This randomized clinical trial uses a nonferiority design to gauge the efficacy of continuous paravertebral block (PVB) and a single-shot intercostal nerve block vs thoracic epidural analgesia in patients undergoing thoracoscopic lung resection.
Key Points
Question
Are continuous paravertebral or single-shot intercostal nerve block noninferior to thoracic epidural analgesia regarding pain and superior regarding quality of recovery?
Findings
In this randomized clinical trial including 450 patients undergoing thoracoscopic lung resection, a single-shot intercostal nerve block was noninferior to thoracic epidural analgesia regarding pain, while continuous paravertebral block remained inconclusive. Quality of recovery was similar across groups.
Meaning
After thoracoscopic anatomical lung resection, single-shot intercostal nerve block is a viable alternative to thoracic epidural analgesia, offering noninferior pain relief and similar quality of recovery, but the risks and benefits of each analgesic technique should be considered in clinical practice.
Abstract
Importance
Effective pain control after thoracic surgery is crucial for enhanced recovery. While thoracic epidural analgesia (TEA) traditionally ensures optimal analgesia, its adverse effects conflict with the principles of enhanced recovery after thoracic surgery. High-quality randomized data regarding less invasive alternative locoregional techniques are lacking.
Objective
To evaluate the efficacy of continuous paravertebral block (PVB) and a single-shot intercostal nerve block (ICNB) as alternatives to TEA.
Design, Setting, and Participants
This randomized clinical trial compared PVB and ICNB vs TEA (1:1:1) in patients undergoing thoracoscopic anatomical lung resection at 11 hospitals in the Netherlands and Belgium, enrolled from March 5, 2021, to September 5, 2023. The study used a noninferiority design for pain and a superiority design for quality of recovery (QoR).
Interventions
Continuous PVB and single-shot ICNB.
Main Outcomes and Measures
Primary outcomes were pain, defined as mean proportion of pain scores 4 or greater during postoperative days (POD) 0 through 2 (noninferiority margin for the upper limit [UL] 1-sided 98.65% CI, 17.5%), and QoR, assessed with the QoR-15 questionnaire at POD 1 and 2. Secondary measures included opioid consumption, mobilization, complications, and hospitalization.
Results
A total of 450 patients were randomized, with 389 included in the intention-to-treat (ITT) analysis (mean [SD] age, 66 [9] years; 208 female patients [54%] and 181 male [46%]). Of these 389 patients, 131 received TEA, 134 received PVB, and 124 received ICNB. The mean proportions of pain scores 4 or greater were 20.7% (95% CI, 16.5%-24.9%) for TEA, 35.5% (95% CI, 30.1%-40.8%) for PVB, and 29.5% (95% CI, 24.6%-34.4%) for ICNB. While PVB was inferior to TEA regarding pain (ITT: UL, 22.4%; analysis per-protocol [PP]: UL, 23.1%), ICNB was noninferior to TEA (ITT: UL, 16.1%; PP: UL, 17.0%). The mean (SD) QoR-15 scores were similar across groups: 104.96 (20.47) for TEA, 106.06 (17.94; P = .641) for PVB (P = .64 for that comparison), and 106.85 (21.11) for ICNB (P = .47 for that comparison). Both ICNB and PVB significantly reduced opioid consumption and enhanced mobility compared with TEA, with no significant differences in complications. Hospitalization was shorter in the ICNB group.
Conclusions and Relevance
After thoracoscopic anatomical lung resection, only ICNB provides noninferior pain relief compared with TEA. ICNB emerges as an alternative to TEA, although risks and benefits should be weighed for optimal personalized pain control.
Trial Registration
ClinicalTrials.gov Identifier: NCT05491239
Introduction
Effective pain control after thoracic surgery is one of the key elements of enhanced recovery after thoracic surgery (ERATS) guidelines.1 Although variation in standard practice prevails, thoracic epidural analgesia (TEA) is considered the reference standard for postoperative pain management following thoracic surgery.2,3 The historical reliance on TEA aligns with traditional thoracotomy procedures, but its adverse effects seem no longer in line with current advances in thoracoscopic surgery and adoption of ERATS principles, resulting in advice against TEA in the Procedure-Specific Postoperative Pain Management (PROSPECT) guideline.4,5 Although there is a strong desire to replace TEA, recent systematic reviews and meta-analyses still lack robust data from randomized clinical trials (RCTs) to definitively favor less invasive regional analgesic techniques after thoracoscopic surgery.5,6,7
Alternative locoregional analgesic techniques are indispensable in multimodal postoperative analgesia as they reduce opioid consumption, asd well as opioid-related adverse effects such as postoperative nausea and vomiting (PONV), and forego epidural-related risks and adverse effects (eg, hypotension and neurogenic bladder dysfunction).8 Currently, several locoregional alternatives are used, with intercostal nerve blocks (ICNB) and paravertebral blocks (PVB) being the most popular.6,9
To reduce scientific uncertainty, and to enable clinical practice to weigh reliable alternatives to TEA in postoperative analgesia after thoracoscopic surgery, the aim of this multicenter RCT was to assess the efficacy of continuous PVB and single-shot ICNB as alternatives to TEA in terms of pain and quality of recovery (QoR).
Methods
This study was conducted in accordance with the Consolidated Standards of Reporting Trials (CONSORT) reporting guideline and approved by the medical research ethics committee NedMec (NL75375.041.20). Patients provided written informed consent.
Trial Design
The study protocol and statistical analysis plan (SAP) are available in Supplement 1 and Supplement 2 (and registered at NCT05491239).10 We conducted a multicenter pragmatic randomized 3-arm trial in 11 hospitals in the Netherlands and Belgium, comparing single-shot ICNB and continuous PVB vs TEA, with a noninferiority design regarding pain and a concomitant superiority design regarding QoR.
