VAS score at 2 hours

A total of three randomized studies reported postoperative VAS scores at 2 h, including 85 patients in the CGNA group and 85 patients in the GA group. The meta-analysis indicated a statistically significant reduction in pain in the CGNA group (MD −1.70 points; 95% CI −3.36 points to −0.04 points; p = 0.045; I2 = 95%; Fig. 2). High heterogeneity remained after leave-one-out sensitivity analysis (Appendix: Supplementary Figure 1). Baujat plot depicts the study by Elsayed et al to be the biggest contributor to the overall heterogeneity (Appendix: Supplementary Figure 2).

Fig. 2

VAS pain scores at 2 h postoperatively improved in the CGNA group. VAS: Visual Analogue Scale; CGNA: Combined general and neuraxial anesthesia; GA: General anesthesia; MD: Mean difference; IV: Inverse variance; CI: Confidence intervals

VAS score at 4 hours

A total of three randomized studies reported postoperative VAS scores at 4 h, including 85 patients in both groups. The meta-analysis showed lower pain scores in the CGNA group (MD −1.35 points; 95% CI −2.55 points to −0.15 points; p = 0.027; I2 = 97%; Fig. 3). After the removal of the study by Luchetti et al., heterogeneity decreased to 0% and the statistical significance of the results remained (Appendix: Supplementary Figure 3). The Baujat plot identified this study as highly influential on the overall pooled effect size (Appendix: Supplementary Figure 4).

Fig. 3

VAS pain scores at 4 h postoperatively improved in the CGNA group. VAS: Visual Analogue Scale; CGNA: Combined general and neuraxial anesthesia; GA: General anesthesia; MD: Mean difference; IV: Inverse variance; CI: Confidence intervals

VAS score at 6 hours

A total of three randomized studies reported postoperative VAS scores at 6 h, with 95 patients in the CGNA group and 95 patients in the GA group. The meta-analysis found a marginally lower pain score in the CGNA group (MD −0.66 points; 95% CI −0.91 points to −0.40 points; p = <0.001; I2 = 23%; Fig. 4).

Fig. 4

VAS pain scores at 6 h postoperatively improved in the CGNA group. VAS: Visual Analogue Scale; CGNA: Combined general and neuraxial anesthesia; GA: General anesthesia; MD: Mean difference; IV: Inverse variance; CI: Confidence intervals

VAS score at 12 hours

A total of three randomized studies reported postoperative VAS scores at 12 h, with 122 patients in each group. The meta-analysis showed lower pains scores in the CGNA group (MD −0.89 points; 95% CI −1.44 points to −0.34 points; p = 0.002; I2 = 55%; Fig. 5). Heterogeneity decreased to 0% after removing the study by Zhu et al. in the leave-one-out sensitivity analysis (Appendix: Supplementary Figure 5). Moreover, the pooled effect remained statistically significant, favoring the CGNA group.

Fig. 5

VAS pain scores at 12 h postoperatively improved in the CGNA group. VAS: Visual Analogue Scale; CGNA: Combined general and neuraxial anesthesia; GA: General anesthesia; MD: Mean difference; IV: Inverse variance; CI: Confidence intervals

Operation time

A total of six randomized studies reported operation time in minutes, with 172 patients in each group. The meta-analysis showed a longer operation time in the CGNA group; however, the results were not statistically significant (MD 1.75 min 95% CI −3.87–7.36 min; p = 0.542; I2 = 62%; Fig. 6). This pattern remained on sensitivity analysis (Appendix: Supplementary Figure 6).

Fig. 6

Operation times were similar between groups. CGNA: Combined general and neuraxial anesthesia; GA: General anesthesia; MD: Mean difference; IV: Inverse variance; CI: Confidence intervals

Anesthesia time

A total of three randomized studies reported anesthesia time in minutes, with 102 patients in each group. After the meta-analysis, no difference was found between the CGNA and GA group (MD 2.29; 95% CI −3.05 to 7.63 min; p = 0.400; I2 = 34%; Fig. 7). The same statistical pattern remained after sensitivity analysis (Appendix: Supplementary Figure 7).

Fig. 7

Similar duration of anesthesia between groups. CGNA: Combined general and neuraxial anesthesia; GA: General anesthesia; MD: Mean difference; IV: Inverse variance; CI: Confidence intervals

Postoperative nausea and vomiting (PONV)

A total of three randomized studies reported the rates of PONV, with 122 patients in each group. The meta-analysis showed a reduction in PONV rates in the CGNA group compared to the GA alone group. However, the results were not statistically significant (OR 0.40; 95 CI%; 0.11 to 1.50; p = 0.176; I2 = 39%; Appendix: Supplementary Figure 8).

Intraoperative heart rate at 30 minutes

A total of four randomized studies and one observational study reported intraoperative HR at 30 min, with 255 patients in the CGNA group and 225 in the GA group. No differences were found between groups (MD −7.93 beats per minute; 95% CI −18.94 beats per minute to 3.08 beats per minute; p = 0.158; I2 = 97%; Appendix: Supplementary Figure 9). The Baujat plot indicated that the study by Sale et al. was highly influential on the overall pooled effect size and the largest contributor to heterogeneity (Appendix: Supplementary Figure 10). However, high heterogeneity persisted even after excluding this study in a leave-one-out sensitivity analysis (Appendix: Supplementary Figure 11).

Intraoperative systolic blood pressure at 30 minutes

A total of two randomized studies and one observational study reported intraoperative SBP at 30 min, with 205 patients in the CGNA group and 175 in the GA group. The meta-analysis showed no difference between groups (SMD −1.65; 95% CI −3.48 to 0.18; p = 0.077; I2 = 98%; Appendix: Supplementary Figure 12). The leave-one-out sensitivity analysis revealed that the study by Zhang et al. was the largest contributor to heterogeneity. Nevertheless, the heterogeneity remained high even after removing this study (Appendix: Supplementary Figure 13).

Intraoperative diastolic blood pressure at 30 minutes

A total of two randomized studies and one observational study reported intraoperative DBP at 30 min, with 205 patients in the CGNA group and 175 in the GA group. The meta-analysis found no difference between groups (SMD −1.53; 95% CI −3.10 to 0.04; p = 0.056; I2 = 98%; Appendix: Supplementary Figure 14). The study by Zhang et al. was identified as the largest contributor to heterogeneity in the leave-one-out sensitivity analysis; however, heterogeneity remained high even after its removal (Appendix: Supplementary Figure 15).

Intraoperative mean arterial pressure at 30 minutes

A total of four randomized studies and one observational study reported MAP at 30 min during surgery, 235 patients in the CGNA group and 205 patients in the in the GA group. No difference was found between groups on primary analysis (SMD −0.51; 95% CI −1.04–0.01; p = 0.054; I2 = 87%; Appendix: Supplementary Figure 16). Leave-one-out sensitivity analysis showed a decrease in MAP when excluding the observational study by Zhang et al (Appendix: Supplementary Figure 17).

Quality assessment

Individual study appraisals are detailed in Figs 8 and 9 for randomized and non-randomized studies, respectively. Among the randomized studies, eight showed some concerns or high risk of bias, while only one study was assessed as having a low risk of bias, with most issues arising from the randomization process. In the non-randomized studies, the primary sources of bias were confounding and missing information. In the GRADE assessment of postoperative pain, the level of certainty ranged from low to moderate, with most downgrades attributed to inconsistency and biases inherent to the study designs.

Fig. 8

Risk of Bias Tool for Randomized Trials (RoB2)

Fig. 9

Risk of Bias in Non-Randomized Studies – of Interventions (ROBINS-I)