Secondary outcome indicators
Surgical duration
Eight studies15-22 described the surgical duration for submandibular gland resection through the retroauricular hairline approach. A total of 330 patients underwent submandibular gland resections; 144 through the retroauricular hairline approach and 186 through the conventional approach. The surgical duration for the retroauricular hairline approach was significantly longer than that for the conventional approach (WMD: 27.26; 95% CI: 6.35–48.18; P = 0.01; Fig. 5), and there was significant heterogeneity among the studies (χ2 = 613.23, df = 7, P = 0.01, I2 = 99%; Fig. 5). Visual assessment of the funnel plot indicated a slight publication bias. However, the Egger’s test was not statistically significant (p = 0.126) (Fig 6).
3.3 Wound length
Four of the studies15-17, 22 described the wound length in 126 patients; 62 and 64 in the retroauricular and conventional approach groups, respectively. The retroauricular hairline group had significantly longer wound lengths compared to the conventional group (WMD: 4.65; 95% CI: 1.94–7.36; P = 0.0008; Fig. 7). There was significant heterogeneity among the studies (χ2 = 120.03, df = 3, P = 0.0008, I2 = 98%; Fig. 7).
3.4 Intraoperative bleeding
Intraoperative bleeding volume was described in two of the studies16, 22, with a total of 88 patients; 43 and 45 in the retroauricular and conventional approach groups, respectively. There were no statistically significant differences between the two surgical approaches (WMD: −7.39; 95% CI: −28.40–43.61; P = 0.49; Fig 8). There was significant heterogeneity among the studies (χ2 = 28.85, df = 1, P = 0.00001, I2= 97%; Fig. 8).
3.5 Postoperative drainage
Postoperative drainage was described in three of the studies18, 19, 22, with a total of 130 patients; 64 and 66 in the retroauricular hairline and conventional approach groups, respectively. Data pooling showed no significant differences in postoperative drainage between the two approaches (WMD: 17.49; 95% CI: −4.36–39.34; P = 0.12; Fig 9). There was significant heterogeneity among the studies (χ2 =28.93, df = 2, P<0.00001, I2 = 93%; Fig. 9).
3.6 Length of hospital stay
Five of the studies reported the duration of hospitalization15, 17-20, 22. A total of 168 patients underwent submandibular gland resection; 83 and 85 through the retroauricular hairline and conventional approaches, respectively. The results showed no significant difference in the number of hospital days between the two groups (WMD: −0.25; 95% CI: −0.99–0.49; P = 0.51; Fig 10).
3.7 Cosmetic effects
Two of the studies [15-16] described cosmetic effects using a visual analogue scale (VAS) in 58 patients, 28 and 30 in the retroauricular and conventional approach groups, respectively. The VAS uses objective content satisfaction scores on a scale of 0–10; the higher the score, the higher the satisfaction level. The retroauricular hairline group had significantly better cosmetic effects than the conventional group. The difference in cosmetic effects between the two surgical approaches was statistically significant (WMD: 3.83; 95% CI: 2.50–5.16; P<0.00001; Fig 11). There was greater heterogeneity among the studies (χ2 =2.62, df = 1, P = 0.11, I2 = 62%; Fig. 11).
3.8 Subgroup analysis
Only two of the studies reported the intraoperative bleeding volume16, 22, which made it difficult to perform a subgroup analysis. Furthermore, postoperative drainage and cosmetic effects were not reported adequately to perform a subgroup analysis.
3.8.1 Open submandibular gland excision Compared to the original analysis, there were no significant differences in the subgroup analysis for surgical durations of open submandibular gland excision through the two approaches (WMD: 0.60; 95% CI: −18.00–19.20; P = 0.95; Fig 5). 3.8.2 Endoscopic submandibular gland excision
Compared to the original analysis, the subgroup analysis was consistent with the original data analysis for surgical duration (WMD: 30.83; 95% CI: 7.39–54.27; P = 0.010; Fig 5), wound length (WMD: 5.76; 95% CI: 5.22–6.29; P < 0.00001; Fig 7), and length of hospital stay (WMD: −0.77; 95% CI: −2.20–0.67; P = 0.30; Fig 10).
3.8.3 Robotic submandibular gland excision
Compared to the original analysis, there were no significant differences in the subgroup analysis for surgical duration between the two groups (WMD: 48.24; 95% CI: −14.27–110.75; P = 0.13; Fig 5). 3.9 Sensitivity analysis
To evaluate the robustness and reliability of our results, we performed a sensitivity analysis. For articles with a significant percentage, we excluded and re-analyzed the data.
3.9.1 Forest plot and meta-analysis for temporary injury to other nerves.
In this study by Carlos et al, there was a significantly higher proportion of the outcome metric of temporary injury to other nerves than in other literature. After excluding this study, the outcome was consistent with the original data analysist (OR: 0.77; 95% CI: 0.24–2.40; P = 0.65; Fig. 12). This was consistent with the original consolidated results and showed that the results were stable.