Adolescent idiopathic scoliosis – to operate or not? A debate article

The Scoliosis Research Society (SRS) definition of AIS is a structural lateral curvature of the spine that creates a thoracic and/or lumbar asymmetry on forward bending (a.k.a. the "Adams forward bend test") combined with a curve of at least 10° as measured by the Cobb technique on a standing radiograph of the spine with associated vertebral rotation. The diagnosis of the curvature must occur after the age of 10 years, and there is no other established etiology for the curve [5]. Given this definition, it has been repeatedly demonstrated that AIS curves can and will progress. Curve progression is defined as an increase in curve magnitude of 5–10° on serial radiographs. The primary factors for AIS curve progression include growth potential and curve magnitude.

The question of untreated AIS causing pain, disability, and negative impact on quality of life in adulthood has demonstrated conflicting results. Early studies reported that untreated scoliosis resulted in increased back pain and disability, negative socioeconomic effects on work and marriage, and early mortality compared to controls [3, 4, 19, 20]. As previously indicated, however, these studies had mixed etiology of scoliosis, including infantile and juvenile idiopathic scoliosis. Recent work by Weinstein and coworkers re-evaluated the amount of disability that adults with untreated AIS incur. Fifty year follow up of 117 patients with untreated AIS (average curve magnitude > 75° for thoracic, thoracolumbar, and double major curve patterns) demonstrated no increase in mortality rates, disabling back pain, and ability to complete daily activities compared to age and gender matched controls [18]. However, acute and chronic back pain, and dissatisfaction with appearance were more prevalent in patients with scoliosis, as was shortness of breath in patients with thoracic curves > 80°. These findings demonstrate that the prognosis for untreated AIS is not as poor as was originally reported, however their findings further demonstrate that patients with untreated AIS are unquestionably dissimilar to matched controls. The authors acknowledge that the clinical significance of their report is to differentiate the natural history of AIS from the more negative prognosis for untreated infantile and juvenile idiopathic scoliosis patients, rather than demonstrate equality between patients with and without AIS. A growing body of literature does indicate that adults with scoliosis demonstrate greater function limitation, greater daily analgesic use, and less satisfaction with their appearance compared to unaffected individuals [19, 21‐23]. In an attempt to quantify the amount of disability associated with scoliosis, Berven et al used established "health-related quality of life" (HRQL) outcomes measures, including the SRS research instrument (SRS-22), to evaluate pain and function in adult patients with and without scoliosis (figure 3). The SRS research instrument was originally developed to evaluate surgical outcomes for AIS in children [24‐26]. This instrument has subsequently been validated as the standard measure for AIS surgical outcomes and disability associated due to AIS, with validated translations in Chinese, Japanese, Spanish, and Turkish. The SRS questionnaire has also demonstrated differences in function, pain, and self-image between children with and without scoliosis, as children with AIS demonstrate lower scores than controls in all domains across cultures [26‐35]. In an attempt to extrapolate these data to adults, Berven and colleagues demonstrated that the SRS-22 is a reliable instrument for measuring disability associated with adult spinal deformity, reporting SRS-22 criteria to be valid in use with the SF-36 [21]. The authors reported that adults with scoliosis scored significantly worse on every SRS-22 and SF-36 domain (including pain, function, self-assessment, and mental health), compared to controls with no scoliosis. These findings have subsequently been supported by Bridwell et al., validating the SRS-22 as an accurate measure of function and disability in adults with scoliosis, and that the SRS-22 demonstrates greater accuracy and consistency for measuring outcomes for the treatment of adult scoliosis compared to previous HRQL measures including SF-12, SF-36, and Oswestry Disability Index (ODI) [29].

The primary surgical goals for treating AIS are to prevent curve progression, obtain solid arthrodesis, and restore spinal balance. A plethora of literature has repeatedly demonstrated that, if established surgical guidelines are followed, surgical arthrodesis prevents AIS curve progression, restores spinal balance, and, when permissible, preserves spinal motion segments via selective fusion of structural curves and indirect deformity correction of compensatory curves [36‐38]. Using modern surgical techniques (i.e. segmental spinal fixation, meticulous posterior element decortication, and bone grafting) solid arthrodesis is achieved in over 95% of AIS cases [39, 40]. Additional goals addressing comesis and self-image concerns are also addressed, as current surgical techniques and have demonstrated improved curve correction and cosmetic outcome compared to early surgical techniques [40‐43].

