**2. Techniques**

RA of the chest wall may be performed at various points along an arch coursing anteriorly from the posterior midline. With TEA as gold standard regarding the breadth of somatic and sympathetic blockade, CWFPB exhibit a variable decrement in their effect as they approach the anterior midline. Autonomic effects are retained proportionally to the extent of LA spread into the epidural space, and the area of sensory loss is inversely related to the distance between the injection spot and spine. A considerable inter-individual variation in the extent and intensity of CWFPB exists, and several reasons may represent the root cause of this: (1) existence of differential sensory blockade [13]; (2) reliance on passive LA spread to achieve analgesia; (3) redundant innervation between peripheral nerve territories, including midline overlapping [14,15].

#### *2.1. Thoracic Epidural Anaesthesia (TEA)*

The role of TEA in cardiac ERAS programs remains an intensely debated topic. TEA produces robust chest wall pain relief, ye<sup>t</sup> it repeatedly failed to improve perioperative morbidity and mortality in CS populations [16]. Potential reasons include the fact that TEA benefits may have a disproportionate impact on CS pain or because TEA side effects and complications may offset its benefits. Notably, pain associated with CS is typically moderate [6], so less intense analgesia (i.e., CWFPB) might suffice. In contrast, adverse events associated with TEA may be clinically relevant (e.g., respiratory depression with epidural opioids and hypotension with epidural LA) and potentially catastrophic (e.g., SEH).

Cardiac sympatholysis was shown to benefit myocardial blood flow [17] but also blunt the heart capacity to cope with hemodynamic challenges, especially within specific subgroups such as those with established pulmonary hypertension [18].

The calculated maximum risks of SEH in CS after TEA were 1:1500 with 95% confidence and 1:1000 with 99% confidence, respectively [19]. In a recent meta-analysis of over 6000 patients, Landoni et al. estimated this risk at 1:3552 (95% CI 1:2552–1:5841) [20]. Placing the epidural one day before surgery could prevent bleeding complications, but such practice patterns would contradict the very essence of ERAS programs.

Overall, minimization of risks outweighs maximization of analgetic potential. Adequate patient selection, risk factors, and anesthesiologist's expertise must be carefully balanced before pursuing TEA or any other type of neuraxial technique. Until more evidence becomes available, the risk-benefit ratio of neuraxial analgesia remains prohibitive.

#### *2.2. Paravertebral Blocks (PVB)*

#### 2.2.1. Mechanism and Clinical Applications

PVB involves LA injection into the thoracic paravertebral space (TPVS) to block the spinal nerve roots as they emerge from the intervertebral foramina. TPVS is a triangularshaped space on both sides of the vertebrae bounded anterolaterally by the parietal pleura, medially by the posterolateral parts of the vertebral body and posteriorly by the superior costotransverse ligament (SCTL) (see also Figure 1). TPVS communicates laterally with the intercostal space, and medially with the epidural space. TPVS is also contiguous with its contralateral counterpart but to a much lesser extent whereas its cranial extension remains ill-defined. Caudal and rostral segmental spread of the LA drug from the injection site generates multilevel ipsilateral somatic and autonomic blockade, with epidural and intercostal LA dispersions likely contributing substantially to analgesia [21]. The clinical effect of single-level PVBLM is highly variable because the LA spread is unpredictable [22]. Consequently, a multiple-injection technique was commonly considered superior to single-injection patterns [23,24]. This theory was first challenged by Renes et al. [25] and Marhofer et al. [26] who used US-guidance to perform PVB (PVBUS). Later, Uppal et al. [27] demonstrated that single- and multilevel PVBUS are equivalent regarding coverage and pain relief duration. Conveniently, the single-level PVBUS are markedly faster and better tolerated by patients, two prerequisites of any ERAS strategy.

