The assessment of the flexibility of the spine in a patient who has scoliosis is important because it provides information regarding the rigidity of the curve, the extent of structural change, the levels to be included in the arthrodesis, the amount of correction that can be achieved safely, and whether a secondary curve should be treated with arthrodesis.
Traditional methods of assessment have included comparison of the Cobb angle measured on radiographs made with the patient standing with that measured on radiographs made with the patient supine4, that on traction radiographs4, that on so-called prone push radiographs7, or that on lateral-bending radiographs made with the patient standing or supine1,8,9. When a lateral-bending radiograph is made with the patient supine, the patient actively flexes the trunk laterally while lying on an x-ray table. Evaluation of the lateral-bending radiograph is one of the most common methods currently used for the preoperative assessment of spinal flexibility.
However, current segmental spinal instrumentation systems3, which are more rigid than the Harrington distraction system, have been found to achieve more correction than would be expected from evaluation of traditional lateral-bending radiographs made with the patient supine9 (Figs. 1-A, 1-B,and 1-C). Moreover, it has been suggested that this increased ability to correct the curve may be a cause of coronal decompensation (the phenomenon whereby spinal balance, as measured with the plumb-line method or on the basis of truncal shifting9-11, is made worse by the operative correction of the scoliosis). This phenomenon has been seen in patients with a type-II curve5 (a double curve in which the thoracic curve is of greater magnitude and is stiffer on lateral bending compared with the lumbar curve) who had selective arthrodesis of the thoracic spine9-11. Coronal decompensation may occur after selective arthrodesis of a type-II curve because, due to the intrinsic stiffness of the lumbar curve, it cannot compensate for the correction obtained by the instrumentation of the thoracic spine.
The indications for an anterior release vary among institutions, and there is no agreed-on standard. We perform an anterior release of the thoracic spine in a patient who has adolescent idiopathic scoliosis when the residual curve is more than 40 degrees on a lateral-bending radiograph. However, in our experience, we have found a discrepancy between the correction that was seen on the preoperative lateral-bending radiograph, made with the patient supine, and the degree of correction achieved after the operation. This finding raised the question of whether it is necessary to perform an anterior release in these patients (Figs. 1-A, 1-B and 1-C). Thus, if the flexibility seen on the traditional lateral-bending radiograph does not accurately reflect the actual flexibility of the scoliosis, a patient may have an anterior release unnecessarily.
We describe a new method for the assessment of spinal flexibility that we believe more accurately reflects the true flexibility of the spine. We compare the results obtained with use of this radiograph, which we call the fulcrum bending radiograph, with those obtained with use of the traditional lateral-bending radiograph made with the patient supine.
*No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. No funds were received in support of this study.
†Department of Orthopedic Surgery, The Duchess of Kent Children's Hospital, The University of Hong Kong, 12 Sandy Bay Road, Hong Kong. E-mail address for Dr. Cheung: cheungmc@hku.hk.
A fulcrum bending radiograph is made with the patient lying on his or her side, over a large, radiolucent plastic cylinder that is padded for comfort (the fulcrum). The fulcrum is available with a diameter of nineteen, twenty-three, or twenty-seven centimeters (Fig. 2). The smallest one that will just lift the patient off the x-ray table should be chosen. The fulcrum is placed directly under the apex of a lumbar curve, and it is placed under the rib corresponding to the apex of a thoracic curve (Fig. 3). Accurate placement of the fulcrum can be achieved by identifying the apex of the curve and its corresponding rib on the anteroposterior radiograph and then marking the apex on the patient after counting cephalad from the most caudad rib. A true lateral position of the patient on the table is determined with reference to the pelvis and the shoulder, which should be perpendicular to the x-ray beam. Once this position has been obtained, sandbags can be used to keep the patient in the position. For thoracic curves, it is important for the fulcrum to be of a sufficient size to allow the shoulder to be lifted off the table so that a maximum passive bending force can be applied to the curve (Fig. 4). Similarly, for lumbar curves, the pelvis should be lifted off the table.
We compared the predictive value of the fulcrum bending radiograph with that of the lateral-bending radiograph, made with the patient supine, by prospectively studying thirty patients who were to be managed operatively for adolescent idiopathic scoliosis. Verbal consent for the additional two radiographs was obtained from the patient and his or her parents. The mean age at the time of the operation was fourteen years (range, eleven to twenty years). The patients were divided into groups according to the five curve patterns described by King5. No patient had a type-I curve (an s-shaped curve in which the lumbar curve is larger and less flexible than the thoracic curve), nineteen patients had a type-II curve (an s-shaped curve in which the thoracic curve is larger and less flexible than the lumbar curve), eight patients had a type-III curve (a single thoracic curve in which no compensatory lumbar curve crosses the midline), one patient had a type-IV curve (a long thoracic curve in which the fourth lumbar vertebra tilts into the thoracic curve), and two patients had a type-V curve (a double thoracic curve). All of the patients were managed with posterior spinal arthrodesis, performed by the senior one of us (K. D. K. L.) with use of Texas Scottish Rite Hospital instrumentation (Sofamor Danek, Memphis, Tennessee).
