Liposomal Bupivacaine Incisional Infiltration for Postoperative Analgesia After Cytoreductive Surgery with Hyperthermic Intraperitoneal Chemotherapy: A Randomized Placebo-Controlled Trial

Abstract

Introduction

Moderate-to-severe postoperative pain is common after cytoreductive surgery (CRS) with hyperthermic intraperitoneal chemotherapy (HIPEC). Evidence that liposomal bupivacaine provides clinically important advantages over conventional local anesthetics is inconsistent, and evidence in CRS + HIPEC remains limited. This trial evaluated the incremental efficacy and safety of liposomal bupivacaine incisional infiltration added to patient-controlled intravenous analgesia (PCIA) compared with saline infiltration plus PCIA.

Methods

In this single-center, randomized, participant- and assessor-blinded, placebo-controlled trial, adults aged 18–65 years scheduled for elective CRS + HIPEC were assigned 1:1 to standardized PCIA plus liposomal bupivacaine incisional infiltration (PCIA-LB) or PCIA plus normal saline infiltration (PCIA-NS). The primary endpoint was the incidence of moderate-to-severe movement-evoked incisional pain (numeric rating scale >  = 4) within 72 h. The prespecified primary efficacy analysis used the per-protocol set (PPS), with a conservative intention-to-treat (ITT) sensitivity analysis. PCIA followed an institutional standard operating procedure. Monitoring was continuous in the intensive care unit (ICU); after transfer to the ward, pulse oximetry was continued through postoperative day 3, with intermittent electrocardiography and noninvasive blood pressure monitoring.

Results

The PPS included 94 patients (PCIA-NS, n = 48; PCIA-LB, n = 46). Moderate-to-severe movement-evoked pain within 72 h was less frequent with PCIA-LB than with PCIA-NS (19.6% versus 70.8%, P < 0.001). The conservative ITT sensitivity analysis supported the same conclusion (26.0% versus 68.0%, P < 0.001). PCIA-LB also produced a modest reduction in cumulative 72-h study-recorded oral morphine milligram equivalents (556.8 [473.4, 600.0] versus 600.0 [557.4, 630.0] mg, P = 0.002) and reduced rescue analgesia (30.4% versus 54.2%, P = 0.020). QoR-15 scores did not differ significantly. Overall analgesia-related adverse events occurred in 23.9% and 10.4% of the PCIA-LB and PCIA-NS groups, respectively (P = 0.143). No participant required naloxone, emergency airway intervention, or ventilatory support for a clinically recognized opioid-related respiratory event, and no local anesthetic systemic toxicity, incisional infection, or fat liquefaction was observed.

Conclusions

Compared with saline infiltration, liposomal bupivacaine incisional infiltration reduced early movement-evoked incisional pain and rescue analgesia and produced a statistically significant but quantitatively modest reduction in 72-h opioid consumption captured by the study records after CRS + HIPEC. It did not improve QoR-15. These findings do not establish superiority over plain bupivacaine, ropivacaine, or catheter-based regional techniques, and the study was not powered to establish safety.

Trial Registration

Chinese Clinical Trial Registry, ChiCTR2500100826. Registered at 00:00 on 15 April 2025, before enrollment of the first participant later the same day.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1007/s40122-026-00876-1.

Plain Language Summary

Patients undergoing cytoreductive surgery with hyperthermic intraperitoneal chemotherapy often experience substantial pain during the first few days after surgery. Opioid-based patient-controlled intravenous analgesia is widely used, but it may not fully control pain and can cause adverse effects. Liposomal bupivacaine is a long-acting local anesthetic formulation designed to release bupivacaine gradually at the wound site. In this randomized placebo-controlled trial, patients received the same institutional patient-controlled intravenous analgesia pathway and were additionally assigned to either liposomal bupivacaine or normal saline infiltration along the abdominal incision. Patients who received liposomal bupivacaine had less moderate-to-severe movement-evoked incisional pain and required fewer rescue analgesic interventions during the first 72 h; the reduction in opioid consumption captured by the study records was statistically significant but modest. Overall recovery quality measured by the 15-item Quality of Recovery score at 72 h was similar between groups. Analgesia-related adverse events were numerically more frequent in the liposomal bupivacaine group, although the difference was not statistically significant, and no severe local-anesthetic-related event was observed. Because the study did not include plain bupivacaine or ropivacaine as an active comparator, it cannot determine whether liposomal bupivacaine is superior or cost-effective compared with less expensive conventional local anesthetics.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1007/s40122-026-00876-1.

Key Summary Points

Why carry out this study?

Moderate-to-severe postoperative pain remains common after cytoreductive surgery with hyperthermic intraperitoneal chemotherapy (CRS + HIPEC) and may impair coughing, mobilization, and early recovery.

This randomized trial asked whether liposomal bupivacaine incisional infiltration could improve early postoperative incisional analgesia when added to an institutional patient-controlled intravenous analgesia (PCIA) pathway.

What was learned from this study?

Compared with saline infiltration, liposomal bupivacaine incisional infiltration reduced moderate-to-severe movement-evoked incisional pain within 72 h and reduced rescue analgesia; the reduction in study-recorded opioid consumption was statistically significant but modest.

The 15-item Quality of Recovery score (QoR-15) at 72 h did not differ significantly between groups, and the study was not powered to establish uncommon safety outcomes or broader recovery benefits.

This placebo-controlled trial does not determine whether liposomal bupivacaine is superior or cost-effective compared with plain bupivacaine, ropivacaine, thoracic epidural analgesia, or single-shot or continuous fascial plane techniques.

Introduction

Peritoneal metastasis is an advanced manifestation of abdominal malignancies, such as gastric, colorectal, and ovarian cancer, and remains a major clinical challenge. Cytoreductive surgery combined with hyperthermic intraperitoneal chemotherapy (CRS + HIPEC) is an important treatment strategy for carefully selected patients with peritoneal malignancy [1]. This approach aims to remove macroscopic tumor deposits and to eradicate residual microscopic disease through heated intraperitoneal chemotherapy. However, CRS + HIPEC is highly invasive. The incision commonly extends from the xiphoid process to the pubic symphysis and is approximately 30–40 cm long, and the procedure is often accompanied by extensive peritonectomy and multivisceral resection. Consequently, postoperative pain is frequently intense, prolonged, and multifactorial [2]. Inadequate control of moderate-to-severe postoperative pain may impair turning, coughing, deep breathing, and early ambulation, and may increase the risk of pulmonary infection, cardiovascular events, and deep venous thrombosis, leading to prolonged hospitalization and increased healthcare burden [3].

Current postoperative analgesic strategies include thoracic epidural analgesia (TEA), patient-controlled intravenous analgesia (PCIA), single-injection or continuous fascial plane techniques (including transversus abdominis plane, rectus sheath, quadratus lumborum, and erector spinae plane blocks), continuous wound or rectus sheath catheters, and local incisional infiltration. Recent guidance supports fascial plane blocks as effective components of multimodal analgesia for open abdominal surgery [4]. TEA remains an important option in centers with experienced acute pain teams when coagulation status and platelet counts are acceptable and no contraindications exist [5–9], but it is not the only regional strategy. Four-quadrant TAP block has been directly compared with TEA in CRS + HIPEC [10], and rectus sheath catheter analgesia has been evaluated after open midline major abdominal surgery [11]. The present study was not designed to refute or replace these techniques. Rather, it evaluated a practical, catheter-free local analgesic strategy for situations in which neuraxial or catheter-based regional analgesia is contraindicated, not selected, or requires supplementation. At the time of study design, local infiltration was a labeled route for liposomal bupivacaine in our setting, whereas TAP and rectus sheath blocks were not included among the labeled regional block indications; this was an additional practical reason for selecting direct-vision incisional infiltration. In our institution, PCIA remains a foundational component of the postoperative analgesic pathway. Combining PCIA with a simple and reproducible local technique may therefore be clinically relevant, although comparative effectiveness must be established against active regional and local-anesthetic comparators.

