Indication and Treatment of Adult Kyphoscoliosis

Overview

Degeneration, iatrogenic-/idiopathic causes and fractures can lead to kypho- scoliotic deformities potentially resulting in pain and loss of function. The surgical strategies rely on surgeon preferences and type of deformity as well as clinical symptoms. The complication rate of surgical treatment is high. The aim of this study is to elucidate the indications for surgical treatment of kyphoscoliosis and evaluate the effectiveness of surgical and non-surgical outcome clinically and radiologically. The primary outcome for the PROMs and x-ray is at 12 months, but the investigators will also follow the patients with the PROMs and x-rays after 2,5 and 10 years.

Detailed description 1 Indications and Treatment of Adult Kypho- Scoliosis (INTRAKS Study) 2. Background:New surgical techniques for osteotomies and promising equipment has led to a renaissance with broader indications, evoking the need for well-founded studies and quality registration of the adult kyphoscoliosis treatment to evaluate the risk of major complication, the benefit for the patients and socioeconomic burden. Prior randomized studies have had high crossover rates to surgical treatment, questioning the result of the studies. As a randomized controlled trial (RCT) will be difficult, the investigators plan to propensity match two cohorts and then compare the correctional surgery and non- surgical treatment at a 1 year follow-up. 2.1 Needs description: 1. Identify predictors for unsatisfactory outcome in adult kyphoscoliosis. This study aims to create a predictive model for unsatisfactory outcome after deformity surgery in adult kyphoscoliosis. A prospective multicenter observational effectiveness study is constructed to detect a \< 30% improvement from baseline in Oswestry Disability Index (ODI) score at 12 months follow up. The predictive model can be a helpful tool in accessing the complexity of this patients and aid in shard decision making and improved surgical planning. Patients are grouped in: 1) unsatisfactory result (UR) \< 30% increase from baseline ODI score, 2) satisfactory result (SR) ≥ 30% increase from baseline ODI score. A predictive model will be calculated for both kyphosis and scoliosis together and separately. The same criteria for inclusion are set in the non-surgical group as well as all the data collected, a predictive model will be created if the difference in outcome is significant. The investigators will also look at the patients with a \> 10% worsening from baseline in ODI score at 12 months follow up. 2. A prospective multicenter observational effectiveness study of adult kyphoscoliosis treatment, surgical and non-surgical outcome with a 12 months follow up. 1. This part of the study will be looking at the effectiveness of surgery vs non-surgical treatment using the PROMS described in the protocol. It is outcome related, the planned follow up time is 12 months, but the investigators have ethical approval (REK) to continue the data recording for 24 months and more for a sufficient follow up time. It exceeds the time limit for the PhD candidates listed, but others will continue the work. It will contain all the patients included. 2. The subgroups of iatrogenic/degenerative kyphosis and scoliosis will be studied separately as it is known that the mainly scoliotic patients do better than the kyphotic once. Around 30-40 patients will be sufficient in each surgical subgroup, the same in the non-surgical. 3. The kyphotic healed high energy fracture. One of the subgroups to be investigated is the high energy fractures healed in kyphosis. High energy fractures are typical males in the age of 25-50. The fractures healed in kyphosis with a regional kyphotic angle of \>20 degrees and a positive sagittal vertical axis of \>4 cm can be included (the rest of the inclusion or exclusion criteria listed will be followed as for the whole study). As this is a subgroup, around 30-40 surgical patients will be sufficient. The investigators hope to have the same number in the non-surgical group to compare with. The main outcome parameter is ODI, but the investigators will also look into the results of the EQ-5D, SRS22 and the numeric rating scale (NRS). As this is a fracture resulting in a secondary deformity, the PROMS listed should be addressing the patient problems. 4. XLIF vs anterior-/transforaminal interbody fusion/pedicle subtraction osteotomy/Smith Petersen/Ponte (ALIF/TLIF/PSO/SP/Ponte). There is a trend towards using XLIF and OLIF procedures for deformity cases. The investigating party will look into the patients operated using the lateral approaches and compare it towards the dorsal applied interventions (TLIF, PSO, Ponte/SP), if possible also ALIF patients. It should be at least 30-40 patients undergoing the lateral intervention to be able to make a comparison that could be statistical significant. 5. Socioeconomic study; surgery vs non-surgical for adult kyphoscoliosis. The correctional surgery is often of long duration with expensive implants and highly complicated with high risk of complications. It is costly and the investigators intend to evaluate the socioeconomic burden. The duration of surgery, cost of implants and duration of hospital stay will be registered, the readmission due to complications will be registered for the first 12 months. As EQ-5D is used as a PROM in this study, it will be used for the socioeconomic study. 6. Identify differences in indication, examination and treatment between a European and Asian population. This is an international multicenter study in cooperation with the Kyoto University Hospital in Japan. The investigators have already detected some vital differences in the surgical approach for the patient group and further comparison will be of interest. The patients, indications, examination as well as the non-surgical and surgical treatment will be examined. 7. Changes in urination and sexual function in a surgical and non- surgical cohort. Few studies address this issue, for spine patients it is of great importance and highly underreported. There are few suitable PROMS, so the investigators will use a PROM from a prior study. It is slightly different for men and women due to the differences in sex and it has no scoring system. It has not been validated, so the outcome will be slightly unsatisfactory, the positive sides convinced us to take use of the non-validated form. The investigators have had it translated into Japanese and Swedish and are interested in the Japanese contribution in regard of their cultural differences. 