The overall objective of this study is to compare the analgesic effects of QLB vs. CB in children undergoing abdominal surgery. The primary outcome is the postoperative rescue analgesia rate, defined as the proportion of patients requiring additional analgesics such as acetaminophen or ibuprofen when their pain score exceeds the predefined threshold within 24 h after surgery. Secondary outcomes include resting pain scores (on a 0–10 scale) at 30 min, 4 h, 12 h, and 24 h post-operatively, the time to first rescue analgesia, the incidence of postoperative nausea and vomiting (PONV), and postoperative complications (such as post-block infection, local anesthetic allergy, hematoma, nerve injury, and local anesthetic toxicity). This systematic review was registered under PROSPERO, ID CRD 42023441447.

The primary literature search was performed in July 2023, and the PubMed, Cochrane Library, Web of Science, and Embase databases were searched from inception until July 2023. To ensure that no newly published article was dropped, the secondary literature search was conducted in August 2023. During the initial literature search, we searched the terms edQL OR quadratus lumborum OR QL block OR QLB OR quadratus lumborum block)” AND “(CB OR caudal block OR caudal epidural blocks OR caudal analgesia OR caudal blockade OR caudal anesthesia OR caudal regional anesthesia OR caudal extradural anesthesia)” AND “(pediatric OR children OR infant OR adolescent OR schoolchild OR preschool OR teens OR youth)”. The secondary literature search was carried out the same as the primary literature search to find new articles.

This meta-analysis was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) (6). The inclusion criteria were as follows: (1) Comparing the effects of analgesia between the QLB and CB for pediatric patients undergoing abdominal surgery; (2) Consideration of the rate of postoperative rescue analgesia as the outcome; (3) Randomized controlled trials (RCTs). The exclusion criteria were as follows: (1) Receiving additional treatment in the control or experimental groups; (2) Duplicate publication; (3) An indeterminate type of study or non-RCTs; (4) Failure to extract valid data; (5) Unavailability of the full text of study; (6) Conference abstracts; (7) Reviews or systematic reviews.

Two independent reviewers reviewed and identified studies based on the previously mentioned strategy. Any discrepancies were resolved by discussing them with a third reviewer.

The data extraction was performed by two researchers. Some studies used graphs to show the outcomes in lieu of data, thus, it was attempted to contact corresponding authors to obtain original data. Nevertheless, if no reply was received from the corresponding authors for the original data, GetData tool was used to extract values for each data point.

The following data were extracted: first author's name, country, date of publication, study design, the sample size of experimental and control groups, participants' characteristics, surgical procedures, analgesic methods (medication type, dose, tube placement, duration), time of follow-up, the rescue analgesia, pain scores [time of evaluation, mean, and standard deviation (SD)], the incidence rates of postoperative complications or by-affects (e.g., PONV). For the quantitative analysis, the numeric rating scale (NRS) or the visual analogue scale (VAS) scores were calculated in the range of 0–10 points (0 = no pain, 10 = extreme pain). Pain scores reported as face, legs, activity, cry, consolability (FLACC), or Children's Hospital of Eastern Ontario pain scale (CHEOPS) were transformed to a scale ranging from 0 to 10 points.

We conducted meta-analysis for all outcomes. The data was analyzed using Review Manager V5.3. (Cochrane Collaboration, Copenhagen). A weighted average difference or standard average difference was used to analyze the continuous data, and 95% confidence intervals (CIs) were used to combine different scales. Data were presented as mean, and SD values were extracted directly. Data were presented as median using the following formula: (a + m + b)/3, where a = Q25 and b = Q75 to convert median to mean, and data were presented as interquartile range (IQR) using the following formula: SD = IQR/1.349 to convert IQR to SD. Data were presented as CI using the following formula: 95% CI = x ± 1.96*SE, where SD = SE* √n to convert CI to SD. The above-mentioned statistical formulas were previously described by Hozo et al. (15). We used I² statistics to quantify heterogeneity. If I² ≤ 50%, we chose a fixed effects model; if I² > 50%, we chose a random effects model.