Participants and Selection Criteria
Consecutive patients (aged >18 years) were eligible if they were referred for anatomical lung resection (pneumonectomy, lobectomy, bi-lobectomy, or segmentectomy for either benign or malignant disease) with the intention of performing thoracoscopic surgery and able to provide informed consent. Exclusion criteria were contraindications for TEA or PVB, allergic reactions to local anesthetics, long-term use of opioids, and a high risk of conversion to thoracotomy. Patients with perioperative conversion to thoracotomy or nonanatomical lung resections were regarded as late exclusions.
Randomization, Stratification, and Blinding
After patients provided informed consent, patient data were entered into a computerized database (Research Manager, 2024, the Netherlands) by the researchers, and with an unchangeable computer-generated number, patients were randomized in variable blocks in a 1:1:1 ratio for 1 of the 3 analgesic strategies: TEA, PVB, or ICNB. As this was a pragmatic trial and slight differences in anesthesiology and surgery protocols between hospitals were permitted, randomization was stratified by hospital to minimize bias. As the analgesic techniques and catheters differ between the treatment arms, blinding was unfeasible.
Analgesia Protocols
Premedication and general anesthesia were applied according to in-house protocols. Patients received dexamethasone, 8 mg, and antiemetics based on local protocols.
The usual care group consisted of TEA and the intervention groups of continuous PVB or single-shot ICNB. Placement of the TEA catheter was performed in the awake patient preoperatively by the anesthesiologist and comprised continuous infusion of local anesthetics and opioids until postoperative day (POD) 2. Placement of PVB and ICNB was performed during general anesthesia, usually by the surgeon. The PVB was positioned under direct thoracoscopic vision with the catheter tip adjacent to the sympathetic trunk, usually at the T4-5 level and at the beginning of surgery, and comprised an immediate perioperative bolus and continuous infusion of local anesthetics without opioids until POD 2. The ICNB was a multilevel (T2-T10) single-shot block under direct thoracoscopic vision, performed with an injection of local anesthetics without opioids as the final step of the operation. Details of the analgesic techniques are described in our protocol.10 Additionally, the protocol dictated in case of TEA failure to perform PVB and in case of PVB failure to perform ICNB in a compulsory order. For more details on the performance of the analgesic techniques, see eTable 1 in Supplement 3.
Outcomes
The primary outcomes were (1) mean proportion of postoperative pain scores at rest of 4 or greater11 measured by the numerical rating scale (NRS; range, 0-10), defined as the number of NRS scores 4 or greater divided by the total amount of NRS scores obtained from POD 0-2 (ie, 8 measurements: at the recovery room, at the ward on the first evening, plus the next 2 PODs in the morning, afternoon, and evening), and (2) mean QoR-15 score of POD 1 and 2, measured by the 15-item QoR questionnaire (range, 0-150; higher value signifies better QoR).
Secondary outcomes were pain during rest and mobilization and QoR from POD 0 through 3 and at follow-up; the cumulative use of opioids by morphine milligram equivalent (MME) conversion factors and nonopioid analgesics at POD 0 through 3 (including both continuous opioids and boluses through analgesic catheters); postoperative complications within 30 postoperative days according to the Clavien-Dindo classification12 and reported as minor (Clavien-Dindo 1 and 2), major (Clavien-Dindo 3 and 4) and mortality (Clavien-Dindo 5); hospitalization days within the first 30 days or until discharge; patient satisfaction on a 5-point Likert scale (not at all satisfied, slightly satisfied, neutral, very satisfied, and extremely satisfied); and degree of mobility during POD 0 through 3 on a 4-point scale (on the bed, to the chair, to the toilet, and outside the hospital room).
Noninferiority and Sample Size Calculation
A pilot study13 comparing TEA with continuous intercostal analgesia showed that a mean (SD) of 17.57% (19.57) and 21.21% (23.33) of the postoperative pain measurements at rest from POD 0 through 3 led to pain scores of 4 or higher, respectively. We considered it clinically acceptable to set the noninferiority upper limit (UL) of the 1-sided 98.65% CI of the difference between means at 17.5% (ie, an increase in pain in about 1 of 6 measurements), given the counterbalancing potential gain in secondary outcomes. To demonstrate noninferiority regarding pain, 64 patients were needed per group to achieve a power of 90% with a 1-sided type I error of 0.0135.14 For superiority regarding QoR to detect a minimum clinically important difference of 8.0 points (SD, 18 points)15 at the time of the study design, 125 patients were needed per group to achieve 90% power with a 2-sided type I error of 0.027. To account for possible nonnormally distributed data with a Mann-Whitney U test (with an asymptotic relative efficiency of 0.955 compared with the t test) and assuming a 12.6% dropout rate due to conversion to thoracotomy,16,17 we aimed to include 150 patients per group.
Because our sample size was based on a pilot study of only 46 patients, a data safety monitoring board evaluated the point estimator and variance of our control group after 50% of TEA inclusions (n = 75) and advised to continue with the predefined sample size.
Statistical Analysis
The SAP is available online.11 As our study design involved multiple comparisons (2 interventions) against a control group, the significance level was adjusted for superiority at P = .03.14 For noninferiority, the UL of the 1-sided 98.65% CI of the difference in means may not exceed 17.5% in both intention-to-treat (ITT) and per-protocol (PP) analyses. The QoR-15 score was compared with a t test for superiority.