Among local analgesic techniques, incisional infiltration is attractive because of its simplicity and direct access to the wound. Conventional local anesthetics, such as bupivacaine hydrochloride or ropivacaine, generally have a shorter duration than the first 72 h of recovery after major laparotomy. Liposomal bupivacaine is designed to release bupivacaine gradually [12]. Studies in several surgical settings have reported potential benefits [13–15], but the evidence is mixed: a Cochrane review found insufficient evidence to establish clinically meaningful superiority over bupivacaine hydrochloride, and a recent meta-analysis of abdominal fascial plane blocks found similar analgesic effectiveness between liposomal and plain local anesthetics [16, 17]. Accordingly, an active-comparator trial is required to determine whether the additional cost of liposomal bupivacaine is justified. The present placebo-controlled trial addressed a narrower question: whether adding liposomal bupivacaine incisional infiltration to the institutional PCIA pathway provides incremental analgesia compared with saline infiltration. CRS + HIPEC involves a long abdominal incision and extensive tissue trauma. We used a two-layer, multipoint infiltration technique along the full incision to increase contact with pain-sensitive tissues while keeping the total liposomal bupivacaine dose within the labeled adult maximum.

The primary objective of this study was to determine whether liposomal bupivacaine incisional infiltration added to standardized PCIA reduces the incidence of moderate-to-severe movement-evoked incisional pain within 72 h after CRS + HIPEC compared with PCIA plus normal saline infiltration. Movement-evoked incisional pain was selected because pain from a long midline incision directly affects turning, coughing, deep breathing, and early ambulation. The trial was not designed to compare liposomal bupivacaine with plain bupivacaine, ropivacaine, TEA, or ultrasound-guided regional techniques. Given the patient-reported nature of pain scores, opioid consumption at 24, 48, and 72 h, cumulative 72-h oral MME, effective PCA activations, and rescue analgesia were prespecified as key secondary outcomes. Additional secondary outcomes included pain at rest, QoR-15, postoperative hospital stay, and analgesia-related adverse events.

Methods

Study Design and Registration

This was a single-center, randomized, participant- and assessor-blinded, placebo-controlled, parallel-group clinical trial.

The study protocol was approved by the Ethics Committee of Beijing Tsinghua Changgung Hospital (approval number: 25038-4-01). The trial was conducted in accordance with the Declaration of Helsinki and CONSORT reporting principles. Written informed consent was obtained from all participants before enrollment, and patient-identifiable data were anonymized.

The trial was registered in the Chinese Clinical Trial Registry on 15 April 2025 at 00:00 (ChiCTR2500100826; https://www.chictr.org.cn). The first participant was enrolled after registration on 15 April 2025; therefore, the trial was prospectively registered. The study was conducted in accordance with the registered protocol and the ethics-approved protocol. The prespecified primary outcome was the incidence of moderate-to-severe movement-evoked incisional pain within 72 h after surgery. Prespecified secondary outcomes included pain at rest, effective PCA activations, rescue analgesia, 72-h MME, QoR-15 score, postoperative hospital stay, and analgesia-related adverse events. The report was prepared with reference to CONSORT reporting recommendations for randomized trials. The trial protocol and statistical analysis plan are available from the corresponding authors on reasonable request. No formal patient or public involvement occurred in the design, conduct, reporting, or dissemination planning of this trial.

Participants

Eligible participants were adults aged 18–65 years scheduled for elective CRS + HIPEC under general anesthesia, with an expected incision from the xiphoid process to the pubic symphysis or an incision length > 30 cm; body mass index 18–28 kg/m²; American Society of Anesthesiologists physical status I-III as specified in the approved protocol; and agreement to use postoperative opioid-based PCIA. No ASA I participant was ultimately enrolled, and all analyzed participants were ASA II or III. Exclusion criteria were severe hepatic or renal dysfunction (estimated glomerular filtration rate < 30 ml/min/1.73 m² or Child–Pugh class C), known hypersensitivity to bupivacaine, any medication used in the study, or any component of the study formulations, coagulation dysfunction (platelet count < 100 × 10⁹/L and/or international normalized ratio >  = 1.5), refusal to participate, history of alcohol or drug abuse, psychiatric disease or impaired mental status, chronic opioid use for more than 3 months, and chronic pain lasting more than 3 months.

Protocol Deviations and Analysis Populations

The following conditions were defined as protocol deviations or reasons for exclusion from the per-protocol set (PPS): participant withdrawal or loss to follow-up; postoperative intensive care unit (ICU) stay > 24 h for surgical reasons, preventing scheduled pain follow-up; failure to use PCIA for surgical reasons; discontinuation of PCIA for surgical reasons; failure to perform study-drug or control infiltration for perioperative safety reasons; or postoperative use of analgesic treatments outside the study protocol. The primary efficacy analysis used the PPS, defined as participants who completed the study intervention, used PCIA postoperatively according to the protocol, and completed the primary outcome assessment. The PPS was selected for the primary efficacy analysis because some randomized patients did not actually receive or complete the prespecified intervention owing to surgery cancellation, failure to receive study-drug/control infiltration, prolonged ICU stay, nonprotocol PCIA use, or inability to complete the primary outcome assessment. The PPS therefore evaluated the treatment effect among patients who completed the intervention and postoperative analgesic management as planned. To assess the influence of post-randomization exclusions and missing primary outcome data, a conservative ITT sensitivity analysis included all 100 randomized participants according to their original randomized allocation. Missing primary outcome data were imputed using the assumption least favorable to the intervention group: missing outcomes in the PCIA-LB group were treated as occurrence of moderate-to-severe movement-evoked pain within 72 h, whereas missing outcomes in the PCIA-NS group were treated as nonoccurrence of the primary outcome. Because patients excluded in the CONSORT flow did not have evaluable postoperative safety data, safety analyses used the evaluable safety population, defined as patients who actually received study-drug or control infiltration and had postoperative safety data; in this study, this population was identical to the PPS.

Randomization and Blinding

Participants were randomized 1:1 to PCIA plus liposomal bupivacaine incisional infiltration (PCIA-LB group) or PCIA plus normal saline incisional infiltration (PCIA-NS group) using a random-number table. The random sequence was generated and stored by an independent researcher who was not involved in postoperative follow-up, outcome assessment, or statistical analysis. Group allocation was disclosed to the surgeon performing infiltration only before abdominal closure, after enrollment had been completed, to minimize early allocation disclosure. Participants, postoperative nurses, pain-management investigators, QoR-15 assessors, and statisticians were blinded to group allocation, and these personnel were not permitted to enter the operating room during infiltration. Because liposomal bupivacaine injection is milky white whereas normal saline is transparent, the surgeons performing the local infiltration could not be blinded. To reduce performance and co-intervention bias, infiltration was performed by two fixed surgeons who received standardized training and followed prespecified injection layers, intervals, and total volume. Intraoperative anesthesia, PCIA formulation, PCIA parameter adjustment, rescue analgesia thresholds, and management algorithms were standardized. Postoperative analgesic management was performed by blinded nurses and pain-management investigators. Any nonprotocol analgesia, additional regional block, PCIA discontinuation, ICU stay > 24 h, or other co-intervention was recorded as a protocol deviation and reported in the CONSORT flow diagram.

