The Introduction of Surgery

The first surgical attempts to treat scoliosis were reported in the mid- to late nineteenth century. Delpech “recorded” surface shape by making plaster casts of his patients,11 and introduced tenotomy in 1818.12 Delpech was the father of French orthopedics. Guerin then became an enthusiast for Delpech’s method and applied it to scoliosis.13 Some thought he carried this to excess. Guerin published his results of 740 patients treated with tenotomy, of whom 358 were completely cured, 287 benefited, 77 did not benefit, and 18 died.13 Malgaigne wrote an editorial on “orthopaedic illusion”14 and Guerin sued him. Malgaigne wrote of Guerin that “It is important to know what to do but no less important to know what not to do.” A fairly vindictive critic, Malgaigne wrote in the Gazette Medical de Paris that “the work of Dr X contains many things both new and good. Unfortunately the good things are not new and the new things are not good.”15

In 1889 Volkman attempted to resect rib deformities, and this is thought to have been the first known scoliosis surgery on bony structures.16 Maas also favored rib resection.17 However, the nineteenth century saw a continued majority opinion that scoliosis was caused by poor posture and therefore could be treated accordingly.18 Lewis Sayre wrote a book on spinal disease and spinal curvature in 1877, describing his methods of suspension and casting—a forerunner of the Boston brace.19 He was also known for the immediate closure of myelomeningocele, and was President of the American Medical Association in 1880 and founder of the Journal of the American Medical Association.

William Adams, apart from realizing that the rotational prominence in scoliosis was made worse by forward bending (the Adams forward-bend test), carefully dissected cadavers with idiopathic scoliosis, and recognized the important lordosis at the curve apex.20 In 1876 he went to the United States with Lister to watch Sayre and was made a Fellow of the American Orthopaedic Association (AOA) in 1898.21

In 1895 Bradford and Brackett developed a horizontal distraction frame that had a “localiser” attachment for curve correction.3 Cast application was then performed.

In 1895 Roentgen discovered X-rays, and they were first used for imaging in surgery in March of 1896.22 Roentgen won the Nobel Prize in 1901. However, it could well be argued that although radiographs of the spine produced beautiful novel pictures, their two-dimensional nature would frustrate further developments in understanding the pathogenesis of idiopathic scoliosis consequent upon Adams’s original dissections and his subsequent statement that “lordosis + rotation = lateral flexion.”20

The Use of Surgical Implants

In the half century after Bradford and Brackett’s work, little progress was made in the nonoperative treatment of spinal deformities, whereas much progress was made in their surgical treatment. Berthold Hadra first applied implants to the spine in the nature of spinous process wiring in 1891.23 Then, in 1902, Fritz Lange implanted metal rods attached to the spinous processes with double slings of silk.24 Both of these early implants were attempts to prevent tuberculous spinal deformity and promote healing.

It would appear that Wreden, in Germany, was the first to apply metal implants to the spine in the treatment of scoliosis.25 He first resected the ribs on each side of the apex, put the patient in an extension bed (the forerunner of halter or halo-extension), and then fixed metal plates to the spinous processes.

For nearly three centuries, osseous defects had been replaced by bone grafts. Perhaps the first surgeon to do so was Meekren in Holland, who in 1682 repaired a defect in the cranium of a soldier with a piece of a dog’s skull.26 The advent of antiseptic surgery allowed William McEwen in 1878 to successfully rebuild a boy’s humeral shaft with bone grafts.27 However, it was not until 1911 that bone grafting was applied to the spine, by Fred Albee28 in the United States and by DeQuervain in Europe.29 Both applied cortical struts to the spine for treating tuberculosis, Albee using a piece of tibial autograft placed between split spinous processes and DeQuervain using the scapular spine instead of the tibia.