Preoperatively, all of the patients had lateral-bending radiographs made while they were supine, fulcrum bending radiographs, and anteroposterior and lateral radiographs made while they were standing. These radiographs were compared with the first postoperative radiograph, which was made with the patient standing one week after the operation. Because most of the patients had a type-II or a type-III curve, arthrodeses were done on only thoracic curves; therefore, only the flexibility of the thoracic curves was studied. All measurements of angles were made with use of the Cobb method. Statistical analyses were performed with use of the paired t test.
Attempts at improving the assessment of spinal flexibility with use of the lateral-bending radiograph have included use of the traction and lateral-pressure radiograph4 and the so-called prone push radiograph7. However, these methods have disadvantages: a physician is needed to apply pressure, the force exerted is difficult to standardize, and the physician is exposed to radiation. Additionally, these methods have been evaluated in relation to correction obtained with the Harrington distraction system6 and not with the newer, more rigid segmental spinal instrumentation systems. The latter instrumentation systems can correct the curve better than the Harrington system, probably by a combination of segmental control and vertebral translation.
As the traditional lateral-bending radiograph made with the patient supine requires active cooperation and effort by the patient, these studies are unreliable for patients who are mentally retarded or have neurological or muscular disorders. The fulcrum bending radiograph is easy to make and the bending force is passive and reproducible, although it may be necessary for a parent to keep the child still. The fulcrum bending radiograph more closely reflects the stiffness of a scoliotic curve; thus, it is a better guide for the prediction of the postoperative correction of the curve as well as for the determination of the need for an anterior release in borderline cases.
King et al.6 recommended selective arthrodesis of the thoracic curve with Harrington instrumentation for a patient who has type-II scoliosis. Subsequently, a number of authors have found that decompensation can occur after selective arthrodesis of the thoracic curve with Cotrel-Dubousset instrumentation9-11. They suggested that the newer instrumentation systems correct the thoracic curve to a degree that cannot be compensated for in the lumbar curve. A reason for this is that spinal flexibility is usually assessed on the basis of a lateral-bending radiograph made with the patient supine, which does not accurately predict the degree of correction achievable by segmental spinal instrumentation systems. Although this may be only one of the factors involved in decompensation after operative management of scoliosis, accurate prediction of the degree of correction possible would nonetheless be very helpful.
In a pilot study of five patients, we found that it is very important both to center the fulcrum under the rib corresponding to the apex of the thoracic curve and to check that the shoulder is lifted off the x-ray table. This ensures that a maximum reproducible passive bending force is exerted. We discovered that the placement of the fulcrum had not been ideal for our first patient, and the angle on the fulcrum bending radiograph was larger than that measured on the lateral-bending radiograph made with the patient supine (Table I). Nevertheless, we included the results for this patient for completeness.
The fulcrum bending radiograph is most useful for patients in whom the curve is relatively stiff—that is, when the angle on the lateral-bending radiograph is greater than 40 degrees. We have found that the angle of the curve in such patients often is smaller on the fulcrum bending radiograph; thus, with the criterion used at our institution, anterior release is not performed in these patients. In the present series, seven patients (Cases 3, 8, 10, 19, 20, 23, and 27) had an angle of greater than 40 degrees on the lateral-bending radiograph, but only one (Case 27) had an angle of greater than 40 degrees on the fulcrum bending radiograph. Thus, an anterior release was avoided in six of the seven patients when the decision was made on the basis of the fulcrum bending radiograph. Case 27 was considered a borderline case, and an anterior release was not performed because the spine was well balanced and the patient did not have a severe rib hump. The over-all correction was only 45 per cent in this group of seven patients, as compared with 60 per cent in the other twenty-three patients. However, the average angle on the postoperative radiograph was only 37 degrees. Although an anterior release would have resulted in better correction in these patients, we did not believe that the additional morbidity was justified, as we consider an angle of 40 degrees or less to be associated with a good clinical and cosmetic result.
Even in the patients who had a more flexible curve, the difference between the angle measured on the lateral-bending radiograph and that on the fulcrum bending radiograph was still significant (p < 0.01). The fulcrum bending radiograph was always more predictive of the final correction obtained with use of our preferred segmental spinal instrumentation system (Texas Scottish Rite Hospital instrumentation). The amount of corrective force that was applied with the system in our patients was based on the experience of the senior one of us, and it seemed to correspond well with the degree of correction that was obtained. Such instrumentation systems are able to take advantage of all of the spinal flexibility detected and seldom correct the curve to a greater degree than the flexibility allows. We believe that this flexibility is best predicted by the fulcrum bending radiograph.
The present study concerned only thoracic curves in patients who had adolescent idiopathic scoliosis. However, we have used this method, with equal effectiveness, to assess the flexibility of lumbar curves and curves with other etiologies, including those associated with congenital and neuromuscular disorders. We do not recommend its use for patients who have soft bones (such as those who have osteogenesis imperfecta), as there is a risk of a fracture of the rib. Also it should not be used when spinal instability is suspected.
We have successfully used the fulcrum bending radiograph in the same manner as we used the lateral-bending radiograph for determining the degree of flexibility and correctability preoperatively. In our practice, the fulcrum bending radiograph has replaced the lateral-bending radiograph in routine preoperative assessment.
NOTE: The authors thank the Department of Diagnostic Radiology, The Duchess of Kent Children's Hospital, for their kind help in carrying out this study.