This review found that there is strong evidence for improvement in patient reported functional outcomes in those who undergo decompressive surgery for LSS. There is relative consistency between studies across multiple timeframes with sustained improvements through to 2-years post-surgery. All studies that examined surgical decompression reported statistically significant improvements in favour of surgery, with moderate to large effect sizes in two studies [10
]. Conversely, there was an overall small initial improvement in patient reported functional outcomes in subjects with LSS who participated in an exercise intervention. These improvements subsided over a 2-year period; however there was limited data for exercise interventions at 2-years.
Given the above findings, it is reasonable to question the use of exercise in the management of patients with LSS. Current guidelines support a trial of conservative management prior to surgery [13
]. Malmivaara's study [10
] reported that only four of 44 subjects conservatively managed proceeded to surgery, while remaining subjects appeared to self-manage their condition. Weinstein and colleagues [33
] also found that the majority of subjects managed without surgery showed small improvements in all outcomes. The selection of exercise for LSS may also be given preference over surgery due to underlying surgical risks, including mortality, particularly in the older population who often present with multiple co-morbidities [35
]. Despite the significant and sustained improvements in patient reported functional outcomes shown with decompressive surgery in this systematic review, self-management may still be a worthwhile option prior to consideration of surgical intervention.
A number of issues within included studies may have influenced the results of this systematic review. There were four surgical studies in this review that only included subjects who had failed conservative management [23
]. These studies may represent samples with more disabling cases of LSS. This creates the possibility of the surgical outcomes being biased towards subjects who had failed conservative management.
The results from the included exercise studies were limited by the implementation of poor quality interventions. Exercise interventions were likely of inadequate duration to demonstrate change in outcome scores. There was also an absence of tailored exercise towards subjects' individual impairments in all but two studies [10
]. Subjects' adherence to exercise also varied between studies.
The moderate methodological quality of included studies was also likely to affect the results of this review. Common issues were small sample sizes, lack of sample size calculations, inadequate description of interventions and numerous co-interventions. There appeared to be a discrepancy in mean age between exercise and surgical intervention arms (59.4 and 66.4 years, respectively); however, testing of homogeneity did not demonstrate a significant bias in sampling.
The availability of only two randomised controlled trials limited calculation of effect sizes. More so, there was only one study that directly compared decompressive surgery and exercise for LSS [10
]. Ideally, this review would have limited its study inclusion criteria to randomised controlled trials and therefore represent a collation of the highest level of evidence. Given the sparseness of high level evidence in this topic, the methodology of this systematic review was modified to include lower levels of evidence. This systematic review of current best available evidence is therefore able to provide useful information to inform clinical practice and future research.
Implications for clinical practice
Due to the heterogeneity of the land based exercise interventions and numerous co-interventions reported in the included literature, this systematic review is not able to provide guidance as to whether certain types of exercise (mode, intensity, duration, location) are more effective in managing patients with LSS. It is however apparent from the randomised clinical trials that investigated numerous exercise interventions that there were statistically significant improvements (p
< 0.05) within each group relative to baseline, but no significant differences between groups when different interventions were compared. The authors would suggest that the literature therefore supports a broad approach to exercise interventions rather than supporting a particular exercise type. This review supports the findings of previously reported literature that a trial of conservative management with land based exercise be considered [10
Implications for research
Further research of land based exercise interventions for LSS would benefit from more accurate descriptions of intervention, including type, duration and intensity. Co-interventions should also be minimised. Future research should also continue to embrace the use of reliable and valid outcome measures.
More consistent follow-up through to 2-years would provide valuable insights as to whether the smaller initial gains reported with exercise are maintained. Analysis of subgroups within exercise interventions as to predictors of good and poor response, such as outcome measure score, age or walking distance, would assist with understanding which LSS patients may more readily benefit from land based exercises. Understanding of subgroups may also assist in determining patients that may benefit from other management options, including surgery.
Ideally further research in this patient group would directly compare decompressive surgery and conservative management with land based exercise, with the addition of a control group. However, as demonstrated by Weinstein's review [33
], the challenges of implementing a randomised controlled study design of sufficient power, with long term follow-up is limited by both subject non-adherence to randomisation and the ethical considerations around use of a true control group.