Anesthesia and Perioperative Management

After arrival in the operating room, standard monitoring was applied, including five-lead electrocardiography, noninvasive blood pressure, and pulse oximetry. Peripheral venous access was established, and radial arterial catheterization was performed for invasive arterial blood pressure monitoring. Anesthesia was induced with intravenous lidocaine 1 mg/kg, propofol 1.5–2 mg/kg, sufentanil 0.2–0.5 µg/kg, and rocuronium 0.6–0.8 mg/kg. After loss of eyelash reflex and adequate neuromuscular relaxation, tracheal intubation was performed. Volume-controlled ventilation was used with a tidal volume of 6–8 ml/kg, oxygen flow 1.5 L/min, inspired oxygen fraction <  = 60%, and end-tidal carbon dioxide maintained at 35–50 mmHg. A central venous catheter was placed after intubation, and central venous pressure was monitored. Anesthesia was maintained with sevoflurane combined with dexmedetomidine (0.2–0.7 µg/kg/h) and remifentanil (0.05–0.15 µg/kg/min). Analgesic and sedative doses were adjusted according to surgical progress and vital signs. Depth of anesthesia was monitored, and the bispectral index was maintained at 40–60. Rocuronium was supplemented intermittently. Intraoperative fluid therapy was individualized according to hemodynamic status, surgical blood loss, urine output, the volume of ascites drainage, the extent of peritoneal stripping, anticipated ongoing fluid and protein losses, and cumulative crystalloid requirement. Lactated Ringer solution was used as the primary replacement fluid. Then, 5% or 20% human albumin was administered selectively when additional intravascular volume expansion or replacement of substantial protein loss was clinically considered necessary; albumin was not administered routinely to every participant. Exact albumin volumes were not collected as study variables. This practice was consistent with CRS + HIPEC consensus and ERAS recommendations supporting individualized goal-directed fluid management and recognizing albumin as the preferred colloid when a colloid is indicated [8, 18, 19]. Mean arterial pressure was maintained at > 60 mmHg or within 20% of baseline. Red-cell transfusion was not initiated solely at a fixed hemoglobin threshold. Because CRS + HIPEC is prolonged and associated with substantial ongoing and anticipated blood loss, when intraoperative hemoglobin decreased to approximately 8 g/dL before completion of major surgical steps, blood was requested from the transfusion service. Actual transfusion could be initiated before hemoglobin fell below 8 g/dL in the presence of active major bleeding or impaired hemodynamic or oxygen-delivery status. The decision and timing of transfusion were individualized according to ongoing and anticipated blood loss, hemodynamics, lactate concentration, evidence of tissue hypoperfusion, expected further bleeding, and cardiovascular comorbidity. Thus, approximately 8 g/dL represented an operational trigger for blood preparation in this setting, not a universal automatic transfusion threshold; current guidelines generally recommend considering transfusion at < 7 g/dL in hemodynamically stable adults while emphasizing the overall clinical context [20]. For prevention of postoperative nausea and vomiting, dexamethasone 5 mg was administered during induction and dolasetron 12.5 mg was administered intravenously at the end of surgery. After surgery, patients were routinely transferred to the ICU. In selected awake patients who remained intubated, a low-dose remifentanil infusion could be administered according to the same ICU standard operating procedure to attenuate endotracheal-tube-related discomfort and airway responses. The exact infusion duration and cumulative remifentanil dose were not available in the study dataset and were therefore not included in the MME calculations. By the ward-based postoperative day-1 assessment, most patients had discontinued remifentanil, undergone extubation, and returned to the ward. After extubation, intravenous sufentanil 5 µg could be administered for rescue analgesia and repeated when clinically required; each documented dose was retained in the postoperative day-1 rescue-analgesia record and included in the 72-h MME calculation.

Interventions

In the PCIA-LB group, 20 ml of liposomal bupivacaine injection (Hengrui; approval number H20223899; 20 ml:266 mg) was mixed with 80 ml normal saline to prepare a total volume of 100 ml, corresponding to a final bupivacaine concentration of 2.66 mg/ml. The expanded volume was selected to provide multi-point coverage along the approximately 35-cm incision without changing the total dose and remained above the minimum diluted concentration stated in the prescribing information [21]. In the PCIA-NS group, 100 ml normal saline was used as the control solution. At the time of the trial, routine local-anesthetic incisional infiltration was not part of the institutional standard analgesic pathway for CRS + HIPEC; both groups otherwise received the same PCIA, monitoring, and rescue-analgesia pathway. After fascial closure, two trained surgeons performed direct-vision incisional infiltration along both sides of the full-length midline incision in two tissue planes: the musculofascial plane immediately superficial to the closed anterior abdominal fascia/rectus sheath and the subcutaneous/subdermal plane. The preperitoneal or peritoneal plane was not infiltrated. At intervals of 1–2 cm, 2–3 ml of study solution was injected at each point, with a total volume of 100 ml. All participants received postoperative PCIA according to the same institutional standard operating procedure. The PCIA solution consisted of sufentanil 200 µg, flurbiprofen axetil 200 mg, dolasetron 37.5 mg, and normal saline to a total volume of 250 ml. Neither intraoperative acetaminophen nor scheduled postoperative acetaminophen was included in the institutional analgesic pathway during the study period. The SOP permitted omission of flurbiprofen axetil when a clinically relevant NSAID contraindication was present; however, no analyzed participant in either group had such a contraindication, and all analyzed participants received the same flurbiprofen-containing PCIA formulation. Because the PCIA solution contained sufentanil 200 µg in 250 ml (0.8 µg/ml), initial pump settings were calculated individually according to body weight. The background infusion was set to deliver sufentanil 0.04 µg/kg/h, equivalent to 0.05 ml/kg/h, and the PCA bolus was set to deliver sufentanil 0.035 µg/kg, equivalent to approximately 0.044 ml/kg per bolus. The calculated doses were converted to milliliters according to the solution concentration, and the initial lockout interval was 20 min. Rescue-related adjustment for inadequate analgesia was limited to shortening a 20-min PCA lockout interval once to 15 min, as described below; no further shortening was permitted, and the PCA bolus volume and background infusion rate were not increased as part of the rescue algorithm. Reduction or discontinuation of the background infusion was permitted only for adverse effects according to the institutional SOP. Monitoring was continuous in the ICU. After transfer to the general ward, bedside pulse oximetry was continued continuously through postoperative day 3, while electrocardiography and noninvasive blood pressure were monitored intermittently. All patients receiving PCIA were managed under this monitoring pathway. In this center, patients undergoing CRS + HIPEC are routinely transferred to the ICU postoperatively and usually return to the ward on the morning of postoperative day 1. Patients who remained in the ICU for > 24 h for reasons unrelated to the study drug were excluded from the PPS because scheduled ward-based pain follow-up could not be completed.