Compared to PVBLM, PVBUS are more reliable and safer [28]. Two assets, equivalent analgesia to TEA but with fewer complications [29–34] and unilateral sympathectomy, favored the resurgence of PVBUS. Still, the latter proves itself ineffectual in CS with sternotomy since this surgery requires bilateral nerve blockade.

**Figure 1.** *Cont*.

**Figure 1.** (**A**) Parasagittal scan of thoracic paravertebral space (TPVS); (**B**) Transverse/oblique scan of TPVS after 75-degree anti-clockwise rotation from A. The needle tip's target is TPVS, which, after probe rotation, appears enlarged and lies anteriorly to superior costotransverse ligament (SCTL)/IIMb (see text).TM, trapezius muscle; RM, rhomboid muscle; ESM, erector spinae muscle; EIM, external intercostal muscle; IIM, internal intercostal muscle; TP, transverse vertebral process; SCTL, superior costotransverse ligament; IIMb, internal intercostal membrane; TPVS, thoracic paravertebral space.

As with TEA, hemorrhagic complications represent a crucial factor to consider. In contrast to TEA, risk quantification of SEH after PVB is less evident and intensely debated. The latest American Society of Regional Anesthesia and Pain Medicine (ASRA) Practice Advisory on RA and anticoagulation maintains the same recommendations for PVB as for any other neuraxial block [11]. Equivocally, ASRA does not differentiate between PVBLM and PVBUS, and between single-shot PVB and PVB with catheters. New data suggests that US guidance during paravertebral blockade could virtually abrogate spinal injury risk even with the large heparin dosing needed in cardiopulmonary bypass (CPB) [35]. El Shora et al. recently compared PVBUS with catheter to TEA to manage pain after on-pump CS [36]. Catheters were placed immediately after induction in both study groups, and LA infusion was started only postoperatively. PVBUS were non-inferior to TEA regarding pain relief, and bleeding complications were not reported in either group.

Future studies will have to address two aspects to maximize the benefits and minimize the potential risks associated with PVBUS. The first is concerned with single-shot PVB being safer than PVB with TPVS catheters because catheter misplacement, including epidurally, is still possible even with US [37,38]. The second aspect is concerned with nerve blockade timing, as suggested by Richardson et al. [39]. Compared to PVBUS established after surgery, preemptive PVBUS may be better tailored to fast-tracking as it would also mitigate the intraoperative opioid consumption.

The best strategy to implement PVBUS has ye<sup>t</sup> to be established. Further research is needed before the routine use of paravertebral blockade in CS is either supported or refuted.

Sonoanatomy and Block Techniques (Figure 1)

PVBUS have superseded PVBLM in every aspect. A comprehensive review described at least nine approaches, all of which share the same three sonoanatomical landmarks circumscribing TPVS—rib, pleura, and transverse process (TP) [40]. At present, formal recommendations on the best way to perform PVBUS do not exist. Instead, personal factors relating to skill, experience and perceived safety seem to play a decisive role. An objective comparative evaluation of currently used PVBUS techniques is essential to enable an informed PVB-based MA.

TPVS scanning breaks down to 4 elements: (1) plane of US beam orientation (i.e., transversal versus sagittal); (2) needling technique (i.e., out-of-plane versus in-plane); (3) direction of angulation (i.e., lateral versus medial, and caudal versus cranial respectively) and (4) safety limit for needle tip (i.e., anteriorly or posteriorly to SCTL) [40]. Choosing between these elements entails a trade-off between two goals, simplicity and accuracy. The latter is advocated in our institute, so we perform an in-plane, lateral to medial, transversal/oblique approach with a safety limit set anteriorly to SCTL (see Figure 1). Based on currently available evidence, catheters are excluded with CPB heparin dosing.

