Abstract

Objectives

Paravertebral block (PVB) has the potential to reduce postoperative pain after breast surgery. The aim of the study was to investigate whether PVB performed immediately before surgery could affect the postoperative morbidities in terms of pain and emesis, and improve the quality of recovery (QoR) in patients after surgery for breast cancer.

Methods

Postoperative data were collected prospectively from two groups of patients undergoing unilateral breast surgery during the study period of 1 month. Forty consecutive patients received either solely general anesthesia (GA group, n = 25) or GA plus ultrasound-guided PVB (GA + PVB group, n = 15) for the surgery. Pain scores and areal distribution of pain were compared between the two groups 1 hour and 6 hours postoperatively and on the midmorning of postoperative Day 1 (POD1). The QoR scores were compared between the two groups 6 hours postoperatively and on the midmorning of POD1. Incidence of postoperative nausea and vomiting and doses of analgesics and narcotics given were also compared.

Results

Pain scores at rest were significantly lower in the GA + PVB group at all designated time points [1 hour (p < 0.0001), 6 hours (p < 0.0001), and on midmorning of POD1 (p = 0.041)]. Pain scores with movements was also significantly lower at all time points in the GA + PVB group (1 hour, p < 0.0001; 6 hours, p < 0.0001; midmorning of POD1, p = 0.0012). Areal distribution of pain at rest and with movement was wider in the GA group 1 hour and 6 hours postoperately but was identical to that of GA+ PVB group on the mid-morning of POD1 [1 hour postoperatively at rest (p < 0.0001), with movement (p < 0.0001); 6 hours postoperatively at rest (p = 0.0018), with movement (p = 0.0048)]. The QoR scores were significantly higher in the GA + PVB group at 6 hours (p < 0.0001) and on midmorning of POD1 (p = 0.0079). The incidences of postoperative nausea and vomiting were significantly lower in the GA + PVB group (p = 0.0004). Doses of postoperative analgesics and narcotics were significantly less in the GA + PVB group (p < 0.0001 and p = 0.001, respectively). Time to first request for analgesics was significantly longer in the GA + PVB group (p = 0.0002).

Conclusions

PVB given before surgery in combination with GA could provide better postoperative analgesia and better QoR than did GA alone in patients undergoing surgery for unilateral breast cancer.

Keywords

breast neoplasms; nerve block: paravertebral; pain, postoperative;

1. Introduction

About 40% of breast surgery patients suffer from clinically concernful acute postoperative pain, the so high occurrence of which indicates that postoperative pain is undertreated.¹ Furthermore, acute postoperative pain is considered as a risk factor for the successive development of persistent chronic postoperative pain in women after breast surgery.1, 2 Besides pain, symptoms, such as nausea and vomiting, are common in the immediate postoperative period3, 4 and may adversely affect patient-oriented outcomes and quality of recovery (QoR). Previous studies have shown that paravertebral block (PVB) is potentially capable of reducing pain and narcotic-related side effects after breast surgery when used in conjunction with general anesthesia (GA) or sedation.5, 6, 7, 8, 9, 10, 11, 12, 13, 14 In Taiwan, PVB has not gained wide acceptance either as an anesthetic technique or a means of postoperative pain control. We hypothesized that PVB could reduce postoperative pain and improve the QoR as measured by the validated Chinese-version questionnaire of QoR score¹⁵ in Chinese-speaking patients in Taiwan. To the best of our knowledge, QoR has not been studied in the immediate postoperative period on breast surgery patients.

We collected data of pain score, distribution of pain, and QoR score from breast surgery patients undergoing GA with or without PVB in April 2011. The aim of the study was to investigate whether PVB could affect the immediate postoperative outcomes in terms of pain, emesis, and QoR in patients undergoing breast cancer surgery patients.

2. Methods

After scrutinization by the hospital ethics committee, this study was approved and regarded as an act of hospital quality assurance concerning surgical care of breast cancer to be performed pursuant to the recommendations of the Helsinki Declaration.