Pain Assessment and Rescue Analgesia

A blinded pain-management investigator assessed pain once daily at 16:00 in the ward after ICU transfer and recorded the patient’s worst pain during the preceding 24 h. Pain was assessed using an 11-point numeric rating scale (NRS), where 0 indicates no pain, 1–3 mild pain, 4–6 moderate pain, and 7–10 severe pain [22, 23]. The NRS was selected because it is validated, practical, and readily completed by postoperative patients. An NRS score >  = 4 was considered moderate-to-severe pain or inadequate pain control [3, 22, 23]. Movement-evoked incisional pain referred to the worst incisional pain during turning, coughing, or ambulation in the preceding 24 h. If intolerable pain occurred at any time, blinded nurses immediately notified the pain-management investigator, who performed bedside assessment. Rescue analgesia was initiated when resting NRS was >  = 4 and the patient requested additional analgesia. According to a prespecified stepwise institutional SOP applied identically to both groups by blinded pain-management personnel, the first-line rescue intervention was a clinician-administered 5-ml loading dose from the PCIA solution, containing 4 µg of sufentanil, 4 mg of flurbiprofen axetil, and 0.75 mg of dolasetron. No more than one 5-ml loading dose was permitted during each postoperative 24-h period. Pain intensity, level of consciousness, respiratory status, pulse oxygen saturation, blood pressure, and electrocardiography were reassessed 10 min after administration. If analgesia remained inadequate and the current lockout interval was 20 min, the interval was shortened once to 15 min; if it was already 15 min, no further shortening was permitted. The PCA bolus volume and background infusion rate remained unchanged. If pain persisted, the ward physician was contacted and 5 mg dezocine was administered intramuscularly. If another episode of inadequate analgesia occurred during the same postoperative 24-h period after the permitted loading dose had already been administered, an additional loading dose was not allowed, and management proceeded through the remaining applicable steps of the algorithm. Each measure was recorded separately as one rescue analgesic intervention: administration of the 5-ml loading dose, shortening of the lockout interval, each intramuscular administration of dezocine, and each documented 5-µg intravenous sufentanil dose administered after extubation in the ICU were each counted as one intervention. The study database retained only the aggregate number of rescue analgesic interventions and did not preserve separate category-level counts for each intervention type. These prespecified or protocol-allowed measures were part of the study analgesic pathway and were not classified as protocol deviations. Any additional nonprotocol analgesic intervention was recorded as a protocol deviation.

Management of PCIA-related adverse reactions was prespecified. For dizziness, the background infusion rate was reduced or stopped and the pump was changed to PCA-only mode, with 10 mg of diphenhydramine intravenously if needed. For nausea or vomiting, the background infusion rate was reduced or stopped, and 12.5 mg of dolasetron intravenously was available as symptom-triggered rescue treatment after clinical reassessment; 1 mg of droperidol intravenously was available as second-line rescue treatment for severe or persistent nausea or vomiting when first-line measures were inadequate. Droperidol was permitted for clinical use at our center but was not used as a routine first-line prophylactic antiemetic. Cumulative exposure to additional rescue dolasetron was not analyzed as a separate study outcome. Mild pruritus required no treatment; clinically significant pruritus was managed by PCIA adjustment, 10 mg of diphenhydramine intravenously, and 0.1–0.2 mg of slow intravenous naloxone if necessary. Nausea, vomiting, dizziness, and pruritus were prospectively recorded in the CRF as present or absent; no validated severity grading scale was prespecified. Sedation, somnolence, drowsiness, respiratory rate, and level of consciousness were not prespecified or retained as structured research outcomes, and no standardized sedation scale was used. Routine clinical monitoring was continued, and clinically recognized opioid-related respiratory events requiring naloxone, emergency airway intervention, or ventilatory support were recorded. Suspected respiratory depression was managed with oxygen, reduction or discontinuation of the PCIA infusion, and 0.1–0.2 mg of naloxone if required.

Outcomes

The sole primary outcome was the incidence of moderate-to-severe movement-evoked incisional pain within 72 h after surgery, defined as NRS >  = 4 during movement at any assessment on postoperative days 1, 2, or 3. Movement-evoked incisional pain was defined as the worst incisional pain during turning, coughing, or ambulation in the preceding 24 h and was assessed at 16:00 each day by a blinded investigator. The NRS >  = 4 threshold was selected because it generally corresponds to at least moderate pain and may interfere with coughing, deep breathing, and early mobilization [3, 22, 23]. The worst pain at rest during each preceding 24-h period was prospectively collected as a prespecified secondary outcome and was not used for the sample-size calculation. Other secondary outcomes were effective PCA activations recorded at 24, 48, and 72 h; rescue analgesia within 72 h; MME at 24, 48, and 72 h and cumulative 72-h MME; QoR-15 at 72 h; analgesia-related adverse events, including dizziness, nausea, vomiting, pruritus, clinically documented opioid-related respiratory events requiring naloxone, emergency airway intervention, or ventilatory support, local anesthetic systemic toxicity, wound infection, and fat liquefaction; and postoperative length of stay. Rescue analgesia was analyzed at both the patient and intervention levels. At the patient level, rescue analgesia was defined as receipt of at least one protocol-defined rescue analgesic intervention within 72 h. At the intervention level, administration of a 5-ml PCIA loading dose, shortening of the PCA lockout interval from 20 to 15 min, each intramuscular administration of dezocine, and each documented 5-µg intravenous sufentanil dose administered after extubation in the ICU were counted separately as one rescue analgesic intervention. The study database retained only the aggregate rescue-intervention count and did not preserve separate counts for each intervention type. Consequently, one patient could contribute more than one intervention during follow-up. The 5-ml loading dose was permitted no more than once during each postoperative 24-h period. The QoR-15 is a multidimensional patient-reported measure of postoperative recovery quality [24]. It contains five dimensions: physical comfort, physical independence, pain, emotional state, and psychological support. Each item is scored on a 0–10 scale, with negative items reverse-scored. The total score ranges from 0 to 150, with higher scores indicating better recovery.

Statistical Analysis

Sample size was estimated using the incidence of moderate-to-severe movement-evoked pain within 72 h after surgery as the primary endpoint. On the basis of reported postoperative moderate-to-severe pain after abdominal surgery and local CRS + HIPEC pain-management data, the control-group incidence was assumed to be 60%, and liposomal bupivacaine infiltration was expected to reduce this incidence to 30%, corresponding to an absolute risk reduction of 30 percentage points. For comparison of two independent proportions with a two-sided alpha of 0.05 and power of 80%, at least 40 participants per group were required. Considering an anticipated post-randomization exclusion or protocol-deviation rate of approximately 20% because of the complexity of CRS + HIPEC, prolonged postoperative ICU stay, PCIA discontinuation, and nonprotocol analgesia, the planned sample size was 100 participants, with 50 in each group. No interim analyses or stopping guidelines were planned.

The primary efficacy analysis used the PPS. The ITT sensitivity analysis included all randomized participants according to original allocation, with missing primary outcomes imputed using the assumption least favorable to the PCIA-LB group. Categorical variables are reported as n (%), and between-group comparisons used the chi-square test or Fisher exact test, as appropriate. Continuous variables are reported as mean ± standard deviation or median (interquartile range [IQR]) depending on distribution, and between-group comparisons used the independent-samples t test or Wilcoxon rank-sum test. Effective PCA activations and the number of rescue analgesic interventions were treated as count variables with right-skewed distributions, summarized primarily as median (IQR), and compared using the Wilcoxon rank-sum test. The proportion of patients receiving at least one rescue analgesic intervention was analyzed as a binary variable. Baseline characteristics after randomization were described in the PPS using descriptive statistics and standardized differences (SMDs), without baseline hypothesis testing. Enrollment, exclusions, and reasons for entry into the PPS were reported in the CONSORT flow diagram. Safety outcomes were considered exploratory; risk differences and risk ratios with 95% confidence intervals were reported where informative. No adjustment for multiple comparisons was applied to secondary outcomes. A two-sided P < 0.05 was considered statistically significant.

