Tilted seat position for non-ambulant individuals with neurological and neuromuscular impairment: a systematic review

Objective: To determine the effects of tilt-in-space seating on outcomes for people with neurological or neuromuscular impairment who cannot walk. Data sources: Search through electronic databases (MEDLINE, Embase, CINAHL, AMED). Discussions with researchers who are active in field. Review methods: Selection criteria included interventional studies that investigated the effects of seat tilt on outcome or observational studies that identified outcomes for those who had used tilt-in-space seating in populations with neurological or neuromuscular impairments. Two reviewers independently selected trials for inclusion, assessed quality and extracted data. Results: Nineteen studies were identified which fulfilled the selection criteria. Seventeen of these were essentially before–after studies investigating the immediate effects of tilting the seating. All studies looked at populations with neurological impairment, and most were on children with cerebral palsy (n = 8) or adults with spinal cord injury (n = 8). Reviewer's conclusion: Posterior tilt can reduce pressures at the interface under the pelvis.


Introduction
Tilt-in-space wheelchairs and seats are increasingly used by people with neurological or neuromuscular impairments who cannot walk. Tilt-inspace systems may be considered for a variety of reasons, including low sitting tolerance or discomfort, a requirement to rest in the seat, and to assist with manual handling. 1 Drawbacks to these systems compared with conventional wheelchairs and seats include purchase costs, size and complexity of equipment. Tilt-in-space wheelchairs are also heavier and less manoeuvrable than more standard wheelchairs due to a longer wheelbase, and this may restrict access to transport. 2 A backwards-tilted sitting position has been suggested to improve head and trunk posture, 3,4 and to reduce the loading under the buttocks [5][6][7] or through the spine. 8 There are concerns that seating that is excessively tilted back limits communication, upper limb function and the ability to stand up from the chair. 9 A forward-tilted sitting position has also been proposed to maintain lumbar lordosis, decrease posterior pelvic tilt, reduce the effect of tight hamstrings on the position of the pelvis and to position a person within reach of the desk or table. 10,11 Forward-tilted positions have been incorporated into some paediatric seating. 12 With no evidence-based criteria or guidelines for provision and use of these systems, practices around the provision of tilt-in-space seating systems vary widely. Tilt-in-space seating may be provided by statutory service in some areas. Systems are also available for purchase directly by the user.
In a qualitative study 2 of severely disabled wheelchair users with multiple sclerosis and significant spasticity themes such as wheelchair size and manoeuvrability, transport difficulties, comfort, pressure ulcers, sitting up during day for prolonged periods and fatigue emerged from in-depth interviews. Seven tilt-in-space and 16 conventional wheelchair users participated.
With this background it was thought that a systematic review on the effects of tilt-in-space seating might inform clinical practice on seating provision and use within these populations, and identify what further research studies on this topic are required in order to establish evidencebased guidelines for provision.

Objective
To identify the effects of seat orientation on physiology; body parts and systems; and on activity for adults and children with neurological or neuromuscular impairments who cannot walk.

Search strategy
A search was carried out in December 2006 of electronic databases including MEDLINE , Embase (1980Embase ( -2006, CINAHL (1982CINAHL ( -2006, AMED (1985AMED ( -2006 using thesaurus terms 'wheelchair', 'wheelchairs', 'seat', 'seating' and free text words 'tilt$' and 'tip$' looking for articles in English on humans. Reference lists in studies and review articles were examined for other appropriate articles. A search for unpublished studies was conducted via contact with experts in the field.

Selection criteria
Studies were identified that investigated the effects of seat tilt on outcome for the seated individual. Experimental studies that compared outcomes at different angles of tilt were included as were observational studies that compared outcomes for those that had used tilt-in-space seating to those that had used a seat in a fixed orientation. A tilt of the seat was taken to be a rotation of the complete seat about a mediolateral axis, and tilt angle is as described in Figure 1. Studies with both randomized and nonrandomized allocation of subjects to seat or seat orientation were selected for review.
Studies included only participants who were non-ambulant and who had a congenital or acquired neurological or neuromuscular condition. Participants could be of any age.
Any outcome was considered that described the effects of seat tilt on physiology; body parts and systems; and on human activity including fulfilment of societal roles.