2.1. Patients

Consecutive patients with American Society of Anesthesiologists physical status Classification I–III, aged 20–70 years, and scheduled for unilateral modified radical mastectomy or breast conserving therapy in our hospital in April 2011 were included in the study. The breast cancer treatment plan was adherent to the National Comprehensive Cancer Network guideline version 2.2011. Criteria for exclusion from analysis included history of allergy to local anesthetics, narcotics, acetaminophen, or Nonsteroidal anti-inflammatory drugs; infection at the thoracic paravertebral injection site; severe spine or chest wall deformity; pregnancy; breast-feeding; severe obesity (body mass index greater than 35 kg/m²); major psychosis; drug and alcohol abuse; chronic pain or long-term analgesic usage; and history of significant neurological, psychiatric, neuromuscular, cardiovascular, pulmonary, renal, or hepatic disease.

2.2. Procedure of anesthesia

The risks and benefits of PVB were explained to every patient at the preoperative evaluation clinic. Patients who consented to have PVB besides GA were allocated to the GA + PVB group. Patients who refused to have PVB received GA only (GA group). Administration of GA in standard procedure was entrusted to one of our six experienced anesthesiologists. Induction of anesthesia was made possible with propofol (2–3 mg/kg), cisatracurium (0.15 mg/kg), and fentanyl (100 μg). After tracheal intubation, anesthesia was maintained with 5–8% deaflurane in O₂, titrated to ideal level according to age, blood pressure, and heart rate. Lactated Ringer’s solution (Hartmann’s injection; Taiwan Otsuka Pharmaceutica Co., Taipei, Taiwan) was given in accordance with restricted fluid management. After intubation and stable anesthetization, PVB was performed only in the patients of the GA + PVB group.

During operation, combination of emetic prophylaxis was given individually with dexamethasone 4 mg, granisetron 1 mg, and droperidol 0.625 mg, according to the patient’s Apfel risk score and consensus guideline.16, 17 Granisetron 1 mg was used only in the postanesthesia care unit or in the surgical ward as the rescue antiemetic medication for postoperative nausea and vomiting (PONV).

After emerging from anesthesia, the patients were transferred to the postanesthesia care unit for 1-hour close observation. Postoperative analgesics were provided to achieve the pain score of less than 4 in Numeric Rating Scale (NRS). Choices of analgesics included intravenous morphine 3–6 mg and intravenous ketorolac 30 mg for patients with moderate to severe pain. Acetaminophen 500–1000 mg was given at patient’s request if the pain was mild or moderate (pain score, 4–7).

2.3. Procedure of PVB

The block was performed by an attending anesthesiologist who was experienced in the in-plane technique as described by Renes et al.¹⁸ A previous study showed that adequate loss of sensation was achieved in 97% of patients with four injections in comparison with only 11% with a single injection.¹⁹ To achieve better analgesia, multilevel injection technique was adopted, and the injections were given at T2–5 levels under ultrasound guidance (GE LOGIQ^® 7 ultrasound, GE Healthcare, Hino, Tokyo, Japan). After induction, the patient in the GA + PVB group was placed in the lateral decubitus position with the side to be blocked facing upward. After skin sterilization, a high-frequency (GE M12L probe, 6–13 MHz, GE Healthcare, Hino, Tokyo, Japan) linear array probe was placed laterally to the spinous process at the level chosen to locate the wedge-shaped paravertebral space. Then, a 22-gauge, 3.5-inch spinal needle (Terumo^®, Kofu Factory of Terumo Corporation, Yamanashi Prefecture, Japan) was inserted in a lateral-to-medial direction from the outer edge of the probe and advanced until the needle tip penetrated the internal intercostal membrane. Depending on the patient’s body mass, 3–5 mL of 0.5% bupivacaine with 1:400,000 epinephrine was injected after negative aspiration of blood or air at each of the four chosen levels. Direct visualization of the needle-tip position and the pleura being pressed ventrally during local anesthetic injection was considered as the endpoint of a successful block.

