Introduction

Perioperative analgesia is essential in cardiac surgery. Ineffective pain control contributes to higher incidence of cardiovascular, pulmonary complications and stress reactions. These events increase in-hospital mortality [1]. Opioid analgesics are traditionally used for postoperative pain relief in cardiac surgery. Unfortunately, opioids can cause unwanted dose-dependent side effects (nausea, vomiting, depression of consciousness and breathing) that can significantly impair recovery [2, 3]. Mitral valve surgery underwent significant changes in the context of fast track postoperative recovery (from traditional sternotomy to thoracoscopic access and minithoracotomy) [4]. Multimodal analgesia with regional anesthesia was used in many studies devoted to perioperative period. These methods of regional anesthesia included local infiltration anesthesia, neuraxial techniques with thoracic epidural anesthesia (EA) and peripheral blocks [5–9].

EA is a highly effective method of pain relief in cardiac surgery. However, one of the potential disadvantages is the risk of epidural hematoma with serious consequences. This risk reaches 0.35% and increases with systemic heparinization. Postoperative hypocoagulation may put the patient at risk of epidural hematoma when the catheter is removed or even make removal dangerous until coagulopathy resolves [10]. The risk of epidural hematoma ranges from 1:150,000 to 1:1500 [10, 11]. However, previous studies and guidelines of the European Society of Anesthesiology and Intensive Care and European Society of Regional Anesthesia and Pain Therapy in patients receiving antithrombotic drugs classify peripheral nerve blockade as interventions with low risk of bleeding and do not require time intervals before and after administration of anticoagulants and antiplatelet agents [10, 12].

Thoracic paravertebral block (PVB) is also a well-established analgesic method, but it is associated with the risk of pneumothorax and inadvertent neuraxial injection [13, 14]. Blockades of thoracic-intercostal fascial plane, thoracic and anterior serratus plane are new types of regional anesthesia. According to certain studies, these methods provide satisfactory analgesia after sternotomy [5, 15]. However, large-scale randomized trials are required to confirm their effectiveness [8, 16].

Another new technique is erector spinae plane blockade (ESP). It is an interfascial block first described by Forero M. et al. in 2016 [17]. This procedure is simple and can be easily realized in perioperative period. Several recent meta-analyses evaluated the efficacy of ESP blockade in various abdominal, thoracic and spinal surgeries and emphasized the benefits of this technique [3, 6, 18–20]. Several reports are devoted to ESP block in cardiac surgery [21–25]. Importantly, the mechanism of sternal analgesia is still unclear despite MR data on transforaminal and epidural penetration of local anesthetics in addition to cephalo-caudal spread [26]. Several studies have shown promising results. However, there is still no convincing evidence of advantages of this blockade over traditional general anesthesia, EA, chest blocks. Some authors reported no parasternal analgesia after ESP blockade [23, 27, 28].

The purpose of this systematic review was to study the clinically significant effect of ESP blockade in adults undergoing cardiac surgery.

Material and methods

Searching strategy and selection criteria

A systematic review and meta-analysis were performed in accordance with the requirements for systematic reviews and meta-analyses (PRISMA, Preferred Reporting Items for Systematic Reviews and Meta-Analyses). No language restriction was set.

We established the following question for a systematic review: "Does ESP blockade cause a significant effect in adults undergoing cardiac surgery?". Inclusion criteria are pre-defined using the PICOS strategy (patient, population or problem (P); intervention (I); comparison (C); outcomes (O) and study design (S)) optimized for development of all steps of systematic review and meta-analysis (Table 1). All authors jointly developed inclusion/exclusion criteria prior to searching for relevant articles. Disagreements were resolved via consensus.

Table 1. Inclusion criteria for systematic review and meta-analysis (PICOS).

We included the studies that adequately represented binary and continuous data (mean/median; interquartile range, standard deviation and/or 95% confidence interval). Anonymized data were extracted only from published reports. Studies of ESP blockade in thoracic surgery were excluded to minimize heterogeneity.

Exclusion criteria: abstracts of conferences, meetings, case reports and case series, technical articles, recommendations, experimental studies. The number of patients was not extremely important.

Searching for appropriate data was performed in the PubMed, Google Scholar, MEDLINE and QxMD databases for the period 2016–2022.