Scanning starts with the linear-array transducer placed in a parasagittal plane to identify the adjacent TP, recognizable as flat, rectangular hypoechoic structures (see also Figure 2). Anti-clockwise rotation to a transversal/oblique plane displays the TPVS. The needle is inserted in-plane, latero-medially and advanced until it reaches the wedgedshaped TPVS. Adequate LA injection pushes the parietal pleura anteriorly. Preemptive bilateral single-shot blocks are performed at the level of the fourth thoracic vertebrae. This alone may provide intraoperative analgesia long enough to sustain most types of CS.

#### *2.3. Chest Wall Fascial Plane Blocks (CWFPB)*

2.3.1. Posterior CWFPB-Erector Spinae Plane Block (ESPB) and other PVB Variants Mechanism and Clinical Applications

Post-mortem data challenge the traditional view that TPVS is a discrete anatomical space and sugges<sup>t</sup> that the SCTL is permeable to LA drugs [41]. Hence, paravertebral blockade of nerves could still be elicited by placing the needle tip outside but close enough to the TPVS.

US guidance has facilitated the emergence of several more superficial needle placement techniques, all collectively labelled as "paraspinal blocks" [42] or "PVB by proxy" [43]. These include the retrolaminar block (RLB) [44,45], midpoint transverse process to pleura block (MTPB) [41], intercostal/paraspinal block (ICPB) [46], rhomboid intercostal and subserratus block (RISS) [47], and erector spinae plane block (ESPB) [48]. Depending on their underlying pathway of LA spread, these novel blocks produce a variable combination of ipsilateral somatic and autonomic blockades, the extent of which remains open for further research. Amongst them, ESPB is the most well characterized to date.

The ESPB target for LA deposition is the plane between the erector spinae muscle (ESM) and the thoracic TP tip. Correct single-level LA injection should lift the ESM off the TP and allow the ipsilateral craniocaudal volume-dependent [49] LA spread across several contiguous dermatomes (i.e., 3 to 7 intercostal spaces) [50]. As with PVB, transforaminal, intercostal and circumferential epidural diffusions likely contribute to its mechanism of action [50,51].

Krishna et al. compared bilateral single-shot ESPB with control (i.e., general anesthesia alone) in CS with sternotomy and found reduced postoperative pain, time to extubation, time to ambulation, opioid usage and total length of intensive care unit (ICU) stay [52]. Interestingly, rescue analgesia was reported in the intervention group only ten hours after extubation compared to six hours in the control group (*p* = 0.0001). Macaire et al. used a before-and-after design to show that in open CS a preemptive strategy with bilateral ESPB catheters is associated with reduced intra- and postoperative opioid consumption. Consequently, several ERAS endpoints were favorably altered, including postoperative adverse events (hypotension, nausea/vomiting and hyperglycemia) and times to chest tube removal and first mobilization. The authors found no differences in extubation time and pain during the first mobilization. Another RCT showed comparable postoperative pain scores between bilateral continuous ESPB and TEA in 50 patients undergoing open CS [53]. Finally, Bousquet at al. endorse the association of bilateral parasternal block with bilateral ESPB [54] given that ESPB alone may sometimes fail to provide adequate parasternal analgesia [55]. This dual blockade significantly reduced the intraoperative sufentanil and postoperative morphine usage in a 20-patient cohort [54]. These four studies did not report any RA related adverse effects, but then again, neither was appropriately powered to detect them.

 **Figure 2.** (**A**) Parasagittal scan—rib level; (**B**) Parasagittal scan—TP level (see text). TM, trapezius muscle; RM, rhomboid muscle; ESM, erector spinae muscle; TP, transverse vertebral process; ESM-to-TP, erector spinae muscle -to-transversus process plane.

Although promising, results from these clinical studies are not generalizable. There is a potential bias concerning the small patient populations, blinding and randomization. Further studies are mandated to fully understand the benefits and extent of incorporating ESPB into routine clinical practice.

