Surgical Treatment of Degenerative Lumbar Kyphoscoliosis

Surgical Treatment of Lumbar Kyphoscoliosis

Surgical Treatment of Degenerative Lumbar Kyphoscoliosis

Kao-Wha Chang, MD, Tsung-Chein Chen, MD, Ku-I Chang, MD

TaiwanSpineCenter and Department of Orthopaedic Surgery

Armed Forces Taichung General Hospital, Taiwan, Republic of China

Address all correspondence and reprint requests to: Kao-Wha Chang, MD

Taiwan Spine Center and Department of Orthopaedic Surgery,

Armed Forces Taichung General Hospital, Taiwan.

No.348,Sec.2, Chung-Shan Rd

TaipingCity, TaichungHsein, Taiwan, Republic of China.

TEL:(8864)23935823; FAX: (8864)23920136

E-Mail:

Abstract

Study Design. Retrospective study.

Objective. To review radiographic and clinical results of patients with degenerative lumbar kyphoscoliosis (DLKS) treated with neurologic decompression, asymmetric closing-opening wedge osteotomy (COWO), instrumentation-assisted correction according to a preoperatively made template, and circumferential fusion with a posterior-only approach.

Summary of Background Data. DLKS causes sagittal and coronal imbalance, posterior sacral migration away from the center of gravity line, muscle weakness, osteoporosis, spinal stenosis and instability, rigid deformities, and implant and adjacent-segment failure after instrumentation-assisted corrective surgery. We know of no reports of the results and complications of current instrumentation and osteotomy techniques to treat DLKS.

Methods. Thirty-one patients with DLKS (mean age, 72.3 years; range, 65–78 years) treated for intractable pain were followed up for a mean of 4.1 years. We assessed their preoperative, 2-month postoperative, and final follow-up radiographs and administered a questionnaire to measure changes in pain, function, self-image, patient satisfaction with surgery, and postoperative complications.

Results. Final radiographs showed increased L1–S1 lordosis from 11.3° to -50.5° (increase of 61.8°), correction of kyphotic deformity from 64.3° to -14.1°, and correction of scoliotic deformity from 48.9° to 8.3°. Sagittal imbalance significantly improved from 68.8 to 27.1 mm, whereas the sacrofemoral distance decreased from 59.3 to -5.1 mm, and the sacral inclination angle increased from 9.7° to 34.3°. Subjective pain was significantly and persistently reduced. Most patients maintained good correction and had good clinical results. No major complication occurred. Eight patients (26%) developed junctional kyphosis.

Conclusions.Thorough neurologic decompression, the best possible correction and restoration of sound associations among the spine, the pelvis, and the center of gravity, are crucial in the surgical treatment of DLKS to obtain satisfactory clinical results. The 3-column release procedure, COWO, and procedures of neurologic decompression and circumferential fusion make DLKS flexible enough to be manipulated adequately from behind. A posterior-only approach minimizes the risk of surgery. Junctional kyphosis remains a significant problem.

Key Words: center of gravity; closing-opening wedge osteotomy; degenerative lumbar kyphoscoliosis;junctional kyphosis;promontory;template.

Key Points

Thorough neurologic decompression, the best possible correction, and restoration of coronal and sagittal spinal balance and of proper lumbopelvic congruity to bring the promontory close to the center of gravity line are crucial in the surgical treatment of DLKS to obtain a satisfactory clinical results.

A lumbosacral curve to bring the promontory close to the center of gravity line can be simulated and a template can be made accordingly.

Asymmetric closing-opening wedge osteotomy,procedures of neurologic decompression and circumferential fusion make DLKS flexible enough to be adequately manipulated from behind to conform the lumbosacral curve to the contour of the template.

Posterior-only approach and staged operation minimize the risk of surgery.

Junctional kyphosis remains a significant problem in DLKS patients treated with instrumentation-assited correction.

Mini Abstract

Thorough neurologic decompression, the best possible correction and restoration of a sound associationsamong the spine, pelvis and the center of gravity are crucial in surgical treatment of degenerative lumbar kyphoscoliosis to obtain satisfactory clinical results.
Introduction

Degenerative lumbar kyphoscoliosis (DLKS), or lumbar degenerative changes, including narrowing of several discs and vertebral wedging or collapse with predominant kyphosis and scoliosis. The curve involves segments from the lower thorax to L5 with an apex at L2 or L3. It may result from degeneration of de novo or preexisting idiopathic kyphoscoliosis,1-3 as well as iatrogenic or traumatic causes.