Search query in the PubMed database: (erector spinae-plane block) OR (ESP) AND (minimal invasive cardiac surgery) OR (cardiac surgery) OR (minimal invasive direct coronary artery bypass) OR (mini-thoracotomy).

Search query in the Google Scholar database: erector spinae-plane block, cardiac surgery.

Search query in the MEDLINE database: Erector Spinae Plane Block, ESP, Cardiac Surgery.

Search query in the QxMD database: Erector Spinae Plane Block, Cardiac Surgery.

The last searching was made on May 23, 2022.

Data extraction and quality assessment

All titles were imported in the database, and duplicate articles were removed manually. After exclusion of duplicates, we analyzed titles and abstracts regarding study objective. After excluding the articles that did not meet the purpose of the study, we selected 21 manuscripts. We found and analyzed full-text versions of these manuscripts regarding inclusion and exclusion criteria. Thus, we selected 13 articles that met the inclusion criteria. We extracted the key data including information about the author(s), country, publication date, characteristics of patients (sample size and age), ESP block technique, pain management, results and statistical analysis.

Methodological quality of randomized trials was assessed using the Cochrane guidelines and Review Manager (RevMan) software version 5.4.1 (The Cochrane Collaboration, 2020). According to the above-mentioned criteria, we analyzed randomized controlled trials (RCTs) using a 5-item checklist (RoB 2, Risk-Of-Bias 2): 1) bias following randomization; 2) bias following deviations from the scheduled interventions; 3) bias following no final data; 4) bias in assessment of the result; 5) bias in reporting.

For a non-randomized controlled trial (non-RCT), the checklist consisted of 7 items (ROBINS-I, Risk Of Bias In Non-randomized Studies-1): 1) confounding bias; 2) patient selection bias; 3) bias in classification of impacts; 4) bias following deviations from the scheduled interventions; 5) data skip bias; 6) bias following effect size measurement error; 7) bias in presenting the results [29]. For each study, we analyzed the risk of bias (high, low or indeterminate).

All measures including searching and selection of studies, data extraction and assessment of their quality were performed by two authors. The third one reviewed the data in case of disagreements. The last ones were resolved via consensus.

The primary endpoints were intraoperative and postoperative opioid consumption. The secondary endpoints were severity of postoperative pain (NRS score), period until emergency analgesia, duration of mechanical ventilation, hospital- and ICU-stay, morbidity.

Effect sizes

Statistical significance of effect sized was assessed using the GRADE approach (Grading of Recommendations Assessment, Development and Evaluation) [30]. The quality was assessed as high, moderate, low or very low. The summary table of results was created using the online program GRADEpro GDT.

Statistical analysis

Statistical analysis was performed using the Review Manager (RevMan) software version 5.4.1 (The Cochrane Collaboration, 2020).

Meta-analysis of binary data was performed considering the differences in odds ratio (OR) with 95% CI. Meta-analysis of continuous data was based on between-group mean difference (MD). Meta-analysis of continuous data presented in different units was performed using standardized mean difference (SMD).

The results of meta-analysis were presented in forest plots. Statistical heterogeneity was assessed using the Pearson's chi-square test (χ²) and heterogeneity index I². Meta-analysis included the following models: random effect model (Random, Rnd) in case of significant heterogeneity (I² > 40%); fixed effect model (Fixed) in case of no significant heterogeneity (p≥0.10 in χ² test and I² ≤ 40%). In case of p≥0.10 and I²>40%, we considered χ² test to select mathematical model for meta-analysis.

Results

Searching results and characteristics of studies

Since ESP blockade is a relatively new technique of regional anesthesia, we searched literature data after 2016. Interestingly, all studies between 2018 and 2021 demonstrate ESP blockade as a state-of-the-art regional anesthesia technique. We initially identified 1,143 articles including 875 manuscripts from the PubMed database, 200 articles from the Google Scholar database, 31 articles from the QxMD database and 37 articles from the MEDLINE database. Then, we excluded 59 duplicates. Analysis of the titles and abstracts established 21 relevant articles. Finally, we enrolled 13 manuscripts after assessment of full-text articles regarding inclusion criteria. Flowchart is presented in Fig. 1.

Fig. 1. PRISMA flow chart.

