Multisurgeon assessment of surgical decision-making in adolescent idiopathic scoliosis: curve classification, operative approach, and fusion levels

STUDY DESIGN

A multisurgeon assessment of curve classification, selection of operative approach, and fusion levels via a case study presentation.

OBJECTIVES

To evaluate the ability of a group of scoliosis surgeons, not involved in the development of a new classification system, to accurately choose the corresponding curve classification of adolescent idiopathic scoliosis (AIS) cases and to evaluate the variability in the selection of operative approaches and both proximal and distal fusion levels in accordance with the new classification system in operative adolescent idiopathic scoliosis.

SUMMARY OF BACKGROUND DATA

Recent evaluations using the King method for classifying AIS has shown poor intraobserver and interobserver reliability. A new, comprehensive classification system of AIS has been developed, but the result of a scoliosis surgeon's ability to apply the objective classification is unknown. In the surgical treatment of AIS, there are three choices for the operative approach (anterior, posterior, or both) and multiple choices for the selection of fusion levels.

METHODS

During an AIS roundtable discussion at a spinal surgery meeting, 28 scoliosis surgeons were presented seven cases of operative AIS via good quality slides. Standard preoperative radiographs and clinical photographs were presented, and the reviewers were asked to classify the cases by a new classification system, choose their preferred surgical approach, and classify both proximal and distal fusion levels.

RESULTS

For the seven cases presented, 84% of the curve types, 86% of lumbar modifiers, and 90% of sagittal thoracic modifiers were classified by the reviewers as described in the new classification. The case study found widely variable operative approaches and fusion levels chosen by the reviewers. There was an average of five different proximal (range, 4-8) and four different distal (range, 3-5) fusion levels chosen by the reviewers for each case.

CONCLUSIONS

This case study assessment found a relatively high rate (84-90%) of agreement in curve classification of the individual components of a new classification system of AIS. This suggests the ability of a group of scoliosis surgeons to identify the specific criteria necessary for this new classification system of AIS. In addition, the high variability in selection of both operative approach and fusion levels confirms the current lack of standardized treatment paradigms. This further reinforces the need for a method to critically and objectively evaluate these variable treatments to determine the "best" radiographic and clinical results.

The inverse effects of load transfer and load sharing on axial compressive stiffness

BACKGROUND CONTEXT

The goal of spinal instrumentation is to stabilize involved motion segments while fusion occurs. Although some degree of load sharing is necessary for fusion, the ability of the instrumentation system to transfer the load may vary.

PURPOSE

The purpose of this study is to formulate a mathematical relationship between load sharing and load transfer of specific spinal instrumentation systems using a well-accepted mechanical model.

STUDY DESIGN/SETTING

Forty-eight American Society for Testing Materials standard ultra high molecular weight polyethylene cylinders were used as per designation F 1717-96, standard test methods for static and fatigue for spinal implant constructs in a corpectomy model.

METHODS

Twenty-four spinal assemblies consisting of anterior plates, anterior rod, and posterior rods were subjected to compression bending tests using a MTS Bionix servo-hydraulic material testing apparatus. Each implant was tested in compression bending with and without the addition of a titanium load-sharing cage. The force applied was the independent variable, and the displacement was the dependent variable. The stiffness was determined for each setup with and without the addition of an anterior load-sharing cage.

RESULTS

The average axial compressive stiffness of a system increased by a factor of 8.5 with the addition of the load-sharing cage. An inverse relationship existed between the compressive stiffness of the construct and its relative increase achieved with the addition of the load-sharing cage. The compressive stiffness of the system with the addition of the load-sharing cage approached that of the anterior device itself as the system flexibility increased. The ability of instrumentation systems to load share or load transfer and their respective stiffness was determined.

CONCLUSIONS

The 5-mm rod screw posterior system was compared with the 7-mm Ti posterior system with the addition of one and two devices for transverse traction (DTTs). The rods with the increased diameter had a stiffness of 1723 n/mm with one DTT and 1815 n/mm with two DTTs. The addition of an anterior cage had little effect on the stiffness of these systems. Anterior plate and screw/rod systems were analyzed and showed similar mechanical behavior to the 5-mm posterior rod/screw systems. A significant increase in stiffness was realized with the addition of an anterior cage. A means to determine the load sharing/transferring properties of a spinal instrumentation system is presented. This technique will allow the amount of load transferred from the fusion mass to the instrumentation to be predicted.

