The Impact of Different Regional Anesthesia Techniques on the Incidence of Chronic Post-surgical Pain in Patients Undergoing Video-Assisted Thoracoscopic Surgery: A Network Meta-analysis

This meta-analysis adhered to the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) guidelines [21] and was registered in the International Prospective Register of Systematic Reviews (PROSPERO, CRD42023480863). Two researchers (Yue Zhao and Yaming Guo) independently conducted searches on PubMed, the Cochrane Library, and Embase for English-language RCTs up to September 12, 2023. Searches were performed using the respective search engines of each database (further details provided in Supplementary Material).

The inclusion criteria were as follows: (1) studies written in English; (2) RCTs; (3) the following types of VATS are included: single-port VATS, multi-port VATS, and video-assisted thoracotomy; (4) studies following up the incidence of chronic pain persisting beyond 2 months after VATS; (Before the ICD-11 definition of CPSP was established, some studies on chronic post-surgical pain following thoracic surgery defined CPSP as pain persisting for at least two months post-surgery. To better analyze the occurrence of CPSP in this meta-analysis, we included RCTs that used 2-month or 3-month cutoffs); (5) studies compared the analgesic efficacy of at least two of the following interventions: control group (without regional anesthesia techniques), PVB, ICNB, SAPB, ESPB, and TEA; (6) surgeries conducted under general anesthesia with non-awake intubation.

The exclusion criteria were as follows: (1) non-randomized controlled trials, retrospective studies, reviews, case reports or animal studies; (2) studies with unreported or insufficiently reported outcomes; (3) studies employing two or more regional anesthesia techniques concurrently on the same patient; (4) studies involving pediatric populations.

Two independent reviewers (Yue Zhao and Yaming Guo) initially screened out irrelevant studies by evaluating titles and abstracts. Subsequently, full-text reading was conducted to exclude studies that did not meet the inclusion criteria. In instances of disagreement between the two reviewers, the opinion of a third reviewer was sought for resolution.

Data extraction from all studies was conducted independently by Yue Zhao and Yaming Guo utilizing a standardized data collection form. If outcome variables in published articles were unavailable or incomplete, the corresponding authors were contacted to obtain original data. Data recorded for each eligible article included: (1) name of the first author; (2) year of publication; (3) country of the trial; (4) age of patients; (5) intervention groups; (6) number of patients in each group; (7) follow-up time; (8) the definition of CPSP; (9) outcome measures; and (10) the incidence of CPSP.

The follow-up time points included 2, 3, 6, and 12 months. However, due to only one study [10] reporting the incidence of CPSP at 12 months after VATS, this time point was not considered a primary outcome. The main outcomes were categorized into two groups: the incidence of CPSP at 2–3 months and at 6 months. In cases where studies separately reported pain at rest and pain during coughing, the incidence of pain during coughing was utilized to calculate the incidence of CPSP.

The risk of bias assessment was conducted using Review Manager (version 5.4). Two researchers (Yue Zhao and Yaming Guo) independently utilized the Cochrane Risk of Bias Assessment Tool [22] to assess the risk of bias. The assessment criteria of the Cochrane Risk of Bias Tool included seven items: random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective reporting, and other bias. Each item was evaluated as low risk of bias, unclear risk of bias, or high risk of bias to evaluate and classify the quality of the studies. In instances of disagreement, an impartial assessment was made by an expert.

R (version 4.2.2, R Project for Statistical Computing, https://​www.​r-project.​org/​) and R Studio were employed for data analysis. A network graph was generated to illustrate the direct or indirect comparisons of regional anesthesia techniques in the included studies. Bayesian analysis was conducted using the Gemtc package [23] in R. The I² statistic and Cochran’s Q test were used to determine the heterogeneity of direct comparisons. If significant heterogeneity was detected (I² > 50% or P < 5%), a random-effect model was employed; otherwise, a fixed-effect model was utilized. Binary variables were compared using the relative risk (RR) and 95% confidence interval (CI). When there was a loop, the consistency between direct comparison and indirect comparison was judged through node-splitting method [24]. P ≥ 0.05 indicated that no significant inconsistency was found. Rank probabilities and the surface under the cumulative ranking curve (SUCRA) [25] were employed to rank the effectiveness of each regional anesthesia technique in reducing CPSP and to identify the optimal analgesic method, with larger SUCRA values indicating better interventions. The small-study effect and publication bias were judged using comparison-adjusted funnel plots [26] with Egger’s test [27]. A P value < 0.05 was considered statistically significant.

This article is based on previously conducted studies and does not involve any new studies with human participants or animals performed by any of the authors.

A total of 716 records were obtained from the initial literature search. Following the screening of titles and abstracts, 57 articles were identified as potentially meeting the criteria. Upon full-text examination, 49 records were excluded for various reasons: one article was withdrawn, one article by Zhao X et al. [28] contained duplicate data found in a subsequent publication by the same author. To prevent redundancy in data analysis, only the later publication [10] was considered. Two studies were not RCTs, 27 studies did not report results, nine studies had a follow-up time of less than 2 months and nine studies involved only one of the following interventions: control group (without regional anesthesia techniques), PVB, ICNB, SAPB, ESPB and TEA or employed two or more regional anesthesia techniques concurrently on the same patient. Ultimately, eight articles [10, 28‐34] met the eligibility criteria and were included in our final NMA. Figure 1 illustrates the literature selection process. These trials, conducted between 2017 and 2023, were from five different countries and involved 785 individuals. Detailed information on participants and intervention measures from the eight studies is provided in Table 1.

