INTRODUCTION
Postoperative pain management during lung transplantation (LT) is an important cornerstone for perioperative support, but represents a real challenge due to the major complexity of lung transplant patients and postoperative care [1]. Effective pain management is essential, as inadequate control of postoperative pain can lead to increased morbidity (atelectasis, pneumonia, etc.) and prolonged recovery times.
Most LT centers apply a multimodal analgesia strategy that combines several drug classes, including opioids, which are associated with numerous side effects [2,3]. In addition, some patients develop long-term chronic pain, some of which is due to suboptimal treatment of acute-phase pain [4]. Chronic pain is associated with increased drug intake, opioid dependence, and impaired quality of life [5,6].
Despite being the analgesic strategy of choice in thoracic surgery, the place of thoracic epidural analgesia (TEA) is less codified, probably because of the anatomical issues associated with the procedure, the emergency nature of this surgery, frequent disorders of haemostasis, major haemodynamic instability, and the frequent need for intraoperative ExtraCorporeal Membrane Oxygenation (ECMO) [7,8]. In addition, the incidence of failed epidural catheter insertion, poor efficacy, or accidental withdrawal is not negligible, ranging from 15 to 30% [9,10]. Nevertheless, in many centers around the world, TEA remains the gold standard for postoperative analgesia.
Intercostal nerve cryoanalgesia (INCA) is a long-standing non-pharmacological approach to reduce both postoperative chest pain and opioid consumption and to optimize the recovery of the forced expiratory volume in one second (FEV1) after thoracic surgery [11,12]. Using this strategy, a major reduction in post-thoracotomy pain has been described, with efficacy up to three months post-operatively [12]. The effectiveness of this technique led to its U.S. Food and Drug Administration approval in 2014. Its use in France has become more widespread in recent years, particularly after thoracotomy or to provide effective analgesia for stepped rib fractures. The objective is to block peripheral nerves to control pain for approximately two months, enabling post-operative recovery and rehabilitation [13,14]. This technique consists of the direct application of cold to the nerve to cool the axon while preserving the entire sheath. Elements such as the perineurium, endoneurium, and epineurium are spared [12].
With respect to LT, there are a few studies evaluating this technique. Salan-Gomez et al. [15] compared INCA with conventional multimodal analgesia (without TEA) and reported that patients who received INCA had lower opioid consumption and, in the medium term, a better FEV1. Another retrospective study involving 102 patients reported relatively similar results [16]. Isaza et al. [17], who compared patients who had received either TEA or INCA, showed no difference between the two therapies before and after adjustment for differences between groups. Finally, a recent meta-analysis revealed a statistically significant effect in favor of INCA versus control (TEA or multimodal analgesia) in terms of pain and opioid consumption. There was no difference between groups in length of stay or in the timing of endotracheal tube removal [18].
We recently implemented this technique in our LT program. Given that there is no randomized controlled prospective study evaluating INCA with TEA, which appears to be the gold standard in LT, we believe that combining the two techniques is important. In addition, as the analgesic effect of INCA is delayed by approximately 24 h—due to transient inflammatory pain related to cryoanalgesia of nerve fibers during the first 12 to 24 h post-procedure—we considered it essential to ensure optimal pain relief during the very early postoperative period.
The aim of our before-and-after study was therefore to compare morphine consumption at 28 days and to assess pain levels in patients with and without INCA.
MATERIALS AND METHODS
Study design
This retrospective before-and-after study was conducted between October 2024 and July 2025 in three departments at the Bichat-Claude Bernard Hospital, Paris, France: the anesthesia-intensive care unit (ICU), the thoracic surgery operating room, and the pulmonology department. The thoracic surgery department began using the INCA routinely in February 2025. In this context, we compared all LT patients who had benefited from TEA between October 2024 and February 2025 (PRE-INCA group) with those who had received both TEA and INCA between February and July 2025 (POST-INCA group).
Study group
All adult patients who underwent single- or double-LT at the Bichat-Claude Bernard Hospital during this period were included. The exclusion criteria were multi-organ transplants, lobar transplants, redo LT, and long-term treatment with morphinomimetics.
Ethics
This study was approved by the French Society of Anesthesiology and Critical Care Medicine Research Ethics Board (CALM study; No. 00010254 ‐ 2025-105, 21/09/2025), who waived the need for informed consent because of the observational nature of the study, according to French law.
Data collection
Patient demographics, diagnoses, types of surgical procedures (single- or double-LT, thoracotomy or clamshell) and clinical data were retrospectively collected. The clamshell incision was analyzed as a separate variable, as it may influence postoperative pain, especially when the patient suffers from sternal override, a complication of the clamshell incision that is not possible when the sternum is not cut.