Results of the Scoliosis Research Society instrument for evaluation of surgical outcome in adolescent idiopathic scoliosis. A multicenter study of 244 patients

STUDY DESIGN

An outcome questionnaire was constructed to evaluate patient satisfaction and performance and to discriminate among patients with adolescent idiopathic scoliosis.

OBJECTIVES

To determine reliability and validity in a new quality-of-life instrument for measuring progress among scoliosis patients.

SUMMARY OF BACKGROUND DATA

Meta-analysis of the surgical treatment of adolescent idiopathic scoliosis determined that a uniform assessment of outcome did not exist. In addition, patient measures of well-being as opposed to process measures (e.g., radiographs) were not consistently reported. This established the need for a standardized questionnaire to assess patient measures in conjunction with process measures.

METHODS

The instrument consists of 24 questions divided into seven equally weighted domains as determined by factor analysis: pain, general self-image, postoperative self-image, general function, overall level of activity, postoperative function, and satisfaction. The questionnaire takes approximately 5 minutes to complete and is taken at predetermined time intervals. A total of 244 of patients from three different sites responded to the questionnaire.

RESULTS

The reliability based on internal consistency was confirmed with a Cronbach's alpha coefficient greater than 0.6 for each domain. In addition, acceptable correlation coefficient values greater than 0.68 were obtained for each domain by the test-retest method on normal controls. Similarly; to establish validity of the questionnaire, responses of normal high school students were compared with that of the patients. Consistent differences were noted in the domains between the two groups with P < 0.003. The largest differences were in pain (control, 29.96 +/- 0.20; patient, 13.23 +/- 5.55) and general level of activity (control, 14.96 +/- 0.20; patient, 12.16 +/- 3.23). Examination of the relationship between the domains and patient satisfaction showed that pain correlates with satisfaction to the greatest degree (Pearson's correlation co-efficient, r = -0.511; P < 0.001), followed by self-image (r = 0.412; P < 0.001).

CONCLUSIONS

This questionnaire addresses patient measures for evaluation of outcome in adolescent idiopathic scoliosis surgery by examining several domains. It also allows for dynamic monitoring of scoliosis patients as they become adults. This is a validated instrument with good reliability measures.

Occipital screw pullout strength. A biomechanical investigation of occipital morphology

STUDY DESIGN

A three-group design with consistent pullout strength measures.

OBJECTIVES

To determine pullout strength of three fixation types (unicortical screws, bicortical screws, wires) and to investigate their correlation with respect to occipital morphology.

SUMMARY OF BACKGROUND DATA

A secured, multidirectional occipitocervical fusion requires internal fixation. Devices secured at occipital protuberance were suggested to offer the greatest pullout strength because of this region's thickness.

METHODS

Twelve fresh human cadaveric occiputs were sketched with a grid delineating 21 fixation sites. Each site was drilled and hand-tapped. Four specimens were instrumented with unicortical screws on one side of the midline and bicortical screws on the other. Another four were instrumented with bicortical screws and wires, and the remaining four were instrumented with unicortical screws and wires. Two points on each specimen were secured with identical fixation to examine side-to-side symmetry. An MTS materials testing apparatus (MTS Systems Corporation, Eden Prairie, MN) was used to displace the fixators. Pullout strengths at different anatomic locations were recorded.

RESULTS

The greatest pullout strength was at the occipital protuberance for all fixation types. The bicortical pullout strength was 50% greater than unicortical. The wire pullout strength was not significantly different from that of the unicortical screw (P > 0.05). Seventy-eight percent of wires broke at 1100 N. Unicortical pullout strength at occipital protuberance was comparable with that of the bicortical screw at other locations.

CONCLUSIONS

Unicortical screw fixation at occipital protuberance offers acceptable pullout strength without the potential complications of bicortical screws or wire fixation.

Technique for the harvesting of human cadaveric spines

STUDY DESIGN

One hundred twelve fresh cadaveric spines were harvested using a newly described technique.

OBJECTIVES

To develop and describe a technique for the expeditious excision of intact human cadaveric spines for biomechanical testing, to educate the dissector on the health and safety issues involved in harvesting spinal specimens, and to review the present recommendations for storage and preservation of spinal segments.

SUMMARY OF THE BACKGROUND DATA

As the need for biomechanical spinal research continues to expand, the demand for fresh human cadaveric vertebral specimens increases. Previous techniques for harvesting are simplistic and sparse. This technique offers a reliable and expeditious method for procurement of spinal vertebral segments of any size.

