Degenerative Spondylolisthesis; Micro-decompression vs Decompression + Instrumented Fusion; Long Term Follow up

It is moderate evidence for that surgical treatment is superior non-surgical treatment for Degenerative Spondylolisthesis (DS). Question regarding the effectiveness of different surgical treatment alternatives is a controversy. The main goal for an operation is to decompress the stenotic neural elements. Based on two influential studies in the early1990's, an additional fusion procedure became more and more common and during the next decades the rate of fusion procedures as well as the complexity of the fusion procedures increased dramatically. In 2011 about 90 % of the surgeries performed in the United States included an instrumented fusion procedure. This practice has been supported by several guidelines and reviews. However, the evidence for this practice has been questioned and as a consequence of the conflicting results from two randomised controlled trial, the role of fusion again has become controversial.

According to the annual reports from Norwegian Spine Registry (NORspine), the majority of the decompression alone surgeries for DS is performed with preserving of the midline structures. These procedures are considered to be biomechanical less extensive and may result in less instability and less progression of the slip compared to a standard laminectomy.

In the now published Norwegian study (NORSTEN-DS; Clinicaltrial.gov Identifier: NCT02051374) DS the participants were randomised into micro-decompression alone and decompression plus instrumented fusion. In the observational non-inferiority cohort study the investigators evaluated the effectiveness of micro-decompression alone and decompression plus instrumented fusion among patients registered in the NORspine registry. The hypothesis was that the clinical result for the micro- decompression group is as good as the results for the instrumented fusion group, at short time (12 months after operation) and at long time (7 to 15 years after operation) for DS .

Statistics For the complete cohort of patients fulfilling the inclusion and exclusion criteria, the baseline demographics and outcome measures will be presented.

To as best as possible adjust for covariate that can bias the allocation to treatment, Propensity score matching (PSM) will be used to make the distribution of observed baseline patient characteristics in the micro-decompression group and the fusion group as similar as possible. The following parameters will be included in the calaculation of the propensity score: Age; Gender; American Society of Anesthesiologists (ASA) grade; Body Mass Index (BMI); Smoking, Oswestry Disability Index(ODI), Numeric Rating Scale for leg pain and for back pain, Euroqol 5-D (EQ5D), foraminal stenosis, degenerative disc disease, predominating back pain, number of levels operated on and neurological palsy. The PSM scores will be derived from a logistic regression model and reflect a patient's theoretical baseline probability for being instrumentally fused. Using the technique of '1:1 matching without replacement', pairs of fused and non-fused patients with a difference in propensity scores less than 0.2 in logit of the standard deviation will be formed.

In order to include data from all three time points and hence make use of all available data, Latent Growth Curve (LGC) models with Full Information Maximum Likelihood will be used for estimating the change scores and the follow-up scores for ODI ,NRS leg pain and NRS back pain. Thus, information from patients will not be excluded if data are missing at one of the follow-ups. Statistical Package for the Social Sciences (SPSS) version 24 will be used for descriptive statistics, tests for distribution of data, cross-tabulations with χ2 test, Student t-tests and Mann-Whitney U tests of group differences and linear and logistic regression modelling Mplus verion 7.3 will be used for analysing LGC models.

To control for biased estimate due to data missingness, a sensitivity analysis will be performed. With use of baseline and follow-up data as predictors for multiple imputation (MI), 10 data sets will be generated with complete follow-up scores for ODI, NRS leg pain and NRS back pain.

The sample size calculation for comparing the effectiveness of the treatments is based on the same assumption as the randomized multicentre trial on the same patient group (ClinicalTrials.gov Identifier: NCT02051374). In a non-inferiority design, with one sided testing, the investigators want to detect whether the responder rate for the micro-decompression group is less than 15 percentages lower than for the fusion group. This has to be tested by forming a 95% confidence interval (CI) for the difference of proportions (i.e. percentage of responders in the fusion group minus percentage of responders in the micro-decompression group) and has to be rejected if the upper limit of the CI is less than 15. If rejected, the conclusion will be that decompression alone is non-inferior, i.e. as good as, decompression with fusion. A 15% difference in responder rate corresponds to a number needed to treat of 7 patients (NNT = 100/15 = 6.67 [12]. Choosing a type 1 error = 0.05 and power = 0.90 gives a sample size of 160. Considering these assumptions, and adding 25% for possible dropouts, a total of 213 patients are required in each group.

To investigate the long term results, we will perform a similar study with data at long term follow-up (5 to 15 years postoperatively) including the same patients as in the primary study (one year follow-up). The patients will receive a similar questionnaire as used in the primary study. The design, primary outcome, secondary outcomes, hypothesis, hypothesis testing, non-inferiority margin and statistical methods will follow the same rules as for the primary study. When analyzing the clinical outcomes (patient reported outcome measurements) the statistician will be blinded for treatment assignment.