Complete cases were predefined according to criteria outlined in the SAP. For missing outcome data, we used multiple imputation using fully conditional specification. For the outcome pain, the proportion of complete cases exceeded 95%, thus data missing completely at random were assumed. Conversely, for the outcome QoR-15, the proportion of complete cases fell below 95% and data missing at random were considered. Accordingly, we included potential confounding variables in the multiple imputation model for QoR-15 identified through an analysis of missing data patterns and reported a separate baseline characteristics table for participants with and without missing data (eTable 2 in Supplement 3).
Because of late exclusions, as well as by chance, potential imbalance in baseline compatibility may occur. For this, careful evaluation of baseline characteristics was performed using standardized mean differences (eTable 3 in Supplement 3). If baseline variables indicated an imbalance greater than 25%, correction for the primary outcomes was performed by linear regression to correct for possible bias (eTable 4 in Supplement 3). For the secondary outcomes, we used the ITT population, and missing data were not imputed. Repeated measurements were analyzed by linear mixed models with random intercepts and baseline scores, time of the measurement, and treatment group as fixed effects. The analyses were performed using the Statistical Package for the Social Sciences version 28.0 (IBM).
Results
From March 5, 2021, to September 5, 2023, we enrolled 450 patients who were randomly assigned to receive TEA (n = 148), PVB (n = 153), or ICNB (n = 149). The flowchart representing the ITT and PP population including all exclusions is shown in Figure 1. The ITT analysis included a total of 389 patients (13.6% late exclusions; mean [SD] age, 66 [9] years; 208 female [54%] and 181 male [46%]); patient characteristics are presented in Table 1. All patients underwent an anatomical lung resection with thoracoscopic surgery.
Figure 1. Enrollment, Random Assignment, and Flow of Study Patients.
ICNB indicates intercostal nerve block; TEA, thoracic epidural analgesia; PVB, paravertebral block.
Table 1. Baseline Patient, Surgical, and Pain Characteristics.
| Characteristic | No. (%) | ||
|---|---|---|---|
| TEA (n = 131) | PVB (n = 134) | ICNB (n = 124) | |
| Age, mean (SD), y | 67 (10) | 65 (10) | 67 (8) |
| Body mass index, mean (SD)a | 27 (12) | 27 (15) | 29 (32) |
| Sex | |||
| Female | 68 (52) | 75 (56) | 65 (52) |
| Male | 63 (48) | 59 (44) | 59 (48) |
| Smoking status | |||
| Smoker | 31 (24) | 30 (23) | 35 (28) |
| Nonsmoker | 21 (16) | 12 (10) | 16 (13) |
| Ex-smoker | 79 (60) | 88 (67) | 72 (59) |
| WHO performance state | |||
| 0-1 | 117 (97) | 118 (98) | 113 (96) |
| 2-3 | 4 (3) | 3 (2) | 5 (4) |
| Comorbidities | |||
| COPD | 42 (32) | 36 (27) | 27 (22) |
| Previous thoracic surgery/radiotherapy | 22 (17) | 14 (10) | 17 (14) |
| Pain syndromes | 4 (3) | 3 (2) | 6 (5) |
| Final surgical indication | |||
| Primary lung cancer | 115 (88) | 120 (90) | 109 (88) |
| Metastasis | 12 (9) | 12 (9) | 12 (10) |
| Benign lesion | 4 (3) | 2 (1) | 3 (2) |
| Pathological TNM stageb | |||
| Tx | 2 (2) | 1 (1) | 5 (6) |
| T0 | 1 (1) | 0 | 0 |
| T1 | 65 (57) | 68 (57) | 64 (59) |
| T2 | 35 (30) | 27 (22) | 27 (24) |
| T3 | 6 (5) | 18 (15) | 13 (12) |
| T4 | 6 (5) | 6 (5) | 0 |
| Nx | 2 (1) | 0 | 1 (1) |
| N0 | 101 (88) | 98 (82) | 97 (89) |
| N1 | 7 (6) | 11 (9) | 7 (6) |
| N2 | 4 (4) | 11 (9) | 4 (4) |
| N3 | 1 (1) | 0 | 0 |
| Baseline NRS score, median (IQR) | 0 (0-0) | 0 (0-0) | 0 (0-0) |
| Baseline QoR-15 score, mean (SD) | 127 (19) | 131 (16) | 129 (16) |
| Surgical characteristics | |||
| Operation technique | |||
| Uniportal VATS | 35 (27) | 34 (25) | 30 (24) |
| Multiportal VATS | 86 (66) | 89 (67) | 85 (69) |
| Multiportal RATS | 10 (7) | 11 (8) | 9 (7) |
| Type of resection | |||
| Segmentectomy | 9 (7) | 8 (5) | 5 (4) |
| Lobectomy | 119 (91) | 122 (91) | 117 (94) |
| Bi-lobectomy | 3 (2) | 2 (2) | 2 (2) |
| Pneumonectomy | 0 | 2 (2) | 0 |
| Operating time, mean (SD), min | 157 (53) | 171 (56) | 155 (49) |
| 1 Chest tube | 130 (99) | 132 (99) | 123 (99) |
| Anesthesiology characteristics | |||
| Preoperative | |||
| Analgesics | |||
| Paracetamol | 100 (76) | 105 (78) | 97 (78) |
| NSAID | 40 (31) | 38 (28) | 39 (32) |
| Benzodiazepine | 22 (29) | 27 (20) | 16 (13) |
| Perioperative | |||
| Induction medication | |||
| Propofol | 127 (97) | 130 (97) | 123 (99) |
| Benzodiazepine | 6 (5) | 4 (3) | 1 (1) |
| Opioids | 118 (91) | 122 (91) | 109 (88) |
| MME, mean (SD), mg | 88 (74) | 102 (118) | 89 (83) |
| Muscle relaxant | 126 (97) | 132 (99) | 121 (98) |
| Intravenous lidocaine | 27 (21) | 16 (12) | 21 (17) |
| Maintenance medication | |||
| Propofol | 92 (70) | 97 (72) | 87 (70) |
| Opioids | 131 (100) | 127 (95) | 120 (97) |
| MME, median (IQR), mg | 290 (55-519) | 128 (23-454) | 181 (28-457) |
| Muscle relaxant | 120 (92) | 125 (93) | 118 (95) |
| Additional anestheticsc | 46 (37) | 6 (5) | 2 (2) |
| Dexamethasone, mean (SD), mg | 6 (2) | 6 (2) | 7 (10) |
| Antiemetic | 95 (73) | 107 (80) | 103 (83) |
| Vasopressives | 120 (92) | 120 (90) | 108 (87) |
Abbreviations: COPD, chronic obstructive pulmonary disease; ICNB, intercostal nerve block; MME, morphine milligram equivalent; NRS, numerical rating scale; NSAID, nonsteroidal anti-inflammatory drug; PVB, paravertebral block; QoR-15, quality of recovery 15 score; RATS, robot assisted thoracoscopic surgery; TEA, thoracic epidural analgesia; TNM, tumor nodes metastasis; VATS, video-assisted thoracoscopic surgery; WHO, World Health Organization.
