The Effect of Preoperative Ultrasound-Guided Erector Spinae Plane Block on Chronic Postsurgical Pain After Breast Cancer Surgery: A Propensity Score-Matched Cohort Study

Being the most frequently diagnosed type of female malignancy, breast cancer accounts for a large number of cancer-related deaths worldwide [1]. Besides, patients that have undergone breast cancer surgery (BCS) are often left with chronic postsurgical pain (CPSP) with a prevalence up to 78% of the survivors [2]. CPSP may gravely impair quality of life and become a heavy burden to BCS patients [3, 4]. The mechanism of CPSP remains unclear but several risk factors including younger age, chemotherapy and radiotherapy, axillary lymph node dissection (ALND), preoperative anxiety and depression, as well as preexisting pain conditions or poorly controlled postoperative pain have been identified in epidemiologic studies [5, 6].

Several regional anesthesia (RA) techniques including paravertebral block (PVB), serratus anterior plane block (SAPB), pectoral nerves (PECS) block, and the more recent erector spinae plane block (ESPB) have been applied for acute pain control after BCS [7‐10]. However, the influence of RA techniques on CPSP has been sparsely investigated and remains controversial. Findings of small studies suggested that PVB might have a preventive effect on the development of CPSP [11, 12]. Unfortunately, in a recently published large randomized controlled trial, no significant impact of PVB on the incidence of CPSP and the rate of neuropathic pain (NP) had been observed [13]. ESPB is a comparatively novel interfascial plane block that targets away from the pleura, aiming to inject local anesthetics between erector spinae muscle and the thoracic transverse process [14]. It has been used as a postoperative analgesia technique in variety of surgical disciplines and shown to provide similar analgesic effects compared to PVB [15, 16]. However, studies evaluating the effectiveness of preoperative ESPB on CPSP after BCS are still scarce.

The practice of preoperative ultrasound-guided ESPB was introduced in 2019 in our center, therefore we conducted this propensity score-matched cohort study to assess its clinical effects. We hypothesized that there would be differences in incidence of CPSP between patients who receive preoperative ultrasound-guided single-shot ESPB combined with general anesthesia (GA) and those who receive GA alone at 1 year following BCS.

A total of 582 patients underwent BCS from October 2019 to June 2020 were investigated for eligibility, of which 361 patients (ESPB + GA group, n = 102; GA group, n = 259) were included in the analysis (Fig. 1).

The demographic characteristics and perioperative data for the total study cohort are presented in Table 1. The sensory block upper level of ESPB was at T2 (T1–T3) and lower level at T6 (T6–T8), whereas the median number of dermatomes blocked was 3 (2–4). Significant statistical differences were detected for age and BMI between the groups before matching. After PSM, 97 patients remained in each group. A good matching balance was achieved with all SMDs of the study variables were less than 10% in the matched cohort (Fig. 2).

The overall incidence of CPSP at 1 year following BCS before matching was 35.4% (128 of 361), with 38.3% (49 of 128) satisfying the ID pain criteria for the diagnosis of NP. CPSP was reported by 74.2% (95 of 128) of the patients to be located in the surgical breast area, whereas 43.8% (56 of 128) in the axilla, 33.6% (43 of 128) in the ipsilateral arm, and 22.7% (29 of 128) in the ipsilateral thorax. The characteristics of pain were described as needle pain (67.5%), burning pain (33.2%), shooting pain (28.5%), and numbness (9.3%). Moderate-to-severe CPSP was reported by 11.7% (15 of 128).

After PSM, the incidence of CPSP were similar (P = 1.000) between the groups (Table 2). There was no difference in the patients who had reported an ID pain score indicative of NP among the matched groups (P = 0.442). The severity of CPSP (P = 0.547) and interference of pain on daily functions (P = 0.376) did not differ between the study groups, as measured by subscale total scores on BPI-SF, respectively. There was no statistical difference (P = 1.000) in the rate of moderate-to-severe CPSP between the two groups (Table 2).

The highest NRS scores of acute postoperative pain within 24 h were significantly reduced in the ESPB + GA group comparing to the GA group (P = 0.043). As shown in Table 2, postoperative rescue analgesia demand within 24 h was statically lower in the ESPB + GA group than in the GA group (P = 0.042). The incidence of PONV was significantly lower in the ESPB + GA group compared to the GA group (P = 0.031). There was no statistical difference (P = 0.244) in the hospital LOS between the two groups (Table 2). No clinical evidence of pneumothorax, hematoma, or other serious complications associated with ESPB were observed.

Considering that data of 167 patients were missing during the matching, and the incidence of CPSP was similar between the matched groups, we performed an additional logistic regression analysis on overall patients. The potential risk factors for the development of CPSP were compared in Table 3. Three significant risk factors were identified by multivariate logistic regression: ALND (OR 3.541, 95% CI: 1.273–9.851, P = 0.015), radiotherapy (OR 1.918, 95% CI: 1.067–3.448, P = 0.029) and acute postoperative pain within 24 h (OR 2.109, 95% CI: 1.097–4.056, P = 0.036), as shown in Table 4.