Reported perioperative complication rates following surgery for AIS range between 5% and 20%, depending upon the surgical approach, and differentiation between major and minor complications [40, 42, 44, 45]. The most recent SRS morbidity and mortality committee update reported a 5.7% surgical complication rate in 6334 patients treated for AIS from 2001–2003 [45]. Recorded complications included pulmonary, infection, neurological deficit, dural tear, deep venous thrombosis, and death, but were not divided into major and minor complications. The incidence of complications following posterior spinal fusion was 5.1% whereas the incidence following combined anteroposterior spinal fusion was 10.2%. Wound complications were the most common complications for PSF (1.4%), and pulmonary problems were the most common for anterior procedures (1.6%). Neurologic complication rates for anterior, posterior, and combined procedures were 0.26%, 0.32%, and 1.75%, respectively. Diab et al, reported 0.69% rate of neurological complications in 1,301 surgically treated cases of AIS [46]. These findings were corroborated by a 1.06% rate of neurological complications following surgery for AIS at a Chinese hospital over a 7 year period [47]. Carreon and colleagues prospectively evaluated the prevalence of non-neurologic complications following surgical treatment of AIS, reporting a 15.4% prevalence in 702 patients treated from 2002–2004 [44]. The reported prevalence of major complications including major blood vessel injury, visceral injury and deep infection was 0%, 0.28% and 0.71%, respectively. The majority of compilations (3.7% prevalence) were labeled "miscellaneous", including transverse process failures, ileus, lateral femoral cutaneous nerve dysesthesia, diarrhea, and allergic reactions to medication. There was no significant association between the prevalence of a non-neurologic complications and curve pattern, curve magnitude, surgical approach, number of spinal levels fused, and type of bone graft used.

Autograft harvest site pain is often reported as a common complication and source of morbidity following spinal arthrodesis, however a separate bone graft harvest might not be necessary when treating AIS. Betz et al. reported equivalent arthrodesis rates for allograft vs. no graft in a prospective, randomized study that evaluated grafting techniques and fusion rates following posterior spinal fusion for AIS [39]. The authors reported a 1.3% pseudarthrosis rate, minimum 2 year follow up, among 76 AIS patients. Patients randomized to no graft demonstrated no pseudarthrosis compared to one patient randomized with allograft, who developed a non-union.

Finally, in order to advocate for surgical treatment of AIS, it must be demonstrated that the consequences of surgical treatment in adolescence are not detrimental in adulthood. Long term studies from Canada and Sweden have indicated that adult patients treated with spinal fusion as adolescents report reduced function and greater frequency of back pain compared to controls [48, 49]. However, the reported severity of reported back pain was mild and daily back pain was uncommon. Additionally, it is difficult to ascertain if these limitations were due to surgical treatment of scoliosis or due to scoliosis itself, as it has been reported that patients treated surgically for AIS indicate greater function and less pain than patients treated non-operatively, however function and pain levels continue to remain inferior to controls without scoliosis [22]. As indicated above, adolescents treated surgically for AIS demonstrate significant improvement in HRQL measures compared to preoperative scores, including improvement in all SRS questionnaire preoperative domains at 2 year follow-up, and report high treatment satisfaction levels [26]. Of note, Upasani et al reported that AIS patients demonstrated significantly increased back pain 5 at year follow-up compared to 2 postoperative values [50]. The etiology of this increase in pain was not determined, nor was the clinical significance of the increase in pain, namely if it reached a so-called "minimally clinical important difference". Even so, pain scores at 5 year follow up were still lower than preoperative values, and patients continued to reported high treatment satisfaction and high self-image, independent of the pain scores.

Surgical treatment of spinal deformity remains and emotional decision for both the patients and their parents. In our modern society, the affected child's parents have to assume the responsibility for making a decision that could influence their child's life forever. It is therefore necessary to define the roles for each treatment option and to define the expected long-term outcomes for each treatment modality. Many adults are now presenting for scoliosis surgery who still remember their childhood years spent in casts and turnbuckle braces. The advancement of general medicines are keeping these patients alive and they are in chronic pain and disabled. They have been able to live a relatively productive life and are now faced with the late effects of untreated scoliosis, quite bitter that they were denied the surgical option by their parents when they were young. It is hard to balance this argument when faced with a young teenager as it is impossible for us to see what future development might bring in term of revolutionizing the disease. It is agreed from both view points that the disease warrants treatment. We therefore suggest that surgical and non surgical treatment option each has their rightful place and that the decision for treatment should be based on the curve type and the risk factories for progression, keeping in mind the end result as well as the expected long-term outcomes (figures 2, 3).

