participate in the study representative of the entire population from which they were recruited?
The study must describe how the participants were recruited. Study would be representative if they comprised the entire source population, an unselected sample of consecutive patients, or a -random sample. When a study does not report the proportion of the source population from which the sample is chosen then the question should be marked as "unable to determine."
12. Were those subjects who were prepared to participate representative of the entire population from which they were recruited?
The proportion of those who agreed should be stated. Validation that the sample was representative would include demonstrating that the distribution of the main confounding factors was the same in the study sample and the source population.
The study should represent some measures to emulate the task analyzed being generalizable to daily functional tasks. For example, self-selected pace of walking or use of standardized stair length and width.
16. If any of the results of the study were based on "data dredging," was this made clear?
Any analysis that had not been planned at the outset of the study should be clearly indicated. If no retrospective unplanned subgroup analyses were reported, then mark "yes."
For studies where the outcome measures are clearly described, and studies which refer to other biomechanical validation studies or other 3D models, the question should be answered "yes."
22. Were cohorts of participants in study group and control group recruited at the same time?
For a study that does not specify the time period over which participants were recruited, the question should be marked "unable to determine."
25. Was there adequate adjustment for confounding in the analyses from which the main findings were drawn?
If the effect of main confounding factors was not investigated or confounding was demonstrated but no adjustment was made in the final analyses, the question should be marked as "no."
Data Collection Process

The following information was extracted based on a standardized form of the Cochrane Collaboration of Systematic reviews [24] by the primary investigator (H.D.) and verified by second reviewer (P.S.): study design, functional task, participant characteristics (age, cause and years since amputation, and type of prosthesis), instrumentation, and outcome measures. For trunk and lumbopelvic joint segments, both total motion from the segment and the data from the amputated and intact sides of a particular segment were extracted. For hip joint, total range of motion from the amputated and intact legs of persons with lower-limb amputation were extracted. The trunk motion was defined as movements occurring only at the thoracic segment including both upper and lower thoracic segments [25]. Only the kinematic data at comfortable walking speed and with a neutral prosthetic alignment were extracted, because walking speed could influence the joint kinematics [26].

Synthesis of Results

The mean difference and 95 percent confidence intervals of kinematic data between the intact and amputated limbs and/or referent limbs of nondisabled controls from individual studies was calculated using Reference Manager 5.2 (Thomson Reuters). For studies without such information, the primary investigator (H.D.) extracted the data from graphs, and where necessary, the authors were contacted via email for additional information. Due to the age of the included articles, this was not always possible. Since we observed considerable variations in presenting the results of kinematic data and limited studies investigating a functional task, pooling study results was not possible. Thus, a descriptive summary of results is presented. The Cochrane Back Review Group rating scale was modified to summarize the level of evidence as strong (consistent findings from multiple high-quality studies), moderate (consistent findings from one high-quality study and one or more moderate- to low-quality studies or multiple moderate- to low-quality studies), limited or conflicting evidence (findings from one high-, moderate-, or low-quality study or inconsistent findings from multiple studies), and no evidence (no studies) [27].

RESULTS
Study Selection

The electronic search strategy resulted in 2,679 articles (Figure). After title, abstract, and full-text screening, 21 articles were included for the final review [6–7,9,28–45]. Five additional articles were included by searching the references of included articles (n = 3) [8,18,46] and from our recent electronic search in February 2014 (n = 2) [47–48]. No additional articles satisfying the inclusion criteria were retrieved from the electronic alerts. Of the final 26 articles, 9 investigated kinematics in persons with TFA [8–9,18,29–30,35,38–40] and 12 in persons with TTA [6,28,31–34,36,41–42,45–47]; 5 investigated both persons with TFA and TTA [7,37,43–44,48].


Figure 1. Search results.

Figure 1.

Search results.

