Intermittent Bolus Versus Continuous Paravertebral Infusion on Postoperative Pain after Unilateral Mastectomy: A Retrospective Cohort Study.

Background: Persistent pain is a common complication after mastectomy for breast cancer, and postoperative analgesia through paravertebral infusion is a promising technique for pain control after mastectomy. Although paravertebral local anesthetic infusion can be performed using continuous or intermittent bolus modes, there is no evidence of superiority of either infusion mode. Methods: We retrospectively studied patients who had undergone unilateral mastectomy with paravertebral catheter placed for postoperative analgesia between October 2017 and September 2018. Logistic regression was created to estimate the relative risk for moderate to severe pain dened by average pain scores greater than 4 using a 0 to 10 Numeric Rating Scale (NRS) within rst 24 hours after surgery. Opioid consumption was compared using a Student T-test for difference in means using the log-transformed cumulative opioid consumption converted to milligrams of oral morphine equivalents (OME) within the rst 48 hours after surgery. Results: 49 patients were included (continuous infusion group: 14; bolus infusion group: 35). Bolus paravertebral infusion was associated with reduction of relative risk of moderate to intense pain (adjusted relative risk: -0.96; 95%CI: -1.81, -0.1; p = 0.029) but there was no association with reduction in opioid consumption (difference of means: -1.4mg of OME; 95%CI: -2.78, 1.42; p = 0.33). Conclusion: Intermittent bolus infusion was associated with reduction of relative risk for moderate to severe postoperative pain for unilateral mastectomies. Therefore, paravertebral catheter infusion modality may inuence postoperative pain experience after unilateral mastectomies.


Background
Breast cancer is the leading cause of cancer-related death in women worldwide 1 . In the United States, breast cancer is the leading form of cancer in women, with a lifetime prevalence of 12%, and represents 30% of all newly diagnosed cases 2 . Mastectomy is an essential part of the treatment for the majority of non-metastatic breast cancer cases and the most common complication after mastectomy is persistent pain with a reported incidence greater than 50% in some studies 3,4 . The principal risk factors associated with persistent pain are younger patients, greater acute postoperative pain, radiotherapy, and axillary lymph node dissection 5 .
Current evidence shows that regional anesthesia techniques, such as paravertebral block and thoracic epidural anesthesia, provide better pain control after mastectomy with subsequent reduction in opioid consumption and fewer opioid-related side effects 6,7 . The paravertebral space allows catheter placement which can prolong analgesia and present a safer option since they are not related to neuraxial hematoma, and infusions are less likely to impact the autonomous nervous system which may cause hypotension. Some studies already concluded that paravertebral blocks are effective for pain after mastectomy 6,8 .
Results from trials in obstetric analgesia demonstrated decreased chances of breakthrough pain in patients receiving intermittent bolus infusion through epidural catheters 9 . However, only few studies have compared infusion modality for paravertebral catheters and current evidence is insu cient to endorse either modality 10,11 . Thus, we tested the primary hypothesis that bolus paravertebral infusion reduces the relative risk of moderate to intense pain de ned as average pain scores greater than 4 using a 0 to 10 Numeric Rating Scale (NRS) in patients undergoing unilateral mastectomy when compared to continuous paravertebral infusion measured by average pain scores. Secondarily, we tested the hypothesis that better analgesia will also reduce opioid consumption.

Methods
With approval from Institutional Review Board (IRB #18-1108), we performed a historical cohort analysis using data extracted from the electronic medical record system at a regional hospital from Cleveland Clinic network. This analysis adheres to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines.
Study population: we included adult female patients who had undergone unilateral mastectomy with ipsilateral paravertebral catheter placement for postoperative pain management before surgery between October 2017 to September 2018. We excluded patients who had bilateral mastectomy, those with paravertebral catheter placement after surgery, and those with evidence of metastatic cancer.

Measurements
We recorded patient baseline characteristics, including age, American Society of Anesthesiologists (ASA) physical status, body mass index, and documented chronic pain before surgery. We also assessed surgical characteristics such as lymph node dissection and pathological strati cation (in situ and early stage or locally advanced).
Moderate to severe pain was de ned as patients who experienced average pain score greater than 4 using a 0 to 10 Numeric Rating Scale (NRS), with 0 de ned as no pain and 10 being the worst possible pain. We calculated the average pain scores by diving the worst score available every four hours divided by the number of observations within 24 hours after surgery. Per hospital policy, pain was assessed at 15 minutes intervals for the rst hour after surgery and then at four-hour intervals.
Opioid consumption was de ned as the total amount of opioid given in the post-anesthesia care unit (PACU) and nursing oor within the rst 48 hours after surgery or until discharge, whichever happened rst. All opioids administered were converted into milligrams of oral morphine equivalents 12 and then divided by the number of days patient stayed in the hospital to standardize daily opioid consumption.