Selected studies have compared ESP blockade with different types of anesthesia. ESP was compared with general anesthesia in 12 studies [12, 21–25, 31–36]. Moll V. et al. [23] and Toscano A. et al. [12] compared ESP block and blockade of the nerves of neurofascial space of serratus anterior muscle. Nagaraja P. et al. [27] compared ESP block and epidural block. In our meta-analysis we assessed effect size in subgroups and overall effect according to study endpoints. ESP was bilateral in 9 studies [21, 23-25, 27, 31, 32, 34, 35]. Borys M. et al. [22] and Sun Y. et al. [36] analyzed unilateral ESP block. In other 2 studies (D'hondt N. et al. [33], Toscano A. et al. l. [12]), we failed to determine the type of block. Overall characteristics of studies are presented in Table 2 [12, 21-25, 27, 31-36].

Table 2. Characteristics of studies [12, 21—25, 27, 31—36]

Note. * — meta-analysis endpoints: ① — intraoperative opioid consumption; ② — postoperative opioid consumption; ③ — severity of pain syndrome (NRS score); ④ — time to emergency analgesia; ⑤ — ventilation time; ⑥ — hospital-stay; ⑦ — ICU-stay; ⑧ — postoperative morbidity; ESP — erector spinae-plane block; SAPB — serratus anterior plane block; EA — epidural anesthesia; CABG — coronary artery bypass grafting; ASD — atrial septal defect; RCT – randomized controlled trial; non-RCT – non-randomized controlled trial, MVR – mitral valve replacement; MIMVS – minimally invasive mitral valve surgery; MIDCAB – minimally invasive direct coronary artery bypass.

Risk of bias

In total, 4 RCTs (Fig. 2) and 9 non-RCTs (Fig. 3) were evaluated according to the Cochrane guidelines. The key areas selected for studies with high risk of bias were confounding, missing data and impact classification error. Studies by P. Macaire et al. [32] and M. Borys et al. [22] had a serious risk of bias that was considered when evaluating the effects and making judgments on the endpoints of the study.

Fig. 2. Risk of bias in randomized controlled trials.

Fig. 3. Risk of bias in non-randomized controlled trials.

Meta-analysis

Meta-analysis of the effect of ESP blockade on intraoperative opioid consumption

There were 8 clinical trials (605 patients including 247 ones who underwent ESP blockade). ESP blockade was compared with general anesthesia in 6 articles [21, 24, 25, 31, 35, 36]. One study analyzed EA compared to ESP [27]. Heterogeneity was high (I²=97%, p<0.001), so random effect model was applied. Such heterogeneity was due to different narcotic drugs for anesthesia and methods of their administration. Comparison of EA and ESP blockade found no differences in intraoperative consumption of narcotic analgesics in cardiac surgery. Pooled results showed lower intraoperative opioid consumption for ESP blockade compared to conventional anesthesia (SMD: -1.50; 95% CI: -2.65 – -0.35; p=0.01) (Fig. 4).

Fig. 4. Meta-analysis of the effect of ESP blockade on intraoperative opioid consumption.

Meta-analysis of the effect of ESP blockade on postoperative opioid consumption

The meta-analysis included 7 studies comparing postoperative opioid consumption [21, 23–25, 31, 35, 36]. A pooled meta-analysis showed significant effect of ESP blockade on postoperative opioid consumption (SMD: -1.85; 95% CI: -2.91 – -0.80; p=0.0006). However, positive effect should be interpreted with caution due to high heterogeneity (I²=97%). Subgroup analyzes revealed lower opioid consumption for ESP blockade compared to general anesthesia (SMD: –2.17; 95% CI: –3.49 – -0.86; p = 0.001). However, there were no differences compared to serratus anterior plane block (Fig. 5).

Fig. 5. Meta-analysis of the effect of ESP blockade on intraoperative opioid consumption.

Meta-analysis of the effect of ESP blockade on severity of pain syndrome

This meta-analysis included 5 studies devoted to pain syndrome throughout a day after general anesthesia and ESP blockade [24, 25, 31–33]. Subgroup analysis was performed in 1, 2, 4, 6, 12 and 24 hours after surgery. There were 3-5 studies per a breakpoint, and heterogeneity ranged (I²) from 0% to 96%. Therefore, random effect model was used (Fig. 6).

Fig. 6. Meta-analysis of the effect of ESP blockade on severity of pain syndrome.