Operative procedure
Regardless of the group, upon arrival in the operating room, each patient underwent TEA between T4 and T8. Once the epidural catheter was inserted, the patient received general anesthesia with multimodal hemodynamic and respiratory monitoring. In some cases, central veno-arterial ECMO was necessary depending on the haemodynamic and/or respiratory status. All patients received standard triple-drug induction immunosuppression associated with induction with basiliximab.
All lung transplants were performed by bilateral anterolateral thoracotomy or Clamshell incision depending on the surgeons’ habits, LT procedures (double or single-lung LT), and intraoperative difficulties. For those patients who underwent INCA, the procedure was performed by a senior surgeon after pneumonectomy and during the preparation of the donor lung allograft on the back table. Under rare circumstances, INCA was performed following reperfusion of the lungs during a period of hemostasis before the initiation of chest closure. Cryoanesthesia was performed at −60 to −80°C for 120 s per nerve level using the AtriCure® Cryoablation System (AtriCure®, Inc.). Under direct visualization, the surgeon identified the intercostal nerves and ablated them at T3 through T7 several centimeters lateral to the sympathetic chain.
After LT, all patients were systematically admitted to the ICU and received standard post-transplant immunosuppressive management, which included tacrolimus, mycophenolate and corticosteroids. Mechanical ventilation was discontinued when patients were awake and demonstrated hemodynamic and respiratory stability in minimal ventilator settings.
Pain management
Postoperatively, intubated patients were sedated with propofol, which was titrated according to the Richmond Agitation and Sedation Scale (RASS), in combination with continuous sufentanil infusions, the administration of which was guided by the Behavioral Pain Scale (BPS). Once clinical, radiological, and laboratory parameters related to hemodynamics, ventilation, and oxygenation were satisfactory, patients were gradually awakened to facilitate early extubation.
In our department, sufentanil is subsequently replaced by Class I analgesics, specifically acetaminophen and nefopam. These medications are administered intravenously and then orally as soon as oral intake is possible. Analgesia is not routinely administered but is assessed using the BPS for intubated and sedated patients and the Numeric Rating Scale (NRS) for extubated patients. The NRS score ranges from 0 (no pain) to 10 (worst possible pain). The recommended dose of acetaminophen is 1 g, with a maximum daily dose of 4 g. The recommended dosage of nefopam is 20 mg, with a maximum daily dose of 120 mg intravenously.
For patients who experience persistent pain despite first- and second-line analgesics after endotracheal tube removal, intravenous morphine is administered via a patient-controlled analgesia device, with bolus doses of 1 mg, a 7-min lockout interval, and a maximum total dosage of 30 mg over 4 h. As patients become more independent, ambulatory, and able to participate in physical and respiratory therapy, and as pain decreases, subcutaneous morphine is introduced. This route is preferred at this stage due to ease of administration and lower risk of adverse effects compared to intravenous morphine. The subcutaneous dose ranged from 5 to 10 mg and was administered 4-6 times daily, depending on patient weight. Upon transfer to the pulmonary medicine ward, patients who are capable of oral intake are switched to oral morphine at 5 to 10 mg, 4-6 times daily, according to weight. Laxatives and antiemetics are prescribed as needed to manage opioid-related side effects.
Epidural analgesia is initiated in the ICU as soon as the patient’s level of sedation permits. A 200 ml solution of ropivacaine (2 mg/ml) was supplemented with sufentanil (50 μg, 0.25 μg/ml). The regimen combines boluses of 2.5 ml (5 mg ropivacaine) with a 20-min lockout interval and a continuous infusion at 5 ml/h (10 mg/h ropivacaine), with a maximum dosage of 100 mg over 4 h. The TEA catheter was routinely removed on Day 5 (D5). This protocol was identical for both the PRE-INCA and POST-INCA groups.
During TEA, sensory and motor blocks are systematically monitored. Sensory block is assessed using a cold test, with the upper sensory level determined bilaterally, aiming for a level above T4. The motor block is evaluated using the Bromage score, which must remain at 0. The Bromage score is classified as follows: 0 - absence of motor block (flexion of hips, knees, and feet); 1 - inability to raise extended legs (flexion of knees and feet possible); 2 - inability to flex knees (flexion of feet possible); and 3 - inability to flex feet.
