Secondary Logo

Journal Logo

Original Research Article

Contact Stress Distribution Under a Lower Thorax Partial Orthosis Worn by Healthy Young Men

Effects of Distraction Force, Sagittal Plane Moment, Respiration Phase, and Thoracic Cross-sectional Shape

Johnson, Daniel D. PhD; Ashton-Miller, James A. PhD; Shih, Albert J. PhD

Author Information
JPO Journal of Prosthetics and Orthotics: January 2013 - Volume 25 - Issue 1 - p 48-57
doi: 10.1097/JPO.0b013e31827bad4e
  • Free


Patients with stable spinal fractures or deformities such as idiopathic scoliosis often have to wear orthoses as part of their treatment. Such orthoses typically apply sustained forces to the trunk and pelvis to immobilize one or more motion segments of the spine1 or apply corrective bending and/or torsional moments.2 In the field of orthotics, braces exert their corrective actions by applying what is termed pressure to the skin over bony landmarks or regions. However, in tissue mechanics, that interface pressure may be decomposed into a contact stress oriented normal to the skin surface (henceforth termed contact stress) and two orthogonal interface shear contact stresses in the plane of the skin. When orthoses are custom-made for each patient,3 orthotists generally seek to reduce normal and shear contact stresses over pressure points by redistributing the stress over a larger area and/or using univalve (for plaster) or bivalve (for fiberglass) designs.4 Sustained or abnormally large contact stress is one of many factors in the etiology of skin abrasion, blisters, and pressure ulcers,5 a serious injury entailing the breakdown of the skin and underlying tissues.6 Even though it is not the sole predictor of the risk of pressure ulcers,7 a better understanding the magnitude of the mechanical load, whether in the form of a normal or a shear stress, or both, that is applied to a unit area of skin by such an orthosis can still be useful in trying to predict and prevent skin complications in current and future orthosis designs.

The contact stress distribution under a thoracolumbosacral orthosis has been measured in patients being treated for idiopathic scoliosis in several studies using a variety of sensor technologies.2,8–11 The recorded stress values showed a wide variation between each study. For example, van den Hout et al.8 observed mean magnitudes varying between 12 and 55 kPa depending on the posture of the patient (supine, prone, sitting, standing, etc.) and location (lumbar padding produced the highest, thoracic the least), whereas Pham et al.10 observed mean stress magnitudes of 7–10 kPa under similar testing conditions. Whether this variation is caused by the technology used to make the measurement, the implementation of the orthosis, or subject selection is not known.

Breakdown of the skin and even subcutaneous tissues under an orthosis is a continual source of concern for the patient and clinician. One mechanism under mechanical stress is ischemia or restriction of the blood supply at the site of loading. This can particularly occur over bony prominences,12 which include the iliac crests, spinous processes, and the ribs. Ischemia is thought to be caused by inadequate perfusion of nutrients and removal of waste products. This is a process that is regulated through a complex series of electrochemical channels that control the relative dilation and, thus, flow rate of capillaries and other adjacent vessels.13 The capillaries near the skin surface in most healthy persons have an average internal pressure of approximately 25–32 mm Hg (3.33–4.27 kPa), and any applied contact stress exceeding this may eventually occlude these vessels,4,6 creating localized ischemia. The skin is able to tolerate higher contact stresses for brief periods of time without any permanent damage; when pressure is removed, tissues that had been under contact stress experience a period of hyperemia, in which the body dilates the blood vessels in the affected area to flush out the accumulated metabolites and restore adequate oxygenation. The time required to fully recover from a sustained applied contact stress varies according to the magnitude and duration of loading.4 Husain14 showed that, in general, the time that skin and muscle tissue can tolerate sustained contact stress has an inverse relationship to the magnitude of the contact stress. The same study also showed that skin and muscle tissue can tolerate contact stresses upward of 100 mm Hg (13.3 kPa) for periods of up to 2 hrs. Although there have been some attempts to better model damage from sustained contact stress in concert with other factors,15 standard prevention still focuses on the temporal threshold (2 hrs of sustained immobility for a patient) and being watchful for risk factors other than contact stress magnitude.6