Data collection and analysis
The two reviewers independently selected trials for inclusion, assessed quality and extracted data.
The methodological strength of each study was evaluated using a commonly used hierarchy of study designs from the NHS Centre for Reviews and Dissemination. 13 Methodological strength was graded on a scale from 1 to 5 where 1 is the highest level (Table 1).
Quality was also assessed in addition to methodological strength. This was based on: whether the study was properly controlled; what methods of randomization or allocation to intervention groups were used; and whether the groups were comparable at baseline. The roles of chance, confounding and bias in the study were also considered. Attempts were made to contact authors to obtain any important data that were missing and necessary for the review.
The studies included in this systematic review were not only randomized control trials. This is because studies have tended to focus on instantaneous outcomes as a result of being tilted compared with upright and alternative designs have often been used (e.g. cross-over trials). However, in appraising such studies particular attention was given to identifying potential sources of bias. Cross-over trials could be rated at levels 1, 2, 4 or 5 depending on samples size and homogeneity; whether the effects of order, timing and knowledge of the intervention on outcome were controlled and validity of outcome measures.
A generally descriptive analysis was selected as most appropriate for the research question, because of the heterogeneity of the studies that were identified. However, a meta-analysis was also carried out involving published and unpublished results from five studies which looked specifically at body/support interface pressure under the ischial tuberosities. A more conservative random effect model was used rather than a fixed effect model due to the presence of heterogeneity across the studies. 14 It was not possible to directly combine all the data in one meta-analysis as two of the studies 15,16 had reported measurements taken from the same participants while sitting on different cushion configurations and therefore the data sets were not considered truly independent of one another. However two separate meta-analyses were carried out, using the results for specific cushions in each study corresponding to the best and worse case scenarios (i.e. most and least pressure reduction).

Results
Of the 389 publications identified in the electronics database searches, only 15 fulfilled the selection criteria (Appendix 1). An additional five publications were identified by other means. Two publications referred to the same study.
Nineteen studies were identified ( Table 2). All of the studies were on populations with neurological impairment. Ten of the studies were on young people: with cerebral palsy (n ¼ 8), neural tube defect (n ¼ 1), or unspecified neurological impairment (n ¼ 1). Nine of the studies were on adults: with spinal cord injury (n ¼ 8) or multiple sclerosis (n ¼ 1).
The seat was tilted anteriorly by up to 30 in three of the studies, was tilted posteriorly by Experimental study (e.g. RCT with concealed allocation) 2 Quasi-experimental study (e.g. experimental study without randomization) 3 Controlled observational study: (a) cohort study, (b) case-control study 4 Observational study without control group 5 Expert opinion based on pathophysiology, bench research or consensus up to 45 in 13 studies and was tilted in both directions in three studies. Several studies included additional interventions. Additional seat configurations and postures were included in the studies of Nwaobi et al. 17 Miedaner, 18 Pellow, 6 Vaisbuch et al., 7 Janssen-Potten et al. 19  Seventeen of the studies were essentially crossover trials comparing seat orientation ( Table 2). Myhr and von Wendt's study 22 can be considered as a series of case reports because of the range of seats and orientations involved in the intervention. In another study 3 a single child was seated at three angles of tilt. In the second part of Chan and Heck's study 4 subjects were randomly assigned to two groups that were tilted back to two different angles of tilt.
In 10 studies the order of tilt was randomized at each measurement session. Two studies 23,24 looked at ordering effects by repeating the measurements in a reverse order and comparing outcomes. In one study the full set of seat positions were measured over multiple sessions 23 . In the other studies all seat positions seemed to be measured in a single session. The measurement period in each position varied between a few seconds to 20 minutes. It was not possible to blind the subject or the researcher to the intervention(s) in any of the studies.
Outcomes included: interface pressure, [5][6][7]15,16,20,25 shear force, 5,20 surface EMG, 17,19,23,26,27 postural measurements, 4,18,21,23,26 change in head position 23,28 , timed upper extremity activity, 24,28 respiratory measurements, 4,29 voice volume 4 and perceived exertion 4 ( Table 3). Figure 2 shows a forest plot for five of the six studies that investigated interface pressure under the ischial tuberosities. 5,7,15,16,25 It was not possible to include Pellow's study 6 as insufficient data were reported and there were only two participants. Spijkerman's 16 unpublished data were used in the analysis and it was necessary to make a conservative calculation of the standard deviation from the reported significance level for Hobson's study. 5,20 Multiple results shown for particular studies 15,16 relate to the use of different seat cushions. An inspection of Figure 2 suggests a reduction in interface pressure when participants were posterior tilted (between 20 and 45 ) compared with upright.