Outcomes
The primary outcome was total opioid consumption from Day 1 (D1) in the ICU to Day 28 (D28), which was measured in morphine milligram equivalent (MME) and calculated using standard conversion factors [19-21]. Sufentanil and morphine doses (intravenous, subcutaneous or oral) during the first 28 days were used for these calculations. Intraoperative sufentanil doses were not included in the overall 28-day analysis. The secondary outcomes were pain measured using NRS on D1, D2, D3, D4, D5, D6, D7, D28, 2 months (M2) and 3 months (M3); FEV1 at 1 month (M1), M2 and M3; duration of mechanical ventilation; length of stay in the ICU; and mortality at M1 and M3. Pain is measured in a patient at rest (six times a day), and the pain reported per day is thus averaged.
Statistical analysis
The distribution of variables was assessed using histograms and QQ plots. Quantitative variables were expressed as medians with interquartile ranges and were compared using the Wilcoxon rank-sum test. Categorical variables were reported as counts and percentages and were analyzed using the chi-square test or Fisher’s exact test, as appropriate. Univariate analyses were performed to examine patients’ baseline characteristics and clinical outcomes in relation to the analgesic protocol. A multivariable linear regression model was constructed, with post-operative D28 MME as the dependent variable. Given the right-skewed distribution of MME, we performed log-transformation of the primary outcome—i.e., log (MME)—prior to multivariable linear regression analysis. Independent variables were selected based on their clinical relevance and presumed clinical relationship with the primary outcome. The results are presented as exponentiated coefficients (exp(β)) with 95% confidence intervals (CIs), representing proportional changes in opioid consumption relative to the reference group. All the statistical analyses and figures were generated using R (version 4.3.3, The R Foundation). A two-sided P value < 0.05 was considered statistically significant.
RESULTS
Study group
Between October 2024 and July 2025, 30 LTs were performed consecutively at our hospital. Two patients were excluded from the analysis, one who underwent liver-LT and one who was treated with morphinomimetics before LT; 14 patients finally included in the PRE-INCA group, and 14 patients were included in the POST-INCA group. All patients had an epidural catheter inserted in the operating room. In the PRE-INCA group, epidural analgesia was not effective in one patient, which led to its removal on D2. In the POST-INCA group, one patient underwent accidental removal of the TEA catheter on D2. The patients’ general characteristics are summarized in Table 1.
Outcomes
The median duration of mechanical ventilation was 5 (2, 16) days, and the ICU and hospital lengths of stay were 21 (13, 34) and 54 (41, 72) days, respectively. The mortality rate at 30 days was 0%, and the mortality rate at 90 days was 3.8%. No differences were found between the groups. The outcomes are shown in Table 2.
The median total opioid consumption during the first 28 days in the overall population was 994 (505, 1,745) MMEs. Patients in the POST-INCA group received significantly fewer opioids during the 28 days post-LT (POST-INCA 543 [153, 1,050] MME vs. PRE-INCA 1,499 [972, 1,909] MME; P = 0.004), as shown in Fig. 1.
During the first 7 days, pain measurements were globally similar between groups, but less pain was perceived on D6 in the POST-INCA group (NRS PRE-INCA = 1.90 (0.00, 4.30) vs. NRS POST-INCA = 0.00 (0.00, 0.00); P = 0.024). At M1, pain was significantly lower in the POST-INCA group than in the PRE-INCA group (NRS PRE-INCA M1 = 2.25 (0.00, 4.50) vs. NRS POST-INCA M1 = 0.00 (0.00, 0.00); P = 0.010). Epidural ropivacaine and sufentanil consumption were similar between the groups. FEV1 measurements at M1, M2, and M3 were similar between groups. These results are expressed in Table 3.
The cumulative consumption of acetaminophen over 28 days was significantly lower in the POST-INCA group (PRE-INCA: 43 [30, 50] g vs. POST-INCA: 8 [6, 15] g; P < 0.001). Cumulative consumption of Nefopam was similar between the two groups (PRE-INCA: 100 [0, 600] mg vs. POST-INCA: 0 [0, 80] mg; P = 0.130). These results are expressed in Table 3. Oral lactulose and oral domperidone consumption at 28 days were significantly lower in the POST-INCA group (oral lactulose: PRE-INCA: 480 [400, 720] mg vs. POST-INCA: 120 [0, 280] mg; P < 0.001, oral domperidone: PRE-INCA: 0 mg [0, 300] mg vs. POST-INCA: 0 [0, 0] mg; P = 0.023). Intravenous ondansetron was similar between groups (PRE-INCA: 0 mg [0, 40] mg vs. POST-INCA: 0 mg [0, 0] mg; P = 0.061). These results are expressed in Table 3.