Knowledge gaps that will be addressed in the present study include documenting the magnitude of the contact stress at the skin-orthosis interface at several locations around the lower thorax under various external loads, including the “damage” thresholds of 32 and 100 mm Hg. This study will also address how the contact stress depends on the distraction force and bending moments applied to the orthosis, the phase of respiration (full inhalation and full exhalation), and, for a given modular orthosis component, the eccentricity (out-of-roundness) of the lower thorax. We tested the null hypothesis that the distribution of the contact stress at the skin-orthosis interface would not be affected by distraction force, sagittal plane bending moment, respiration phase, or the geometry of the lower thorax in healthy young men.



Twenty healthy young male adults between 20 and 30 years of age and with body mass indices (BMIs) between 19.47 and 28.41 kg/m2 were recruited. The exclusion criteria included history of medical treatment for vertebral fractures, spondylolysis, spondylolithesis, congenital abnormalities of the spine, scoliosis, kyphosis, osteoporosis, recurrent back pain, or disc herniation. All participants gave written informed consent and all procedures were approved by an institutional review board.



The design of a thin and flexible transducer suited to measuring the normal contact stress at the skin-orthosis interface over the lower rib cage was developed by modifying a force-sensitive resistor (FSR) described by Perner-Wilson.16 Two linear arrays of seven piezoresistive sensors each were created (Figures 1 and 3). Each sensor consisted of a core of conductive polymer (Velostat®; 3M, St Paul, MN) sandwiched between two layers of conductive fabric (ArgenMesh™; Less EMF, Inc, Albany, NY), all of which were sewn and glued together in a polyester sleeve. The active sensor area encompassed where the Velostat and ArgenMesh layers overlapped entirely, approximately 38.1 mm × 19.1 mm. Two standard-sized 16-nap fasteners were used for electrical connections.

Figure 1:
Left, A single custom-made piezoresistive contact stress sensor, exploded view. The sensor layers are glued and sewn together in final assembly. Center, Example of sensor calibration data with resulting linear-fit calibration function. Right, Example of load cell calibration data with resulting linear-fit calibration function.

The electrical resistance of each sensor decreased as the contact stress applied to the active sensor area increased. To measure the change in resistance, each sensor was placed in series with a 10 kΩ resistor as part of a voltage divider circuit (Figure 2). Each array of seven sensors (and their corresponding voltage divider circuits) shared a single driving voltage input of +5 V direct current and a single local ground. Each sensor/circuit signal was sampled by an analog input channel at 1 kHz via a 16-bit DAQCard-6024E (National Instruments, Inc, Austin, TX) analog-to-digital converter board and notebook PC running LabVIEW™ (version 9.0, National Instruments). Note that each sensor was numbered according to its corresponding analog input channel on the DAQ (channels 2–15).

Figure 2:
Contact stress sensor circuit layout. Each sensor used one analog input channel on the data acquisition unit (DAQ). Each sensor array (seven sensors) used one +5 V and one ground channel, shared across each of its sensors.
Figure 3:
Test orthosis segments (shown in white), line attachment points, and sensor array placement, exploded view. Not shown: adjustable buckles and straps connecting each adjacent segment. Adhesive tape secured each sensor array during testing.

Before each session with a participant, each sensor array was laid on a flat surface and each sensor was calibrated individually using a sequence of nonordinal applied loads ad modum Brimacombe et al.17 First, an aluminum plate of 12.54 g sized to fit the active sensor area (called the “base plate”) was applied to the sensor. Next, a known reference mass was applied on top of the base plate. After a few seconds, data were recorded for 5 seconds; the mean of the recorded voltage readings was stored as the calibration point for that load. After a reading was taken, the sensor was offloaded except for the base plate and then loaded with the next mass in the sequence. A total of six masses were used to calibrate each sensor in this manner, in the following sequence: 1.0, 0.7, 1.5, 0.9, 1.7, and 1.2 kg (resulting in applied pressures of 13.6, 9.6, 20.4, 12.3, 23.1, and 16.3 kPa, respectively). See Figure 1 for an example sensor calibration graph.