2.4. QoR score

Our primary endpoint was QoR. QoR is a validated and easy-to-use instrument to serve as an outcome variable to assess the general QoR in the immediate postoperative period.15, 20 A QoR summary score, ranging from 0 to 18, was obtained by asking the patient questions regarding degree of general well-being, support from others; general mental function; ability to perform personal hygiene; bowel/bladder function; ease of respiration; presence of headache-backache-myalgias, emesis, and pain. QoR scores 6 hours after operation and at 10 AM of POD1 were collected.

2.5. Assessment of pain and adverse events

We collected the acute postoperative pain scores using NRS 1 hour postoperatively (just before discharge from the recovery room), 6 hours postoperatively, and at 10 AM of POD 1. At each time point, patients’ NRS pain scores were recorded both at rest and with movement (moving the ipsilateral arm until the body and the arm lie at an angle of 90°). The extent of pain in the following four specific regions, that is, the area of the breast (defined as the area where either the breast was surgically affected or from which the breast was removed); the area of axilla; the arm; and the affected side of the body, was assessed at each time point both at rest and with movement. The amounts of narcotics and analgesics required throughout the postoperative period and the time to first request of rescue analgesics were recorded. We also collected the data of adverse events (accidental vascular puncture, pneumothorax, nerve damage, local anesthetic toxicity, pain in injection site, puncture of the subarachnoid space, and Horner/Harlequin syndrome) and the number of patients with PONV throughout the postoperative period.

2.6. Statistical analysis

Normally distributed data were presented as mean ± standard deviation. Non-Gaussian data were presented as median (25^th percentile and 75^th percentile). Maximum and minimum levels were also presented. Nonparametric Wilcoxon two-sample test and Fisher’s exact test were used to test differences between the two groups of patients. All of the analyses were carried out using SAS 9.1 software (SAS Institute INC, Cary, NC, USA). We used normal approximation method for the nonparametric analysis. A two-tailed p value <0.05 was considered statistically significant.

3. Results

A total of 40 patients were included in the analysis, all of whom were women. We had 25 patients in the GA group and 15 in the GA + PVB group. We compared the demographic and surgical data between the two groups and found that they were statistically similar in terms of age (50.4 ± 8.71 years vs. 54.2 ± 9.41 years); body mass index (23.16 ± 2.72 kg/m² vs. 23.74 ± 3.43 kg/m²); American Society of Anesthesiologists physical status; stratified risk of PONV (low risk denoted Apfel risk score of 0–1, medium risk denoted Apfel risk score of 2–3, and high risk denoted Apfel risk score of 4); duration of operation (160.2 ± 39.78 minutes vs. 164.67 ± 41.17 minutes); fluid management; and type of operation (Table 1). However, the use of antiemetic was significantly less in the GA + PVB group (Table 1).

Table 1. Baseline characteristics of the patients in the two groups.

Download full-size image

The NRS pain scores revealed significantly less pain at resting 1 hour postoperatively (GA 8 [6, 9] vs. GA + PVB 1 [0, 3], p < 0.0001); 6 hours postoperatively (GA 3 [3, 6] vs. GA + PVB 1 [0, 1], p < 0.0001); and in midmorning of POD1 (GA 2 [1, 3] vs. GA + PVB 1 [0, 1], p = 0.041) in the GA + PVB group (Table 2). The NRS pain scores also signified less pain with movement in the GA + PVB group at all of the three time points (GA 9 [7, 10] vs. GA + PVB 1 [0, 3], p < 0.0001); GA 5 [4, 7] vs. GA + PVB 1 [0, 1], p < 0.0001; and GA 3 [1, 6] vs. GA + PVB 1 [0, 1], p = 0.0012).