Lumbar muscles can be fatigued due to overwork against a center of gravity far in front of the lumbosacral junction, an important cause of pain in DLKS.4 Correction of the deformity need to restore coronal and sagittal spinal balance and to reconstruct lumbopelvic congruity to bring the promontory near the center of gravity.

We reviewed the results of DLKS treated with neurologic decompression, asymmetric closing-opening wedge osteotomy (COWO), instrumentation-assisted correction according to a preoperatively made template, and circumferential fusion with a posterior-only approach and determined factors influencing satisfactory outcomes.

Materials and Methods

Patients

We reviewed 37 patients with DLKS undergoing surgery in 2000–2004. Two died, and 4 dropped out; therefore, 8 men and 23 women (mean age at surgery, 72.3 years; range, 65–78 years) were followed up for a mean of 4.1 years (range, 2–6.3 years). DLKS related to causes other than degeneration, such as iatrogenic or revision surgery was not included in this study. Their clinical records were reviewed for demographic data, surgical times, intraoperative blood loss, and complications.

Evaluations

Preoperative, 2-month postoperative, and final follow-up radiographs were analyzed.

Sagittal measurements were made on a 36-in. standing lateral views of the entire spine and upper femur obtained with the hips and knees fully extended. Measurements included curvatures and Cobb angles (negative for lordosis), T1–T12 kyphosis, L1–S1 lordosis (negative values), inclination of the upper surface of the sacrum5 (SIA, positive for anterior inclination), and sacrofemoral distance (SFD, distance between plumb lines through the hip axis6,7 and sacral promontory; positive values for femora anterior to the promontory). SFD was a function of the inclination of the upper sacral surface.4,8 Sagittal offset was the horizontal distance between the C7 sagittal plumb line and the posterior superior corner of S1. Because the posterosuperior aspect of the S1 body was the reference, the normal neutral range for sagittal balance was 0–4㎝from this point (plumb line through the L5–S1 disc).

Coronal measurements included curvatures and Cobb angles. Coronal balance was assessed on posteroanterior radiographs as deviation of the C7 plumb line from the median sacral line. Magnetic resonance imaging was used to confirm spinal stenosis and identify neural compression (retropulsed bone or disc).All patients received the standard method of measuring bone density via dura-energy radiograph absorptiometry (DEXA) scan.Ten patients wereosteopenic ( T-scores between -1.0 and -2.5 ) and twenty-one patients were osteoporotic (T-scores less than -2.5).

At last follow-up, patients completed a modified 46-item questionnaire9 regarding demographics, function (8 questions; maximal = 5, minimal = 1), pain (3 questions; 0 = none, 9 = severe), self-image (scored similar to function), and satisfaction.

Data were compared with the Mann-Whitney U test with significance set at 0.05.

Surgery

Patients were positioned prone with padding at the iliac crests, knees, shoulders, and chest. The abdomen was free to reduce intraoperative bleeding. The osteotomy site (usually L2) was over the hinge in the table so that, as the osteotomy was closed, the table could be moved from the neutral to V position. A standard posterior midline incision was made (usually from proximal level identified as the stable, neural and horizontal vertebra with a stable suprajacent disc in the coronal and sagittal planes to the sacrum). The spine was bilaterally exposed to the tip of the transverse processes with a strictly subperiosteal approach to reduce bleeding. Pedicle screws were inserted (usually from T9 to the sacrum except at the osteotomy level).

Wide posterior decompression and formal lateral-recess decompression and foraminotomy of the involved stenotic levels were usually necessary to treat neurogenic claudication and pain.

The vertebral pedicles for osteotomy were decorticated with a rongeur to facilitate guide-pin insertion. They were entered with a small-diameter curette guided with intraoperative fluoroscopy and radiography. The ideal path was from the lateral side of the medial pedicular wall to the anteromedial wall of the vertebral body midway between the endplates. The lamina and vertebral pedicles were removed. A blunt-end cage trial was hammered to penetrate the anterior cortex of the vertebral body transvertebrally and bilaterally. The path in the middle column was enlarged by pushing the cancellous bone up and down. The posterior cortex of the vertebral body was carefully removed by using a curette or rongeur on both sides. The posterior and middle columns were completely released.