To accurately measure the external forces that were imposed on each test participant, two load cells (TLL-500; Transducer Techniques, Inc, Temecula, CA) were used in tension as part of the apparatus (see below). Each load cell was connected to a digital signal conditioner (DPM-3; Transducer Techniques), which then output a DC voltage updated at a frequency of approximately 60 Hz that varied in proportion to the applied tensile force. As with the contact stress sensors, the load cells were calibrated before each session with a test participant using a series of known applied loads, with data recorded for 5 seconds at each load and the mean of the readings used as the calibration value (as for the contact stress sensors). First, the digital signal conditioner was balanced to discount the mass of the load cell, attachments, and cabling from any subsequent readings. The load cells were then calibrated by loading them in sequence with masses of 0.5387, 5.537, 4.538, and 9.5369 kg. See Figure 1 for an example load cell calibration graph.


Once each participant arrived, he was first asked to remove any shirts, jewelry, and others, from the upper body and was then fitted with a shirt made from a section of tubular stockinette (tg® Tubular; Lohmann & Rauscher, Inc, Topeka, KS). Participant height and mass were then measured using a medical scale. Next, the lower edge of the ribcage (encompassing the 11th and 12th, or “floating,” ribs) was located by palpation while the participant held a full inhalation. Using a flexible measuring tape, a single measurement of the circumference of the lower edge of the ribcage was then taken at both a full inhalation (maximum lung volume) and full exhalation (minimum lung volume). Next, calipers were used to record single measurements of the lateral and midsagittal diameters of the lower edge of the ribcage, also at both full inhalation and full exhalation.


Each participant was assisted in donning the testing apparatus (Figures 3 and 4), which consisted of an adjustable harness (the “test orthosis”) lined with the two arrays of contact stress sensors and attached at two points (anterior and posterior midline) to wire rope cables anchored in an overhead gantry. Loads were applied via two cables (the “load lines” shown in Figure 4), each of which contained an inline load cell. Two additional cables (the “safety lines” shown in Figure 4) were also attached to the test orthosis, which could catch and hold the participant upright above the floor in case he fainted or the load lines failed. Each load line could be attached to a combination of up to four constant force springs mounted in the overhead gantry: three springs were rated at approximately 44 N and one at approximately 22 N.

Figure 4:
Test orthosis and sensor placement over stockinette shirt on a test subject. Left, Anterior view. Center, Lateral view. Right, Lateral view with testing apparatus components labeled for visibility.

The test orthosis (shown with subcomponents labeled in Figure 3) consisted of four flexible plastic segments: two flat padded plates over the sternum and spine (each with an attachment point for cabling) and two semicircular segments, each of which encircled one side of the ribcage. The segments were connected by nylon webbing and buckles, which allowed the separation between the segments to be adjusted. The test orthosis was adjusted for each participant such that the semicircular segments were positioned to cover the floating ribs (centered on each side) and each plate was centered on the midline.


For the first group of tests, approximately equal forces were applied to the two load attachment points of the test orthosis to apply a cranial distraction force to the thorax. Corresponding front and back loads were varied by changing which of the constant force springs were connected to each load line, with increasing loads from 22 to 154 N in 22-N increments. At each applied load, each participant was asked to fully inhale and hold his breath for 5 seconds while data were recorded. After taking several normal breaths, each participant was asked to fully exhale and hold his breath, again for 5 seconds, as data were recorded.


The second battery of tests consisted of applying two sequences of increasing moments to each test participant at the midsagittal plane. To produce each sequence of moments, one load line remained connected to the lowest-rated coil spring (22 N) as the load on the other load line was increased from 44 to 154 N. As in the distraction tests, data were recorded in 5-second intervals as participants held a full inhalation followed by a full exhalation. Once the first moment sequence was complete, the sequence was repeated for the other load line.


