To evaluate in a pragmatic study if surgical selection criteria and variation in use of arthrodesis in three Scandinavian countries can be linked to variation in treatment effectiveness.
An observational study based on a combined cohort from the national spine registries of Norway, Sweden, and Denmark.
Patients aged 50 and higher operated 2011-2013 for LSS were included.
Patient-reported outcome measures (PROMs) Oswestry disability index (ODI) (primary outcome), numeric rating scale (NRS) for leg pain and back pain, and health-related quality of life (EQ-5D). Analysis included case-mix adjustment. In addition, we report differences in hospital stay.
Analyses of baseline data were done by analysis of variance (ANOVA), Chi-square, or logistic regression tests. The comparisons of the mean changes of PROMs at one-year follow-up between the countries were done by ANOVA (crude) and analyses of covariance (ANCOVA, case mix adjustment). There are no conflicts of interest. Funding was received from the Danish Society of Spinal Surgery ($5,925), the Northern Norway Regional Health Authority ($5,925) and from Swedish Association of Local Authorities and Regions ($11,885). The sponsor had no role in the acquisition of data, analysis, or preparation of the manuscript.
Out of 14,223 included patients, 10,890 (77%) responded at one-year follow-up. Apart from fewer smokers in Sweden and higher comorbidity rate in Norway, baseline characteristics were similar. The rate of additional fusion surgery (patients without, with spondylolisthesis) was: Norway 11% (4%, 47%), Sweden 21% (9%, 56%) and Denmark 28% (15%, 88%). At one-year follow-up the mean improvement for ODI (95%CI) was: Norway 18 (17 to 18), Sweden 17 (17 to 18), and Denmark 18 (17 to 19). Patients operated with arthrodesis had prolonged hospital stay.
Real life data from three national spine registers showed similar indications for decompression surgery, but significant differences in the use of concomitant arthrodesis in Scandinavia. Additional arthrodesis was not associated with better treatment effectiveness.
Introduction
Low back pain is the leading specific cause for years lived with disability worldwide
[1]. Narrowing of the spinal canal, known as lumbar spinal stenosis (LSS) is the most common indication for spine surgery in the elderly population. LSS typically causes symptoms of low back pain, lower extremity pain and numbness due to nerve root compression, resulting in walking disability
[2]. Decompression of the spinal canal is the key objective of surgery and is considered superior to non-surgical treatment for patients with moderate to severe LSS
[3]. Often, there is a coexisting degenerative spondylolisthesis, i.e. a slip of one vertebra in relation to another. Traditionally, this radiological finding has been regarded as a sign of segmental instability. Although this interpretation has been disputed, adding surgical fusion between the two vertebrae (arthrodesis) in addition to decompression has been recommended to prevent persisting back pain
4 ;
5. However, several recent studies found no effect of additional arthrodesis surgery
6;
7 ;
8. Due to lack of uniform guidelines in this field, there is a large and possibly unwarranted practice variation in the use of additional arthrodesis
9 ;
10. In a recent study fusion rate (with, without spondylolisthesis) was considerably lower in university hospitals of Norway (44%, 6%) compare to Boston, US (95%, 29%)
[11]. In the US, rising costs connected to arthrodesis of the lumbar spine have attracted the attention of health providers and policy makers. In 2011 spinal fusion accounted for the highest aggregate hospital costs of any surgical procedure performed in U.S. hospitals ($12.8 billion)
[12].
The higher cost connected to arthrodesis surgery should be justified by better patient-reported outcome. In 2015, the International consortium for health outcome measurement (ICHOM) recommended a set of patient-reported outcome measures (PROMs) for evaluating surgical treatment of degenerative conditions in the lumbar spine to facilitate clinical studies across nations and centers
[13]. The national spine surgery registries of Norway, Sweden, and Denmark were among the collaborators. Scandinavian countries are characterized by a genetically homogenous population, similar social security systems, and public based health care and health insurance systems, facilitating comparative studies
[14]. The incidence of surgically treated lumbar spinal stenosis is similar (30-35/100 000/year) based on imputed numbers from the registries. Clinical registries collecting data from everyday practice can evaluate different treatment strategies by linking practice-based variation to patient-reported outcomes in a pragmatic trail. Unlike randomized controlled trials, registry-based studies allow for surgeons and patients preferences to be included in the process prior to surgery, as in the “real world” of clinical practice, and adds external validity to already published data from randomized controlled trails
[15]. Such information may aid in guideline development and resource allocation.