Calculated as weight in kilograms divided by height in meters squared.
In case of primary lung cancer.
Mostly given through epidural or paravertebral catheter, or intravenously.
Primary Outcomes
In ITT analysis, the proportion of NRS scores 4 or greater at rest was 20.7% (95% CI, 16.5%-24.9%) for TEA, 35.5% (95% CI, 30.1%-40.8%) for PVB (∆14.8%; UL 1-sided 98.65% CI, 22.4%; noninferiority P = .21), and 29.5% (95% CI, 24.6%-34.4%) for ICNB (∆8.8%; UL 1-sided 98.65% CI, 16.1%; noninferiority P = .004) (Figure 2A). In the PP analysis, the proportion of NRS scores 4 or greater at rest was 20.4% (95% CI, 15.6%-25.1%) for TEA, 34.7% (95% CI, 28.5%-40.9%) for PVB (∆14.3%; UL 1-sided 98.65% CI, 23.1%; noninferiority P = .21), and 29.7% (95% CI, 24.8%-34.6%) for ICNB (∆9.3%; UL 1-sided 98.65% CI, 17.0%; noninferiority P = .009) (Figure 2A).
Figure 2. Graphical Representation of the 2 Co-Primary Outcomes.
A, Co-primary outcome of pain as indicated by mean score on the numerical rating scale (NRS) for the intention-to-treat (ITT) and per-protocol (PP) populations. The differences between the intervention groups vs thoracic epidural analgesia (TEA) as the control are demonstrated with their respective upper limits (ULs) of the 98.65% CI for noninferiority. A indicates the difference (∆) was 14.8% (UL, 22.4%); B, ∆8.8% (UL, 16.1%); C, ∆14.3% (UL, 23.1%); and D, ∆9.3% (UL, 17.0%). Error bars in this panel indicate 95% CI. B, Co-primary outcome of quality of recovery (QoR) as indicated by mean score on the 15-item QoR questionnaire for the ITT population. For the TEA (control group) vs paravertebral block (PVB) group, P = .64; for the TEA group vs intercostal nerve block (ICNB) group, P = .47. Error bars in this panel indicate SD.
aNoninferiority P = .004 for B and P = .009 for D.
The mean (SD) QoR-15 score in the ITT analysis was 104.96 ( 20.47) in the TEA group compared with 106.06 (17.94) in the PVB group (P = .64 for that comparison) and 106.85 (21.11) and ICNB group (P = .47 for that comparison) (Figure 2B).
Correction for imbalance of baseline variables can be found in eTable 4 in Supplement 3, with no meaningful changes in the primary outcome analysis.
Repeated Measurements for Pain and QoR
While comparing repeated NRS measurements from POD 0 through 3, TEA demonstrated significantly lower pain scores immediately after surgery until the evening of POD 1 at rest (eFigure A and eTable 5A in Supplement 3) and until the morning of POD 2 during mobilization (eFigure B and eTable 5B in Supplement 3). Thereafter, PVB and ICNB showed similar pain scores at rest and mobilization until POD 3 and follow-up. During POD 0 through 3 and at follow-up, there was no significant difference in QoR-15 between TEA and the intervention groups (eFigure C and eTable 5C in Supplement 3). In addition, a subgroup analysis was performed for each domain of the QoR-15 questionnaire and can be found in eTable 6 in Supplement 3.
Additional Analgesia
On POD 0, the TEA group used significantly fewer nonsteroidal anti-inflammatory drugs (NSAIDs) than the PVB group (54 patients [41%] vs 75 [56%], respectively; P = .02) and on POD 1, less than the ICNB group (69 patients [53%] vs 83 [67%], respectively; P = .02). Patients receiving TEA consumed significantly more opioids during POD 0 through 3 compared with PVB and during POD 0 through 2 compared with ICNB (Table 2).