The greatest effect of ESP blockade was observed within 6 postoperative hours. Pain syndrome after ESP blockade was lower in 1, 2, 4 and 6 hours compared to general anesthesia. Between-group differences disappeared after 12 and 24 hours.

A pooled meta-analysis revealed lower pain syndrome after ESP blockade compared to general anesthesia and conventional pain relief (MD: -1.55; 95% CI: -2.00 – -1.09; p < 0.001). Heterogeneity was also too high (I²=100%, p< 0.001) (Fig. 6). A separate analysis of ESP blockade and EA found the effect of ESP blockade after 24 and 48 hours [27]. However, more studies are required for definite conclusions. We decided not to present only 1 study devoted to this comparison.

Meta-analysis of the effect of ESP blockade on the period until emergency analgesia

ESP blockade in addition to conventional anesthesia reduced pain syndrome and prolonged the period until additional emergency analgesia (MD: 274.11; 95% CI: -194.37 – -353.84; p = 0, 0006). This meta-analysis included only 2 studies with a group of 80 patients (40 ones underwent ESP blockade) [24, 31]. Heterogeneity in random effect model was high (I²=92%, p< 0.001) (Fig. 7).

Fig. 7. Meta-analysis of the effect of ESP blockade on the period until emergency analgesia.

Meta-analysis of the effect of ESP blockade on duration of mechanical ventilation

Meta-analysis of the effect of ESP blockade on mechanical ventilation time was based on 11 studies comparing ESP blockade with general anesthesia [21–25, 31–36] and one study devoted to EA [27]. Analysis included 780 patients. The overall effect of reducing the time of mechanical ventilation after ESP blockade added to conventional anesthesia was shown (MD: -0.87; 95% CI: -1.42– -0.33; p=0.002). Subgroup analysis revealed that the main contribution was made by comparing ESP blockade with general anesthesia. Duration of mechanical ventilation was significantly less in the EA group compared to ESP blockade (MD: 1.15; 95% CI: 0.38-1.92; p=0.003). This result should be interpreted with caution due to high heterogeneity of reports (I²=93%) (Fig. 8).

Fig. 8. Meta-analysis of the effect of ESP blockade on duration of mechanical ventilation.

Meta-analysis of the effect of ESP blockade on ICU-stay

A pooled meta-analysis included 10 studies, In 9 articles, the authors compared ESP blockade with general anesthesia [21, 22, 25, 31–36], one trial compared ESP blockade with serratus anterior plane blockade [12] and one study compared ESP and EA [27]. A pooled result showed similar length of ICU-stay (MD = -9.84; 95% CI -21.77 – -2.08; p = 0.11). Heterogeneity of overall effect was high (I²=100%, p<0.001) (Fig. 9). Subgroup analysis revealed differences in ICU-stay between ESP blockade and serratus anterior plane blockade (MD= -8.64; 95% CI -15.13 – -2.15; p=0.009). However, this finding was obtained only in 1 study.

Fig. 9. Meta-analysis of the effect of ESP blockade on ICU-stay.

Meta-analysis of the effect of ESP blockade on hospital-stay

Meta-analysis included 9 studies; 7 ones compared ESP blockade with general anesthesia [12, 21, 23, 33–36], 2 study – ESP blockade with serratus anterior plane blockade [12, 23]. A pooled result showed that addition of ESP blockade did not reduce hospital-stay (MD= -0.27; 95% CI -0.72 – -0.18; p=0.11). Subgroup analysis of ESP blockade and serratus anterior space block regarding hospital-stay revealed the same effect as on the length of ICU-stay (MD= -0.38; 95% CI -0.71 – -0.05; p=0.03), and heterogeneity was low (I²=0%). Heterogeneity of the overall effect was high that indicated possible distortion of overall effect (I²=64%, p= 0.005) (Fig. 10).

Fig. 10. Meta-analysis of the effect of ESP blockade on hospital-stay.

Meta-analysis of the effect of ESP on postoperative morbidity

This meta-analysis included 2 subgroups. Four studies devoted to postoperative nausea and vomiting were selected from the 1^st subgroup [24, 31–33]. The 2^nd subgroup included all other complications. There was only 1 study by Macaire R. et al. [32] in this subgroup. The authors analyzed the incidence of episodes of acute heart failure. Subgroup analysis revealed that ESP blockade reduced the incidence of postoperative nausea and vomiting but did not affect other complications. There were no complications in other four studies. In other articles, the authors did not mention complications. A pooled effect of meta-analysis indicated that additional ESP blockade reduced postoperative morbidity (OR=0.29; 95% CI 0.14-0.60; p=0.0009). Heterogeneity was low in fixed effect model (I²=11%, p=0.34) (Fig. 11).

