3.1 Study selection
The two authors(Chao Xu and He Zhu) retrieved a total of 843 articles, 32 of which were viewed in full text, 13 articles that met the inclusion criteria and were analyzed for data[10, 14, 16, 18, 23-31]. Fig. 1 details the search, inclusion and exclusion of the literature. A total of 751 patients (372 patients underwent caesarean section[10, 16, 18, 29-31] and 379 patients underwent laparoscopic surgery[14, 23-28] in the lower abdomen ) were finally included in the data analysis, 373 in the QLB group ( 215 patients were performed with posterior QLB[10, 14, 16, 23, 26, 30, 31] and 130 patients were performed with transmuscular QLB[24, 25, 27-29], and 28 patients were performed lateral QLB[18] ) and 380 in the control group. All QLB were guided by ultrasound and all included patients undergoing lower abdominal surgery were adults. Most trials had demonstrated random sequence generation and 1 trial did not demonstrate this, but there is no evidence to prove the low quality of it [14]. Strict Blinding was performed on all evaluators, ensuring minimal performance bias. However, patients in 4 trials were not blinded and may result in a detection bias of pain scores[10, 17, 18, 22].
Additional file 1 presents the outcomes of included articles, Additional file 2 presents the results of the meta-analysis and Fig. 2 presents the results of quality assessment completed by two authors; the characteristics of the included articles were presented in Additional file 3. Sensitivity analysis was performed on all data synthesis, and no trials were found to reduce I2 to 50%. Since less than 10 articles were included, the Begg’s test was not performed.
3.2 Interval opioid consumption
We synthesized the total consumption of opioids at 12h, 24h and 48h after surgery to quantitatively evaluate whether QLB could reduce the need for opioids in patients after surgery. Cumulative opioid consumption was transformed to intravenous morphine equivalents (morphine 10 mg i.v. = morphine 30mg p.o.= tramadol 100 mg i.v. = ketobemidone 8 mg i.v. = fentanyl 1mg i.v. = sufentanil 0.01mg i.v.).
Compared with control group, QLB group can effectively reduce 6, 12 and 24h cumulative opioid consumption, but the data are highly heterogeneous (I2=92%, 93% and 93%,respectively), after the subgroup analysis, we found: QLB group can effectively reduce 24 and 48h cumulative opioid consumption 10.1mg (95%CI: -13,-7.2; P<0.00001) and 16.22mg (95%CI: -19.39,13.03; P<0.00001) in patients with cesarean section, For patients underwent laparoscopic surgery , QLB group can reduce effectively cumulative opioid consumption at 6, 12 and 24h with 4.91mg (95%CI: -9.68,-0.14; P=0.04), 3.19mg (95%CI: -6.23,-0.16; P=0.04) and 7.95mg (95%CI: -15.88,-0.01; P=0.05) ,respectively. Posterior QLB group and transmuscular QLB group reduced 24h cumulative opioid consumption 4.03mg ( 95%CI: -7.89, -0.19; p=0.04) and 12.44mg ( 95%CI: -20.2, -4.68; p=0.002), respectively (Additional file 2 and Fig. 3).
The cumulative morphine consumption in the laparoscopic subgroup at 12h and in the cesarean section subgroup at 24 and 48h showed low heterogeneity (I2=41%, 18% and 0%,respectively). The rest of the morphine consumption data showed high heterogeneity, and the sensitivity analysis still could not effectively reduce the I2 value.
3.3 Rest pain
Because the included articles used visual analogue scale (VAS)[10, 23, 26, 27, 30, 31] or numerical rating scale scoring (NRS)[14, 16, 18, 29] to assess postoperative rest pain, we synthesized the data using the same scoring method at the same time. Because only one trial[18] evaluated the resti NRS score at 36h after surgery, we calculated pain scores at 6, 12, 24, and 48h postoperatively.
At 6h postoperatively,there was no statistically significant between QLB group and control group in rest VAS score.
At 12h postoperatively, QLB can reduce rest VAS score and NRS score 1.66cm ( 95%CI: -2.07, -1.26, P<0.00001; I2=61%) and 1.1cm ( 95%CI: -1.65, 0.55, P<0.0001; I2=20%) , respectively. The laparoscopic subgroup(I2=64%), cesarean section subgroup(I2=20%) and posterior QLB subgroup (I2=61%) all showed statistical differences compared with control group.
At 24h postoperatively, QLB can reduce rest VAS score and NRS score 2.05cm ( 95%CI: -3.39, -0.7, P=0.003; I2=97%) and 2.34cm ( 95%CI: -3.22, -1.45, P<0.00001; I2=94%) , respectively. Compared with control group, QLB subgroup can reduce NRS score 1.07cm ( 95%CI: -1.51, -0.62, P<0.00001; I2=0%) in patients underwent cesarean section (Additional file 2 and Additional Fig. 4).
At 48h after operation, QLB can only reduce rest VAS score 0.81cm ( 95%CI: -1.56, -0.03, P=0.04; I2=0%) but can not statistically reduce NRS score.
3.4 Dynamic pain
Six trials[10, 14, 16, 29-31] evaluated postoperative dynamic pain scores, 3 trials[10, 30, 31] used VAS, and 3 trials[14, 16, 29] used NRS. [10, 30, 31]. Because few trials could be included, we analysed pain scores at 6, 12 and 24h after operation.
Only at 24h postoperatively, QLB can effectively reduce dynamic NRS scores 1.76cm ( 95%CI: -2.29, -1.22, P<0.00001; I2=83%),the dynamic NRS of patients after cesarean section was reduced by 0.89cm (95%CI: -1.67, -0.12, P=0.02; I2=0%) (Additional file 2).
3.5 Opioid-related adverse effects
Seven articles [10, 15-19, 22] reported the incidence of nausea and vomiting 24 hours after surgery, and only three of these articles reported nausea and vomiting[15, 17, 19]. Two articles [14, 16] used NRS scores to measure the degree of nausea and vomiting. We synthesized the incidence of nausea and vomiting 24 hours after surgery. There was no significant difference in the incidence of nausea and vomiting between the QLB group and the control group.
Few observations were reported about the incidence of itching and sedation[10, 18, 22], therefore statistical synthesis was not performed. Only one article[10] observed the incidence of itching and sedation (one case of itching and one case of sedation), and it was not possible to determine whether a significant difference exists between the QLB group and the control group.
3.6 Additionalanalgesic
Three articles[14, 18, 29] reported the time interval to the first additional analgesic ,and 5 articles[14, 18, 23, 26, 28] reported the numbers of patients who used additional analgesic. No statistical differences were found after subgroup analysis (Additional file 2).
3.7 Opioid-related adverse effects
Eight articles[10, 14, 16, 18, 26, 28, 29, 31] reported the incidence of postoperative nausea and vomiting (PONV) within 24 hours. We found that statistical differences only exist in the transmuscular QLB subgroup with an odds ratio of 0.29 (95%CI:0.12,0.70,P=0.005, I2=0%) (Additional file 2).
Four articles[10, 18, 26, 31] reported the incidence of pruritus within 24 hours after surgery, and no statistical differences were found after analysis (Additional file 2 and Fig.5 ).
3.8 Patient satisfaction
Three trials[23, 28, 29] performed patient satisfaction surveys, but different scales were used( Bruggemann comfort scale (BCS) scores, Quality of Recovery 40 (QoR-40) questionnaire score and obstetric quality-of-recovery scoring tool( ObsQoR -11)), We did not perform data synthesis.
Zhu et al[28] reported that the recovery quality of patients in the QLB group was significantly higher than that of the control group. No statistical difference was found in the other two articles[23, 31].