Table 2. Secondary Outcome Measures.
| Outcome | TEA (n = 131) | PVB (n = 134) | P valuea | ICNB (n = 124) | P valueb |
|---|---|---|---|---|---|
| NRS scores ≥4, median (IQR), % | |||||
| Mobilizing on POD 0-3 | 38 (32-43) | 55 (49-61) | <.001 | 51 (45-57) | .002 |
| At rest on POD 0-3 | 21 (17-25) | 32 (27-37) | .001 | 28 (23-32) | .04 |
| Follow-up NRS score, median (IQR) | 0 (0-1) | 0 (0-1) | .66 | 0 (0-1) | .48 |
| Follow-up QoR-15 score | 122 (20) | 122 (22) | .91 | 119 (21) | .32 |
| Analgesics and opioids at POD 0 | |||||
| Paracetamol, No. (%) | 123 (94) | 130 (97) | .22 | 121 (98) | .15 |
| NSAID, No. (%) | 54 (41) | 75 (56) | .02 | 67 (54) | .04 |
| Opioids MME, median (IQR), mg | 97 (75-125) | 18 (12-38) | <.001 | 24 (15-35) | <.001 |
| Analgesics and opioids at POD 1 | |||||
| Paracetamol, No. (%) | 128 (98) | 134 (100) | .12 | 119 (96) | .49 |
| NSAID, No. (%) | 69 (53) | 82 (61) | .16 | 83 (67) | .02 |
| Opioids MME, median (IQR), mg | 195 (135-216) | 23 (15-38) | <.001 | 25 (15-38) | <.001 |
| Analgesics and opioids at POD 2 | |||||
| Paracetamol, No. (%) | 125 (95) | 130 (97) | .54 | 113 (91) | .17 |
| NSAID, No. (%) | 70 (53) | 75 (56) | .68 | 64 (52) | .77 |
| Opioids MME, median (IQR), mg | 90 (30-180) | 23 (15-39) | <.001 | 30 (15-40) | <.001 |
| Analgesics and opioids at POD 3 | |||||
| Paracetamol, No. (%) | 102 (78) | 106 (79) | .81 | 81 (65) | .03 |
| NSAID, No. (%) | 38 (29) | 40 (30) | .88 | 27 (22) | .19 |
| Opioids MME, median (IQR), mg | 38 (15-75) | 20 (15-30) | <.001 | 30 (15-45) | .08 |
| Postoperative complications, No. (%) | 23 (18) | 30 (22) | .36 | 33 (27) | .08 |
| Clavien-Dindo Classification | |||||
| Minor | 16 (12) | 19 (14) | 18 (15) | ||
| Major | 4 (3) | 9 (7) | 14 (11) | ||
| Mortality | 3 (2) | 2 (1) | 1 (1) | ||
| Postoperative nausea and vomiting score, mean (SD)c | 8 (2) | 9 (2) | .01 | 9 (2) | .01 |
| Hospitalization, median (IQR), d | 4 (3-5) | 4 (2-5) | .43 | 3 (2-5) | .04 |
| Time to removal of chest tube, mean (SD), d | 1.5 (0.8) | 1.6 (0.8) | .89 | 1.5 (0.7) | .62 |
| Presence of urinary catheter, mean (SD), d | 1.5 (1.1) | 0.5 (0.9) | <.001 | 0.4 (0.7) | <.001 |
Abbreviations: ICNB, intercostal nerve block; MME, morphine milligram equivalent; NRS, numerical rating scale; NSAID, nonsteroidal anti-inflammatory drug; POD, postoperative day; PVB, paravertebral block; QoR-15, quality of recovery 15 score; TEA, thoracic epidural analgesia.
P value comparing TEA and PVB groups.
P value comparing TEA and ICNB groups.
From the QoR-15 questionnaire (range, 0-10; 0 = always; 10 = never).
Secondary Outcomes
Postoperative complications were found in 23 of 131 patients (18%; 95% CI, 11%-24%) for TEA vs 30 of 134 patients (22%; 95% CI, 15%-30%; P = .36) for PVB and 33 of 124 patients (27%; 95% CI, 19%-35%; P = .08) for ICNB (Table 2). Complications classified as Clavien-Dindo 3 or 4 are depicted in eTable 7 in Supplement 3 and were predominantly prolonged air leak (PAL) that required chest tube reinsertion (TEA, 3; PVB, 1; and ICNB, 6 patients) and thoracic empyema (TEA, 1; PVB, 2; and ICNB, 2 patients). In the ICNB group, 2 patients underwent re-VATS because of middle lobar torsion.
The median length of stay (LOS) was 4 days (IQR, 3-5) for TEA vs 4 days (IQR, 2-5) for PVB (P = .43) and 3 days (IQR, 2-5) for ICNB (P = .04) (Table 2). Within 30 POD, there were 2 readmissions in both the TEA and ICNB groups, while the PVB group had no readmissions. According to the QoR-15 question about PONV, patients in the TEA group experienced significantly more PONV compared with those in the PVB and ICNB groups (Table 2). Patients in the TEA group were significantly more (extremely) satisfied at the day of the operation as compared with those in PVB and ICNB groups, corresponding to 100 patients (94%) vs 85 patients (80%; P < .001) and 84 patients (79%; P = .002), respectively. During POD 1 through 3, patient satisfaction was equal among all groups (Figure 3A).
Figure 3. Patient Satisfaction and Mobilization Comparing Thoracic Epidural Analgesia (TEA) vs Paravertebral Block (PVB) and TEA vs Intercostal Nerve Block (ICNB).
Patients receiving TEA were significantly more bed-bound compared with the PVB group during POD 1 and 2 and to ICNB during POD 0 and 1. Inversely, patients receiving ICNB were significantly more mobile from the bed to the chair or to the toilet on POD 0, compared with those in the TEA group. On POD 1, patients in the PVB and ICNB groups were significantly more mobile outside of the patient room compared with the TEA group (Figure 3B).