Table 2. Comparison of outcomes between the two groups.

Download full-size image

As regarding the areal distribution of pain at rest and with movement, more sites were involved in the GA group 1 hour (resting, p < 0.0001; moving, p < 0.0001) and 6 hours (resting, p = 0.0018; moving, p = 0.0048) postoperatively but not in midmorning of POD1 (Table 3). The QoR scores (0–18) were statistically different between the two groups 6 hours postoperatively (GA 12 [10, 15] vs. GA + PVB 18 [16, 18], p < 0.0001) and in midmorning of POD1 (GA 16 [14, 18] vs. GA + PVB 18 [17, 18], p = 0.0079) (Table 2).

Table 3. Distribution of pain.

Download full-size image

Cumulative postoperative analgesic consumption was 4 [3, 4] doses in GA group and 1 [1, 2] dose (p < 0.0001) in GA + PVB group, respectively. Cumulative postoperative narcotic consumptions were 1 dose [0, 1] in GA group and 0 [0, 0] dose (p = 0.001) in GA + PVB group (Table 2).

The time from the end of the operation to the first request of analgesic differed significantly between the two groups (GA 30 [18, 50] minutes vs. GA + PVB 435 [240, 1110] minutes, p = 0.0002). The time from induction of anesthesia to first request of analgesics was also statistically different (GA 199 [170, 235] minutes vs. GA + PVB 605 [400, 1250] minutes, p = 0.0002). The incidence of PONV was significantly higher in the GA group (64%) than that in the GA + PVB group (6.67%) (p < 0.001, Table 2).

None of the patients in the GA + PVB group had clinical signs or radiological evidence of pneumothorax. One patient in the GA + PVB group complained of mild soreness at the PVB injection site, which was relieved by 500-mg oral acetaminophen. Local anesthetic toxicity, nerve injury, hematoma, infection, puncture of the subarachnoid space, and Horner/Harlequin syndrome were not observed in any of these patients.

4. Discussion

In this prospective study, we found that different anesthetic techniques could result in different postoperative pictures. The GA + PVB group presented with significantly lower postoperative pain scores, lesser regions involving pain, and higher QoR scores throughout the immediate postoperative period. The greater pain scores in the GA group signified severe pain, which was further exacerbated by movement 1 hour postoperatively. Six hours postoperatively, the severity of pain decreased to moderate level in the GA group. However, only mild pain was observed, either at rest or with moving, in the GA + PVB group 1 hour and 6 hours postoperatively. The pain scores were low in most patients of the two groups at the time of discharge on POD1. Nevertheless, some patients in the GA group still experienced considerable pain at rest and with movement, which caused prolongation of hospital stay; none in the GA + PVB group did so. The analgesic effect lasted as long as 24 hours after initial nerve block in some patients.

The use of antiemetic was less in the GA + PVB group. The overall dosage of antiemetics was also less in the GA + PVB group. We are of the opinion that there could be bias in existence because of the fact that the study design was unblinded. The practitioners might expect the postoperative opioid use and the likelihood of PONV related to postoperative use of opioid in the GA + PVB group to be less. They might have consequently used prophylactic antiemetics to a lesser extent so as to avoid the side effects of droperidol (QT prolongation and torsades de pointes) and dexamethasone (glucose intolerance and stomach upset). If the use of analgesics or antiemetics is less in the GA + PVB group, we might expect to see a higher incidence of pain or emesis. However, the incidence of PONV was low, and the pain was less severe in this group. The consumption of narcotics was significantly less in the GA + PVB group. This could probably explain why the incidence of PONV was significantly low in the GA + PVB group when every patient was given proper prophylaxis based on individualized Apfel risk score during operation.

