Study selection and characteristics
According to the search strategy, a total of 130 studies were searched and 3 additional studies were added in from other sources. 57 duplicate studies were excluded; 42 studies were excluded by reading the titles and abstracts; 34 relevant studies were assessed by reading the full text. 18 non-randomized controlled trials and 6 studies reporting the same data were excluded, and 10 studies were finally included. Nine studies[23–31] were published in English and one study[32] was published in Chinese. All studies reported no differences in the basic conditions of the patients and preoperative outcomes to be researched between the two groups. The search process and results were shown in Fig. 1, and the characteristics of the studies were shown in Table 1.
Study quality and risk of bias
The quality of the included studies was assessed by the Cochrane risk of bias assessment criteria and the results were shown in Fig. 2.
Analysis results
The risk of reherniation
Six studies[23, 26, 27, 29, 30, 32] reported the risk of postoperative reherniations, and 897 patients were included. The result was (RR: 0.42, 95%CI [0.30, 0.58], P < 0.00001), and the heterogeneity test was I2 = 0%, P = 0.47. It was suggested that there was no heterogeneity between the studies, suggesting the result was stable, so a fixed effect was used for the analysis. The results showed that the DAFR group significantly reduced the occurrence of postoperative disc reherniations (Fig. 3).
The risk of reherniation-related reoperation
Six studies[23, 24, 27, 28, 30, 32] reported the risk of postoperative reoperations, and 1509 patients were included. The result was (RR: 0.63, 95%CI [0.46, 0.87], P = 0.005), and the heterogeneity test was I2 = 0%, P = 0.61. It was suggested that there was no heterogeneity between the studies, suggesting the result was stable, so a fixed effect was used for the analysis. The results showed that the DAFR group significantly reduced the occurrence of postoperative reherniation-related reoperations (Fig. 3).
Oswestry Disability Index (ODI)
Eight studies[23–26, 28–30, 32] reported the postoperative ODI, and 2117 patients were included. The heterogeneity test was I2 = 87%, P < 0.00001. It was suggested that there was heterogeneity between the studies, so a sensitivity analysis was performed. No significant data deviation and no source of heterogeneity were found, suggesting that the results were relatively stable with low sensitivity, so a random-effects model was used. The result was (MD: -0.65, 95%CI [-2.34, 1.14], P = 0.48). The results showed no difference in postoperative ODI between the two groups (Fig. 4).
Visual Analogue Score of back pain (VAS-back pain)
Eight studies[23–26, 28–30, 32] reported postoperative VAS-back pain, and 2067 patients were included. The heterogeneity test was I2 = 70%, P = 0.001. It was suggested that there was heterogeneity between the studies, so a sensitivity analysis was performed. The analysis revealed that a study reported by Li[28] had a greater effect on heterogeneity, and the heterogeneity test performed after excluding this study was I2 = 32%, P = 0.19, suggesting that there was no heterogeneity, so a fixed-effect model was used. The result was (SMD: -0.06, 95%CI [-0.15, 0.02], P = 0.15). The results showed no difference in postoperative VAS-back pain between the two groups (Fig. 4).
Visual Analogue Score of leg pain (VAS-leg pain)
Seven studies[23–26, 28–30] reported postoperative VAS-leg pain, and 2019 patients were included. The heterogeneity test was I2 = 82%, P < 0.0001. It was suggested that there was heterogeneity between the studies, so a sensitivity analysis was performed. The analysis revealed that a study reported by Li[28] had a greater effect on heterogeneity, and the heterogeneity test performed after excluding this study was I2 = 0%, P = 0.42, suggesting that there was no heterogeneity, so a fixed-effect model was used. The result was (SMD: -0.06, 95%CI [-0.15, 0.03], P = 0.18). The results showed no difference in postoperative VAS-leg pain between the two groups (Fig. 4).
Operation time
Five studies[25, 26, 29, 31, 32] reported the operation time, and 736 patients were included. The heterogeneity test was I2 = 94%, P < 0.00001. It was suggested that there was heterogeneity between the studies, so a sensitivity analysis was performed. No significant data deviation and no source of heterogeneity were found, suggesting that the results were relatively stable with low sensitivity, so a random-effects model was used. The result was (MD: 11.66, 95%CI [2.77, 20.55], P = 0.01). The results showed that the DAFR group increased the operation time (Fig. 5).
Short-form health survey scale (SF-scale) and Disc height
Three studies[24, 26, 29] reported the postoperative SF scale, and 892 patients were included. The result was (SMD: -0.05, 95%CI [0.19, 0.08], P = 0.45), and the heterogeneity test was I2 = 0%, P = 0.55. Two studies[26, 29] reported postoperative Disc height, and 165 patients were included. The result was (MD: 0.95, 95%CI [0.53, 1.38], P < 0.00001), and the heterogeneity test was I2 = 0%, P = 0.37. It was suggested that there was no heterogeneity between the studies, so a fixed effect was used for the analysis. The results showed no difference in the postoperative SF scale between the two groups, while the disc height of the DAFR group was higher (Fig. 5).
The subgroup of the risk of reherniations and reherniation-related reoperations based on different annular fibrosus repair methods.
Three repair methods were included in the reherniation subgroup, and four repair methods were included in the reoperation subgroup. The analyses showed that the results of the Barricaid repair method were statistically significant (P < 0.0001), whereas the other repair methods did not (Annular Stapler (P = 0.05), Amniotic Membrane (P = 0.19) in the subgroup-reherniations. The subgroup-reherniation-related reoperations showed similar results (Barricaid (P < 0.01), Amniotic Membrane (P = 0.19), Annular Stapler (P = 0.19), Xclose (P = 0.5)). The results showed that the Barricaid repair method was more effective in reducing the risk of reherniations and the risk of reherniation-related reoperations (Fig. 6).