Low-dose remifentanil infusion administered to selected patients while they remained intubated in the ICU was not included because the exact infusion duration and cumulative dose could not be reliably retrieved. Accordingly, the reported MME represents study-recorded opioid consumption rather than total postoperative opioid exposure. Conversion factors were prespecified before analysis and were used only for group-level comparison of study-recorded postoperative opioid consumption, not for clinical opioid rotation or dose interchange. Intravenous sufentanil was converted using the published perioperative conversion factor of 1 µg of intravenous sufentanil = 3 mg oral morphine, or 3 MME [33]. For intramuscular dezocine, we used evidence that 10 mg of intramuscular dezocine produced postoperative analgesia not significantly different from 10 mg of intramuscular morphine [35], together with a conservative 1:3 parenteral-to-oral morphine conversion [34], yielding 5 mg of intramuscular dezocine = 15 mg of oral morphine, or 15 MME.

In the PPS population with complete postoperative day 1–3 pain data, generalized estimating equation (GEE) analyses were additionally performed for repeated binary outcomes of moderate-to-severe movement-evoked and resting pain (NRS >  = 4). The GEE models used a binomial distribution with a logit link and an exchangeable working correlation structure, with patient as the clustering unit. Group and postoperative time were included in the model, and a group-by-time interaction term was added to evaluate whether pain trajectories differed between groups. The GEE analyses were used to account for within-patient correlation across repeated measurements; between-group comparisons at individual time points remained descriptive results within the PPS primary analysis.

Results

Participant Flow and Baseline Characteristics

Recruitment was conducted from April 2025 to August 2025, and the trial ended after the planned sample size had been reached. A total of 100 patients scheduled for CRS + HIPEC were randomized to the PCIA-LB group (n = 50) or the PCIA-NS group (n = 50). No ASA I participant was enrolled; all analyzed participants were ASA II or III. In the PCIA-NS group, two patients were excluded from the PPS: one discontinued PCIA in the ICU on postoperative day 1 because of impaired consciousness and was unable to complete protocol-defined pain follow-up, and one developed supraventricular tachycardia during anesthesia induction, resulting in cancellation of surgery and no study intervention. In the PCIA-LB group, four patients were excluded from the PPS: one experienced intraoperative anaphylactic shock, did not receive liposomal bupivacaine infiltration, had postoperative circulatory instability, remained in the ICU for > 3 days, and did not use PCIA; one had intraoperative leakage of intestinal contents with incision contamination, and infiltration was withheld for safety reasons; one remained in the ICU for > 48 h and could not complete postoperative pain follow-up; and one was lost to follow-up. The final PPS included 94 patients (PCIA-NS, n = 48; PCIA-LB, n = 46; Fig. 1). All 100 randomized patients were included in the ITT sensitivity analysis. The evaluable safety population was identical to the PPS because the excluded patients did not have evaluable postoperative safety data. Baseline characteristics in the PPS were described using SMDs without hypothesis testing (Table 1). Surgical characteristics are presented in Table 2.

Fig. 1.

CONSORT flow diagram. PCIA patient-controlled intravenous analgesia, LB liposomal bupivacaine, NS normal saline, PPS per-protocol set, ITT intention-to-treat, ICU intensive care unit

Table 1.

Baseline characteristics in the PPS

Characteristic	PCIA-NS (n = 48)	PCIA-LB (n = 46)	SMD
Age, years, mean ± SD	51.6 ± 9.4	52.9 ± 10.2	0.13
Female, n (%)	22 (45.8%)	23 (50.0%)	0.09
BMI, kg/m2, mean ± SD	23.3 ± 3.8	23.7 ± 3.6	0.11
ASA physical status, n (%)	–	–	0.11
II	17 (35.4%)	14 (30.4%)	–
III	31 (64.6%)	32 (69.6%)	–

PPS per-protocol set, PCIA patient-controlled intravenous analgesia, NS normal saline, LB liposomal bupivacaine, SMD standardized mean difference, BMI body mass index, ASA American Society of Anesthesiologists. Baseline characteristics are presented descriptively without hypothesis testing

Table 2.

Surgical characteristics in the PPS

Variable	PCIA-NS (n = 48)	PCIA-LB (n = 46)	Comment
PCI, median (Q1, Q3)	23 (10, 35.25)	23 (9.25, 35.75)	Descriptive
Incision length, cm, median (Q1, Q3)	35 (35, 35)	35 (35, 35)	Descriptive
Operative time, h, median (Q1, Q3)	7.2 (6.23, 9.75)	8.5 (7, 9.5)	Descriptive
Postoperative hospital stay, d, median (Q1, Q3)	9 (8, 12)	10 (8, 12)	Descriptive
Tumor origin
Primary peritoneal malignancy	13	12
Secondary peritoneal malignancy	35	34
Appendiceal origin	6	8
Gastric origin	5	3
Intestinal origin	14	16
Uterine origin	4	2
Ovarian origin	3	4
Unknown origin	3	1

PPS per-protocol set, PCI peritoneal cancer index. Surgical characteristics are presented descriptively

Primary Outcome

In the PPS primary analysis, the incidence of moderate-to-severe movement-evoked pain (NRS >  = 4) within 72 h after surgery was significantly lower in the PCIA-LB group than in the PCIA-NS group (19.6% versus 70.8%, P < 0.001; absolute risk difference, −51.3 percentage points [95% CI  −68.5 to −34.0]; risk ratio, 0.28 [95% CI 0.15–0.51]). On postoperative days 1, 2, and 3, the incidence of moderate-to-severe movement-evoked pain was also lower in the PCIA-LB group than in the PCIA-NS group (all P < 0.05; Table 3). The distribution of moderate-to-severe movement-evoked NRS scores is provided in Supplementary Table S1. In the conservative ITT sensitivity analysis, the four missing outcomes in the PCIA-LB group were imputed as occurrence of the primary outcome and the two missing outcomes in the PCIA-NS group were imputed as nonoccurrence. The incidence of moderate-to-severe movement-evoked pain within 72 h remained lower in the PCIA-LB group than in the PCIA-NS group (13/50 [26.0%] versus 34/50 [68.0%]; chi-square = 17.70, P < 0.001), supporting the robustness of the primary finding to post-randomization exclusions and missing primary outcome data.

Table 3.

Incidence of movement-evoked NRS score >  = 4 at each postoperative time point in the PPS

Time point	PCIA-NS (n = 48)	PCIA-LB (n = 46)	Chi-square	P value
Within postoperative 3 days	34 (70.8%)	9 (19.6%)	24.9	< 0.001
Postoperative day 1	34 (70.8%)	9 (19.6%)	24.87	< 0.001
Postoperative day 2	23 (47.9%)	5 (10.9%)	15.42	< 0.001
Postoperative day 3	14 (29.2%)	3 (6.5%)	6.67	0.010

NRS numeric rating scale. Values are n (%)

In the PPS GEE analysis, after adjustment for postoperative time and within-patient correlation, the overall risk of moderate-to-severe movement-evoked pain on postoperative days 1–3 was significantly lower in the PCIA-LB group than in the PCIA-NS group (OR = 0.11, 95% CI 0.04–0.28, P < 0.001). The main effect of time was statistically significant (P < 0.001), indicating a decline in moderate-to-severe movement-evoked pain over time. The group-by-time interaction was not statistically significant (P = 0.713), suggesting no clear difference in the temporal trend between groups.

