Clinically-significant positive outcomes for the surgical treatment of adult spinal deformity have proven elusive for researchers, with many studies limited by the non-randomisation of participants. A new study from the European Spine Study Group, however, has found better clinical results at two years for patients treated operatively than for their non-operative counterparts in a propensity-matched cohort. The research was presented at the Annual Meeting of the Scoliosis Research Society (21–24 September 2016; Prague, Czech Republic) by lead author, Mitsuru Takemoto (Kyoto, Japan).
Takemoto reported that no randomised controlled trials have taken place to address the question of whether surgery for adult spinal deformity is “really better” than non-surgical treatment. Seeking to address the paucity of non-biased data comparing outcomes for one treatment method over the other, the team used a propensity-adjusted analysis to compensate for selection bias in their observational study.
“There are few high quality studies that compare clinical results of surgery versus non-surgical treatment for adult spinal deformity,” Takemoto told Spinal News International. “In some observational studies, baseline patients’ characteristics are quite different between groups. Usually, surgical patients have more severe symptoms at baseline. In such studies, non-surgical treatment patients are at a potential disadvantage for reaching minimal clinically important difference (MCID) after treatment. Therefore, we wanted to know whether surgical treatments are really better, even in a matched cohort.”
The researchers used the prospective, multicentre European Spine Study Group database to select participants according to age (≥18 years), adult spinal deformity diagnosis and follow-up (≥2 years). From an initial cohort of 1,449 patients (834 treated operatively, 615 treated non-operatively) 208 non-surgically treated patients and 139 surgically treated patients had completed follow-up. When questioned as to why the numbers had reduced so dramatically, Takemoto emphasised how young the database was. Most patients who appeared on the surface to be lost at two-year follow-up, had simply not yet reached two postoperative years. “This is one of the limitations of this study,” Takemoto told Spinal News International, “We have to wait for our registry to grow up.” The current results, however, he affirmed were still significant. Two hundred and thirty-two non-operative patients and 145 operative patients were excluded due to incomplete baseline data.
The authors calculated propensity scores using baseline characteristics of the population, according to the likelihood of their being treated surgically or non-surgically. Commenting on the importance of the propensity-matched study design, Takemoto told Spinal News International that this “allows us to match baseline patients’ characteristics of selected variables,” enabling the researchers to compensate for variation in symptom intensity, for example, between groups. Considering variables including age, sex and body mass index, radiographical parameters and health-related quality-of-life (HRQoL) scores, the researchers found that those patients treated operatively tended to be older (mean age 50.8 years versus 43.7 in the non-operative group), with “worse sagittal alignment and disability”. The authors paired 158 patients on a one-to-one basis according to their propensity score similarity for further analysis.
The groups were considered “well-matched”, with similar baseline characteristics between the propensity-matched non-operative and operative cohorts, including average age (45.2 years and 45.6 years, respectively), sagittal vertical axis (15mm and 18.2mm, respectively) and Oswestry Disability Index (31 and 30, respectively).
After propensity matching, the researchers observed statistically significant differences (p<0.001) at two year follow-up between the non-operative and operative groups in Coronal Cobb degree (47.1 to 47.5 degrees versus 47.6 to 26.6 degrees, respectively), Oswestry Disability Index (31 to 31.9 versus 30 to 21.3, respectively), leg pain (numerical rating scale, 3.3 to 3.9 versus 2.9 to 1.9, respectively) and back pain (numerical rating scale, 5.8 to 5.7 versus 5.5 to 3.3, respectively).
The researchers also found that the propensity-matched operative group displayed significantly better Short Form-36 mental (43.4 versus 48.2) and physical (38.6 versus 44) component score improvements, as well as Scoliosis Research Society-22 (SRS-22)self-image (2.9 versus 3.6) and mental (3.2 versus 3.7) component score improvement at two years. When quizzed on whether this amounted to clinical significance beyond statistics, Takemoto told Spinal News International, “Yes, many HRQOL measures reached MCID in surgical patients—SRS-22 function, mental, pain, self-image, and NRS back pain. As for Oswestry Disability Index and Short-Form-36 PCS, they did not reach MCID. We think that this is because patients with worse scores, who apparently receive benefit from surgery and could easily reach MCID, are discarded during propensity score-matching.”
The study was not without its limitations. Takemoto explained, “Propensity score-matching cannot match hidden confounders, such as social condition, education, etc.” He also pointed out that the cost-effectiveness of treatments was not evaluated by this study—an important factor for the decision to operate. Future research, Takemoto said, should focus on more detailed analyses of “What types of patients (age, gender, occupation, life-style, etc.) and activities can receive benefit or be restricted by surgical treatment.” Cost-effectiveness, he asserted, should also be a focus of future study.
According to the authors, “surgery for adult spinal deformity is really better than non-surgical treatment in terms of pain, function, cosmesis, and mental health.” Given the strength of their data, Takemoto told Spinal News International, “We can now recommend surgical treatment to adult spinal deformity patients with confidence.”