Discussion
Our trial demonstrated that ICNB was noninferior to TEA regarding pain relief whereas PVB showed inconclusive results. Despite more moments of pain, patients receiving PVB or ICNB reported similar QoR-15 scores and lower opioid consumption, less PONV, and improved mobility, leading to a shorter LOS in the ICNB group.
In a meta-analysis comparing TEA vs PVB in patients undergoing VATS, TEA also provided significant lower pain scores during the first 24 hours, albeit with clinically unimportant short-term benefits.18 Although surgeon-performed PVB has been proven noninferior to ultrasound-guided PVB,19 possible explanations for the overall suboptimal results of PVB may be the absence of opioids in the analgesic infusion and variability in paravertebral spread.20,21 A recent RCT showed superiority of TEA regarding pain compared with continuous PVB, albeit with more adverse effects from TEA.22 Continuous PVB relies on longitudinal spread within the paravertebral space, even though the catheter tip is positioned at a single vertebral level. Despite correct catheter placement, 23% of patients may still experience inadequate pain control,23 possibly due to an insufficient extent of paravertebral spread. The PVB can also be provided as a multilevel single-shot technique; however, this is not feasible for placement of a catheter for continuous infusion.19 The beneficial effects of continuous PVB, when compared with continuous TEA, can be explained by fewer adverse effects of PVB (reduced hypotension, PONV, limb weakness, and urinary bladder dysfunction).
For single-shot ICNB, given at multiple intercostal levels (T2-10), we found slightly more moments of pain compared with TEA, but the extent of pain remained within our noninferiority margin. Whereas a recent RCT24 also demonstrated more pain by ICNB compared with TEA within the first 24 hours, a meta-analysis (even including thoracotomy cases), suggested that ICNB may be noninferior to TEA, positioning it as potential alternative to TEA or PVB in at least thoracoscopic surgery.25
Our findings on the counterbalancing benefits of ICNB and PVB align with current developments toward ERATS but seem not entirely captured by the QoR-15 questionnaire, which was selected as co-primary outcome. Based on a Delphi consensus, QoR is recommended as 1 of 6 important outcomes for evaluating perioperative patient comfort through standardized end points.26 Next to QoR, measures such as pain at less than 24 hours, PONV, and time to mobilization are also depicted as important indicators of patient comfort and were included as secondary outcomes in our trial. A drawback of QoR as co-primary outcome to counterbalance pain is that pain-related items are embedded in the QoR-15 tool. Consequently, the gain in QoR was attenuated, potentially obscuring superiority of the intervention techniques in ERATS outcomes such as mobility and LOS, which were not reflected in the mean QoR scores (eTable 8A and 8B in Supplement 3). Another study chose QoR-15 as the primary outcome and demonstrated noninferiority, rather than superiority, in patients undergoing continuous erector spinae block compared with TEA.27 To date, superiority in QoR has only been shown in patients with vs without a regional block in thoracic surgery.28 No studies have described superiority of locoregional techniques over TEA in terms of QoR.
When evaluating ERATS principles, postoperative complications are as important as mobility and LOS. Our analyses revealed no statistically significant differences in complications between the analgesic techniques. The most notable complication was PAL contributing to a slight difference in major complications between the TEA and ICNB groups. Although we attribute the higher incidence of PAL in the ICNB group to chance, we cannot exclude that improved patient mobility may influence any persistence of air leakage. We know that early mobilization is a strong predictor of reduced LOS in ERATS populations.29 Despite a lower incidence of PAL after TEA, it was associated with a longer LOS, probably due to reduced mobility.
We also emphasize the importance of patients’ preferences in defining to what extent postoperative pain is acceptable. Although we selected a debatable noninferiority margin for increased pain from a clinician’s perspective, patients may prioritize the benefits of locoregional analgesia in exchange for even more moments of pain. Notably, 41 patients declined participation because of strong opposition to TEA compared with only 9 who preferred it. It is of utmost importance to use our results as a guidance in shared decision-making during preoperative counselling.
Limitations
Our study encountered limitations. The pragmatic nature allowed flexibility of local in-house protocols, not mandating standard administration of opioids in the intervention groups, but only when needed. To the opposite, administration of opioids in the epidural solution is standard of care, which may have contributed to the lower pain scores but higher opioid consumption in the TEA group. Patients who used long-term preoperative opioids were excluded from participation as they were considered to require TEA for adequate analgesia. Next, approximately 40% of patients in the PVB and ICNB groups did not receive standard NSAIDs as part of a multimodal analgesic protocol, probably because of contraindications. Furthermore, the dropout rate slightly exceeded expectations (13.5% instead of 12.6%), potentially causing a minimal loss of statistical power for the QoR-15 outcome. Lastly, local in-house protocols regarding postoperative analgesia and ERATS implementation were important factors influencing our outcomes. To account for these variations, we applied a robust stratification method.
Conclusions
In patients undergoing thoracoscopic anatomical lung resection, ICNB was noninferior to TEA regarding pain, while results for PVB were inconclusive. Moreover, QoR scores were comparable across groups. Both PVB and ICNB were associated with reduced opioid consumption and PONV, as well as improved mobility. Additionally, ICNB was linked to a 1-day reduction in LOS. These findings suggest that ICNB may be a viable alternative to TEA; however, the risks and benefits of each analgesic technique should be weighted by clinicians and patients for adequate personalized pain control. A planned economic evaluation will demonstrate which strategy is more cost-effective.