A total of 41 trials were identified in the primary literature search, and there were 21 of the records after duplicates were removed. Based on the inclusion and exclusion criteria, 9 trials were excluded, and 12 trials were further screened, of which 5 trials were removed (Figure 1) (13, 16–21). Characteristics of the included studies are listed in Table 1. A summary of findings is reported in Table 2. The risk of bias (RoB) assessment is reported in Figures 2, 3, with the randomization process and selection of reported results being the most common sources of potential bias.

Six trials (13, 16–19, 21) that enrolled 392 patients reported the rate of postoperative rescue analgesia. Compared with caudal analgesia, QLB showed a significant reduction in the rate of postoperative rescue analgesia. [relative risk (RR) = 0.37; 95% CI = 0.26 to 0.51; P < 0.01; [I² = 50%, P = 0.07], Figure 4]. Among these trials, only five of them (13, 16, 17, 19, 21) utilized posterior QLB, whereas the trial by İpek et al. (18) employed lateral QLB. In order to exclude the influence of different QLB approaches on the results, subgroup analysis was conducted, and the final result was similar. [RR = 0.33; 95% CI = 0.23 to 0.47; P < 0.01; (I² = 49%, P = 0.10), Figure 4].

The postoperative pain scores at 0.5, 4, 12, and 24 h were investigated, and all studies reported pain scores at 0.5, 4, and 24 h. However, only 3 studies reported pain scores at 12 h (16, 20, 21), except for the trial conducted by Stato. M et al. (13) However, data from only four studies (13, 16, 20, 21) were analyzed in the forest plot. Because data conducted by Ashoor et al. (17) and Ipek et al. (18) were presented as graphs rather than specific values, and no original data could be obtained. In addition, the pain scores in the study conducted by Öksüz et al. (19) were reported in the form of a median, which could not be converted to the form of mean and standard deviation due to the skewed distribution of the data.

At 30 min after surgery, there was no significant difference in pain scores between the two methods (SMD = 0.24; 95% CI = −0.15 to 0.63; P = 0.22; Figure 5), and this result was influenced by real heterogeneity (I² = 75%, P = 0.007).

At 4 h after surgery, the pain score of the QLB group was lower than that of the CB group, and the results were statistically significant [SMD = −0.11; 95% CI = −0.21 to −0.01; P = 0.02, (I² = 0%, P = 0.41); Figure 6].

At 12 h after surgery, there was no significant difference in pain scores between the two methods (SMD = −0.76; 95% CI = −2.03 to 0.51; P = 0.24; Figure 7), and this result was influenced by real heterogeneity (I² = 98%, P < 0.01).

At 24 h after surgery, there was no significant difference in pain scores between the two methods (SMD = 0.12; 95% CI = −0.30 to 0.54; P = 0.57; Figure 8), and this result was influenced by real heterogeneity (I² = 87%, P < 0.01).

All of the trials (13, 16–21) investigated the incidence of postoperative complications or by-effects, such as nausea, vomiting, hypotension, bradycardia, and urinary retention. Gözen Öksüz et al. (19) reported one case of nausea in the CB group. Celal Bulut İPEK et al. (18) identified 3 cases of postoperative urinary retention in the CB group. Tarek M. Ashoor et al. (17) detected 5 cases of urine retention, 3 cases of nausea and vomiting, 3 cases of hematoma, 2 cases of hypotension, and one case of bradycardia in the QLB group, while 9 cases of nausea and vomiting, 3 cases of urine retention, and one case of hypotension in the CB group. Yue Zhang et al. (21) reported 3 cases of nausea or vomiting in the CB group and 2 cases in the QLB group. Amin M. Alansary et al. (16) reported 7 cases of PONV in the CB group and 6 cases in the trans-incisional QLB (TiQLB) group. Makoto Sato et al. (13) and Safaa Gaber Ragab et al. (20) reported no complications. No significant difference was found between QLB and CB in the complications or side effects (RR = 0.94; 95% CI = 0.59 to 1.48; P = 0.77). In addition, there was no significant heterogeneity among the trials (I² = 0%, P = 0.56; Figure 9).