Discussion
The restriction of the search to papers written in English may have limited the findings of the review. The comparative difficulty in identifying unpublished studies compared to published work may also have limited the findings.
Wide search criteria were used in the systematic review because there was not thought to be much evidence available on the effects of tilted positions. Therefore selection included studies on a range of populations, interventions, experimental methodologies and outcomes.
Studies on different populations (spinal cord injury and neural tube defect) and at different tilt angles were included in the meta-analysis. There were insufficient data to rigorously test the validity of this strategy.
Studies included in the meta-analysis were randomized 7,16 and non-randomized 5,15,25 trials where participants acted as their own controls. Ideally the order of tilt orientation should have been randomized in all the studies that were included in the meta-analysis as this would remove a potential source of bias.

Experimental design
Most of the studies were cross-over trials looking at the immediate effects of seat orientation on the seated person. With a cross-over experimental design there is potential for tilt order to affect outcome due to fatigue and other physiological responses. Tilt order will also affect outcome if there are changes to the baseline sitting posture during the experimental procedure due to sliding in the seat. In a cross-over study it is feasible to control the effects of order of tilt through experimental design. Measurement at different tilt angles may take place in different sessions. Alternatively, the tilt sequence may be randomized across the sample or the measurements at each tilt angle may be repeated in a different order. These approaches are recommended in future cross-over studies.
It is also possible for knowledge of the seat orientation during the experimental protocol to affect the outcome. Unfortunately it is not practical to blind the subject or experimenter to the orientation of the seat.
The quantitative studies which compared outcomes on different seats (or at different tilt angles) involved small samples of fewer than 20 people. There is potential for actual differences between tilt angles or seats not to be identified as significant because of the distribution of data within the small samples (a type II error). As the number of reported results increase, there will be scope for additional meta-analysis.
In some of the studies on the effects of an anterior seat tilt the intervention comprised a forward tilt of the seat base without additional support about the pelvis or trunk. 18,23,26  The variation in findings between studies may be because protocols did not control for other influences on posture.
The number of seating systems on the market providing an anterior tilt is limited, and such seating is not widely used. This may be due to difficulties using these systems in vehicle transport and using them with desks and powered mobility systems. For this reason an investigation into the effects of forward tilt may not be the highest priority for the next stage of research.
No cohort studies were identified which investigated longer term effects of tilt-in-space usage with a quantitative methodology. This approach may be worth considering for future work.

