Seeking to “go beyond the evidence presented by randomised control trials”, a group of researchers has used the international Spine Tango registry to compare the results of randomised control trials with those of real-world patients treated by either single-level total disc arthroplasty or anterior cervical interbody fusion procedures for degenerative disc disease or disc herniation. Their research, published in The Spine Journal, confirmed borderline significant results in favour of arthroplasty, trending towards superiority. Given that significance was not reached in patients beyond the inclusion criteria of randomised control trials, the results suggest that the clinical superiority of arthroplasty could be negligible.
The authors—who won an “Outstanding Paper” award from the North American Spine Society—conducted research in three stages. They began with a 1:1 matched study, comparing results from nine randomised control trials with that of a cohort of the Spine Tango registry (n=987 of 75,890) who met standard trial exclusion criteria (n=739). The registry data catalogues surgical records from a number of European hospitals, as well as outcomes data in the form of self-reported Core Outcome Measures Index (COMI) survey results, which include graphic scales for recording neck and arm pain. A propensity score-matched analysis of total disc arthroplasty vs. anterior interbody fusion was conducted, with all arthroplasty cases (n=190) matched, and 359 fusion patients remaining unmatched. Neck and arm pain results (mean difference (MD): 0.6 and 0.7 points, respectively) did not differ significantly between the two treatments, whilst significant scores in favour of total disc arthroplasty were found for COMI scores (MD: 0.8). When compared in terms of postoperative changes in each variable, COMI scores (MD: -1) were again significantly greater for total disc arthroplasty. Neck (MD: -0.5) and arm pain (MD:-0.7) changes did not differ significantly.
The second stage of research involved the study of both procedures in patients from the registry who would normally fail the exclusion criteria of randomised control studies. This cohort included, for example, patients over the age of 60 and those with pre-existing spondylosis. Within this group, 27 had undergone total disc arthroplasty, whilst 221 had been treated by anterior cervical interbody fusion (n=248). No significant differences were noted in the outcomes data between these two groups. However, it was observed that those patients undergoing arthroplasty were likely to be younger than fusion patients, and less likely to have undergone C7–T1 surgery.
The third element of this research took the form of a long-term follow-up study. Two-year follow-up data was available for certain patients (n=149) within the Spine Tango registry cohort. Again, the arthroplasty patients were generally younger than the fusion patients. Whilst short-term data had revealed greater improvements and lower pain scores in the total disc arthroplasty group, the long-term data failed to reveal significant differences between the treatment groups.
Overall, the results showed a trend towards superiority for total disc arthroplasty, with significantly better treatment outcomes observed in patients meeting strict inclusion criteria. The authors suggest that this may reflect “tighter indications for surgery”. Whilst other research has found long-term improvements in neck pain associated with arthroplasty, this was not borne out by the Spine Tango data. Long-term results were shown to be relatively stable over time, with both procedures bearing sustained improvements in treatment outcomes up to five years.
The authors note that their research was limited in a number of ways.
Firstly, participation in the registry was voluntary, which could leave it open to selection bias. Furthermore, participating clinics were not required to record data for all patients treated at their centres, nor can it be guaranteed that data recorded is complete. Patient-reported results may also be subject to expectation bias. It is possible that patients may have reported more favourable outcomes for those treatments they perceive to be “novel”.
The researchers noted that the randomised control trial model is the “gold standard” study design for the analysis of treatment efficacy. Given this data, they note that further studies could be required to assess treatments according to terms outside of whether or not they meet or surpass current standards. “Although these trials provide evidence about the relative efficacy of treatments in selected patients, additional studies may be needed to measure the relative effectiveness in broader populations,” write the authors. Documenting patients efficiently in-hospital can provide data which reaches beyond that of randomised control trials, the authors claim, stating, “The need for cost-effective, multi-source, and widely-shareable data collection has never been greater”.