METHODS

Human cadaveric spines were harvested using an adaptation of previous posterior spinal approaches. Techniques for sectioning each vertebral region were developed. Detailed description of these techniques was meticulously documented. The procured spinal segments have been used for multiple biomechanical investigations.

RESULTS

The technique has been used successfully in more than 100 spinal harvests. Approximate time required is 30 minutes. The harvested segments have been reliable biomechanical specimens in many published studies.

CONCLUSIONS

A new technique for the rapid extraction of human cadaveric spines has been developed. Dissectors may benefit from the recommendations offered for sectioning of each region.

Substitution of transpedicular screws by hook claws in a vertebrectomy model

At times one pedicle screw cannot be used in a posterior construct because of morphometric reasons or when the pedicle is fractured intraoperatively. It is not clear from the literature whether and to what extent a construct's stability is compromised by substituting a hook claw for one pedicle screw. A synthetic vertebra model using a hook claw in lieu of a pedicle screw in a compressive and torsional mechanical study is evaluated. Isola screw-based constructs varying in the number of pedicle screws and hook claws were used. To recreate a realistic clinical scenario, transverse connectors and a Harms cage were used as well. In compression, substituting screws with claws did not substantially change the stiffness, whereas substituting one or two pedicle screws with one or two hook claws significantly reduced torsional stiffness. We conclude that using one or two hook claws is a viable alternative for pedicle screws when the latter cannot be used in a clinical situation, especially as an adjunct when the anterior column is reconstructed.

The mechanical role of laminar hook protection of pedicle screws at the caudal end vertebra

Biomechanical studies have shown hooks to be superior to pedicle screws in pull-out, especially in osteoporosis. This study evaluates the possible increase in stiffness of pedicle screws provided by laminar hooks while applying non-destructive forces to a vertebrectomy model assembled with Compact Cotrel Dubousset (CCD) instrumentation. Synthetic vertebrae were employed in a three-level vertebrectomy model. CCD screw-based three-level constructs with and without sublaminar hooks in the caudal element were tested in flexion, extension, compression, lateral bending, and torsion. There was no statistically significant advantage in adding inferior laminar hooks to a caudal end vertebra that had bilateral pedicle screws in any of the testing modes. Torsional stability, however, was augmented, but not significantly. Torsional instability and osteoporotic bone may be the clinical justifications for adding laminar hooks below screws in the caudal end vertebra.

Occipital morphology. An anatomic guide to internal fixation

STUDY DESIGN

The authors present the results of an anatomic study of the human occiput to delineate appropriate screw placement sites.

OBJECTIVES

Occipital bone morphologic characteristics were evaluated to determine whether significant variability exists and to determine the position of greatest bone thickness for safe and effective internal fixation.

SUMMARY OF BACKGROUND DATA

New instrumentation and techniques for occipital fixation are being developed in response to concerns about occipital bone variability. Thirty cadaveric occiputs were evaluated to determine if such variability exists and the location of greatest bone thickness. Radial thickness, occipital locations, and gender differences, were determined.

METHODS

Twenty-six skulls were sectioned sagittally to determine the contributions of the inner, middle, and outer tables to overall occipital thickness. The angle required to gain maximal cortical purchase was determined. Mean values and variance were analyzed statistically to determine variability and thickness. Data was plotted in three dimensions. Variability in morphologic features was minimal.

RESULTS

The internal occipital protuberance-external occipital protuberance was thickest at 17.55 mm (SD = 3.18 mm) and was consistently located on the superior nuchal line 43 degrees from the horizontal skull base line. Bone thickness decreased radially from the central internal occipital protuberance position. Bone thickness above the superior nuchal line exceeded that below by 2.74 mm (P < 0.05) vertically and at the oblique positions (P < 0.05). Bone to the right of the midline was only 1 mm thicker than that to the left. Gender differences were minimal. The inner table contributed only 10% to overall occipital thickness. As occipital thickness decreased, the optimal purchase angle increased.

CONCLUSIONS

Unicortical purchase at and above the superior nuchal line is warranted with a low risk of intracranial venous penetration. Internal fixation devices developed in response to occipital bone variability should be considered with respect to occipital bone thickness distributions. Attention to cervical morphologic characteristics should result in higher success rates in occipitocervical arthrodesis.