Fig. 11. Meta-analysis of the effect of ESP blockade on morbidity.

Significant of evidence

Evidence validity was assessed using GRADE approaches for results with significant effects (Table 3). The following outcomes were important: effect of ESP blockade on intraoperative opioid consumption, postoperative opioid consumption, severity of pain syndrome (NRS score), ventilation time and postoperative morbidity. We found very low quality of evidence for the effect of ESP blockade on intraoperative and postoperative opioid consumption, low quality regarding severity of pain and ventilation time, as well as moderate quality for postoperative morbidity. Importantly, evidence quality was reduced due to high heterogeneity of results (I²>50%).

Table 3. Evidence quality (GRADE method)

Note. CI — confidence interval; MD — mean difference; OR — odds ratio; SMD — standardized mean difference.

Discussion

To date, ESP blockade mechanism is still discussed. Schwartzmann A. et al. [26] revealed MR signs of deeper spread of local anesthetics with contrast agent after ESP blockade (into paravertebral region and epidural space). On the contrary, other studies including analysis of cadaveric material did not confirm these results and found no spread to the paravertebral space or ventral spinal roots [37–39].

Nevertheless, ESP blockade has been used in various surgeries for postoperative pain relief [1, 3, 6, 7] since the first description of ESP blockade for thoracic neuropathic pain by Forero M. et al. [17]. To date, this approach is used in cardiac surgery since technique of blockade is simple and incidence of complications is lower compared to EA and paravertebral blockade [6, 8, 15].

We have studied 8 main effects. Pooled data revealed significant benefit of additional ESP blockade for 6 effects (intraoperative and postoperative opioid consumption, NRS score of pain syndrome, time to emergency analgesia, ventilation time, postoperative morbidity). This study demonstrated that ESP blockade effectively controlled acute pain after cardiac surgery, as evidenced by significantly less opioid consumption and pain scores compared to general anesthesia. The most pronounced effect was observed within 12 postoperative hours. On the contrary, ESP blockade was less effective than EA within 12 postoperative hours, but pain syndrome was lower after 24 and 48 hours.

A systematic review with meta-analysis of ESP blockade in thoracic surgery showed that this technique can provide effective analgesia in postoperative period. Considering lower opioid consumption and less pain syndrome, Koo C.H. et al. [6] believe that ESP blockade is not inferior and even superior to traditional general anesthesia. ESP blockade was also not inferior in pain relief compared to paravertebral and serratus anterior plane blockades.

There are particular concerns about the risk of complications associated with EA and paravertebral blockade [40]. Mustafa M.A. et al. [41] reported hematoma as a common complication (13% after paravertebral blockade and 0% after ESP blockade). This is an important aspect due to high risk of hematoma in cardiac surgery because of anticoagulation and antiplatelet therapy. Moreover, success rate of anesthesia was higher for ESP blockade (100%) compared to paravertebral blockade (77.8%) that indicates technical simplicity of ESP blockade. In our study, we found no data on hematoma in any report. The most common event was postoperative nausea and vomiting with higher incidence after general anesthesia.

Conclusion

1. Additional ESP blockade can reduce intraoperative and postoperative opioid consumption, as well as severity of pain syndrome in early postoperative period. We recognized these significant effects as important despite low evidence quality.

2. ESP blockade reduces ventilation time compared to general anesthesia (MD: -0.87; 95% CI: -1.42 – -0.33; p=0.002).

3. ESP blockade does not affect the length of ICU- and hospital-stay.

4. ESP blockade reduces the incidence of postoperative nausea and vomiting (OR=0.17; 95% CI 0.07-0.45; p=0.0003), but does not affect other complications.

Author contribution:

Concept and design of the study – Shchegolev A.V., Lakhin R.E.

Collection and analysis of data – Shapovalov P.A., Uvarov D.N.

Statistical analysis – Stukalov A.V., Shapovalov P.A.

Writing the text – Shapovalov P.A., Tsvetkov V.G.

Editing – Shchegolev A.V., Lakhin R.E.

The authors declare no conflicts of interest.