Secondary Outcomes

In the PPS, the incidence of moderate-to-severe pain at rest (NRS >  = 4) was lower in the PCIA-LB group than in the PCIA-NS group at all postoperative time points (all P < 0.05; Table 4). By postoperative day 3, the incidence of moderate-to-severe pain at rest in the PCIA-LB group had decreased to 2.2%. The distribution of moderate-to-severe resting NRS scores is provided in Supplementary Table S2. In the PPS GEE analysis of resting pain, the overall risk of moderate-to-severe resting pain was significantly lower in the PCIA-LB group (OR = 0.12, 95% CI 0.05–0.31, P < 0.001). The main effect of time was significant (P < 0.001), whereas the group-by-time interaction was not (P = 0.316). The daily incidences of movement-evoked and resting pain are summarized graphically in Fig. 2.

Table 4.

Incidence of resting NRS score >  = 4 at each postoperative time point in the PPS

Time point	PCIA-NS (n = 48)	PCIA-LB (n = 46)	Chi-square	P value
Within postoperative 3 days	29 (60.4%)	7 (15.2%)	20.31	< 0.001
Postoperative day 1	28 (58.3%)	7 (15.2%)	18.68	< 0.001
Postoperative day 2	13 (27.1%)	1 (2.2%)	9.61	0.002
Postoperative day 3	10 (20.8%)	1 (2.2%)	6.21	0.013

NRS numeric rating scale. Values are n (%)

Fig. 2.

Daily incidence of moderate-to-severe movement-evoked and resting pain (NRS >  = 4) on postoperative days 1–3 in the PPS. PCIA patient-controlled intravenous analgesia, NS normal saline, LB liposomal bupivacaine, NRS numeric rating scale, POD postoperative day, PPS per-protocol set

Effective PCA activations were treated as skewed cumulative count data and analyzed using the Wilcoxon rank-sum test. At 24 h, the PCIA-LB group had fewer effective PCA activations than the PCIA-NS group (3.0 [1.0, 5.0] versus 4.0 [2.0, 9.3], P = 0.041). At 48 h, the cumulative number was numerically lower in the PCIA-LB group but did not reach statistical significance (7.0 [3.0, 11.8] versus 10.0 [3.0, 18.0], P = 0.083). No significant difference was observed at 72 h (P = 0.870; Table 5).

Table 5.

Cumulative effective PCA activations at 24, 48, and 72 h in the PPS

Time point	PCIA-NS (n = 48)	PCIA-LB (n = 46)	U statistic	P value
Postoperative day 1	4.0 (2.0, 9.3)	3.0 (1.0, 5.0)	1373.0	0.041
Postoperative day 2	10.0 (3.0, 18.0)	7.0 (3.0, 11.8)	1333.5	0.083
Postoperative day 3	10.0 (5.0, 12.0)	10.0 (4.0, 14.8)	1126.0	0.870

Values are median (IQR) cumulative counts recorded at 24, 48, and 72 h. PCA patient-controlled analgesia, PPS per-protocol set

The 72-h cumulative study-recorded oral MME was significantly lower in the PCIA-LB group than in the PCIA-NS group (556.8 [473.4, 600.0] versus 600.0 [557.4, 630.0] mg, P = 0.002). The difference of 43.2 mg relative to the control-group median represented an approximately 7.2% reduction, indicating a statistically significant but quantitatively modest reduction in study-recorded opioid use. Day-specific analysis showed significantly lower MME on postoperative day 1 in the PCIA-LB group, whereas differences on postoperative days 2 and 3 were not statistically significant. Fourteen patients in the PCIA-LB group received at least one rescue analgesic intervention, and a total of 16 individual rescue analgesic interventions were recorded; the patient-level rescue analgesia rate was 30.4%. In the PCIA-NS group, 26 patients received at least one rescue analgesic intervention, and 37 individual interventions were recorded; the patient-level rescue analgesia rate was 54.2%. The rescue analgesia rate was significantly lower in the PCIA-LB group (chi-square = 5.412, P = 0.020; Table 6). Because each loading dose, lockout-interval adjustment, intramuscular dezocine administration, and documented post-extubation intravenous sufentanil dose was counted separately, the intervention totals could exceed the numbers of patients receiving rescue analgesia. Component-specific frequencies could not be reported because the study database retained only the aggregate intervention count. These findings show a reduction in rescue analgesic demand, but the magnitude of the reduction in study-recorded opioid consumption was smaller than the difference in pain incidence.

Table 6.

Postoperative study-recorded oral MME and rescue analgesia in the PPS

Variable	PCIA-NS (n = 48)	PCIA-LB (n = 46)	Statistic	P value
72-h MME, mg, median (Q1, Q3)	600.0 (557.4, 630.0)	556.8 (473.4, 600.0)	U = 1505.0	0.002
Day 1 MME, mg, median (Q1, Q3)	178.2 (153.6, 229.35)	166.8 (137.4, 206.4)	U = 1365.5	0.048
Day 2 MME, mg, median (Q1, Q3)	206.4 (172.2, 266.4)	194.4 (154.2, 241.8)	U = 1268.0	0.216
Day 3 MME, mg, median (Q1, Q3)	183.6 (138.0, 219.0)	168.0 (142.2, 204.6)	U = 1212.5	0.414
Patients receiving rescue analgesia, n (%)	26 (54.2%)	14 (30.4%)	chi-square = 5.412	0.020
Mean rescue analgesic interventions per patient	0.77	0.35	Z = -2.529	0.011

MME oral morphine milligram equivalents. MME calculations included PCIA background infusion, PCA boluses, PCIA loading doses, each documented 5-µg of intravenous sufentanil rescue dose administered after extubation, and intramuscular dezocine rescue doses; nonopioid components, such as flurbiprofen axetil, were not included. Low-dose remifentanil infusion administered to selected patients while they remained intubated in the ICU was excluded because cumulative-dose data were unavailable. Rescue analgesia was analyzed at both the patient and intervention levels. The following were each counted as one rescue analgesic intervention: administration of a 5-ml PCIA loading dose, shortening of the PCA lockout interval from 20 to 15 min, each 5-mg intramuscular dose of dezocine, and each documented 5-µg post-extubation intravenous sufentanil dose. Only the aggregate intervention count was retained for PCIA loading doses, lockout-interval adjustments, and intramuscular dezocine; component-specific frequencies for these PCIA-related steps were unavailable. Documented post-extubation intravenous sufentanil doses were available and were included in both the rescue-intervention count and the MME calculation. Although the aggregate rescue-intervention count did not preserve separate categories for all PCIA-related rescue steps, opioid-dose data required for MME calculation were obtained from PCIA delivery records and medication-administration records where applicable. The 5-ml loading dose was permitted no more than once during each postoperative 24-h period

The 72-h QoR-15 score did not differ significantly between groups (PCIA-NS: 93 (IQR 91, 96) versus PCIA-LB: 94 (IQR 91, 96); Z = 0.35, P = 0.73).