Figure 2 illustrates the risk of bias for eight eligible RCTs. Risk levels are denoted by colors: green indicates low risk, yellow indicates unclear risk, and red indicates high risk. Six studies demonstrated low risk regarding random sequence generation by detailing methods such as computer-generated list of random numbers, block randomization, or Research Randomizer. Three studies achieved low risk in allocation concealment by employing consecutively numbered, sealed, opaque envelopes for participant assignment. Four RCTs were deemed to have low risk in blinding of participants and personnel, as participants were blinded to group assignments by receiving the regional anesthesia technique under midazolam-induced sedation or after induction of general anesthesia. Seven studies maintained low risk in blinding of outcome assessment, ensuring outcome assessors remained unaware of group assignments. In terms of incomplete outcome data, six RCTs showed low risk, with balanced dropout rates and reasons across groups. One study was deemed high risk due to a notable loss of participants during follow-up. Regarding selective reporting, one article reporting all predefined outcomes was rated low risk, while four articles were rated high risk due to unreported predefined primary outcomes and non-predefined reported outcomes. Three articles were categorized as unclear risk due to lack of registration information and unreported predefined secondary outcomes. All articles were considered to have low risk in terms of other bias, with no significant additional bias identified.

Visual network geometries (Fig. 3A, B) were utilized to illustrate each arm of the analysis. Each treatment was represented by a unique node, with the size of the node proportional to the number of patients receiving that treatment.

Our NMA for CPSP at 2–3 months included six analgesic methods (Fig. 3A). PVB was the most frequently studied intervention, with five studies (n = 230) included. Control group followed closely, with four studies (n = 193) included. ESPB was evaluated in two studies (n = 83), while ICNB and SAPB were each studied in one trial (n = 50 for ICNB and n = 46 for SAPB). TEA was also evaluated in one study (n = 50).

In the heterogeneity tests for CPSP incidence at 2–3 months, no heterogeneity was observed among the included studies (I² = 0%, P = 0.2603). Therefore, a fixed-effect model was employed for the analysis. A closed loop was formed by control group, PVB and ESPB and the inconsistency test was conducted using the “node-splitting” method with P > 0.05, indicating no inconsistency.

For CPSP at 2–3 months, PVB (RR = 0.716, 95% CrI = 0.557–0.901), ICNB (RR = 0.589, 95% CrI = 0.407–0.806) or TEA (RR = 0.642, 95% CrI = 0.454–0.862) showed significantly lower incidences compared to control group. Additionally, the incidence of CPSP was significantly lower with PVB (RR = 0.539, 95% CrI = 0.286–0.937), ICNB (RR = 0.442, 95% CrI = 0.211–0.854) or TEA (RR = 0.480, 95% CrI = 0.233–0.926) compared to SAPB. No significant difference in the incidence of CPSP was observed in the comparison of the remaining interventions (Fig. 4A).

The rank probability results and Surface Under the Cumulative Ranking Curve (SUCRA) are presented in Figs. 5A and 6A, respectively. In Fig. 5A, the probability of ICNB being the most effective technique is 0.595, with a 0.000 probability of being the least effective. For SAPB, the probability of being the most effective is 0.003, while the probability of being the least effective is 0.790. The SUCRA results indicate that ICNB has the highest probability (0.8853) of being the most effective intervention, followed by TEA (0.789), PVB (0.641), ESPB (0.405), control group (0.211) and SAPB (0.068).

For CPSP at 6 months, a total of seven studies were included, incorporating the same six analgesic methods mentioned previously (Fig. 3B).

In the heterogeneity tests for CPSP incidence at 6 months, no heterogeneity was observed among the included studies (I² = 32.1%, P = 0.1158). Therefore, a fixed-effect model was employed for the analysis. Two closed loops were formed by control group, PVB and ESPB as well as control group, PVB and TEA. The inconsistency tests were conducted using the “node-splitting” method with P > 0.05, indicating no inconsistency.

The NMA results indicated that PVB (RR = 0.695, 95% CrI = 0.525–0.904), ICNB (RR = 0.600, 95% CrI = 0.407–0.832) or TEA (RR = 0.608, 95% CrI = 0.427–0.830) exhibited significantly lower incidences than control group. No significant difference in the incidence of CPSP was observed in the comparison of the remaining interventions (Fig. 4B).

The rank probability results and SUCRA are presented in Figs. 5B and 6B, respectively. In Fig. 5B, the probability of ICNB being the most effective technique is 0.429, with a 0.000 probability of being the least effective. For SAPB, the probability of being the most effective is 0.099, while the probability of being the least effective is 0.491. The SUCRA results suggest that ICNB has the highest probability (0.815) of being the most effective intervention, followed by TEA (0.804), PVB (0.622), ESPB (0.312), SAPB (0.266), and control group (0.1802).

The corrected funnel plots revealed no significant asymmetry in the intergroup comparison of the incidence of CPSP at 2–3 months and 6 months (Fig. 7A, B). Furthermore, the P values of Egger’s test were > 0.05 (P = 0.7875 for 2–3 months CPSP, P = 0.1836 for 6 months CPSP).