All data were initially processed using Excel (version 2007; Microsoft Corp, Redmond, WA). First, the height and mass of each participant were used to calculate his BMI. Next, the eccentricity of the ribcage of each participant (approximated as an ellipse for the purposes of this study) was calculated by specifying the frontal plane and midsagittal diameters of the lower thorax as the major and minor axes of the ellipse. The eccentricity was calculated at both the full inhalation and exhalation and then averaged to produce a single average thorax eccentricity value for each participant (used as a factor in the statistical analysis, below). Finally, the calibration data (applied contact stress vs. measured voltage) for each sensor were fitted to its own linear function. The data for each sensor were then converted from a measured voltage into a contact stress value (in kPa).


The postprocessed data were subjected to statistical analysis using IBM SPSS Statistics (version 19.0.0; International Business Machine Corp, Armonk, NY). Each sensor was analyzed individually using a mixed linear model (for repeated measures), with the recorded contact stress chosen as the single dependent variable. Participant breath status (fully inhaled or fully exhaled), average thorax eccentricity, normalized distraction force, and normalized moment were chosen as the dependent factors to be analyzed for significance, using a 95% confidence interval. A p value < 0.05 was considered significant for the test of each hypothesis.


The mean quasistatic contact stress magnitude data for each pair of corresponding left-right sensors from the distraction and moment test batteries are shown in Figures 5 and 6, respectively. The results of the F tests for statistical significance of all factors in the mixed linear model constructed for each sensor are shown in Table 1. Individual data points were excluded only if the calculated contact stress reading (produced from the measured voltage reading and linear calibration function) was negative.

Table 1:
Mixed linear model results
Figure 5:
Distraction test battery results, all subjects, by sensor pair (right and left). Horizontal and vertical bars denote SEM.
Figure 6:
Moment test battery results, all subjects, by sensor pair (right and left). Horizontal and vertical bars denote SEM.


The distraction and moment tests produced different ranges of observed contact stresses. In the distraction test battery, the recorded contact stress varied from a minimum of 8.2 kPa (sensors 8 and 9, 22 N/22 N front/back loading, exhaled; Figure 5G) to a maximum of 26.5 kPa (sensors 7 and 10, 154 N/154 N front/back loading, inhaled; Figure 5F), which corresponds to a range of approximately 61.7–198.3 mm Hg. In the moment test batteries, the recorded contact stress varied from a minimum of 13.0 kPa (sensors 2 and 15, 22 N/44 N front/back loading, exhaled; Figure 6A) to a maximum of 27.8 kPa (sensors 7 and 10, 154 N/22 N front/back loading, inhaled; Figure 6F), which corresponds to a range of approximately 97.7–208.7 mm Hg.

The normalized applied distraction force was a significant factor affecting the recorded contact stress for every sensor except sensor 15 (Table 1). Among every other sensor, the largest fixed effect was observed at sensor 7 (+34.6 kPa per unit of normalized distraction force), whereas the smallest was observed at sensor 12 (+14.1 kPa per unit of normalized distraction force). All sensors exhibited a positive proportional relationship between normalized applied distraction force and recorded contact stress.


Normalized applied moment was also found to be a significant factor affecting the contact stress recorded at most sensors, with some exceptions (see Table 1): sensor 4 (p = 0.083), sensor 13 (p = 0.304), sensor 14 (p = 0.083), and sensor 15 (p = 0.392). Among those sensors where the normalized applied moment was found to be significant, the largest effect for a positive moment (flexion) was observed at sensor 2 (+448.9 kPa per unit of normalized moment), whereas the largest effect for a negative moment (extension) was observed at sensor 7 (−444.3 kPa per unit of normalized moment). The smallest magnitude estimated effect from the normalized applied moment was observed in sensor 12 (−49.1 kPa per unit of normalized moment). It should also be noted that the sign of the estimated effect from normalized applied moment reversed at one location in each sensor array: between sensors 4 and 5 (attached to participant right) and between sensors 12 and 13 (attached to participant left).