In this study, we found no significant association between preoperative single-shot ESPB and a lower incidence of CPSP at 1 year after BCS. Proportion of NP, as well as severity and interference of CPSP, were also similar for patients receiving ESPB combined with GA and GA alone.

The incidence of CPSP after BCS varied from 25 to 78% in previous studies [2]. This may partially be attributed to different definitions of chronic pain. Our study showed that the overall incidence of CPSP was 35.4% at 1 year after BCS, which was within the range of previously reported incidences. Although the mechanism of CPSP remains unclear, evidence has suggested that peripheral and central sensitization may play an important role in the development of chronic pain after surgery [17]. Theoretically, RA could alleviate acute postoperative pain by blocking peripheral nociceptive input into the spinal cord, which contributed to spinal sensitization. Previous studies have reported that PVB and PECS could reduce incidence of chronic pain after breast surgery [22‐24]. In a recent meta-analysis, the use of single-shot PVB has been shown to be protective against CPSP at 6 months following BCS, but the studies included are weakened by methodological limitations [12]. However, our study found that the effect of a single-shot ESPB might be limited on prevention of CPSP at 1 year after BCS. The contrasting results may be partially explained by different RA techniques used in our study and the previous ones. In fact, differences in analgesic effects between the block techniques for BCS patients have been found in some studies [25‐27], which indicates that ESPB and PVB may have different mechanisms of action [28, 29]. The most possible mechanism of analgesia for ESPB is a direct effect of local anesthetics via diffusion to the erector spinae muscle within the costotransverse foramen region with possible spread to the paravertebral space which is the main target of PVB [30]. However, the extent of craniocaudal spread to the adjacent tissue compartments and the sensory blockade of ESPB varied greatly. Our study showed a median number of anesthetized dermatomes of 3. which was comparatively smaller than other studies that reported significant local anesthetic distribution with a single ESPB injection [14].

The high risk of NP after BCS might be another possible explanation for our findings on the primary outcome. In agreement to previous studies, we found that 38.3% of the patients who developed chronic pain after BCS described a neuropathic component [31, 32]. Moreover, our study also confirmed that a substantial proportion of patients reported CPSP as moderate to severe in nature, which is in accordance with previous study [2]. Due to the cancer-related multimodal treatment, chronic pain after BCS is often described as a mixed pain, which consists predominantly of a neuropathic nature and partially related to intraoperative nerve injury [33, 34]. Although RA may prevent chronic pain after certain types of surgeries with a low risk of NP, the situation may be different for surgeries with a high risk of NP, such as BCS [23].

Although no association of ESPB and reduced incidence of CPSP was found in our study, we still confirmed that ESPB markedly improve postoperative analgesia and reduce rescue analgesics after BCS, which is in line with the existing literature [25, 35]. In addition, the rate of PONV and postoperative rescue analgesia were significantly differed between the two study groups, which may partly demonstrate the opioid-sparing effect of ESPB and its potential to facilitate recovery. PVB has been demonstrated to be one of the most effective RA techniques for acute postoperative pain management. However, it is also challenging to perform the block proximate to the pleura and central neuraxial system. ESPB is a relatively safer method which target away from the pleura. In our study, no episodes of pneumothorax, epidural administration and other serious complications that related to nerve block were observed, suggesting the safety of ESPB under ultrasound guidance with reduced risks associated with conventional thoracic epidural analgesia (TEA) and PVB.

The etiology of CPSP is multifactorial. Consistent with previous studies exploring predictors of chronic pain after BCS, our study suggests that ALND, radiotherapy, and acute postoperative pain appear to be the risk factors of CPSP [6, 36]. Thus, multimodal approaches including both surgical and anesthetic techniques should be utilized for the purpose of minimizing the occurrence of CPSP. The effects of perioperative pregabalin, ketamine therapy, and lidocaine infusion on prevention of CPSP after BCS have also been investigated in some randomized trials, which showed promising results [37‐39]. Therefore, further studies based on multimodal strategies will be needed, considering the high rates of CPSP and the high proportion of NP after BCS.

There are several limitations to our study. First, due to the non-randomized design and monocentric nature of this study, the results only indicate association between the covariates of interest and may not be representative in larger population. Second, although PSM method could help lower the risk of confounding bias and selection bias, we could only account for confounders of which relevant data were documented. For instance, psychological status as well as pain threshold should be taken into consideration. Third, different techniques of ESPB may change the effects of the block. That said, the number of injection levels, type, volume, and concentration of local anesthetics might affect the sensory blockade and analgesic effects. Fourth, we did not compare the effects of ESPB with other blocks on CPSP after BCS. Fifth, due to the complex sensory innervation of the breast and nipple area, nerves other than upper thoracic spinal nerves may be involved, making a single block with ESPB less effective. Conversely, the main strength of this study was being the first study to evaluate effects of ESPB on chronic pain at 1 year in patients undergoing BCS.

In conclusion, our results revealed that a preoperative single shot of ESPB combined with GA was not associated with lower incidence of CPSP at 1 year compared with GA alone after BCS. ALND, radiotherapy, and acute postoperative pain within 24 h were independent risk factors for the development of CPSP after BCS. However, the limitations of the study do not permit us to be conclusive on this argument. Further well-conducted randomized controlled studies should be needed to confirm the results.