CB is the most frequently utilized regional anesthesia technique among children, particularly suitable for surgical interventions situated beneath the umbilicus (T10 dermatome) (26, 27). This method offers safe access to the epidural space in pediatric patients (28, 29). However, a single CB may have a relatively short duration of analgesia, only around 4 to 6 h, which can often be considered insufficient (30, 31), while placing catheters in the tail area increases the risk of infection and prevents early activity (32). The reason for its short duration is the rich distribution of blood vessels, leading to the rapid absorption of the local anesthetics (33). Furthermore, it comes with potential drawbacks such as the risk of accidental dural puncture, bladder dysfunction, and so on. Due to the above disadvantages, the application of CB is limited.

QLB is not only applicable to pediatric abdominal surgeries but it has also been proven effective in lumbar disc herniation surgeries (34), total hip replacement surgeries (35), as well as adult abdominal surgeries such as laparoscopic cholecystectomy (36) and cesarean section (37). A cadaver study (38) involving the performance of anterior QLB, using a dye, has demonstrated that QLB can induce analgesia from T10 to L4, as evidenced by the colored lumbar nerve roots and occasionally some nerves in the TAP. Additionally, other studies have indicated that anterior QLB may extend cephalad beyond previous levels, reaching T7–T12 spinal nerve roots, which could be the reason for the effectiveness of QLB (7). In adults, the mean effective duration of sensory block was 14.1 h after unilateral anterior QLB at the L4 level with 20 ml of 0.375% ropivacaine (39). Previous research (40) reported that QLB had safety features with minimal complications and/or side effects and demonstrated that QLB was the most effective technique for providing postoperative analgesia in pediatric patients undergoing lower abdominal surgery. Compared with the conventional CB or opioid analgesics in pediatric patients, QLB had a better analgesic effect and minimal side effects (41), which is consistent with our results. Compared with CB, QLB reduced the rate of rescue analgesia.

In terms of adverse reactions, although we concluded that there was no statistical difference between the QLB group and the CB group, we should pay more attention to block-related complications rather than general postoperative complications, but unfortunately, the RCTS we included did not indicate it.

The aggregated data clearly shows that QLB leads to a significantly lower rate of postoperative rescue analgesia compared to CB. As such, it is imperative to delve into the clinical significance of this difference and explore its effects on perioperative care and the overall quality of recovery. The observed trends, such as improved pain scores and a decrease in the number of patients needing additional pain relief, point to crucial research areas. This knowledge has the potential to significantly influence the development, application, and success of enhanced recovery after surgery protocols (42–44).

Our summary analysis is rooted in a well-considered and systematic search process, which involves independent verification and data extraction by two authors. This enables the analysis of seven studies, involving a total of 444 patients. The results derived from this large cohort are significant, as it is challenging to conduct large-scale pediatric trials required to address issues related to recovery characteristics.

However, the included studies and sample sizes were small. Small study effect bias and unpublished bias may exist. Through our results, although the difference is firm between the two groups, more studies are still needed in this area. The results of this study exhibited varying degrees of heterogeneity, prompting us to conduct a sensitivity analysis using the “one-by-one exclusion” method. Our meta-analysis revealed that a significant portion of this heterogeneity could be attributed to the article by Amin M. Alansary et al. (16). This could potentially stem from the unique approach employed in this RCT, where the author administered ultrasound-guided TiQLB to pediatric patients undergoing open renal surgery. The timing of the block, the type of surgery performed, and the specific technique used for the block all differed from those in other studies, contributing to the observed heterogeneity. In addition, in our article, the pain scores at most time points failed to reach statistical significance. This observation can likely be attributed to the comprehensive pain management protocols employed in these trials, which effectively mitigated pain levels among participants.