Statistical methods
To assess the effectiveness of bolus versus continuous paravertebral infusion for pain management after mastectomy, we calculated the relative risk for moderate to intense pain using logistic regression to predict average pain scores greater than 4 in a 0 to 10 NRS. Since opioid consumption is usually skewed a log transformation was performed to achieve a normal distribution. The signi cance level was 0.05 for primary analysis. Statistical Analysis System (SAS) software version 9.4 (SAS Institute, Cary, NC) was used for all analyses.

Sample size calculation
Data from a previous publication shows that up to 60% of patients experience severe postoperative pain after mastectomy 13 . Then we estimated that our study required at least 34 patients receiving bolus infusion and 14 patients receiving continuous infusion to achieve 80% power to test a relative risk of 1.2 or higher.

Results
From 65 patients who underwent mastectomies at Fairview hospital during the study period: 16 patients were excluded (the majority due to bilateral mastectomy) and 49 were analyzed. Among these, 71% (n = 35) received bolus infusion and 29% (n = 14) received continuous infusion in the paravertebral catheter. Demographic, baseline, and surgical characteristics are reported in Table 1.  Abbreviations: NRS, numeric rating system; CI, con dence interval; OME, oral morphine equivalents a: Moderate to intense pain was de ned as an average pain score greater than 4 and considered as dichotomous outcome for relative risk estimation. Adjustment was done by a multivariable log-linear regression b: Average and difference in opioid consumption was obtained by the exponentiation of the normally distributed log-transformed data c: Difference in means using Student T-test for unadjusted data and least-square means for adjusted data

Discussion
We compared two thoracic paravertebral catheter infusion modalities (continuous versus intermittent bolus) for postoperative pain management after unilateral mastectomies. We found that the bolus paravertebral infusion was associated with lower relative risk of moderate to severe postoperative pain when compared to continuous infusion (adjusted relative risk: -0.96; 95%CI: -1.81, -0.10; p = 0.029). However, bolus infusion was not associated with decreased inpatient opioid consumption (difference in means: -1.4mg of OME; 95%CI: -2.78, 1.42; p = 0.33).
Our results regarding reduction of risk of severe postoperative pain are congruent with the ndings of previously published studies. One trial showed the bene cial effects of paravertebral bolus infusion concerning pain scores, although the effect itself was not clinically relevant, and reduced use of local anesthetic rescue boluses in comparison with continuous paravertebral infusion 14 . A possible explanation for this clinical result is the high pressure achieved by modern pumps during bolus infusions 15 , which allow better spread of the infusion solution. This bene t has already been proven in one cadaveric study studying dye dispersion in the paravertebral space with bolus infusion 16 . Another evidence of this effect was established in the study by Hida et al., who found a wider dermatomal spread of the paravertebral block with paravertebral bolus infusion after video-assisted thoracoscopy but noted no differences in pain scores 10 . Given the large number of dermatomes affected during mastectomy we believe that the wider block dispersion associated with bolus infusion contributes to better postoperative pain control.
Since opioid consumption and average pain scores are intrinsically correlated most studies will also present similar results from our secondary outcomes. In a trial conducted by Taketa  Our primary outcome was the relative risk for severe postoperative pain which is based on an average of pain assessments done in the rst 24 hours after surgery. However, pain assessment was not performed systematically, and infusion rates were not standardized (90% of the sample received 12 mL per hour in the continuous group and 6 mL every 30 minutes in the bolus group). Also, the imbalance in sample size of the two infusion modalities could decrease the study power, however, the post hoc analysis showed a satisfactory power with the established relative risk. As for any single-center study, results will likely differ elsewhere. Most importantly, since our institution is a regional hospital our sample of patients present fewer comorbidities and complexity compared to specialized oncologic institutions, and patients with lower risk will presumably bene t less from regional anesthesia techniques.

Conclusions
In summary, after unilateral mastectomy, postoperative bolus paravertebral infusion was associated with reduction of relative risk for moderate to severe postoperative pain. In contrast, we could not establish an association between paravertebral infusion modality with postoperative opioid consumption. Given the lack of de nitive evidence, further prospective studies are required to further determine possible differences in opioid consumption and its side effects regarding paravertebral infusion modality. In an era of evidence-based medical practice, such research is crucial to establish best practices, especially in rapidly evolving areas such as regional anesthesia 19 Availability of data and materials: The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.