Click Image to Enlarge. View as PowerPoint Slide

Study Characteristics
Participants

Table 2 and 3 present summaries of included studies investigating persons with TFA and TTA, respectively. Most of the included studies (n = 23) adopted a cross––sectional study design, with the exception of three studies that had a case series design [7,9,30]. Most of the included studies (n = 17) recruited a minimum of 10 participants for the study [8–9,18,28,31–35,37,40–43,46–48]. The majority (82%) of amputations were due to either trauma or tumors.


Table 2. 
Age, yr
(mean ± SD
or range)
Cause of Amputation
Time Since Amputation, yr (mean ± SD)
Bae et al., 2007 [29]
Cross-
sectional
7-camera Vicon (30 markers,
3 DOF)
Sagittal, frontal
Case series
Ottobock (3), USMC 24800 (1)
Flexfoot (2), M+IND SLF 135 (2)
4-camera motion (12 markers,
3 DOF)
Sagittal, frontal
Goujon-
Pillet et al., 2008 [8]
Cross-
sectional
Trauma (24), tumor (2), congenital (1)
Monoaxial (19), polycentric (8)
12-camera Vicon (46 markers,
6 DOF)
Trunk,
pelvis
Sagittal, frontal,
horizontal
Hendershot & Wolf, 2014 [48]
Cross-
sectional
23-camera Vicon (30+ markers,
6 DOF)
AMP vs CON, AMP vs INT
Sagittal, frontal,
horizontal
Case series
Monoaxial (3), polycentric (8)
Electrogoniometers (trunk, hip, and knee angles)
Sagittal, frontal
Michaud
et al., 2000 [7]
Case series
Flexfoot (2), Seattle foot (1)
CODA 3 system (2 markers, 1 DOF)
Morgenroth et al., 2010 [18]
Cross-
sectional
Trauma (14), tumor (1), congenital (1),
vascular (1)
10-camera Vicon (41 markers,
6 DOF)
Lumbar spine
Sagittal, frontal,
horizontal
Rabuffetti
et al., 2005 [35]
Cross-
sectional
4-camera system (12 markers, 3 DOF)
Schmalz
et al., 2007 [37]
Cross-
sectional
Stair descending
4-camera
PRIMAS (8 markers, 3 DOF)
Sjödahl
et al., 2002 [38]
Cross-
sectional
Pneumatic (6), mechanical (3), hydraulic (2)
Flexfoot (6), Seattle foot (2), multiflex ankle (1)
5-camera Vicon (21 markers,
3 DOF)
AMP vs INT, AMP vs CON
Sjödahl
et al., 2003 [39]
Cross-
sectional
Pneumatic (6), mechanical (3), hydraulic (2)
Flexfoot (6), Seattle foot (2), multiflex ankle (1)
5-camera Vicon (21 markers,
3 DOF)
AMP vs INT, AMP vs CON
Frontal,
horizontal
Tazawa, 1997 [40]
Cross-
sectional
6-camera Vicon (12 markers,
6 DOF)
Sagittal, frontal,
horizontal
Vrieling et al., 2007 [43]
Cross-
sectional
Obstacle crossing
Trauma (4), tumor (3),
vascular (1)
Electrogoniometers (hip, knee, and ankle joints)
AM P vs INT, AMP vs CON
Vrieling et al., 2008 [44]
Cross-
sectional
Ramp
walking
Trauma (4), tumor (3)
Graph-lite (3), C-Leg (1), 3R60 (1), Safelife (1), Total Knee (1)
Multiflex (3), SACH (2), C-Walk (1)
Electrogoniometers (hip, knee, and ankle joints)
AM P vs INT, AMP vs CON