Meta-analysis of surgical outcome in adolescent idiopathic scoliosis. A 35-year English literature review of 11,000 patients

STUDY DESIGN

Meta-analysis of the English literature on the surgical treatment of adolescent idiopathic scoliosis.

OBJECTIVE

To gather comparable data from a number of different sources and combine the data to create a larger, more statistically significant pool of information for the analysis of surgical outcome.

SUMMARY OF BACKGROUND DATA

Meta-analysis is a technique of scientific literature review used in outcome evaluation of medical treatment. This technique has been applied to the surgical outcome of adolescent idiopathic scoliosis.

METHODS

A structured literature review was performed that cross-referenced English literature articles pertaining to the surgical treatment of adolescent idiopathic scoliosis with a focus on patient-based outcomes. Measures of patient satisfaction were compared with process measures of care.

RESULTS

A number of patients (10,989) were reviewed in 139 patient populations. Unspecified curve types (9424) and King curve types (1565) were reviewed over a 35-year period from 1958 to 1993. Of the patients, 87.32% were studied retrospectively and 12.70% prospectively. Effect-weighted follow-up was 6.8 years. Only studies with complete process and patient data for unspecified or King curve types were included for satisfaction correlation calculations. Pearson product moment correlation for n = 33 studies, n' = 2926 patients revealed a positive r' = 0.628 correlation between degree of curve correction and percent satisfaction per study. To determine the degree of curvature correction resulting in patient satisfaction, a stepwise multiple linear regression analysis was performed with level of confidence (P < or = 0.05). Of significance was that the degree of curvature corrected accounted for all the satisfaction variance predicted. A significant correction exists between degree of curve correction and percent of patients satisfied. The percent of correction and the Group type (either unspecified or King classified), did not significantly alter this prediction. The best predictor of satisfaction appears to be degree of curve correction according to these data.

CONCLUSION

Patients appear to be more satisfied by the magnitude of curve correction rather than the percent of curve correction. The degree of curvature before surgery did not predict patient satisfaction. Pearsons r' = 0.045. Satisfaction appears to be best predicted by the degree of correction only and not by the percent curve correction, the curve magnitude before surgery, nor the Group type (King, unspecified). Patient satisfaction is subjective. It does not reflect the benefits of surgery with respect to the future preservation of pulmonary function in thoracic curves nor the prevention of osteoarthritis in lumbar curves.

Computed tomographic evaluation of the normal adult odontoid. Implications for internal fixation

STUDY DESIGN

Computed tomography scans of the dens were performed on patients who had no atlantoaxial pathology.

OBJECTIVES

To determine whether one or two screws is optimal for fracture fixation and whether two screws can always negotiate the intramedullary odontoid cavity.

SUMMARY OF BACKGROUND DATA

Fixation of Type II dens fractures traditionally has used C1-C2 posterior wiring and fusion. Two screws placed across an odontoid fracture as a method of rigid internal fixation also has been described. However, it is not known whether two screws can always negotiate the odontoid canal.

METHODS

Ninety-two consecutive computerized tomography scans of the dens were performed on adults who had no atlantoaxial pathology. Measurements were taken from the scan and compared with the cross-sectional diameter of two odontoid screws.

RESULTS

The critical diameter for the placement of two 3.5-mm cortical screws with tapping was 9.0 mm. This dimension was present in 95% of the patients studied.

CONCLUSIONS

Correct orientation of the computerized tomography scanner is critical for accurate measurements. Two 3.5-mm screws can be used in internal fixation of Type II dens fractures in 95% of the patients if the inner cortex is tapped.

The role of the lumbar facet joints in spinal stability. Identification of alternative paths of loading

STUDY DESIGN

The role of the facet joints in spinal stability was analyzed using long-segment cadaver spines.

OBJECTIVES

To establish the existence of alternate paths of loading when the facet joints and anulus are compromised.

SUMMARY OF BACKGROUND DATA

It has been reported that facet arthrosis and degeneration never occur without the presence of adjacent disc degeneration. This suggests that intact discs protect the facets from severe loading and degeneration. Based on these studies of spinal mechanics, the authors devised an experiment to further explain the relationship of the disc and the facet in sharing compressive loads.

METHODS

Ten human cadaver spines were placed unsupported in an Instron Model 4206 Tensile Testing Machine. Compressive loads of 1000 N then were applied to the specimens at a rate of 20 N per minute. The facet joints, anterior anulus, and lateral anulus then were destroyed sequentially at L3 and the loading cycles were repeated. The load-deflection curves for each cycle were evaluated and compared with the intact specimen.