We simulated a lumbosacral curve that brought the promontory close to the center of gravity line and made a template.

The lumbopelvic portion of the standing lateral radiograph was magnified to life size and printed on transparent paper, which was divided into the hips, lumbopelvic portion caudal to the COWO, and thoracolumbar portion cephalic to the COWO. We located the hip axis and rotated and translated the paper with the lumbopelvic portion to a position with the original pelvic-radius length and pelvic radius–S1 angle (constants for each individual)6,7and with an SFD of 0 mm.(The center of gravity is normally directly under the promontory.10)

We translated the paper with the thoracolumbar portion to continue with the lumbopelvic portion and rotated the former with a hinge at the pedicular base of the osteomized vertebrae to simulate the closing and opening wedges for COWO. The posterior superior corner of L1 (point L1) was on the extension of a curve connecting the posterior superior corner of S1(point S), the posterior edge of each lumbopelvic vertebral body, and the pedicular base of the osteomized vertebrae. The curve between S and L1 was the temporary template. One end of a rod of appropriate length was bent to conform to the template. Points S and L1 were marked on the rod, which was locked at 2 points with 2 pedicle screws.

Legaye et al11postulated a predictive equation for lumbar lordosis based on pelvic parameters. According to the manipulated figure, pelvic parameters included pelvic incidence, overhang of S1, sacral slope, and pelvic tilting were measured to predict the lumbar lordosis to be created.11 Approximately two-thirds of L1–S1 lordoses are below L4.12 For L1–L5 lordosis, 40% are at L4–5 in subjects ≥70 years.13 Total L1–S1 lordosis was estimated accordingly.

The contour of the rod segment responding to L4–S1 was used to approximate the lordosis between the pedicle screws to the estimated L1–S1 lordosis.In theory, the promontory can be brought near the center of gravity line if the lumbosacral curve can be reconstructed accordingly.

Deformity Correction

The rod was connected to the pedicle screws on the convex side marked S and L1. The operating table was slowly moved to a V position to facilitate correction and to provide space for lumbopelvic sagittal translation and rotation around the hip axis. The surgeon rotated the convex rod to correct scoliosis, and then pushed the rod at the osteotomy site, and compressed the pedicle screws immediately above and below the osteotomy to correct kyphotic deformity and create the lumbosacral lordosis. Fracturing of the anterior cortex and anterior longitudinal ligament was sometimes heard. We thus created an asymmetrical closing wedge of the posterior and middle columns at the convex and posterolateral sides of the osteomized vertebrae and an opening wedge of the anterior column at the concave and anteromedial side of the osteomized vertebrae.

Procedures of neurologic decompression and circumferential fusion could provide adequate release at L4-S1 segment to conform the segment to the contour of the rod responding to the L4-S1 segment.

The lateral mass was closed down tightly. The correction was fixed with another concave rod and fused with autogenous bone grafts. The roots and dura were checked to ensure that no residual compression by centrally enlarging the canal. A Woodson elevator was passed up and down the canal through the area of central decompression to detect dorsal neural compression created by osteotomy closure. We performed wake-up tests.

Bilateral iliac screws were used to protect the sacral screws and sacrum from failure and fracture. Circumferential fusion with posteriorly placed wedge-shaped cages or strut grafts, bone, and bone substitutes were used for anterior-column support and fusion at L5–S1. For patientswith T-scores less than-2.5, we augmented the most cephalad adjacent vertebrae with 2 intrabody cages (with the thickest sizethat the vertebrae can fit) in the anterior and middle columns to prevent the segment failure.

We stopped the operation when blood loss was >5000 mL and resumed it 1–2weeks after the patient recovered.Patients ambulated 48 hours later and used custom-made thoracolumbar orthoses for 6 months.

Results

Mean estimated blood loss was 3650 mL (range, 2519–9362 mL). Mean operating time was 245 minutes (range, 191–313 minutes).

No perioperative deaths or neurovascular complications occurred. Two patients had postoperative pneumonia, 1 had a superficial infection. All were successfully treated.

