There were 2947 patients enrolled in the 17 studies.Overall, 1436(48%) patients with LSS received decompression surgery alone, compared to 1511(52%) patients who underwent decompression plus fusion surgery. The number of females and males was available in 15 studies containing 1837 females and 861males.Two studies do not give accurate gender data. While the likely dominant age was approximately 65years, females were higher in numbers and average onset age than males. Characteristics of the 17 studies are shown in Table 1.
Primary outcomes
Back pain
Eleven studies reported changes in back pain between the two subgroups,including 6 based on VAS, 3 based on JOA and 2 based on NRS.Because of the high reliability of randomized controlled trials, we divided the included studies into two subgroups. One is RCT, and the other is observational research.With regard to the overall results of heterogeneity testing, there was a statistically significant difference between the two groups(P = 0.0002, I2 = 71%),and a random effects model was applied for meta-analysis(Fig. 4).Through the statistical analysis of preoperative and postoperative back pain, it was found that there was significant statistical significance between decompression and fusion groups in the subgroup analysis of RCT.The improvement of back pain in fusion group is better than that in decompression group.But in the subgroup analysis of comparative observational studies,there was no statistical significance between the two surgical methods .[RCT subgroup, SMD=-0.42, 95% CI (–0.60, -0.23), Z = 4.31 P < 0.0001; comparative observational studies subgroup, SMD=-0.04, 95% CI (–0.23, 0.15), Z = 0.44, P = 0.66].
Leg pain
In the study we included, 5 articles used VAS to score leg pain, 2 articles used JOA to score leg pain, and 2 articles used NRS to score leg pain.A heterogeneity test indicated no statistically significant difference between the two groups (P = 0.12, I2 = 37%).Although the combined effect of the study shows that the study is homogeneous, because the pain scoring system is used differently and the research type is different, in order to increase the credibility of the data analysis, we use the random effect modelfor meta-analysis(Fig. 5).In the RCT subgroup, SMD=-0.27,95% CI (–0.,55, 0.01), Z = 1.88, and P = 0.06;in the comparative observational studies subgroup, SMD = 0.04, 95% CI (–0.05, 0.13), z = 0.95, and P = 0.34. These results demonstrated that the differences in pre- and post-operative leg pain were not significantly different between the two groups.
ODI
Eight studies reported ODI in the two groups. The heterogeneity test showed a statistically significant difference between the two groups (P = 0.001, I2 = 71%), and a random effects model was applied for meta-analysis(Fig. 6).There was no significant difference in ODI between the decompression group and the decompression plus fusion group(RCT subgroup, MD = 2.90, 95% CI (–3.24, 9.03), z = 0.93 P = 0.35; comparative observational studies subgroup, MD = 1.58, 95% CI (–1.84, 5.00), Z = 0.90, P = 0.37).
EQ5D
Four studies reported EQ-5D in the two groups. There was no statistically significant heterogeneity between the two groups (P = 0.68, I2 = 0%), and a fixed effects model was applied for meta-analysis(Fig. 7). No statistically significant difference was identified between the two groups.(RCT subgroup, MD = 0.06, 95% CI (–0.04, 0.16), z = 1.17 P = 0.24,comparative observational studies subgroup, MD = 0.01, 95% CI (–0.01, 0.02), Z = 0.71, P = 0.48).
Walking ability
Two comparative observational studies reported walking ability in the two groups. The heterogeneity test showed no statistically significant difference between the two groups (P = 0.29, I2 = 11%), and a fixed effects model was applied for meta-analysis (Fig. 8). There was no significant difference in walking ability between the decompression group and the decompression plus fusion group (MD = 0.01, 95% CI (–0.54,0.56), Z = 0.03, P = 0.98).
Secondary outcomes
Intraoperative blood loss
Six studies reported the Intraoperative blood loss in the two groups. There was statistically significant heterogeneity between the two groups (P < 0.00001, I2 = 99%), and a random effects model was applied for meta-analysis (Fig. 9). The Intraoperative blood loss was demonstrated that the blood loss in the decompression alone group was significantly less than in the decompression plus fusion group(RCT subgroup, MD=-361.97, 95% CI (–637.16, -86.77), Z = 2.58 P = 0.010, comparative observational studies subgroup, MD=-322.99, 95% CI (–396.40, -249.58), Z = 8.62, P<0.00001).
Length of hospital stay
Five studies reported the length of hospital stay in the two groups.Statistically significant heterogeneity was found between the two groups (P < 0.003, I2 = 76%). A random effects model was applied for meta-analysis (Fig. 10).The length of hospital stay was statistically different between the decompression alone group and decompression plus fusion group. (RCT group MD=-2.06, 95% CI (–3.03, -1.08), Z = 4.13 P<0.0001.Comparative observational studies group(MD=-3.06, 95% CI (–4.25, -1.88), Z = 5.06 P<0.00001).The Length of hospital stay in fusion group is much longer than decompression group.
Duration of operation
Six studies reported the duration of operation in the two groups. There was statistically significant heterogeneity between the two subgroups (P < 0.00001, I2 = 96%). A random effects model was applied for meta-analysis (Fig. 11), which indicated that the decompression plus fusion group underwent more operative time than the decompression alone group.(in RCT subgroup MD=-86.53, 95% CI (–167.68, -5.38), Z = 2.09 P = 0.04.In comparative observational studies group MD=-81.90, 95% CI (–102.25, -61.55), Z = 7.89 P<0.0001).
Major complications
Nine studies reported major complications in the two groups with a heterogeneity test showing statistically significant heterogeneity between groups (P༜0.00001, I2 = 83%). A random effects model was applied for meta-analysis (Fig. 12),which indicated that there was no significant difference in Major complications between the decompression group and the decompression plus fusion group(RCT subgroup,RR = 1.62, 95% CI (0.77, 3.41), z = 1.27 P = 0.20,Comparative observational studies subgroup, RR = 0.47, 95% CI (0.17, 1.27), z = 1.49, P = 0.14).
Number of reoperation
Eight studies reported reoperation numbers in the two groups. Low heterogeneity was identified between the two groups (P = 0.23, I2 = 25%).Because the article is a two-category variable and the research types of the two subgroups are different,so a random effects model was applied for meta-analysis (Fig. 13). There was no statistically significant difference between the two groups[RCT subgroup, RR = 1.52, 95% CI (0.94, 2.46), Z = 1.72 P = 0.09)].Comparative observational studies subgroup[RR = 0.65,95% CI (0.,1, 2.61), Z = 0.61, P = 0.54)].
Clinically satisfaction
Six studies reported the clinically atisfactions in the two groups. A significantly different heterogeneity was found between the two subgroups (P༜0.00001, I2 = 84%), and a random effects model was applied for meta-analysis (Fig. 14). The clinical satisfaction in the decompression plus fusion group was not better than in the decompression alone group[RR = 0.83, 95% CI (0.60,1.14), Z = 1.15 P = 0.25].
Sensitivity analysis
For the studies with I2༞50%, we used the random effect model to analyze the heterogeneity, and we used the article-by-article elimination literature method to analyze the indicators with heterogeneity.Because of the statistical differences in back pain scores between the RCT subgroup and the observational subgroup, we focused on the source of their heterogeneity.There was no significant change in heterogeneity after excluding the sensitivity of literature analysis.The analysis of ODI, Major complications, Duration of operation and Clinically satisfaction showed that the slight change of heterogeneity was not statistically significant.This proves that our research results are more reliable.