Trial protocol
Statistical analysis plan
eTable 1. Performance of analgesic techniques in the ITT population
eTable 2. Baseline patient, surgical and pain characteristics for patients with and without missing data on the primary outcome QoR-15
eTable 3. Standardized mean differences for baseline characteristics in Table 1
eTable 4. Differences in primary outcomes between TEA and intervention groups after correcting for baseline characteristics
eTable 5. Mean pain scores at rest, during mobilization, and comparing between-group variation
eTable 6. Mean QoR-15 scores comparing the 5 domains of the QoR-15 questionnaire with a linear mixed model for repeated measurements
eFigure. Linear mixed model depicting mean predicted values of repeated measurements during postoperative day 0-3 and follow-up for pain scores in rest (S1a) and during mobilization (S1b)and QoR-15 scores (S1c) for the intervention groups (PVB and ICNB) and the control group (TEA).
eTable 7. Type of Clavien-Dindo 3 and 4 complications per group
eResults. Secondary outcomes not in manuscript
eTable 8. QoR-15 scores POD 1 and POD 2 per analgesic technique
Data sharing statement
References
- 1.Batchelor TJP, Rasburn NJ, Abdelnour-Berchtold E, et al. Guidelines for enhanced recovery after lung surgery: recommendations of the Enhanced Recovery After Surgery (ERAS®) Society and the European Society of Thoracic Surgeons (ESTS). Eur J Cardiothorac Surg. 2019;55(1):91-115. doi: 10.1093/ejcts/ezy301 [DOI] [PubMed] [Google Scholar]
- 2.Steinthorsdottir KJ, Wildgaard L, Hansen HJ, Petersen RH, Wildgaard K. Regional analgesia for video-assisted thoracic surgery: a systematic review. Eur J Cardiothorac Surg. 2014;45(6):959-966. doi: 10.1093/ejcts/ezt525 [DOI] [PubMed] [Google Scholar]
- 3.Spaans LN, Bousema JE, van den Broek FJC. Variation in postoperative pain management after lung surgery in the Netherlands: a survey of Dutch thoracic surgeons. Br J Anaesth. 2022;128(3):e222-e225. doi: 10.1016/j.bja.2021.12.005 [DOI] [PubMed] [Google Scholar]
- 4.Obuchi T, Yoshida Y, Moroga T, Miyahara N, Iwasaki A. Postoperative pain in thoracic surgery: re-evaluating the benefits of VATS when coupled with epidural analgesia. J Thorac Dis. 2017;9(11):4347-4352. doi: 10.21037/jtd.2017.09.133 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Feray S, Lubach J, Joshi GP, Bonnet F, Van de Velde M; PROSPECT Working Group *of the European Society of Regional Anaesthesia and Pain Therapy . PROSPECT guidelines for video-assisted thoracoscopic surgery: a systematic review and procedure-specific postoperative pain management recommendations. Anaesthesia. 2022;77(3):311-325. doi: 10.1111/anae.15609 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Harky A, Clarke CG, Kar A, Bashir M. Epidural analgesia versus paravertebral block in video-assisted thoracoscopic surgery. Interact Cardiovasc Thorac Surg. 2019;28(3):404-406. doi: 10.1093/icvts/ivy265 [DOI] [PubMed] [Google Scholar]
- 7.Spaans LN, Bousema JE, Meijer P, et al. Acute pain management after thoracoscopic lung resection: a systematic review and explorative meta-analysis. Interdiscip Cardiovasc Thorac Surg. 2023;36(1):ivad003. doi: 10.1093/icvts/ivad003 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Sertcakacilar G, Tire Y, Kelava M, et al. Regional anesthesia for thoracic surgery: a narrative review of indications and clinical considerations. J Thorac Dis. 2022;14(12):5012-5028. doi: 10.21037/jtd-22-599 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Lin J, Liao Y, Gong C, et al. Regional analgesia in video-assisted thoracic surgery: a bayesian network meta-analysis. Front Med (Lausanne). 2022;9:842332. doi: 10.3389/fmed.2022.842332 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Spaans LN, Dijkgraaf MGW, Meijer P, et al. ; OPtriAL study group . Optimal postoperative pain management after VATS lung resection by thoracic epidural analgesia, continuous paravertebral block or single-shot intercostal nerve block (OPtriAL): study protocol of a three-arm multicentre randomised controlled trial. BMC Surg. 2022;22(1):330. doi: 10.1186/s12893-022-01765-y [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Gerbershagen HJ, Rothaug J, Kalkman CJ, Meissner W. Determination of moderate-to-severe postoperative pain on the numeric rating scale: a cut-off point analysis applying four different methods. Br J Anaesth. 2011;107(4):619-626. doi: 10.1093/bja/aer195 [DOI] [PubMed] [Google Scholar]
- 12.Dindo D, Demartines N, Clavien PA. Classification of surgical complications: a new proposal with evaluation in a cohort of 6336 patients and results of a survey. Ann Surg. 2004;240(2):205-213. doi: 10.1097/01.sla.0000133083.54934.ae [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Bousema JE, Dias EM, Hagen SM, Govaert B, Meijer P, van den Broek FJC. Subpleural multilevel intercostal continuous analgesia after thoracoscopic pulmonary resection: a pilot study. J Cardiothorac Surg. 2019;14(1):179. doi: 10.1186/s13019-019-1003-y [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Dunnett CW. A multiple comparison procedure for comparing several treatments with a control. J Am Stat Assoc. 2021;50(272):1096-1121. doi: 10.1080/01621459.1955.10501294 [DOI] [Google Scholar]