One patient had a pulmonary embolus complicated by congestive heart failure, which resolved after medical management. One patient had wound dehiscence 3 weeks after initial surgery. All cultures were negative, and after the wound was surgically repaired, the patient did well. Ten patients developed paralytic ileus, which resolved after a Levin tube was inserted and their oral intake restricted.

No sacral screw pullout occurred.One patient was found to have the most cephaladscrew pullout and required extension to the upper thoracic spine. Two rods in 2 patients broke at L5–S1 due to pseudoarthrosis, and 2 dural tears occurred during surgery; all were managed uneventfully.

Eight of 31 patients (26%) develop progressive junctional kyphosis at the cephalad end of the construct. One developedlatecompression fracture of the most cephalad vertebral body included in the construct. A patient without preventive cage augmentation developed a compression fracture immediately above the instrumented level. Two patients had failure 2 segment cephalad from the last instrumented vertebrae. All leading to kyphosis and required extension to the upper thoracic spine.Four did not have radiographic evidence of any compression fracture but still seemed to become progressive kyphotic and two required extension to the upper thoracic spine. Surgery was staged in 5 patients (Table 1).

Radiographic Results

Six patients had normal preoperative sagittal C7 plumb-line distance but substantially increased SFDs compared with norms.4,6The preoperative estimated and postoperative mean L1-S1 lordoses was -43.3o±9.8 o (range -36 - -61 o) and -52.3 o±12o (range -35 - -63 o) respectively , a significant difference (P = 0.04).No patient had notable loss of correction between 2-month and final follow-up except in thoracic kyphosis and sagittal balance (Table 2). Figures 1 and 2 show representative radiographs.

Questionnaire

Mean pain scores were 7.9 ± 1.1 before and 2.3 ± 1.5 after surgery (P=0.01). Four (13%) reported equal postoperative pain. In 27 (87%), surgery directly decreased their pain.

Median preoperative and postoperative function scores were 9 and 15, respectively (P = 0.03). Function improved in 16 (52%) patients, was unchanged in 13 (42%), and decreased in 2 (6%).

Medianself-image scores improved from 4.5 before to 9.5 after surgery (P = 0.03), improving in 28 (90%) patients, not changing in 3 (10%), and worsening in none.

Twenty-nine (94%) patients were extremely or somewhat satisfied with the results; 2 (6%), somewhat dissatisfied; and none, extremely dissatisfied. Twenty-seven (87%) felt much better or better after surgery, and none felt much worse. Twenty-nine (94%) would choose treatment again.

Discussion

DLKS is increasingly prevalent and predominantly affects the elderly, who may be frail because of qualitative and quantitative weakening of the bones or osteoporosis, which makes instrumentation and fusion difficult. DLKS canexacerbate disc degeneration, spinal stenosis, and facet arthropathy, increasing spinal rigidity and making treatment difficult. Risks of instrumentation failure, nonunion, and adjacent-segment failure are considerable.

Severe deformities can increase surgical morbidity and mortality.14–18Therefore, conservative measures must be attempted first.2,19,20 Goals of surgery are to resolve intractable low back pain or radicular pain that interferes with activities of daily living or to make it controllable with medications, to reduce the drug load, and to fuse the spine in an anatomic position as normal as possible. Cosmesis and progression of deformity are not indications for surgery.

In DLKS, increased extremity radicular pain and claudication are due to neural compression; this is relieved with thorough neurologic decompression. Degenerative pain is caused by disc and facet instability and degeneration due to deformity and imbalance; this is relieved with deformity correction, fixation, and fusion. Myogenic pain is caused by lumbar muscular fatigue and is relieved by restoring sagittal and coronal balance and approximating the promontory to the center of gravity line.4

Spinal stenosis aggravates underlying spinal stenosis attributable to DLKS. Asymmetric collapse, rotation, frequent rotatory and lateral listhesis, and spondylolisthesis or retrolisthesis all compromise the neural elements. Neural impingement can occur centrally and in the lateral recess and foramina. Facet-joint hypertrophy and pedicular kinking caused by disc-height collapse are commonly present. Thorough neurologic decompensationis crucial to treat neurogenic claudication and pain.