The averaged recorded contact stresses show that for nearly each pair of sensors at every test point, the stresses recorded when the participants held a full inhalation were higher than those when they held a full exhalation (Figures 5 and 6). The most obvious exception to this occurred in sensors 2 and 15 in the moment tests, which shows the full-exhalation contact stress to become greater than the full-inhalation contact stress for small negative (extension) normalized moments (Figure 6A). It should also be noted that many readings overlap within the bounds of standard error.

When the output of each sensor is examined individually via the mixed linear model analysis, participant respiration phase (full inhalation or full exhalation) was found to have a significant effect on the recorded contact stress for nearly every sensor except sensors 14 and 15 (p = 0.738 and p = 0.489, respectively). The effect of switching between the two states was calculated using a full exhalation as the default state for the mixed linear model of each sensor; thus, the effect estimates listed in Table 1 are the estimated changes in contact stress in each sensor as the participant fully inhales. For each sensor where respiration phase was found to be significant, the recorded contact stress was found to increase when the participant fully inhaled; the largest estimated effect was found to occur in sensor 2 (+5.0 kPa) and the smallest was found to occur in sensor 5 (+1.5 kPa).


Average thorax eccentricity was found to vary between 0.405 and 0.787. Using the mixed linear models for each sensor, it was found to be a significant factor affecting recorded contact stress for only sensors 10, 12, and 13 (p = 0.022, 0.046, and 0.023, respectively), which were all on the sensor array attached to the participant left. The estimated effect average thorax eccentricity produced on skin contact stress was similar in magnitude (but not necessarily sign) for each of the three sensors: −9.0 kPa (sensor 10), +6.2 kPa (sensor 12), and +7.8 kPa (sensor 13) (Table 1). For the other sensor array (sensors 2–8), slight positive trends were observed for three sensors (2, 3, and 6, with p = 0.089, 0.087, and 0.074, respectively), but thorax eccentricity did not approach significance for any other sensor on either array.


This is the first study that describes the contact stress distribution developed between the skin and a partial thoracic orthosis for a given applied external distraction force and moment, as well as other contributing factors. The results show that the contact stress was significantly affected by applied external distraction force, applied external moment, and respiration phase at nearly every location tested. The effect of the eccentricity of the lower thorax was not significant for all but a small minority of measurement sites. For all test conditions, recorded contact stresses exceeded 32 mm Hg (4.27 kPa), the average contact stress required to occlude the capillaries near the surface skin. In nearly all testing conditions, the contact stresses exceeded the 100 mm Hg (13.3 kPa) threshold for eventual damage first suggested by Husain.14 The range of contact stresses observed is similar to that observed by van den Hout et al.8 in their study, which also featured an orthosis applying sustained forces (for correction of scoliosis), albeit at different load application sites from this study.

The mean recorded contact stress values for every sensor increased with applied normalized distraction load across the entire testing range, generally outside the bounds of standard error (as shown in Figure 5). Any external load imposed on the test orthosis was ultimately counterbalanced by contact stress distribution applied to the thorax of each participant. The lack of any noticeable decrease in the mean contact stress at all sites suggests that there are no secondary loading mechanisms that arise at any loading threshold within the testing range. This is reasonable given that the modular orthosis essentially acted as a ring halfway up the lower torso (shaped like an inverted truncated cone) so as to lift it. The shape of the loading curve likely depends on the shape of the thorax, the compliance of the bony thorax, the thickness of subcutaneous fat and muscle over the ribs, the muscle tone in the intercostals muscles, and the mechanical behavior of the tissues being loaded. Apart from perfusion away from the site of application (blood, lymph, interstitial fluid), skin undergoes several structural changes when subjected to contact stress, including a reorientation of the collagen fiber bundles constituting much of the intercellular support structure along with elastin fibers.5,18 Such time-dependent changes in the tissue produce creep effects in the load/displacement behavior of skin under load.4 Finally, skin and other tissues require time to regain full integrity after the contact stress is removed, with the time required increasing with both the magnitude and duration of the applied contact stress.14