- 15.Myles PS, Myles DB, Galagher W, Chew C, MacDonald N, Dennis A. Minimal clinically important difference for three quality of recovery scales. Anesthesiology. 2016;125(1):39-45. doi: 10.1097/ALN.0000000000001158 [DOI] [PubMed] [Google Scholar]
- 16.Beck N, Hoeijmakers F, van der Willik EM, et al. National comparison of hospital performances in lung cancer surgery: the role of case mix adjustment. Ann Thorac Surg. 2018;106(2):412-420. doi: 10.1016/j.athoracsur.2018.02.074 [DOI] [PubMed] [Google Scholar]
- 17.Ten Berge M, Beck N, Heineman DJ, et al. Dutch lung surgery audit: a national audit comprising lung and thoracic surgery patients. Ann Thorac Surg. 2018;106(2):390-397. doi: 10.1016/j.athoracsur.2018.03.049 [DOI] [PubMed] [Google Scholar]
- 18.Liang XL, An R, Chen Q, Liu HL. The analgesic effects of thoracic paravertebral block versus thoracic epidural anesthesia after thoracoscopic surgery: a meta-analysis. J Pain Res. 2021;14:815-825. doi: 10.2147/JPR.S299595 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Chenesseau J, Fourdrain A, Pastene B, et al. Effectiveness of surgeon-performed paravertebral block analgesia for minimally invasive thoracic surgery: a randomized clinical trial. JAMA Surg. 2023;158(12):1255-1263. doi: 10.1001/jamasurg.2023.5228 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Ardon AE, Lee J, Franco CD, Riutort KT, Greengrass RA. Paravertebral block: anatomy and relevant safety issues. Korean J Anesthesiol. 2020;73(5):394-400. doi: 10.4097/kja.20065 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Marhofer D, Marhofer P, Kettner S, et al. Magnetic resonance imaging analysis of the spread of local anesthetic solution after ultrasound-guided lateral thoracic paraverterbal blockade. Anesthesiology. 2013;118:1106-1112. doi: 10.1097/ALN.0b013e318289465f [DOI] [PubMed] [Google Scholar]
- 22.Lai J, Situ D, Xie M, et al. Continuous paravertebral analgesia versus continuous epidural analgesia after video-assisted thoracoscopic lobectomy for lung cancer: a randomized controlled trial. Ann Thorac Cardiovasc Surg. 2021;27(5):297-303. doi: 10.5761/atcs.oa.20-00283 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Luyet C, Siegenthaler A, Szucs-Farkas Z, Hummel G, Eichenberger U, Vogt A. The location of paravertebral catheters placed using the landmark technique. Anaesthesia. 2012;67(12):1321-1326. doi: 10.1111/j.1365-2044.2012.07234.x [DOI] [PubMed] [Google Scholar]
- 24.Shen L, Ye Z, Wang F, Sun GF, Ji C. Comparative analysis of the analgesic effects of intercostal nerve block, ultrasound-guided paravertebral nerve block, and epidural block following single-port thoracoscopic lung surgery. J Cardiothorac Surg. 2024;19(1):406. doi: 10.1186/s13019-024-02877-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Guerra-Londono CE, Privorotskiy A, Cozowicz C, et al. Assessment of intercostal nerve block analgesia for thoracic surgery: a systematic review and meta-analysis. JAMA Netw Open. 2021;4(11):e2133394. doi: 10.1001/jamanetworkopen.2021.33394 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Myles PS, Boney O, Botti M, et al. ; StEP–COMPAC Group . Systematic review and consensus definitions for the Standardised Endpoints in Perioperative Medicine (StEP) initiative: patient comfort. Br J Anaesth. 2018;120(4):705-711. doi: 10.1016/j.bja.2017.12.037 [DOI] [PubMed] [Google Scholar]
- 27.van den Broek RJC, Postema JMC, Koopman JSHA, et al. Continuous erector spinae plane block versus thoracic epidural analgesia in video-assisted thoracoscopic surgery: a prospective randomized open-label non-inferiority trial. Reg Anesth Pain Med. 2025;50(1):11-19. doi: 10.1136/rapm-2023-105047 [DOI] [PubMed] [Google Scholar]
- 28.Xu M, Zhang G, Tang Y, Wang R, Yang J. Impact of regional anesthesia on subjective quality of recovery in patients undergoing thoracic surgery: a systematic review and meta-analysis. J Cardiothorac Vasc Anesth. 2023;37(9):1744-1750. doi: 10.1053/j.jvca.2023.05.003 [DOI] [PubMed] [Google Scholar]
- 29.Rogers LJ, Bleetman D, Messenger DE, et al. The impact of enhanced recovery after surgery (ERAS) protocol compliance on morbidity from resection for primary lung cancer. J Thorac Cardiovasc Surg. 2018;155(4):1843-1852. doi: 10.1016/j.jtcvs.2017.10.151 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Trial protocol
Statistical analysis plan
eTable 1. Performance of analgesic techniques in the ITT population
eTable 2. Baseline patient, surgical and pain characteristics for patients with and without missing data on the primary outcome QoR-15
eTable 3. Standardized mean differences for baseline characteristics in Table 1
eTable 4. Differences in primary outcomes between TEA and intervention groups after correcting for baseline characteristics
eTable 5. Mean pain scores at rest, during mobilization, and comparing between-group variation
eTable 6. Mean QoR-15 scores comparing the 5 domains of the QoR-15 questionnaire with a linear mixed model for repeated measurements
eFigure. Linear mixed model depicting mean predicted values of repeated measurements during postoperative day 0-3 and follow-up for pain scores in rest (S1a) and during mobilization (S1b)and QoR-15 scores (S1c) for the intervention groups (PVB and ICNB) and the control group (TEA).
eTable 7. Type of Clavien-Dindo 3 and 4 complications per group
eResults. Secondary outcomes not in manuscript
eTable 8. QoR-15 scores POD 1 and POD 2 per analgesic technique
Data sharing statement