#### Sonoanatomy and Block Tachnique (Figure 2)

Scanning starts with the linear-array transducer set 5–6 cm away from the dorsal midline in a parasagittal orientation. The ribs are then displayed as rounded acoustic shadows with an interceding hyperechoic pleural line (see Figure 2A). Sliding the transducer medially along the short axis allows visualization of the TP as flat, squared-off acoustic shadows (see Figure 2B). Additionally, the pleural line is more in-depth and ill-defined. A too medial position identifies the thoracic laminae as a continuous flat hyperechoic line with regularly interspersed notches representing the facet joint interfaces. Needle insertion follows an in-plane approach, either craniocaudal or vice versa, to contact the ESM-to-TP plane. Real-time imaging guarantees correct LA hydro-dissection beneath the ESM and catheter placement whenever continuous pain relief is warranted. Single-level injection ESPBs (i.e., at the 5th thoracic vertebrae), as initially described by Forero et al. [48], continue to be the norm but this view has recently been challenged by Tulgar et al. who propose a bilevel approach to ensure a more homogeneous LA spread [56].

#### 2.3.2. Anterolateral CWFPB—Pectoral Blocks and Serratus Plane Block Mechanism and Clinical Application

Anterolateral CWFPB provide ipsilateral somatic anesthesia of the upper anterolateral hemithorax but may spare the anterior branches of the intercostal nerves and hence do not consistently provide anesthetic coverage to the ipsilateral parasternal region [12]. This theoretically hinders their use in CS with sternotomy. Established techniques include the serratus anterior plane block (SAPB) [57] and the pectoralis block type I (PECS I) [58] and II (PECS II) [59]. Whilst PECS I and SAPB are distinct blocks, targeting two separate musculofascial planes, PECS II merely represents an attempt to achieve both PECS I and SAPB during a two-staged single needle pass (see also Figure 3).

A 40-patient RCT compared PECS II with no block as part of a postoperative MA strategy in patients undergoing CS with sternotomy. PECS group patients were extubated earlier, had lower pain scores and fewer episodes of rescue analgesia [60].

SAPB was studied in minimally invasive heart valve surgery (MIHVS) with right thoracotomy and minimally invasive direct coronary artery bypass (MIDCAB) with left thoracotomy. Berthoud et al. compared postoperative single-shot deep SAPB to continuous wound infiltration (CWI) and reported significantly lower morphine consumption, reduced length of ICU stay and improved pain control during the first 48 h following MIHVS [61]. Another group of authors compared pre-incisional single-shot and postoperative catheterbased deep SABP against parenteral morphine [62]. The intraoperative opioid usage remained unaffected, but the combined regional nerve blockade significantly spared the postoperative morphine consumption. Nevertheless, this did not change the postoperative course, that is, ICU and hospital lengths of stay and ventilator-free days. According to one study, SAPB appears well suited for MIDCAB thoracotomies [63] but remains inferior to PVB in terms of analgesic coverage and intensity [64]. Lastly, SAPB and PECS II showed an equivalent analgesic effect in an RCT conducted on pediatric patients undergoing CS with thoracotomy without CPB [65].

Anterolateral CWFPB have an excellent safety profile that will allow their ongoing integration in cardiac ERAS pathways. Their impact relies markedly on adequate timing (i.e., pre- versus postoperative blockade) and indication.

**Figure 3.** (**A**) Parasagittal scan along the medioclavicular line-2nd rib level; (**B**) Oblique scan after a slight medial tilt with inferolateral sliding towards the midaxillary line-4th rib level (see text). PMAJOR, pectoralis major muscle; PMINOR, pectoralis minor muscle; AxA, axillary artery; AxV, axillary vein; red arrows, thoracoacromial artery and vein; SAM, serratus anterior muscle; IM, intercostal muscle; TTM, transversus thoracic muscle; P1, PECS I plane; P2, superficial plane for SAPB/PECS II; P3, deep plane for SAPB/PECS II. To elicit an adequate SAPB coverage, P2 or P3 need to be targeted at the 4th or 5th rib level.