8. Quality control of the Norwegian Quality Register for Spinal Surgery. The investigators will be conducting x-ray measurements (sagittal vertical axis and pelvic parameters, lumbar lordosis) and compare it with the values that the surgeons have entered into the Norwegian quality register for spinal surgery. The measurements will be conducted by an independent radiologist at Haukeland or Oslo University Hospital in Norway and by the PhD student in Norway. It will be an inter-and intra-observer study. The values mentioned above have been used for the last year in the register and a quality control would be of interest. 9. Spinal stiffness after spinal fusion using the Lumbar spine disability index (LSDI). The Lumbar spine disability index (LSDI) is developed by our Japanese colleges at the Kyoto University Hospital and they are motivated to test it on the deformity patient. It is the second (2/2) non validated PROMS in this study and have been translated from Japanese to Norwegian and Swedish. It has been published in peer review papers by our Japanese colleges. Most likely it will be of most interest to our Japanese colleges and they will go forward with the publishing of the result. 4\. Project methodology 4.1 Project design, method selection and analyses Patients to be operated for the correction of spinal deformity during the following two-tree years at Oslo University Hospital Ullevål in Norway, Haukeland University Hospital in Bergen Norway, Malmö University Hospital in Sweden, Örebro University Hospital in Sweden and Kyoto University Hospital in Japan. A randomized controlled trial (RCT) will be difficult, so the investigators plan to propensity match the two cohorts and compare correctional surgery and non- surgical treatment at 1 year follow-up. During the 1 year follow-up period the non-surgical patients will attend an intensive and structured physio-logical training program. The patient inclusion will follow at the polyclinic by the surgeon; the patient will be presented to the "Intraks.org" homepage for the study. An information leaflet will be handed out and the patient will be given time to decide. If the patient is eligible to the study and agrees to join, a study number will be created using google Sheet, then the patient will be introduced to the informed consent form, paper form in Sweden and Japan, in Norway it can be signed using "Nettskjema" an "Bank ID". Nettskjema is an internet form used to create questionnaire, it is secure, but not secure enough for sensitive patient data, so the data from the forms will be imported directly to "service for sensitive data" (TSD). TSD is a service from Oslo University, secure for all sensitive data and in use for several prospective studies. In the homepage, the patients will, in addition to the informed consent, find the patient reported outcome measures (PROMS). The home page can be found using smartphone, tablet or computer and all forms can be filled out using the homepage. 12 months after surgery or inclusion to non-surgical treatment, the patient need to fill out the forms again. For the non-surgical treatment the homepage (Intraks.org) contains specific text, illustrations and videos with training instruction for the non-surgical treatment, developed with dedicated physiotherapists from Oslo University Hospital. The training-program consists of 3 levels of muscle core training and stretching to be exercised 1-3 times a week for 12 months. They will have a training diary to follow up the compliance that will be controlled at the 3 months follow up. The exercises are available at intraks.org as illustration, text and video. See the protocol for detailed data collection. 4.2 Project management See protocol. 4.3 Plan for implementation See protocol, in addition: The "intraks.org" homepage will be updating the public on published articles and other activities. 4.4 Statistical analysis Altogether 75-100 patients are expected eligible for inclusion in the surgical group of WP2 annually. For the non-surgical treatment, approximately the same number will be needed. By using alfa=5% and power=80% to detect 30% ODI improvement from baseline and 30% improvement in proportion of patients that improved and calculating for 10% crossover, 10% mismatch in the propensity score matching and sufficient power for the subgroup analyzes, the study will need a total of 300 patients. Considering drop outs, the study would have enough patients after 2-3 years of inclusion. Predictive analysis: Frequency analyses for categorical variables will be conducted via Pearson's c2 analysis. All analysis will be conducted using commercially available software (SPSS version 24, IBM Inc.) and the level of significance is set to p\<0.05 in a two-sided test. For the predictive model, missing values within the database will be imputed using standard techniques such as mean and median imputation. Once a complete data set is constructed, an ensemble of decision trees will be constructed with a binary target variable that includes patients with: 1) unsatisfactory result (UR) \< 30% increase from baseline ODI score (code = 0), 2) satisfactory result (SR) ≥ 30% increase from baseline ODI score (code = 1). The decision-tree algorithm is C5.0 and 5 different bootstrapped models will be built. Internal validation is accomplished via a 70/30 data split for training and testing the model, respectively. Final overall predictions from the models will be combined and chosen by voting with random selection for tied votes. Overall accuracy and the area under the receiver operating characteristic (AUROC) curve will be calculated as well as predictor importance as determined by the model. The model will be built using commercially available software (SPSS Modeler version 24, IBM Inc.). A sensitivity analysis will be performed 1-2 years into the study period, the MCID/"satisfactory result" and power will be adjusted accordingly