Spinal fusion surgery, which is recommended when the magnitude of curvature exceeds 40–45 degrees, has been used as a treatment for nearly a century [51‐53]. The aims and goals of surgery have varied widely. The early belief was that spinal fusion could be used to leave the patient with a mild residual deformity but this is not the case as one third of patients lost all postoperative correction within one to ten years post surgery [54]. Expectations have been revised to the more modest goals of preventing progression, restoring 'acceptability,' and reducing curvature. In spinal fusion, the vertebrae are accessed by posterior, anterior, or thoracoscopic incision. The main principle of these surgical techniques is to use the spine as a structural scaffold, cementing the parts onto this via a bone paste, giving it an overall straighter shape. [53, 55‐57]. These surgical methods are based on the assumption that this will heal well and remain sturdy for patient's lifespan. Metal rods, screws, wires etc. have been used to reinforce the stability of the spinal fusion [56‐61] and the choice of instrumentation is based upon the preferences of the surgeon [62‐66]. Failure of spinal fusion requires re-operation to restore curvature correction [62].

What specific evidence is there to support scoliosis surgery? The signs and symptoms of scoliosis obviously cannot be changed by scoliosis surgery and long-term beneficial effects have not been reported yet [67, 68] and, in addition to this, there are no studies presenting the long-term risks [67]. Moreover, no report of the long-term surgical outcome (balance, rate of fusion, rib hump correction, cosmetic correction, pain, and patient satisfaction) is available for any study series. Further prospective studies including these parameters will be required to determine the real benefit of such procedures for the patient [67].

There are in fact prospective controlled studies comparing the outcome of patients with AIS treated conservatively with a series of patients treated surgically [48, 69‐73]. Nevertheless no study is available comparing surgery to the natural history prospectively [74‐76]. The Gothenburg's papers do not offer any evidence that the long-term outcome of surgery is superior to the long-term outcome of patients treated conservatively [48, 69‐73]. The studies relating to HRQL/SRS-22 questionnaires do not demonstrate differences between the two groups of patients [70], pain and function do not differ [48, 73], nor does degeneration [71], sexual function [72] or restrictive ventilation disorder [69]. As early as 1973, Paul Harrington envisioned in the future a common database or registry of all Scoliosis Research Society (SRS) members' patient's treatment results [51]. Unfortunately the SRS failed to follow this vision until recently. Instead of achieving long-term evidence for surgical treatment on a higher level and addressing the problems after surgery to attempt to improve patient safety, the surgical community is presenting large numbers of papers describing HRQL after surgery and related research [26, 77‐81].

The problem with such studies is that they lack validity as they do not investigate the actual signs of scoliosis or the post-surgery symptoms of the patient [82]. Those studies containing psychological questionnaires may be compromised by the dissonance effect [74‐76, 82, 83]. This implies that all situations which include important decisions to be made. Cognitive dissonance occurs most often in the situations where an individual must choose between two incompatible beliefs or actions and there is a tendency for individuals to seek consistency among their cognitions. Unable to face an inconsistency, such as being dissatisfied with a surgical procedure, an individual will often change an attitude or action. Surgery is impossible to reverse, but subjective beliefs and public attitude can be altered more easily. The clinical significance of this is that a patient dissatisfied with surgical treatment may not necessarily admit this publicly, as the findings of the following studies:

"Slim objective favorable outcomes correlate with high post-surgical patient satisfaction, while a considerable share of patients with whom a highly favorable outcome has been attained express relatively low post-surgical patient satisfaction. This paradoxical trend may be well understood when applying Cognitive Dissonance Theory. The whole pattern of results point again at highly complex and powerful psychological processes, some of them seemingly irrational" [82].

"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" [84].

"Radiographic and physical measures of deformity do not correlate well with patients' and parents' perceptions of appearance. Patients and parents do not strongly agree on the cosmetic outcome of AIS surgery" [85].

From searching all of the studies based on questionnaires within this review, no evidence can be derived that supports the assumption that patients have experienced benefits from undergoing surgery, as none were able to rule out the cognitive effect of dissonance. Without being able to rule out such effects on the post-operative experience these outcomes do not appear to be valid for the group of scoliosis patients treated surgically [74‐76].