Despite the emphasis on measuring patient-oriented outcomes in many fields of health care and acceptance of these outcomes as a valid endpoint in clinical trials,21, 22, 23 few studies have examined the effect of postoperative pain, nausea, and other symptoms on patient-oriented outcomes in the immediate postoperative period. Wu et al²⁴ had demonstrated close correlation among patient-oriented outcome; QoR; and the common postoperative symptoms, such as nausea and pain, in patients undergoing elective radical retropubic prostatectomy. In this study, we demonstrated a significant impact of PVB on reduction of the postoperative pain and on improvement of QoR 6 hours postoperatively and at midmorning of POD1 (Table 2) in patients who underwent breast cancer surgery.

We used surrogate measures to confirm successful blockade as the PVB was performed in the operating room after the patients had lost sensation after induction of GA. It was not possible to confirm the outcome of the blockade by assessing sensation to pinprick or cold at this moment. Instead, we used ultrasound image for direct visualization of needle-tip position and ventral compression of the pleura by the anesthetic solution at the time of injection, both of which served as circumstantial evidence of a successful block. From the excellent reduction of postoperative pain score in the GA + PVB group, we considered that the act of performing the block under the ultrasound guidance could serve as a reliable surrogate for blockade confirmation by sensation assessment.

Although we did not count the costs of the additional time and operating-room personnel in performing PVB as done in previous studies,5, 8 it could be deduced by general consent that more time could be involved in performing the PVB than in inducing the GA. If the PVB could be performed in the preoperative area with monitors where space and facility could sustain, it might be a way to solve this problem. Although there is a possible way of cost saving as suggested by Weltz et al,⁵ PVB has not yet gained general acceptance in Taiwan. This could be partly because of PVB being a sweated job along with poor reimbursement by National Health Insurance.

Complications after PVB, as reported in Naja’s study,²⁵ where nerve stimulation technique was adopted, included inadvertent vascular puncture (6.8%), hypotension (4.0%), hematoma (2.4%), pain at the site of skin puncture (1.3%), signs of epidural or intrathecal spread (1.0%), pleural puncture (0.8%), and pneumothorax (0.5%). In contrast, only one patient in the GA + PVB group in our study complained of pain at the injection site (6.67%). No other complications related to PVB were noted in our patients using the ultrasound-guided technique. The failure rate reported was 6.1% in that study.²⁵ We achieved the endpoint of a successful PVB (direct visualization of needle-tip position and the pleura being pressed ventrally during local anesthetic injection) in every patient in the GA + PVB group. The results of significant reduction of pain scores in the GA + PVB group also suggested a high success rate. However, further study with a larger sample size is necessary to prove the superiority of the method¹⁸ we used in this study.

Time from induction of GA to first request of analgesic could give us a hint of the duration of intensive analgesia of PVB because PVB was done within half an hour after the induction of GA in all of our patients. The median duration of 605 [400, 1250] minutes suggested that the intensive analgesic effect could last for about 10 hours in most patients.

There are some limitations in our study. First, it was not a randomized study; thus, it did not allow us to establish a causal relationship between the use of PVB and the QoR. Nonetheless, given the similarity of the baseline patient characteristics in the two groups, our results still implied a clinically significant benefit of PVB in terms of postoperative pain control and QoR for breast cancer surgery. Second, the number of our patients was small. It was difficult to demonstrate the real correlation between pain/emesis and QoR in each group. Third, we did not collect baseline data for evaluation of QoR. Although the QoR scores just before discharge on POD1 in both groups were much higher than those on POD0, there was a statistical difference between the two groups. Thus, we cannot say from our results if patients had regained their baseline general condition at the time of discharge.

Our results conclude that PVB could be carried out in the wake of induction of GA in patients undergoing unilateral breast cancer surgery to achieve better postoperative pain control, lessen the extensiveness of pain, lower the incidence of PONV, and provide better QoR in the immediate postoperative period in comparison with GA alone.

Acknowledgment

The authors wish to thank Yong Alison Wang MD, PhD, for revision, and Chen-Fang Hung, professional statistician, for statistical help.