Outcomes
Outcomes measures in most of the studies were related to abnormality of anatomical structure or function (impairment). There was little consideration of the importance of any differences that were identified to the health or social participation of the user.
Six studies reported that tilting the seat back reduced the pressure under the ischial tuberosities in a range of conditions. However the sample sizes involved in the above studies were relatively small and the methods of statistical analysis and levels of significance (when reported) varied noticeably. Pooling data across five of these studies in a meta-analysis produced more robust evidence of a statistically significant reduction in pressure under the ischial tuberosities when participants are tilted backward compared to when upright.
Hobson's 5,20 finding of reduced frictional shear stress underneath the seat base with a 200 posterior tilt, is consistent with a generalized biomechanical analysis of a seated person. 30 Loading at the interface with the seat is likely to influence susceptibility to pressure ulcers and comfort during sitting. A cohort study on pressure ulcer prevalence in tilt-in-space wheelchair users compared with in a control group of conventional wheelchair users would identify whether the reduction in loading when tilted backwards results in reduced pressure ulcer prevalence for tilt-in-space users.
Studies in different muscle groups and in the cerebral palsy and spinal cord-injured populations have reported that EMG activity in some muscle groups is affected by tilt. 19,26,27 In populations and muscle groups where raised activity restricts functional movements and leads to the  development of contractures, decreased activity may be advantageous. Reduced muscle activity may also be associated with reduced effort during movements or with the maintenance of position. However in other populations and circumstances, increased muscle activity may be associated with increased functional movements and improved posture. Overall, the effect of seat tilt on EMG activity and how that affects functional outcomes has not been established. Postural measurements were either between anatomical markers, or between an anatomical marker and the seat surface and were focused on trunk and head position in the sagittal plane. In the studies where measurements were taken from photographs or video frames 21,28 there was potential for error from neglected out-of-plane components of position. The postural results overall were inconclusive (Table 3).
Head control has been assessed from measurements 23,28 of head position over time. Sochaniwskyj 23 used a potentiometric linkage, however the measurements were not set into a functional context. Head control has also been assessed from observations of head position over time, 3,22 but in Myhr and von Wendt's study 22 the inter-rater reliability of the observers was reported for only two of six positions and ranged from 0.9 to 0.31 using Spearman's rank correlation coefficient.
Ability to perform an activity from the seat is a key aspect of any study on the effects of seat tilt. Nwaobi 24 used timed switch operation and McClenaghan et al. 28 used timed tasks as measures of upper extremity function. Myhr and von Wendt 22 evaluated hand and arm function using observational techniques and a rating scale. Respiratory measurements were included as an outcome in only two of the studies identified by this review. Reid and Sochaniwsky 29 made indirect measurements of tidal volume via plethysmography whereas Chan and Heck 4 took measurements of vital capacity using lung function spirometry. Additional studies on capabilities in tilted postures for specific populations would be worth while.
No studies on ability to transfer into and out of the seat were identified in the populations of interest. Studies on other more ambulatory populations 31 have suggested that ability to independently transfer may be reduced by a posteriorly tilted position. However many people within the populations that are covered by this review have to use a hoist to transfer into and out of the seat, so the effect should be investigated separately.

Effects within populations
No studies were identified on the effects of seat tilt on people with progressive neuromuscular conditions (e.g. muscular dystrophy). This population would benefit from study, as the question of whether to provide a tilt facility on a wheelchair is a common clinical issue.
The studies with cerebral palsy were on young people and tended to measure posture and muscle activity. However it was not possible to identify consistent finding from these studies due to variation in interventions, outcome measures and heterogeneity of the population. The use of the Gross Motor Function Classification System 32 in future investigations to identify the participants' level of physical ability would enable clinicians to judge the advantages and disadvantages of varying angles of tilt for specific children.
Most of the studies in populations with spinal cord injury and neural tube defect were on the effects of seat tilt on interface loading. This is an important outcome in populations that are prone to pressure ulcers.
The only quantitative study that was identified was one on people with multiple sclerosis by Chan and Heck. 4 Themes which emerged from in-depth interviews 2 with this population included prolonged sitting up during day and fatigue. Chan and Heck 4 attempted to identify the immediate effects of a change in orientation on fatigue using Borg's Rating of Perceived Exertion scale. However, the increase in fatigue with tilt that was identified is likely to be affected by their protocol, which involved a fixed order of tilt.
Previous cohort studies on how fatigue, duration of sitting, other health and social factors are affected by long-term use of a tilted position have not been identified. Future cohort studies on tilt-in-space seat usage, compared with standard seat usage would greatly inform clinical practice.

Conclusions
Results from studies on populations with spinal cord injury and neural tube defect suggest that a posterior seat tilt of 20 or more reduces pressures under the pelvis.
Overall there is a lack of quality evidence to support and guide the use of the tilted position in seating for populations with neurological and neuromuscular impairment. Current evidence is weakened by mixed interventions and confounding factors. Outcome measures, participants and interventions need to be determined more rigorously to ensure that confounders do not reduce the quality and usefulness of future studies.
A priority area for future studies might be effect of posterior seat tilt on functional activity and seat use, in populations with progressive neuromuscular conditions.