The aims of this observational multinational register study were to compare practice-based variation in surgical treatment of LSS by; (1) surgical selection criteria (preoperative patient characteristics), (2) type of surgery (decompression only or decompression plus arthrodesis), and (3) to assess if practice-based variations were associated to different patient-reported outcomes (crude and case mix adjusted), in a large combined registry cohort from three Scandinavian countries.
Methods
This observational study reviews data from the national spine registries of Norway (NORspine), Sweden (Swespine), and Denmark (DaneSpine). Eligible patients were aged 50 or older with no history of previous lumbar spine surgery, operated for LSS during 2011, 2012, or 2013. At baseline, the surgeon recorded diagnosis and treatment according to standardized questionnaire. The diagnosis of LSS was based on the surgeons’ clinical judgment and assessment of magnetic resonance imaging, MRI. Concomitant spondylolisthesis is defined as a visible slip, 3 mm or more, of one vertebra in relation to another. All patients received surgical decompression, some with concomitant arthrodesis.
The registers
All three national spine registries are designed for quality control and research. The participation is voluntary for the surgical departments as well as the patient. At admission for surgery (baseline), the patient reports data on demographics, risk factors, and PROMs. During the hospital stay, the surgeon records diagnosis, type of surgery, and perioperative complications. At one-year follow-up, questionnaires are distributed from the central national registry office, completed at home by the patients, and returned in pre-stamped envelopes. The treating hospitals are not involved in follow up. The oldest registry, Swespine, has included patients since 1998. Swespine covers approximately 95% of the surgical units in Sweden. Completeness, the proportion of operated patients reported to Swespine, was approximately 75%. NORspine is based on the concept of Swespine, and was founded in 2007 (coverage 95%, completeness of 65%). DaneSpine was acquired by the Danish Spine Society from the Swedish Society of Spinal Surgeons in 2009 and has successively been implemented (coverage 80%, completeness 62%).
Patient-reported outcome measures (PROMs)
We used the ICHOM recommended set of PROMs
[13]. The primary outcome was the Oswestry Disability Index (ODI, version 2.1), a standard for measuring back pain related disability, ranging from 0 (no disability) to 100 (bedridden)
[16].
Secondary outcome measures were numeric rating scales (NRS) for back and leg pain, ranging from 0 (no pain) to 10 (worst conceivable pain). Health-related quality of life was measured with the Euro-Qol-5D (EQ-5D) ranging from -0.596 to 1, with higher scores indicating better quality of life.
NORspine used the NRS for leg and back pain, while Swespine and DaneSpine used the Visual Analogue Scale (VAS), ranging from 0-100. Conversion to NRS was done by dividing the VAS score by ten with a stochastic approximation of decimals to the closest integer.
Discussion
To our knowledge, this represents the worlds’ largest observational study of patients operated for LSS, and the first comparison across countries using the ICHOM-recommended core data set. Even though the selection criteria for surgery in terms of demographic characteristics, pain intensity and disability were similar, we found a significant practice variation, i.e. use of additional arthrodesis surgery was almost three times higher in Denmark and two times higher in Sweden as compared to Norway (
Figure 2). This demonstrates that even in homogenous populations with similar health care systems the treatment traditions can vary considerably. We observed longer hospital stay among patients operated with additional arthrodesis, which, together with the implants used, indicates higher cost but no better treatment effectiveness.
Our findings are in accordance with a recent Swedish randomized controlled trial (RCT) by Försth et al. of 247 patients showing that additional arthrodesis neither reduced reoperation rates nor improved clinical outcomes (ODI)
[6]. A randomized controlled trial from the US by Ghogawala et al. involving 66 patients found that additional arthrodesis surgery for LSS with mild spondylolisthesis reduced the risk for reoperation and gave larger improvement of physical health–related quality of life (generic SF 36) than laminectomy alone
[7]. For all other outcomes, including the disease specific ODI, no difference was found. This study has been heavily criticized, also because reoperation rate during follow-up was remarkably high
[21]. Higher frequency of reoperations in the US may however reflect potential cultural differences in patient expectations, difference in treatment traditions and incentives for arthrodesis surgery driven by health insurance and reimbursement programs compare to those found in countries like Sweden.
A Swedish non-randomized registry study of 5390 LSS patients with or without spondylolisthesis operated between 1998 and 2008, found no benefit of additional arthrodesis after two years
[8]. Similar results were shown in a Swiss multicenter study from 2017 of 185 patients with LSS and spondylolisthesis after three years
[22]. A recent Norwegian pragmatic comparative effectiveness study showed marginally better improvement (less than MCIC), of back pain among LSS patients with spondylolisthesis receiving decompression plus arthrodesis. No such association was found for ODI
[23].