For the moment tests, it was expected that a more pronounced difference between the contact stresses recorded at corresponding positive and negative moments would appear in sensors further from the midline (sensors 5 and 12), the approximate center of rotation. Sensors at the posterior of the orthosis (such as sensors 8 and 9) would be expected to record much higher average mean contact stresses when a positive (flexion) moment was applied to the orthosis than when a negative (extension) moment was applied, and vice versa. These large expected differences did not appear in the data; in fact, a slight trend toward the opposite appears at those sensors near the anterior and posterior of the test orthosis. This is most likely because of how moments were generated in the moment test batteries. All of the tests in this study used combinations of vertical distraction loads applied at the anterior and posterior of the test orthosis. Thus, the tests involving an applied moment also included a large distraction force applied simultaneously. Any large differences in recorded contact stress between positive and negative moments from the moment alone may have been masked by the balanced contribution of the distraction force component of the loading. As for the slight reverse trend in observed loading differences, this may have arisen from participants not remaining in an entirely erect posture, especially at the higher load configurations. The participant may have partially shifted toward the opposite surface for comfort/balance, thus redistributing more contact stress to the sensors opposite the point of load application (i.e., for a large negative moment, participants may have assumed a partially extended posture and shifted some of their weight to the posterior of the test orthosis).

Respiration phase (full inhalation or full exhalation) was expected to have a pronounced effect on the recorded contact stress levels, and this result is clearly seen in the relatively large separation in readings produced by the subjects at each breathing phase for corresponding loading configurations. The effect is especially pronounced at lower load configurations: approximately less than 0.20 normalized distraction force and between ±0.002 normalized applied moment. As noted above, the highest recorded contact stresses were observed in tests where the subjects were at maximal inhalation, whereas the lowest contact stresses were recorded when subjects were at maximal exhalation. This compares with the study by Pham et al.10 of an orthosis worn to treat idiopathic scoliosis; respiration was found to be significant in contributing to recorded contact stress at the orthosis-skin interface, but the effect produced smaller changes in contact stress (approximately 1.3 kPa at most) than those observed here. It should be noted that the range in contact stress values recorded between inhalation and exhalation was not constant across all loading configurations. In general, as the overall external applied loads increased in magnitude, the difference between inhalation and exhalation contact stresses shrank: for the distraction tests, the range varied between approximately 9.2 kPa (sensors 8 and 9, 22 N front/back loading; Figure 5G) and 0.4 kPa (sensors 7 and 10, 154 N front/back loading; Figure 5F). As suggested above, the shape of the loading curves (and specifically, the separation between those at maximal inhalation and exhalation) may be influenced by creep effects that appear in the tissue under sustained loading. Had all imposed external loading been removed for some period of time between each test, giving the loaded tissues time to recover full integrity (and elastic response), the separation in recorded contact stresses at full inhalation and exhalation might have remained more constant.

