We found functional outcome to be better in older patients who were operated with instrumented fusion as compared to non-instrumented fusion. As it was not a randomized study, several possibilities for bias exist. Patients in the instrumented group were younger and had a slightly longer follow-up. However, after controlling for age and sex by adjusting with norm scores, the SF-36 results were still in favor of the instrumented group; nor did analyzing outcome scores according to length of follow-up change the fact that the results were better in the instrumented group. Furthermore, inclusion of drop-out patients in the comparisons did not change the results with respect to differences seen between the two groups. Thus, the likelihood of bias introduced by differences in drop-out rate appears to be small. Age was the only demographic variable that differed between the two groups, and it could not explain the differences in SF-36 scores observed. Despite this, it might still represent selection of patients for the instrumented procedure and selection bias favoring instrumentation cannot be ruled out entirely. In general, however, the results obtained in both patient groups are similar to what has been published previously (Glassman et al. 2007
, Rampersaud et al. 2008
). Thus, the differences cannot be explained by the non-instrumented group being exceptionally poor.
The randomized studies comparing fusion with and without additional pedicle screw fixation have mainly been performed in patient samples with an average age well below that of this series. One exception is the study by Fischgrund et al. (1997)
in which the average ages in the instrumented and non-instrumented groups were 69 and 66 years, respectively. In the original study there was a statistically significant difference in fusion rates, but not in patient-assessed outcome. In a later long-term follow-up on this study, Kornblum et al. (2004)
compared patients with solid arthrodesis to those with pseudoarthrosis and could demonstrate better outcome in those who were solid-fused. They interpreted this in favor of instrumentation, although the study involved a smaller patient material than the original and the data were not analyzed according to the original assigned treatment groups. In a cohort study comparing laminectomy alone to laminectomy with non-instrumented or instrumented fusion, Katz et al. (1997)
could not demonstrate any beneficial effect of instrumentation (relative to non-instrumented fusion) in patients older than 50 years. We observed the same tendency of better outcome in those patients who achieved a solid fusion as did Kornblum et al. We did, however, only use plain radiographs for fusion assessment; thus, it is likely that the fusion rate is overestimated—as it has been shown to be reduced by the use of more detailed diagnostic modalities (Brodsky et al. 1991
). However, the uncertainty in determining fusion rate does not affect our main observation, which was the difference in functional outcome between the two groups. Other documentation for any relation between outcome and achievment of solid fusion has been somewhat controversial. In a meta-analysis, Mardjetko et al. (1994)
could not demonstrate any relationship between fusion rates and patient satisfaction. In a historical study on pedicle screw fixation, Yuan et al. (1994)
found higher fusion rates and better outcomes in patients fused with pedicle screw instrumentation than in patients with uninstrumented fusions.
What argued against the use of instrumentation was the higher number of additional spine surgeries in this group, as additional spine surgery after the primary procedure was associated with poorer outcome. Several studies have investigated the rate of complications associated with spinal surgery in this age group (Deyo et al. 1992
, Carreon et al. 2003
, Ragab et al. 2003
, Cassinelli et al. 2007
), but few have related the presence of complications or additional surgeries to functional outcome. In the Maine lumbar spine study, additional spine surgery over an 8–10-year follow-up period was associated with smaller improvement and less satisfaction as compared to those who had only undergone the primary intervention, which, however, rarely involved fusion but only decompression (Atlas et al. 2005
). Tokuhashi et al. (2008)
reported a high degree of independence 10 years after instrumented fusion in patients over 70 years. They did not, however, report on the influence of complications on outcome. In a study similar to ours, Glassman et al. (2007)
reported inferior results in patients older than 65 years who required revision surgery after a primary lumbar fusion, as compared to patients only operated once. Despite the higher number of additional spinal surgeries in the instrumented group, outcome was still better than in the non-instrumented group. Thus, the poorer results associated with additional surgery could not outweigh the better outcome achieved in the instrumented group in general.
One study investigating the long-term results of decompressive surgery has shown a deterioration in improvement with time (Jonsson et al. 1997
). In fusion surgery, the stability of the improvement in outcome achieved has varied between studies (Ekman et al. 2005
, Andersen et al. 2008
). In the current study, the improvement in both groups was stable and the long-term effect of the fusion procedure appears to be preserved also in this patient category.
In summary, we have found that superior outcomes can be achieved in selected patients over 60 years of age who have been treated with instrumented spinal fusion using allograft, as compared to non-instrumented fusion. The study suggests that the achievement of a solid fusion was one of the explanatory factors for this finding. However, pedicle screw instrumentation was associated with a larger number of additional surgeries, which resulted in inferior outcomes. Thus, the selection of procedure for the older patient requiring spinal fusion remains a balancing act, but instrumentation should not be discarded just because of the age of the patient. Future research should concentrate on determining the most efficient fusion procedure in elderly patients.