We also found a large difference in the use of additional arthrodesis in patients without spondylolisthesis in 2011 – 2013. This treatment strategy has been discussed among spinal surgeons for many years, and is not in accordance with guidelines from 2013, where “decompression alone is suggested for patients with leg predominant symptoms without instability”
2;
4 ;
9. The term “instability” is poorly defined, but has been linked to low back pain, a frequent symptom in LSS. This may explain the practice variation, also shown in a previous study where the arthrodesis rate in cases without spondylolisthesis was 29% in Boston (US), compared to only 6% in Norway
[11]. We observed a rising rate of arthrodesis from Norway, via Sweden, to Denmark across the countries (
Figure 2), but no corresponding trend (dose-response effect) in terms of higher treatment effectiveness (
Table 3). In fact, the mean improvement of back pain in the spondylolisthesis group was somewhat higher in Norway (3.6) than in Denmark (2.7), which had the highest rate of arthrodesis (
Figure 3). Hence, this study does not support the argument that arthrodesis prevents low back pain related to instability in spinal stenosis patients. The different frequency of multiple level surgery was small, and can neither explain the difference in the fusion rate, nor the lack of difference in outcome.
We did both crude analysis and case mix analysis. Crude data shows small, not clinical relevant difference in the outcome between those with spondylolisthesis having decompression and fusion, but these differences vanished after the case mix adjustment (
Table 4).
Fox et al. concluded in 1996 that radiological instability was common after decompression for degenerative LSS without spondylolisthesis, but correlated poorly with clinical outcome (back pain)
[24]. The quality of some earlier studies advocating additional arthrodesis routinely is low due to small sample sizes, weak design, and outcome based on radiological findings
[25]. Moreover, a change towards using more minimally invasive decompression techniques may have reduced the risk for postoperative instability
[26]. Previous studies show that arthrodesis adds higher risk of major complications, and even mortality
[27]. Like Ghogawala et al., we found no association between the use of concomitant arthrodesis and surgeon reported complications
[7].
Comorbidity rate in NORspine was physician-reported and higher compared to the patient-reported rate in Swespine and DaneSpine. However, outcomes were similar, also when adjusting for comorbidity (
Table 3). Between countries with larger diversity in demographic, socio-economic and cultural features, case mix adjustment may be more important.
Even if the differences in effects sizes were smaller than considered as clinically relevant, subgroups of patients may benefit from additional arthrodesis. This should be investigated further in studies utilizing more precise data on radiological findings and with long term follow-up to assess reoperation rates.
Quality assurance
Loss to follow-up may bias the results. Two Scandinavian studies found that a loss to follow-up of as high as 23% would not bias conclusions about overall treatment effects
28 ;
29. They found, similar to our results, that non-responders were younger and more likely smokers. Therefore, it would be reasonable to assume that loss to follow up did not bias our results.
Strength and limitations
Register-based studies in general have advantages such as large sample sizes and high external validity, but also limitations due to lack of randomization, lower follow-up rates, and lower internal validity compared to closely monitored clinical trials. In contrast to RCTs, this study allows surgeons and patients preferences to be included in a shared decision-making process prior to surgery, like in the “real world” of clinical practice. Still, there is increasing evidence in the literature that observational studies, conducted according to STROBE check list, report corresponding results similar to those found in RCTs
[30].
There are limitations associated with this work. Even though registry data were collected prospectively for quality control and research, the hypotheses were decided on in retrospect. In addition, we did not have exact data on reoperation rates and only one-year follow-up. Reoperation rates may be as high as 20% at long term (3 to 5 years)
[6], but previous studies have shown that clinical outcomes are stable up to 5 years
[6].
“In Scandinavia it is recommended to try conservative treatment prior to surgery for lumbar spinal stenosis. Previous studies show that the content of non-operative care is hard to define
[31], and the effects of different conservative treatment alternatives are ambiguous. Since no uniform Scandinavian guidelines for such treatment exist, the type of preoperative conservative treatment was not recorded in the registries, only duration of symptoms.
The use of the ICHOM concept and adding case mix analyses makes comparisons more credible, but a relative small set of baseline variables has been used for case mix adjustment.
Conclusion
Real life data from three national spine registers showed similar indications for decompression surgery, but significant differences in the use of concomitant arthrodesis in Scandinavia. Additional arthrodesis was not associated with better treatment effectiveness.
Acknowledgement
The authors thank all the patients and surgeons contributing with data to the spine registers in Sweden, Denmark and Norway.