There are several limitations to this study. First, perhaps the greatest factor influencing the accuracy and precision of the data is the reliability of the contact stress sensors used in this study. Each sensor array was constructed by hand for this study, and there are likely to have been many inconsistencies in the dimensioning and performance of each individual sensor because of this. In addition, the conductive layers within each sensor are only held together by the polyester sleeve and the glue applied to the edges during assembly; they can separate from one another within the sleeve under certain conditions (e.g., a sensor is held vertically, unsecured by any portion of the active sensor area). If separation occurs, the electrical resistance of a sensor increases above the nominal maximum that occurs for zero applied contact stress, thus resulting in a nonsensical negative contact stress reading. These sensors also share many of the same limitations as other so-called FSRs: hysteresis, drift, requiring a “break-in” period before behavior becomes stable enough to be consistent.19 Certain countermeasures taken in this study were suggested by Brimacombe et al.17 and can help mitigate these effects, but an even more thorough construction methodology might improve the validity and reliability of the contact stress measurements, including 1) adding an extra calibration period after each subject completes all of the test batteries (which would capture any effect of drift over the testing period), 2) completely offloading the orthosis in between each test (as in the calibration phase), and 3) randomizing the testing configuration (rather than conducting the test batteries in the same sequence). Second, the fit of the orthosis most likely greatly influenced the results from each subject and may throw any analysis of the effect of the average thoracic eccentricity into doubt. The test orthosis, although made of a soft polystyrene and broken into segments to improve its ability to conform to varying geometries, still developed some resistance to bending in the lateral strips where each sensor array was mounted. This may have made sensor readings near the anterior and posterior poles of those with more elliptical thorax geometries appear much lower than would appear in a well-fitted orthosis. Any lack of proper fit may also have led to shifting of the test orthosis over the course of the testing period, which could cause one or more sensors to move from intercostal tissue onto a rib (thus creating a sudden stress concentration) or vice versa. The ideal testing scenario would use a custom test orthosis for each subject, a scenario that was cost-prohibitive for this feasibility study. Third, this study did not take into account the element of time. Both the tissues being loaded and the sensors used in this study are known to exhibit some degree of variation in their responses to load with time (i.e., creep effects). Although each testing period was short (5 seconds per test configuration, and total subject time in-harness limited to approximately 45 minutes), those tests conducted last might have demonstrated pronounced creep effects, and this statistical analysis would be unable to separate from other factors. Future work may require set recovery time periods between each loading configuration, to ensure full tissue recovery between testing, as well as tracking loads with time for each sensor (rather than averaging, as was done here). Fourth, the loading protocol required the use of “impure” moments, which were largely determined by the nature of the test orthosis: constant distraction forces were required to maintain contact with each subject. To more effectively analyze the effect of imposed moments, a more rigid test orthosis would have to be subjected to torque without any distraction forces. Fifth, more precise anthropometry could be provided by optical scanning systems to trace the thoracic contours and geometries.20,21 Sixth, although subjects were instructed to inhale and exhale fully during testing, the exact phase of respiration was not measured during testing. As external loading increased, subjects may not have been able to successfully reach maximum inhalation, meaning that the effect of respiration may be underestimated for higher load configurations. To produce a more accurate characterization of the effect of respiration phase, measurements of relative lung air volume should be used in future work to ensure consistent respiration phases throughout testing. Finally, this study examined the brace effects only in a modest population of young, healthy men. The results cannot be extrapolated to women, obese individuals, the elderly population, or those with musculoskeletal impairments without further research.

There is no single threshold of maximum “safe” contact stress that can be applied indefinitely to the skin of a specific person, with the possible exception of dermal capillary pressure. Thus, knowing the precise magnitude of the applied contact stress at any moment may not be as important as awareness of other risk factors for contact stress–related injury, such as duration of loading. Contact stress monitoring can be used in existing orthoses as a means of reminding the patient/caregiver that the patient needs to be repositioned (i.e., have the contact stress load on his or her skin redistributed) at regular intervals to prevent damage. Monitoring the levels of contact stress applied to the skin by various orthoses is also a useful design tool when it is used to assess the relative contact stress magnitudes created by various orthosis configurations. In general, it can be said that applying lower contact stress magnitudes at the skin is an improvement, as long as the required clinical effects (immobilization, corrective forces/moments) are still satisfactory.

As a final note, it should be remembered that this study did not include measurements of the two orthogonal interface shear contact stress components that contribute to the overall mechanical loading created at the surface of the skin when it contacts another body. However, some conclusions about the overall shear stress (and its consequences) can still be made. As part of their literature review, Sanders et al.5 state that the duration that the skin can tolerate shear stress decreases as the magnitude of the shear stress increases. Of course, shear stress on the surface of the skin is produced by friction, which is a product of surface characteristics and normal (contact) stress. Thus, contact stress applied to the skin contributes not only to pressure-related injuries (i.e., pressure ulcers) but also to shear-related injuries (i.e., blisters). Reducing the magnitude of applied contact stress in an orthosis (again, through repositioning and/or design) thus has the potential to reduce the likelihood of both forms of injury.


The authors acknowledge the University of Michigan Department of Neurosurgery for supporting the research presented in this article.