Albee, and his colleague Kusher, went on to describe his spinal fusion operation in the treatment of scoliosis.30 He used his tibial strut graft on the curve concavity and anchored the apical vertebrae transversely with bone keys ( Fig. 1.2 ). He also propped up the lower ribs on the concave side of the pelvis with graft material. Albee carpentered grafts with his newly developed power saw, and likened callus to cabinetmaker’s glue.31

Spinal Fusion

Russell Hibbs, in New York, changed the face of fusion surgery, using his fusion procedure between 1914 and 1919 to treat 59 patients, most of whom were polio patients who had undergone preoperative correction through head-pelvic traction.32 He dissected subperiosteally right out to the facet joints and base of the transverse processes, and excised the facet joints. Using a gouge and bone forceps, he then raised flaps that he turned up and down so that adjacent vertebrae would be conjoined with bone graft. He next closed the periosteum over the fusion area. His operative technique is precisely the same as that used today. As Hibbs stated, “the dissection may be made in a practically dry field without injury to the muscles if it is sub-periosteal and if free use is made of gauze packs. Only in an operative wound that is free from hemorrhage can the operator see to exercise the care necessary for thorough work. Not only the baring of the bones may be complete, but the periosteum may be separated from them in a practically unbroken sheet and without disturbance of its relation to the surrounding tissues and blood supply. The greatest care should be exercised in the dissection as by its extent and thoroughness the area of fusion is measured.” This was an extraordinary concept of a real biological approach to surgery for scoliosis, and not surprisingly there was only a 2% mortality rate.

In 1931, Hibbs, along with Joe Risser and Albert Ferguson, went on to report on 360 cases treated surgically over a 13-year period.33 The purpose of the surgery was to prevent progression, and this was achieved in almost 50% of the cases, with about 30% having an increase in deformity because of too short a fusion or inaccurately selected fusion areas. Notwithstanding, the 50% rate of good results was encouraging, and occurred in association with meticulous preoperative and postoperative care. Risser, along with Hibbs, designed a turnbuckle cast, which they began using in 1920 with traction and bending forces applied over the 2 to 4 weeks before fusion. Disadvantages included the obligatory prolonged bed rest along with the possibility of pressure sores.

During this early part of the twentieth century, mixed results were reported for spinal fusion, with Arthur Steindler in 1929 giving up spinal fusion in the face of a 60% rate of pseudarthrosis or failure to obtain or maintain correction.34 However, Howorth, in 1943, reported 600 cases with only a 14% pseudarthrosis rate.24

In the early 1950s Risser developed his localizer cast, which was applied using a special head and pelvic traction frame, and exerted pressure over the back of the rotational prominence.35 In some cases three casts were used before surgery. As was standard practice, a window was cut out in the back of whatever form of preoperative immobilization device had been applied, so that the fusion operation could be performed and the position maintained postoperatively for a minimum of 6 months. Risser also noted that spinal growth appeared to correlate with the development, migration, and fusion of the apophysis of the iliac crest, and this is referred to as the Risser sign.36 Unfortunately, the spine often grows until the late teens or early twenties, when the vertebral endplate epiphyses eventually fuse.37 Not surprisingly, 2 cm of spinal growth (seen as an increase in sitting height) occurs after the apophysis of the iliac crest has fused.38

Meanwhile, in 1941 the AOA conducted a multicenter review of the treatment of scoliosis, and 425 cases were examined, half of which were treated with spinal fusion.39 The rate of pseudarthrosis was 28%, with an even greater rate of complete loss of correction. Among those treated nonoperatively, 60% had an increase in their deformity. The overall end results were a discouraging figure of almost 70% of outcomes rated as fair or poor, with only 30% rated as good or excellent. It was concluded that correction with a cast or turnbuckle followed by fusion produced better results.

The great polio epidemics of the 1940s and 1950s motivated further development of scoliosis treatment. John Cobb was an active proponent of spinal fusion and in 1952 reported on 672 cases treated over a 15-year period, with only a 4% rate of pseudarthrosis.40 He emphasized the need for additional bone graft material to supplement Hibbs type fusions, and used autologous, donor, or cadaveric bone. Cobb insisted on a 6- to 9-month period of bed rest postoperatively.

Cobb also devised, in 1948, the method of measuring the size of a curve on a frontal X-ray film, which is still widely used today.41

The Introduction of Bracing

Walter Blount in Milwaukee developed his brace, which was designed for postoperative support of the collapsing spine in poliomyelitis.42,43 It initially sought to prop up the occiput and chin against the pelvis by distraction, but dental problems44 led to the use of a choker and then to the realization that superstructure wasn’t required, and that scoliosis could be treated nonoperatively with an underarm brace originally devised by John Hall in Boston.45

Notwithstanding the good results with spinal fusion reported by Cobb and Risser, these were not the norm, and indeed, Blount and colleagues, among 87 patients treated with spinal fusion and immobilization in a brace, reported a pseudarthrosis rate of almost 40%.46 However, he did not routinely use facet joint fusions or an adequate period of immobilization.

