Total disc arthroplasty versus anterior cervical interbody fusion: use of the Spine Tango registry to supplement the evidence from randomized control trials

Abstract

Background Context

Several randomized controlled trials (RCTs) have compared patient outcomes of anterior (cervical) interbody fusion (AIF) with those of total disc arthroplasty (TDA). Because RCTs have known limitations with regard to their external validity, the comparative effectiveness of the two therapies in daily practice remains unknown.

Purpose

This study aimed to compare patient-reported outcomes after TDA versus AIF based on data from an international spine registry.

Study Design and Setting

A retrospective analysis of registry data was carried out.

Patient Sample

Inclusion criteria were degenerative disc or disc herniation of the cervical spine treated by single-level TDA or AIF, no previous surgery, and a Core Outcome Measures Index (COMI) completed at baseline and at least 3 months' follow-up. Overall, 987 patients were identified.

Outcome Measures

Neck and arm pain relief and COMI score improvement were the outcome measures.

Methods

Three separate analyses were performed to compare TDA and AIF surgical outcomes: (1) mimicking an RCT setting, with admission criteria typical of those in published RCTs, a 1:1 matched analysis was carried out in 739 patients; (2) an analysis was performed on 248 patients outside the classic RCT spectrum, that is, with one or more typical RCT exclusion criteria; (3) a subgroup analysis of all patients with additional follow-up longer than 2 years (n=149).

Results

Matching resulted in 190 pairs with an average follow-up of 17 months that had no residual significant differences for any patient characteristics. Small but statistically significant differences in outcome were observed in favor of TDA, which are potentially clinically relevant. Subgroup analyses of atypical patients and of patients with longer-term follow-up showed no significant differences in outcome between the treatments.

Conclusions

The results of this observational study were in accordance with those of the published RCTs, suggesting substantial pain reduction both after AIF and TDA, with slightly greater benefit after arthroplasty. The analysis of atypical patients suggested that, in patients outside the spectrum of clinical trials, both surgical interventions appeared to work to a similar extent to that shown for the cohort in the matched study. Also, in the longer-term perspective, both therapies resulted in similar benefits to the patients.

Introduction

The neurologic symptoms and neck pain associated with disc degeneration of the cervical spine can be treated surgically using either the established standard procedure of anterior interbody fusion (AIF) or the more recently introduced total disc arthroplasty (TDA). The evidence regarding the comparative effectiveness of these surgical procedures has been summarized recently in a Cochrane Review [1]. The review included nine randomized controlled trials (RCTs), five of which were considered to have a low risk of bias [2], [3], [4], [5], [6]. It identified evidence for a small but statistically significant difference between the treatments in the alleviation of arm pain, neck-related function, and neurologic outcome, all in favor of arthroplasty [1]. Although statistically significant, the effect sizes were not clinically relevant. Based on these results, the reviewers concluded that at this point in time both treatments could be considered valid therapeutic options. They also cautioned that, as the available RCTs only included follow-up periods of up to 2 years, the hypothesis that TDA results in a reduced incidence of secondary symptoms at adjacent levels would require evaluation in future updates of the review once the results of long-term studies were available.

Randomized control trials have an important limitation in that they often define narrow admission criteria which may not be applicable to daily clinical practice. For instance, some studies only recruited patients younger than 60 years [7], [8], limited the treated segments to C3–C7 [2], [3], [7], [8], [9], [10], [11], [12], or excluded marked spondylosis [2], [10], [11], trauma [5], [7], [8] or spondylolisthesis, which are sometimes considered contraindications for TDA. Such restrictions are common for RCTs in an effort to examine treatment effects under ideal circumstances. Although these trials provide evidence about the relative efficacy of treatments in selected patients, additional studies may be needed that measure the relative effectiveness of the treatments in broader populations [13], [14]. In doing so, they serve to assess the external validity of the results of RCTs.

The purpose of this study is to demonstrate how data from a surgical registry can be used to supplement the evidence from RCTs with regard to the outcomes after different types of surgery for degenerative cervical spine disease. The Spine Tango registry collects real-world data of patients undergoing surgery of the spine, without the restrictions inherent in RCTs, and some participating clinics follow their patients for a long period of time. Using such prospective cohorts, well designed and carefully executed observational studies can help to answer clinical questions that go beyond the evidence presented by RCTs [15], [16]. Spine Tango collects detailed clinical data of patients undergoing TDA and AIF, among other treatments. With the careful application of matching algorithms, a quasi-experimental comparison of the two surgical groups can be achieved to confirm (or refute) RCT results [17]. Registry-based observational studies can also serve to describe the surgical outcome of a broader spectrum of patients outside the narrow admission criteria of RCTs. Finally, the registry gives us the opportunity to perform longer-term effectiveness studies of the surgical treatments.

The aims of this paper were threefold. In a first study we applied the admission criteria used in the available RCTs and conducted a propensity score-adjusted outcome analysis to evaluate whether this design produces short-term results similar to those of the RCTs, for the comparison of TDA and AIF. We then analyzed the comparative clinical outcome (for TDA vs. AIF) of patients typically excluded from the RCTs published to date, but present in the registry. Third, we evaluated the mid to long-term effects of TDA versus AIF in all patients with follow-up longer than 2 years.