In principal, all kinds of complications may occur in all scoliosis aetiologies [67]. However, in the otherwise healthy subjects with AIS the incidence of major complications may not be as high as in neuromuscular disorders [67, 86]. Risks of spinal fusion include those occurring in any major surgery such as severe blood loss; urinary infections due to catheterization; pancreatitis; and obstructive bowel dysfunction due to immobilisation during and after surgery [67, 86]. The frequency of specific complications, including death is unknown. This is due to problems in reporting such as; mandatory reporting, definitions, interpretation of complications and variations within compliance [67]. Information is based on voluntary reporting by clinicians. Other risks of scoliosis surgery, as listed in recent reviews [67, 86] are summarised below:

From the patient's perspective, the preferred plan of action would likely be based upon avoiding unnecessary risk i.e. avoiding surgery, or to keep it as the final option, once all conservative measures have failed. Under this premise, every effort should be undertaken to improve non-operative treatments for AIS, the most common form of scoliosis, which is regarded to be relatively benign [18]. In view of the fact that there is no evidence that health related signs and symptoms of a scoliosis can be changed by spinal fusion in the long-term [67, 74‐76], a clear medical indication cannot be derived from most scoliosis conditions [74‐76]. In the light of an actual publication on adolescent idiopathic scoliosis with a prospective design [44], showing the short-term risks of scoliosis surgery to be more than 3 times higher than previously expected from retrospective reviews, the matter of surgical indications at present should be investigated more closely in order to improve the safety for patients, however, as has been shown [86], still today there are more open questions, than answers regarding the outcome of spinal fusion surgery for AIS patients.

In consideration of the questions generated by research [67, 74‐76, 86], the lack of measurable medical benefit [67, 90] and the high amount of short and long-term risks of the surgical procedures applied, the decision to have surgery does not rely on any valid evidence to support it. The informed patient perhaps should make the final decision after being provided with all the objective facts available. There is evidence on physical therapy to be found [91] justifying a grade A recommendation while conservative management [75] to be a grade B recommendation and spinal fusion surgery is considered as only a grade C recommendation [67, 74‐76, 90]. Some spine surgeons still do not acknowledge the role of conservative management although evidence exists [92], reasons for this are perhaps based in an underlying conflict [93]. Today's "best practice" bracing technology enables to provide significant clinical and radiological improvements (figures 5, 6) [94, 95], which clinically are comparable to what is achievable by applying spinal fusion surgery (figure 5, compared to figure 4). The current development of scoliosis braces seems to have a less negative impact on quality of life than braces of previous standards [96] using less physically restrictive material but meeting the same in-brace correction effects (Figure 6) [95]. Physical therapy due to present evidence has to be regarded to be effective in preventing curvature progression in smaller curves [75, 91]. Therefore, spinal fusion surgery seems not to be indicated medically, while in bigger curvatures the patients themselves should understand the complications and decide whether their deformity is having too much of a negative psychological impact to resort to such invasive treatment. Spinal fusion surgery for AIS patients should be performed only after describing all the adverse effects the operation might have [67, 86] to enable the patient to weigh the risks against the benefits in the long term [67]. Conservative management involves very different skills from surgery, which may sway surgical opinions towards a negative perspective on this subject [92]. However, anyone claiming to be a specialist in the treatment of scoliosis should advise patients according to evidence based practice and should offer a high quality conservative management to the patients.

AIS is a relatively uncommon condition, occurring in approximately 2–3% of children. The number of children that require treatment for AIS, is even smaller, approximately 0.1–0.3%, and those that receive surgical treatment is smaller still²⁷. However, progressive AIS is not a benign condition and should not be observed. Children with progressive AIS report lower self-image and worse HRQL scores than controls with no scoliosis. These low scores translate into further disability and dissatisfaction into adulthood. Although recent studies have indicated that the prognosis for untreated AIS in adults is not as dire as was originally thought, adult patients with untreated scoliosis demonstrate greater disability and lower HRQL scores compared to controls and those treated surgically. When progressive AIS has been identified, surgical intervention reliably stops curve progression and restores spinal balance, with minimal perioperative complications. Postoperatively, children surgically treated for AIS consistently demonstrate improved HRQL scores, including improved function, improved self-image and decreased pain from preoperative values. These findings support surgical treatment for progressive AIS, as non-operative measures will only allow the condition to worsen.

Health-related signs and symptoms of idiopathic scoliosis cannot be changed by surgery. Long-term beneficial effects of spinal fusion are not yet revealed and the long-term risks of surgery for scoliosis simply are not known. HRQL scores cannot be regarded as being valid considering the dissonance effect. Therefore the indication for spinal fusion surgery is for cosmetic reasons, only. When recognizing, that todays standard of conservative treatment can in fact lead to the same cosmetic effects as spinal fusion surgery, the latter treatment is rarely needed in patients with AIS.