In the evaluable safety population, 11 patients in the PCIA-LB group experienced analgesia-related adverse events (23.9%), including nausea in five patients (10.9%), dizziness in five patients (10.9%), and pruritus in one patient (2.2%); no vomiting occurred. In the PCIA-NS group, five patients experienced adverse events (10.4%), including nausea in three patients (6.3%), vomiting in one patient (2.1%), and dizziness in one patient (2.1%). The overall risk difference was 13.5 percentage points (95% CI  −1.6 to 28.6), and the risk ratio was 2.30 (95% CI 0.86–6.10; P = 0.143). No participant received droperidol rescue treatment. No participant required naloxone, emergency airway intervention, or ventilatory support for a clinically recognized opioid-related respiratory event. No local anesthetic systemic toxicity, wound infection, fat liquefaction, or other severe adverse event was observed. The prespecified analgesia-related adverse events are presented in Table 7. The wide confidence intervals, absence of structured sedation and ventilation data, and small safety population do not establish equivalence in adverse-event risk.

Table 7.

Analgesia-related adverse events in the evaluable safety population

Adverse event	PCIA-NS (n = 48)	PCIA-LB (n = 46)	Chi-square	P value
Dizziness	1 (2.1%)	5 (10.9%)	2.182	0.140
Nausea	3 (6.3%)	5 (10.9%)	0.225	0.635
Vomiting	1 (2.1%)	0 (0.0%)	-*	1.000
Pruritus	0 (0.0%)	1 (2.2%)	-*	0.488
Overall incidence	5 (10.4%)	11 (23.9%)	2.149	0.143

*Fisher exact test. Excluded patients in the CONSORT flow did not have evaluable postoperative safety data; therefore, safety analyses used the evaluable safety population, which was identical to the PPS. Nausea, vomiting, dizziness, and pruritus were prospectively recorded as present or absent without a validated severity grading scale. No participant received droperidol rescue treatment. Sedation, somnolence, respiratory rate, and level of consciousness were not retained as structured study outcomes

Discussion

This single-center, randomized, participant- and assessor-blinded, placebo-controlled trial evaluated the incremental effect of liposomal bupivacaine incisional infiltration on top of an institutional PCIA pathway after CRS + HIPEC. In both the prespecified PPS analysis and the conservative ITT sensitivity analysis, PCIA-LB reduced the incidence of moderate-to-severe movement-evoked incisional pain. It also reduced rescue analgesia and produced a statistically significant but modest reduction in cumulative 72-h study-recorded MME. QoR-15 did not improve, and overall analgesia-related adverse events were numerically more frequent in the PCIA-LB group, although the difference was not statistically significant. The high incidence of moderate-to-severe pain in the control group is consistent with broader evidence that postoperative pain remains insufficiently controlled after major surgery and abdominal surgery [25, 26]. These findings should therefore be interpreted as placebo-controlled evidence for an incision-focused adjunct and not as evidence that liposomal bupivacaine is superior to plain local anesthetics, TEA, or fascial plane and catheter techniques.

The potential prolonged analgesic effect of liposomal bupivacaine may be related to its liposomal carrier-mediated sustained-release profile. Liposomal vesicles can form a local drug depot and gradually release active bupivacaine [27]. The two-layer, multi-point technique delivered the solution along the long midline incision and may have reduced nociceptive input from skin, subcutaneous tissue, and the abdominal wall. However, incisional infiltration primarily targets somatic pain and is not expected to provide comprehensive visceral analgesia. The primary endpoint was deliberately restricted to movement-evoked incisional pain rather than global abdominal pain. The large between-group difference may reflect the importance of the approximately 35-cm incision during turning and coughing, as well as the use of saline rather than an active local-anesthetic comparator. It should not be extrapolated to an effect on all components of pain after CRS + HIPEC.

Analgesia after CRS + HIPEC should be individualized and multimodal. TEA remains an important technique, and real-world implementation of epidural analgesia varies substantially across institutions [28], but current practice also includes TAP, rectus sheath, quadratus lumborum, and erector spinae plane blocks, as well as continuous wound or rectus sheath catheters. Recent practice guidance strongly recommends fascial plane blocks for adults undergoing open abdominal surgery to reduce early pain and/or opioid requirements [4], and a randomized trial has evaluated rectus sheath catheter analgesia against TEA after open midline major abdominal surgery [11]. The present trial did not compare these techniques. Direct-vision incisional infiltration was selected in part because local infiltration is a labeled route for liposomal bupivacaine, whereas TAP and rectus sheath blocks are not included among the labeled block indications in the prescribing information [21]; this regulatory consideration does not imply that those techniques are ineffective. The trial also did not include plain bupivacaine or ropivacaine. This is important because evidence comparing liposomal with conventional local anesthetics has been inconsistent; a Cochrane review and a recent meta-analysis did not demonstrate a consistent clinically important advantage of liposomal bupivacaine over plain local anesthetics [16, 17]. Plain local-anesthetic infiltration might plausibly have improved early postoperative pain, particularly on postoperative day 1. For contextual cost comparison, one vial of liposomal bupivacaine cost CNY 360 at our institution, whereas an institutionally estimated ropivacaine infiltration regimen using up to two vials cost approximately CNY 66.6. Both drugs were eligible for reimbursement under the local medical-insurance pathway, and surgeon-performed incisional infiltration did not incur a separate procedural charge. These institution-specific figures do not constitute a formal economic evaluation. Because no active comparator or formal health-economic analysis was included, the clinical value and cost-effectiveness of the additional acquisition cost of liposomal bupivacaine remain uncertain.

The 72-h cumulative study-recorded oral MME was reduced by approximately 7.2%, whereas the improvements in pain incidence and patient-level rescue analgesia were more pronounced. This difference is clinically plausible because the background PCIA infusion contributed a large fixed component of opioid exposure in both groups, thereby limiting the maximum between-group difference attributable to PCA demand. Effective PCA activations and day-specific MME were significantly lower only at 24 h; differences at 48 and 72 h were not statistically significant. A cumulative 72-h endpoint can nevertheless become statistically significant when smaller directional differences accrue over time. The rescue outcome was analyzed both as the proportion of patients receiving any rescue intervention and as the total number of individual intervention steps; therefore, a patient could contribute more than one intervention when the loading dose, lockout-interval adjustment, dezocine administration, or a repeated 5-µg post-extubation intravenous sufentanil dose was recorded. Low-dose remifentanil infusion administered to selected awake patients while they remained intubated in the ICU to attenuate endotracheal-tube-related discomfort and airway responses may have affected day-1 pain reporting and actual postoperative opioid exposure. Because cumulative remifentanil doses were unavailable, they were not included in MME calculations; by contrast, each documented 5-µg intravenous sufentanil rescue dose administered after extubation was included in both the postoperative day-1 rescue-analgesia record and the 72-h MME calculation. Most patients had discontinued remifentanil and undergone extubation before the ward-based postoperative day-1 assessment. Overall, the reduction in study-recorded opioid use should be described as modest rather than large.

Although the PCIA-LB group showed improvements in movement-evoked incisional pain, resting pain, rescue analgesia, and cumulative MME, the 72-h QoR-15 total score did not differ significantly between groups. QoR-15 reflects physical comfort, independence, pain, emotional state, psychological support, sleep, gastrointestinal function, nausea, fatigue, and activity [24]. Early recovery after CRS + HIPEC is also influenced by systemic inflammation, bowel function, drains, nutritional status, and the ICU experience. Improvement in one component of incisional analgesia may therefore be insufficient to alter the total score. The sample size was calculated for the pain-incidence endpoint and may have been underpowered to detect a clinically important QoR-15 difference. Claims regarding broader recovery should consequently remain modest.