Sonoanatomy and Block Technique (Figure 3)

PECS I targets the lateral (C5–C7) and medial (C8–T1) pectoral nerves travelling within the fascial plane between the pectoralis minor and major muscles. SAPB targets the plane either above or below the serratus anterior muscle (SAM). Although some authors favor the latter [66], the differences between these two juxtaposed fascial planes have not ye<sup>t</sup> been elucidated. SAPB blocks the lateral cutaneous branches of the intercostal branches and, when superficially performed, the long thoracic (C5–C7) and thoracodorsal nerves (C6–C8). A single needle pass may secure both blocks (i.e., PECS II) and achieve ipsilateral anesthesia of the anterolateral hemithorax and axilla. Scanning is carried out craniocaudally along the midclavicular line, sliding laterally to intersect the midaxillary line at the fourth and fifth ribs level. Needle insertion follows an in-plane, mediolateral approach (see Figure 3).

#### 2.3.3. Anteromedial CWFPB—Parasternal Block Variants

#### Mechanism and Clinical Applications

These blocks complement the anterolateral CWFPB by providing anesthesia confined to the parasternal region [67]. Depending on where the anterior branches of the intercostal nerves are blocked, anteromedial CWFPB consist of two interrelated approaches: the pecto-intercostal fascial plane block (PIFB) [68] and transverse thoracic muscle plane block (TTMPB) [69]. The former is the injection of LA between the external intercostal and pectoralis major muscles. The latter targets a deeper fascial layer between the inner intercostal and transverse thoracic muscles. Some authors promote PIFB because of a potentially superior safety profile [70,71] and others inform that the transverse thoracic muscles may be too thin to identify with US [72].

Both parasternal variants have been evaluated in CS with sternotomy. Two small RCTs looked at bilateral single-shot PIFB as part of a postoperative MA regimen. Adverse effects were not recorded, and pain scores were significantly reduced in both trials [73,74]. There was a trend towards reduced cumulative opioid consumption, but this reached statistical significance in only one trial [73]. Anecdotal evidence supports the combination of PIFB with other fascial plane blocks as clinically required [75]. Furthermore, such an approach may be readily generalizable to all CWFPB and lend itself to an individualized USRA.

Preemptive single-shot bilateral TTMPB was compared with placebo in an RCT of 48 adult patients undergoing CS with median sternotomy. Several ERAS-specific outcomes were significantly improved, including first 24 h opioid requirement, rescue analgesia, pain scores, and ICU discharge time [76]. Similar findings have been reported by several pediatric RCTs in CS via midline sternotomy [77,78], with one trial using a combination of TTMPB with rectus sheath block [79].

#### Sonoanatomy and Block Technique (Figure 4)

The linear-array probe is placed in the parasagittal plane, 1 cm lateral from sternum's edge in the fourth or fifth intercostal space (see Figure 4). Structures to be identified include the pectoralis major muscle, intercostal muscle, thoracic transversus muscles and rib shadows with the intervening pleural line. The internal thoracic artery and vein run longitudinally and share the same plane with TTMPB (i.e., superficial to the thoracic transversus muscle). Perforating branches may cross the intercostal muscles to reach the sternum. Careful scanning in two orthogonal planes is thus mandated before needle insertion to avoid inadvertent vascular puncture. To this end, some authors recommend a transversal approach with lateral to medial needle advancement [72]. Regardless of probe orientation, one or both target planes can then be selected to deposit LA using an in-plane approach.

**Figure 4.** (**A**) Sagittal parasternal scan; (**B**) Sagittal parasternal scan with markings. Note that TTM appears as a hypoechoic band folding over the hyperechoic pleura. PMAJOR, pectoralis major muscle; EIM, external intercostal muscle; IIM, internal intercostal muscle; INNIM, innermost intercostal muscle; TTM, thoracic transversus muscle; P1, target plane for PIFB; P2, target plane for TTMPB.