Subsequently, John Moe, who developed the world-famous Twin Cities Scoliosis Treatment Center in Minneapolis, reported performing 266 spinal fusions in 1958, replicating Hibbs’ method of careful dissection and facet joint fusion augmented with supplemental bone grafting.47 Moe advocated fusion from neutral vertebra above to neutral vertebra below the scoliotic curve in the spine as his fusion levels, and with this the proportion of failed fusions fell to only 14%. However, cast correction and spinal fusion continued to mean 6 to 9 months of bed rest and hospitalization approaching a year, still without insignificant rates of fusion failure, infection, and loss of correction. In response to this, various attempts at using internal fixation methods were reported.

Allan in 1955 reported using an expandable jack-type device placed between the transverse processes,48 and Gruca implanted springs on the convex side of the scoliotic curve fastened to the transverse processes at the end of the curve.49

The Harrington Revolution

The development in 1955 by Paul Harrington in Houston, Texas, of his distraction and compression instrumentation was the most significant milestone in the development of effective scoliosis surgery.50 For the first time there was a reliable means of obtaining and maintaining maximal deformity correction ( Fig. 1.3 ). The driver for this was again the growing polio population, which did not well tolerate cast correction. Harrington conceived of his instrumentation as a means of halting curve progression, and regarded this as “dynamic correction” unaccompanied by spinal fusion. However, the early results were disappointing, with metalwork cut out and failure prompting the routine addition of spinal fusion to this procedure.

In 1966 Moe and Valuska published the results for 173 scoliosis patients treated by Harrington instrumentation and posterior fusion as compared with those for 100 patients treated with Risser localizer casting and fusion.51 The instrumented group had greater correction (61% vs. 54%), was ambulatory sooner (2 ½ months vs. 5 ½ months), and was immobilized for a shorter period (7 months vs. 10 months). The pseudarthrosis rate in the two groups was similar (17% vs. 13%), but the instrumented group had more complications, with metalwork displacement occurring in 15%, and a greater rate of infection. Despite the apparent advantages of Harrington instrumentation, Moe said at the 1966 Paris meeting of the Société Internationale de Chirurgie Orthopédique et de Traumatologie (SICOT) that the general conclusion was that a good result of surgical treatment was an identical degree of curvature at the end of growth as when treatment had begun.

Harrington continued to improve his results, and this was demonstrated in a report of almost 600 cases in 1973.52 He recommended a long fusion, from one vertebra above to two below the upper and lower end-vertebrae of the scoliotic region of spine, respectively. It was important that these levels fell within what Harrington described as the “stable zone” when parallel lines were drawn upward from the lumbosacral facet joints. Interestingly, despite Harrington’s enormous improvement in spinal instrumentation, his early reports were published in the local writings of the Texas Institute of Rehabilitation and Research53 because his articles were often turned down by the leading orthopedic journals, presumably because it was felt that his treatment was still somewhat revolutionary.

In 1973, Paul Harrington, then President of the Scoliosis Research Society (SRS) in Gothenburg, Sweden, recommended a common database or registry of all scoliosis surgeons to document their treatment results.6 His lead was not followed by this group, but was taken up enthusiastically by surgeons who implanted hip and knee replacements.54

Some modifications were made of Harrington’s original instrumentation, and so as to maintain a lumbar lordosis, Moe developed the square-ended rod-and-hook configuration for a better sagittal contour.55 However, it was Harrington’s original design that the distraction rod on the concave side be complemented by a compression system on the convex side that conferred considerable stability to the construct.56,57 Nevertheless, many practitioners ignored the compression system, which was one reason for the greater pseudarthrosis rates among their patients than among Harrington’s.

Clearly, greater curves tended to be instrumented more often than lesser ones in the early years, and there was concern over neurological injury being caused by the rapid stretching of these rigid curves, which in turn led to the development of preoperative distraction devices. In 1959 Nickel and Perry designed the halo,58 and various halo-traction devices were then developed, including those for halo-femoral traction by Moe,55 halo-pelvic traction by DeWald,59 and halo-wheelchair traction by Stagnara.60