RESULTS

Unilateral and bilateral facetectomies had little affect on the ability of the specimen to support a physiologic load. Facetectomies in combination with anterior anulus destruction showed a significant change in the ability of the specimen to support a load with an extension moment applied.

CONCLUSIONS

The facet joints of the lumbar spine are not the principle support structures in extension. With destruction of the facets, an alternate path of loading is established. The alternate path of loading transfers axial loads to the anulus and anterior longitudinal ligament to support the spine. Although facet joint destruction will not produce acute instability, it will transfer the loads to the adjacent disc and conceivably accelerate its degeneration.

Biomechanics of the spine in sports

An understanding of spinal mechanics is necessary for the treatment of athletic injuries. Recognizing and isolating the mechanism of injury through noninvasive techniques will lead to specific treatment for that injury. Because the prevention of physical injury is the goal of all health care professionals, the authors hope that the information will be helpful.

Stress-riser fractures of the hip after sliding screw plate fixation

Fractures occurring after fixation of intertrochanteric femur fractures have been described previously in the literature. Terms such as "stress-riser fracture" and "Young's modulus fracture" have been applied. The prevalence of these fracture types has increased, and so has use of the sliding screw plate device for fixation of intertrochanteric hip fractures. The object of this paper is to describe, by case examples, types of stress-related fractures of the proximal femur in association with the sliding screw plate and to define each biomechanical type in review.

Instantaneous axis of rotation as a function of the three columns of the spine

A knowledge of the rotatory motion of the vertebral bodies is needed to understand the normal biomechanical behavior of the spine. The aims of this investigation were 1) to define the instantaneous axis of rotation of the lumbar spine in rotation; and 2) to study the effect of the loss of the anulus, facet joints, and ligamentous structures on the location of the instantaneous axis of rotation. The instantaneous axis of rotation was found in 10 human cadaver thoracolumbar spines by the method of Reuleaux from superimposed serial photographs. Long-segment specimens were tested to minimize the effect of the imposed axis of the testing device. The instantaneous axis of rotation was consistently posterior to the anulus in the intact spine. With isolated destruction of the columns of the spine, the instantaneous axis of rotation migrated to the remaining intact structures. Anterior releases enhance derotation by removing the primary rotatory stabilizer. Ultimate control of a rotatory deformity or instability lies in the recognition that the anterior structures have a mechanical advantage in resisting torsion.

The effect of the three columns of the spine on the instantaneous axis of rotation in flexion and extension

Instrumentation designed for stabilization and correction of spinal deformities must limit the amount of motion in flexion and extension. In flexion or extension, the vertebral bodies move about a specific point called the instantaneous axis of rotation. The ability of the implant to limit this motion is a function of its relation to the axis of rotation of the spine. The goal of this study was threefold: 1) to define the instantaneous axis of rotation of the spine in flexion and extension; 2) to study the effect of the loss of the three columns of the spine on the location of the instantaneous axis of rotation; and 3) to determine how the above parameters relate to the choice of anterior or posterior instrumentation. Ten human cadaver spines were subjected to compressive loads in flexion and extension. The columns of the spine were then destroyed in sequence at L3. The instantaneous axis of rotation for each vertebral body was found by the method of Reuleaux, and the effect of the compromise of the columns on the location of the instantaneous axis of rotation was noted. Understanding the exact location of the instantaneous axis of rotation after a specific injury would allow the clinician to objectively choose the best surgical approach and the appropriate instrumentation.

The contribution of the three columns of the spine to spinal stability: a biomechanical model

The load carrying capacity (LCC) of the human spine was evaluated in 10 human cadaver spines. The specimens consisted of segments from T11 to S1 with markers placed on the specimens at each vertebral level in both Ap and lateral planes. The specimens were loaded to 1250 N and spinal deflections were recorded and photographed at 125 N intervals during the loading cycle. In 5 specimens, axial and flexion loads were applied to the intact spine. The anterior and middle columns were destroyed in sequence at L2 and the loading process repeated. In the remaining 5 specimens, axial and extension loads were applied with the spine intact and after the posterior and middle columns were destroyed in sequence at L2. Load deflection curves were generated for each test and comparisons were made between intact spines and spines with single and double column destruction.