Multivariate analysis
A multivariable linear regression model was constructed, with post-operative D28 opioid consumption expressed in log(MME) as the dependent variable. After adjusting for surgical protocol, type of incision, and the number of transplanted lungs, INCA was significantly associated with an 87% reduction in opioid consumption on D28 (exp(β) = 0.13, 95% CI: 0.03-0.55; P = 0.01) (Table 4).
DISCUSSION
In this retrospective study, we reported that adding INCA to TEA significantly reduced opioid consumption on D28 and postoperative pain following LT.
Numerous retrospective studies on LT have shown decreased opioid consumption [15-17]. All of these studies compared this device with conventional anesthesia techniques involving TEA and/or the use of intravenous opioids. We considered that, in the absence of a randomized prospective study comparing TEA vs. INCA, it did not seem ethical to propose INCA without TEA. Our work is innovative because INCA significantly reduced post-operative pain and opioid consumption, even though all patients received TEA, which is regarded as the gold standard in many LT centers. The synergy between the two techniques probably makes it possible to truly reduce pain and thus decrease opioid consumption. In addition, the TEA was left in place for 5 days, with a secondary switch to opioids. The residual effect of INCA for M2 to M3 would thus allow for an effective reduction in opioid consumption after these 5 days of TEA. Interestingly, the use of cryoanalgesia also resulted in a significant reduction in acetaminophen doses at 28 days, underscoring the effectiveness of this analgesic technique. Furthermore, although adverse effects related to morphine use (such as constipation, nausea, and vomiting) were not directly collected, we indirectly observed reduced use of laxatives and antiemetics in the POST-INCA group. These findings support the potential benefit of reduced morphine requirements associated with the use of INCA.
With regard to the other parameters, our study is consistent with other studies, particularly with similar outcome data between groups. Interestingly, despite the reduction in pain between D6 and D28 in the POST-INCA group compared with the PRE-INCA group, no significant differences were observed in the duration of mechanical ventilation or length of stay. Several explanations may account for this. First, pain is not the primary determinant of outcomes in the ICU, where postoperative complications following LT—such as infections, acute rejection, or anastomotic dehiscence—may delay discharge. Second, difficult weaning from mechanical ventilation is common in this population due to infections, atelectasis, fluid overload, and other factors, meaning that improved pain control does not necessarily translate to a shorter ventilation duration. Finally, the small sample size (n = 28) likely limits the statistical power of our analysis. In addition, we did not observe any difference between groups in terms of FEV1. Studies are rather disparate for this parameter, with some studies showing an increase in this parameter compared with the control group and other studies finding no difference. In addition to the small number of participants in our study, it should be noted that this measurement is difficult to perform postoperatively because many independent factors must be considered, such as the presence of atelectasis, postoperative pulmonary infection, or pleural drains, and these factors make interpretation complex.
Interestingly, in our center, only two out of 28 (7%) epidural administrations did not work or were accidentally removed. This is lower than the 20-30% reported in the literature, probably due to a high level of expertise in this technique in our center. Nevertheless, this 7% reduction in effectiveness should still be considered, as it potentially increases opioid consumption.
Our study has several limitations. First, this was a retrospective analysis with a small sample size. Given the susceptibility of linear regression models to limited sample sizes, the results of the multivariable analysis should be considered exploratory and interpreted with caution. Nevertheless, despite the limited number of patients, both univariate and multivariable analyses consistently suggested a beneficial effect of INCA supplementation. Secondly, we did not measure neuropathic pain in this study because we did not systematically record it in the electronic health records at our center. This information could have been very informative. Thirdly, we did not measure analgesia data or long-term outcomes (6 months, one year, etc.). However, as the effective duration of INCA is M2 to M3, we did not consider it relevant to measure pain or opioid consumption in the longer term. In addition, given the small number of patients, we did not stratify by the surgeons who performed the INCA or the anesthesiologists who performed the TEA. Although this would have been an interesting concept, we did not consider it ethical. Finally, cryoanalgesia may induce inflammatory pain during the first 24 h of treatment, likely due to inflammation following nerve fiber injury (manufacturer data), which may necessitate additional analgesic therapy during this period. Given that all the patients in our study underwent TEA, we were unable to assess this aspect in our analysis.
In conclusion, our study suggests that adding INCA to TEA significantly reduces opioid consumption on D28 and decreases postoperative pain following LT. Although our study adds to the body of evidence, only a prospective randomized controlled trial comparing INCA to TEA will allow us to conclude whether INCA alone is beneficial.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation
