Introduction:
It is an ongoing debate whether fusion surgery is superior to non-operative treatment for non-specific low back pain (LBP) in terms of patient outcome [1; 2]. Further, the evidence for how signs of intervertebral disc (IVD) degeneration on magnetic resonance imaging (MRI) correlate with patient outcome is insufficient. Longitudinal studies of LBP patients are thus of interest for increased knowledge. The current aim was to investigate the long-term self-reported patient outcome and longitudinal MRI appearance in LBP patients, in a group of patients in whom discography had been performed as diagnostic guidance at baseline. .

Methods:
30 LBP patients, prospectively enrolled year 2007-2010 in a comparative discography-MRI study (1.5T, sagittal T2/T1-weighted) were asked to participate in this long-term follow-up. The reason for discography at baseline was LBP (>6 months), severe enough to consider surgery. Those accepting to participate in the follow-up performed MRI year 2021 and answered validated questionnaires (VAS/ODI/EQ5D). MRIs (Th12-S1) at baseline and follow-up were evaluated, blinded, according to Pfirrmann classification, Endplate Classification Score (EPS) and HIZ. Each IVD segment was also classified into if surgery (any type of fusion) had been performed between baseline and follow-up. Numbers of IVDs with Pfirrmann≥3, IVDs with HIZ and segments with EPS≥4 per individual were calculated and fused segments allocated in the high score Pfirrmann and EPS groups.

Results:
17 patients (6 male/mean 58.5 years; range 49-72), accepted participation, at this 11-14 year follow-up. Reasons for not participating were; 2 deceased, 5 unreachable, 1 had obtained a non-MR compatible device and 5 declined participation. Between baseline and follow-up, 10 patients (27 IVDs) had undergone fusion surgery, including all lumbar levels in 2 patients and 1-3 IVDs in 8 patients. No significant differences were found regarding VAS/ODI/EQ5D between the surgical patients (mean;51/32/0.54) and the non-surgery patients (mean;50/37/0.40) (0.77<p>0.65). Neither were any significant differences for MRI parameters at baseline nor at follow-up found, except for significantly more HIZ IVDs at follow-up in the non-surgery group (p<0.05) (Fig1). Comparing the groups regarding longitudinal changes of MRI parameters, no differences were found, except for more segments with EPS≥4 in the surgery group (p<0.05) (Table 1).

Discussion:
In this group of chronic patients with LBP, severe enough to be considered for fusion surgery, no differences in self-reported patient outcome could be established at the long-term follow-up between patients treated with or without surgery. Further, over a period of more than 10 years, only minimal differences were found between the groups in terms of in changes of MR parameters. Thus, this study confirms the complexity of LBP and that based on MRI appearance at baseline, neither long-term self-reported patient outcome, nor progression of degeneration can be anticipated.