1. Flanagan P, Gavin TM, Gavin DQ, Patwardhan AG. Spinal orthoses. In: Lusardi M, Nielsen C, eds. Orthotics and Prosthetics in Rehabilitation. Woburn, MA: Butterworth-Heinemann Medical; 2000: 231–252.
2. Chase AP, Bader DL, Houghton GR. The biomechanical effectiveness of the Boston brace in the management of adolescent idiopathic scoliosis. Spine 1989; 14: 636–642.
3. Bernardoni GP, Gavin TM. Comparison between custom and noncustom spinal orthoses. Phys Med Rehabil Clin N Am 2006; 17: 73–89.
4. Remaley D, Jaeblon T. Pressure ulcers in orthopaedics. J Am Acad Orthop Surg 2010; 18: 568–575.
5. Sanders JE, Goldstein BS, Leotta DF. Skin response to mechanical stress: adaptation rather than breakdown—a review of the literature. J Rehabil Res Dev 1995; 32: 214–226.
6. Leigh IH, Bennett G. Pressure ulcers: prevalence, etiology, and treatment modalities—a review. Am J Surg 1994; 167: 25S–30S.
7. Reenalda J, Van Geffen P, Nederhand M, et al.. Analysis of healthy sitting behavior: interface pressure distribution and subcutaneous tissue oxygenation. J Rehabil Res Dev 2009; 46: 577–586.
8. van den Hout JAAM, van Rhijn LW, van den Munckhof RJH, van Ooy A. Interface corrective force measurements in Boston brace treatment. Eur Spine J 2002; 11: 332–335.
9. Périé D, Aubin CE, Lacroix M, et al.. Biomechanical modelling of orthotic treatment of the scoliotic spine including a detailed representation of the brace-torso interface. Med Biol Eng Comput 2004; 42: 339–344.
10. Pham VM, Houilliez A, Schill A, et al.. Study of pressures applied by a Chêneau brace for correction of adolescent idiopathic scoliosis. Prosthet Orthot Int 2008; 32: 345–355.
11. Lou E, Hill DL, Raso JV. A wireless sensor network system to determine biomechanics of spinal braces during daily living. Med Biol Eng Comput 2010; 48: 235–243.
12. Black J, Baharestani M, Cuddigan J, et al.. National Pressure Ulcer Advisory Panel’s updated pressure ulcer staging system. Dermatol Nurs 2007; 19: 343–349.
13. Wywialowski EF. Tissue perfusion as a key underlying concept of pressure ulcer development and treatment. J Vasc Nurs 1999; 17: 12–16.
14. Husain T. An experimental study of some pressure effects on tissues, with reference to the bed-sore problem. J Pathol Bacteriol 1953; 66: 347–363.
15. Gefen A. Risk factors for a pressure-related deep tissue injury: a theoretical model. Med Biol Eng Comput 2007; 45: 563–573.
16. Perner-Wilson H, Buechley L, Satomi M. Handcrafting textile interfaces from a kit-of-no-parts. In: Proceedings of the Fifth International Conference on Tangible, Embedded, and Embodied Interaction (TEI ’11). New York, NY: ACM; 2010:61–68.
17. Brimacombe JM, Wilson DR, Hodgson AJ, et al.. Effect of calibration method on Tekscan sensor accuracy. J Biomech Eng 2009; 131: 034503–1–4.
18. Edsberg LE, Natiella JR, Baier RE, Earle J. Microstructural characteristics of human skin subjected to static versus cyclic pressures. J Rehabil Res Dev 2001; 38: 477–486.
19. Hall RS, Desmoulin GT, Milner TE. A technique for conditioning and calibrating force-sensing resistors for repeatable and reliable measurement of compressive force. J Biomech 2008; 41: 3492–3495.
20. Fourie Z, Damstra J, Gerrits PO, Ren Y. Evaluation of anthropometric accuracy and reliability using different three-dimensional scanning systems. Forensic Sci Int 2011; 207: 127–134.
21. Susato S. Development and application of portable manual non-contact-type anthropometric instruments for measuring human anatomical longitudinal parameters. J Physiol Anthropol 2011; 30: 55–67.

thorax; skin; interface; contact stress; pressure; respiration; distraction

© 2013 American Academy of Orthotists & Prosthetists