Table 3. 
Age, yr
(mean ± SD)
Cause of Amputation
Time Since Amputation, yr (mean ± SD)
Prosthesis
Foot Type
Alimusaj et al., 2009 [28]
Cross-
sectional
Stair
climbing
Trauma (13), tumor (3)
Vicon (number of cameras and markers = NA, 6 DOF)
Cross-
sectional
Ramp
walking
Trauma (13), tumor (3)
Vicon (number of cameras and markers = NA, 6 DOF)
Cross-
sectional
20-camera system (55 markers, 6 DOF)
AMP vs CON, AMP vs INT
Hill et al., 1997 [46]
Cross-
sectional
Obstacle crossing
Flexfoot (4),
Re-Flex Pylon (3), Seattle foot (2), Safe foot (1)
5-camera Vicon (14 markers, 3 DOF)
Trunk, hip, knee, ankle
Cross-
sectional
Dynamic foot (7), Greissenger (2), Flexfoot (2)
8-camera system (markers = NA, 6 DOF)
Sagittal,
frontal
Michaud et al., 2000 [7]
Case series
Flexfoot (2),
Greissenger (1), Seattle foot (1), Carbon Copy II (1)
CODA 3 system (2 markers, 1 DOF)
Cross-
sectional
Stair
climbing
Trauma (8), vascular (2)
6-camera Vicon (23 markers, 6 DOF)
Pelvis, hip, knee, ankle
Molina Rueda et al., 2013 [47]
Cross-
sectional
Trauma (12), tumor (3)
Flexfoot (7), Ceterus (2),
Variflex (2), SACH (1), Talux (1), Trias (1), Quantum (1)
8-camera Vicon (23 markers, 6 DOF)
Sanderson & Martin, 1996 [36]
Cross-
sectional
Single video camera (markers = NA, 3 DOF)
AMP vs INT, AMP vs CON
Sanderson & Martin, 1997 [6]
Cross-
sectional
Single video camera (10 markers, 3 DOF)
AMP vs INT, AMP vs CON
Schmalz et al., 2007 [37]
Cross-
sectional
Stair
climbing
Trauma (7), tumor (1)
4-camera PRIMAS (8 markers, 3 DOF)
Cross-
sectional
Trauma (6), vascular (3), congenital (2)
Multiflex (7),
Variflex (2), Dynamic (1), Ceterus (1)
10-camera Qualysis (28 markers, 6 DOF)
AMP
(fallers) vs AMP (non-
fallers)
Sagittal,
frontal
Trauma (6), vascular (3), congenital (2)
Multiflex (7),
Variflex (2), Dynamic (1), Ceterus (1)
10-camera Qualysis (28 markers, 6 DOF)
AMP
(fallers) vs AMP (non-
fallers)
Obstacle crossing
Trauma (6), tumor (4),
vascular (2)
Electrogoniometers (hip, knee, and ankle joints
AMP vs INT, AMP vs CON
Trauma (5), vascular (1)
8-camera Vicon (24 markers, 6 DOF)
Outcome Variables

Overall, 16 studies investigated kinematics during walking [6–9,18,29,32–33,35,38–41,45,47–48] and 4 studies investigated kinematics of stair climbing [28,34,37,42], with few studies investigating the kinematics of other functional tasks, i.e., obstacle crossing (n = 2) [43,46], ramp walking (n = 2) [31,44], and running (n = 2) [30,36]. During walking, overall 11 studies investigated pelvis (n = 7) [7–8,29,35,38–40], lumbar spine (n = 1) [18], and trunk (n = 4) [8–9,40,48] kinematics in participants with TFA. Only three studies investigated pelvis (n = 1) [7] and trunk (n = 2) [47–48] kinematics in participants with TTA. For functional tasks such as stair climbing, ramp walking, obstacle crossing, and running, only lower-limb kinematics at the hip, knee, and ankle joints were investigated.

Risk of Bias Within Studies

Table 4 presents the scores of included studies. Most of the studies (n = 20) were classified as moderate quality [6–9,28–34,36–39,44–48], with four classified as high quality [19,41–43] and two classified as low quality [35,40]. The total quality assessment score was 12.0 ± 2.8 (mean ± standard deviation). The percentage agreement scores between the two reviewers were good (Cohen kappa: 0.67). Only four studies scored >50 percent in both external validity and internal validity sections [18,33,41,43]. None of the included studies reported power calculation and all scored poorly for this question.


Table 4.