RESULTS

When the axis of loading was anterior to the posterior longitudinal ligament (PLL), destruction of the anterior and middle columns reduced the LCC by 46% and 68% respectively and destruction of the posterior and middle columns reduced the LCC by 30% and 63% respectively. There was minimal change in the LCC when the axis of loading was posterior to the PLL and the anterior and middle columns were destroyed. Two column destruction of the spine reduced its load carrying capacity for flexion loads by 70%. In thoracolumbar spinal fractures where flexion loads are predominant and anticipated, the authors conclude that surgical stabilisation is indicated with double column failure.

The contribution of the three columns of the spine to rotational stability. A biomechanical model

A function of the spine as a structural column is its ability to resist torsion. The goal of this study was to evaluate the contribution of the columns of the spine to torsional rigidity. Ten human cadaver spines were harvested and frozen. The specimens, consisting of segments from T11 to S1, were subjected to torsional loads of up to 20 N-m. Rotation was recorded throughout the loading cycle. The columns of the spine were destroyed at the L2-3 interspace in a predetermined fashion and loading was repeated. The data were plotted as torsion versus rotation in degrees. The curves of each loading cycle were compared and the ratios of the intact and compromised specimens were calculated. The contribution of each column of the spine to torsional rigidity was determined. In an intact lumbar spine, the anulus was the most effective structure in resisting torsion. Experience gained in this study supports the following conclusion: Flexion-distraction injuries of the thoracolumbar and lumbar spine involving soft tissue destruction of the anterior column and anulus result in rotational instability.

Bilateral rupture of extensor pollicis longus. A case report

Bilateral rupture of the extensor pollicis longus tendon is a rare entity. Most case studies in the literature have been reported in patients with an underlying systemic condition such as rheumatoid arthritis or following an episode of trauma. An interesting and unusual case is presented and theories concerning etiology, various operative techniques, and recent cases in the literature are discussed.

Neuroblastoma presenting as a paravertebral mass and painful scoliosis in a 21-year-old woman

Idiopathic scoliosis is a disease of unknown etiology. It presents in growing children, progresses slowly with time, and is painless. Any variation from the usual presentation warrants investigation. Extra dural pressure, tethering of the cord, tumors, and progressive neurologic lesions must be excluded as causative factors.

A case report of calcinosis universalis

Calcium deposits in soft tissues without previous trauma may represent calcinosis universalis, a condition without systemic manifestations and with laboratory values of blood and urine that are consistently normal. Radiographs reveal areas of calcification with homogenous density, and microscopic examination of these deposits shows foreign body giant cell reaction without capsule formation. In this case of calcinosis universalis, wide excision and lavage of the calcium deposits were temporarily beneficial, but the deposits recurred in this patient.

The release of flexion contractures as a prerequisite for the treatment of pressure sores in multiple sclerosis: a report of ten cases

Ten patients with multiple sclerosis who had severe flexion contractures of the lower extremity ranging from 120 to 180 degrees were treated for large pressure sores. Each patient had at least one pressure sore in the sacral, greater trochanteric, or ischial areas. It was necessary to perform a tendon release procedure of the knee joint prior to addressing the problem of the pressure sores. Prior to treatment all patients were unable to utilize a wheelchair owing to their condition. Subsequent treatment, including development of skin flaps and skin grafts, postoperative casting, and immobilization in a special lamb's wool sling, allowed all 10 patients to improve to the point where they could utilize a wheelchair, with the result that their self-images improved.

Abnormalities of the spine in relation to congenital upper limb deficiencies

A retrospective study was performed to determine the incidence of spinal abnormalities in patients with skeletal defects of the upper extremities. The incidence of scoliosis in the patient population studied was 16%, and spinal abnormalities of all types were present in 18% of the population. Patients with bilateral amelia had a 100% incidence of scoliosis. Those patients with unilateral amelia had a 50% incidence of scoliosis. Patients with ulnar and radial hemimelias had an incidence of scoliosis of 20 and 13%, respectively. The transverse complete and partial hemimelics had an 18% incidence of scoliosis and the phocomelic group had an incidence of 11%. The patient population studied had an increased incidence of scoliosis compared with the general population. The patients with amelia had a particularly high incidence of scoliosis. There was no relationship between the side of the deficiency and the direction of the curve. The age of onset was not found to be related to the progression of the curve. The results of bracing were poor, due to patient rejection of the brace. All patients with upper limb deficiencies warrant close observation throughout growth for the development of scoliosis.