Regarding safety, the overall incidence of analgesia-related adverse events was numerically higher in the PCIA-LB group, with a risk ratio of 2.30 but wide confidence intervals crossing unity. No local anesthetic systemic toxicity or wound complication was observed, consistent with prior pharmacokinetic and cardiac safety data for liposomal bupivacaine local infiltration within recommended dosing [29, 30], and no participant required naloxone, emergency airway intervention, or ventilatory support for a clinically recognized opioid-related respiratory event. PCIA was delivered under an institutional SOP. Monitoring was continuous in the ICU; after transfer to the ward, pulse oximetry was continuous through postoperative day 3, while electrocardiography and noninvasive blood pressure were monitored intermittently. Nevertheless, absence of a severe event in 94 evaluable patients cannot establish safety. Contemporary opioid-monitoring guidance emphasizes systematic assessment of sedation and respiratory status in hospitalized adults receiving opioids [31]. Pulse oximetry monitors oxygenation rather than ventilation and may not detect early hypoventilation, particularly during supplemental oxygen administration. Sedation, somnolence, respiratory rate, and level of consciousness were not retained as structured study outcomes, and the study did not use a validated sedation scale, capnography, or formal severity scales for nausea, vomiting, or pruritus. It was also not powered for uncommon respiratory or local-anesthetic adverse events. The safety findings should therefore be regarded as descriptive.

Limitations

This study has several limitations. First, because liposomal bupivacaine is milky white and normal saline is transparent, the surgeons performing infiltration could not be blinded. Although postoperative outcome assessment, PCIA management, and statistical analysis were blinded, and infiltration volume, injection intervals, and layers were standardized, performance bias related to surgeon awareness of allocation cannot be excluded.

Second, the control group received saline rather than plain bupivacaine or ropivacaine. The placebo design permits estimation of the incremental efficacy of adding liposomal bupivacaine to PCIA, but it does not answer the more clinically relevant question of whether liposomal bupivacaine provides additional benefit over less expensive conventional local anesthetics. The study also did not include a formal cost-effectiveness analysis. The results should not be used to claim superiority over active local-anesthetic infiltration.

Third, this was a single-center study and did not directly compare liposomal bupivacaine infiltration with TEA, TAP, rectus sheath, quadratus lumborum, or erector spinae plane blocks, or with continuous wound or rectus sheath catheters. Institutional experience, coagulation management, ICU pathways, and ward monitoring may limit generalizability. Accordingly, this trial cannot determine the efficacy or safety of liposomal bupivacaine when used outside the labeled route for TAP, rectus sheath, or other fascial plane blocks in this setting.

Fourth, the prespecified primary efficacy analysis used the PPS rather than ITT, and post-randomization exclusions occurred in two PCIA-NS and four PCIA-LB participants. Such exclusions may introduce bias. A conservative ITT sensitivity analysis using assumptions least favorable to the PCIA-LB group yielded a consistent result, but missing outcome data cannot be fully eliminated as a source of uncertainty.

Fifth, pain was assessed once daily at a fixed time using a 24-h retrospective report. This pragmatic method may introduce recall bias, does not capture short-term fluctuations or immediate postoperative onset, and precludes a pain-intensity area-under-the-curve analysis. Movement stimuli were not standardized beyond turning, coughing, or ambulation. Moreover, the intervention was directed at somatic incisional pain and cannot be assumed to control visceral pain.

Sixth, the institutional PCIA protocol used a background infusion and a prespecified stepwise rescue algorithm. Although the same algorithm was applied to both groups by blinded personnel, monitoring was continuous in the ICU, and ward pulse oximetry was continuous through postoperative day 3, use of a background infusion and protocol-allowed lockout-interval adjustment may affect opioid consumption and limits generalizability to PCA-only pathways. Low-dose remifentanil could be infused in selected awake patients while they remained intubated in the ICU to attenuate endotracheal-tube-related discomfort and airway responses. This practice followed the same ICU SOP in both groups but was not part of the randomized intervention. The exact infusion duration and cumulative remifentanil dose could not be retrieved and were therefore excluded from MME calculations, potentially underestimating actual postoperative opioid exposure. Remifentanil administration may also have influenced day-1 pain reporting, although most patients had discontinued the infusion, undergone extubation, and returned to the ward before the day-1 assessment. By contrast, each documented 5-µg intravenous sufentanil rescue dose administered after extubation was included in the postoperative day-1 rescue-analgesia record and the 72-h MME calculation. In addition, only the aggregate number of rescue analgesic interventions was retained; component-specific frequencies for PCIA loading doses, lockout-interval adjustments, intramuscular dezocine, and post-extubation intravenous sufentanil could not be reconstructed. Sedation, somnolence, respiratory rate, and level of consciousness were not retained as structured research data; no standardized sedation scale or continuous capnography was used, and mild transient sedation or hypoventilation cannot be excluded. Cumulative exposure to additional rescue dolasetron was not analyzed separately. The institutional pathway also did not include scheduled acetaminophen, which may limit generalizability to more comprehensive contemporary multimodal pathways. Urinary retention, constipation, and postoperative ileus were not systematically captured using prespecified definitions.

Finally, the sample was not powered for uncommon adverse events, QoR-15 was assessed only at 72 h, and follow-up was limited to the first 72 h for pain-related outcomes. The study did not evaluate mobilization metrics, return of bowel function, chronic postsurgical pain, post-discharge recovery, or long-term functional outcomes [32]. Larger multicenter active-comparator studies should include repeated pain and recovery assessments, comprehensive opioid and ICU medication accounting, formal safety monitoring, and economic evaluation.

Conclusions

In this single-center randomized placebo-controlled trial, liposomal bupivacaine incisional infiltration added to an institutional PCIA pathway reduced early movement-evoked and resting incisional pain and rescue analgesia after CRS + HIPEC and produced a statistically significant but modest reduction in cumulative 72-h study-recorded opioid consumption. QoR-15 did not improve. No local anesthetic systemic toxicity or clinically recognized opioid-related respiratory event requiring naloxone, emergency airway intervention, or ventilatory support was documented, but the study was not powered to establish safety. Because the comparator was saline, the findings do not demonstrate superiority or cost-effectiveness relative to plain bupivacaine, ropivacaine, or regional catheter techniques. Multicenter active-comparator trials with standardized opioid accounting and broader recovery and safety outcomes are needed.

Supplementary Information

Below is the link to the electronic supplementary material.

Author Contributions

Haijing Zhang coordinated study implementation, performed the statistical analysis, and drafted the initial manuscript. Zhexuan Gong conducted postoperative follow-up. Huan Zhang was responsible for data quality control and manuscript revision. Siyi Yan was responsible for study conception and design and manuscript revision. Bing Li, Yang Yu, Gang Liu, Zhonghe Ji, and Yan Li performed the incisional infiltration. All authors reviewed and approved the final manuscript. Haijing Zhang and Zhexuan Gong contributed equally to this work and share first authorship. Huan Zhang and Siyi Yan are co-corresponding authors.

Funding

No funding or sponsorship was received for this study or publication of this article. The Rapid Service Fee was funded by the authors.

Data Availability

The datasets generated and/or analyzed during the current study are available from the corresponding author on reasonable request, subject to institutional and ethical restrictions related to participant privacy. The trial protocol, statistical analysis plan, and statistical code are available from the corresponding authors on reasonable request.

Declarations

Conflict of Interest

Haijing Zhang, Zhexuan Gong, Yan Li, Bing Li, Yang Yu, Gang Liu, Zhonghe Ji, Huan Zhang, and Siyi Yan have nothing to disclose.

Ethical Approval

The study protocol was approved by the Ethics Committee of Beijing Tsinghua Changgung Hospital (approval number: 25038-4-01). The trial was conducted in accordance with the Declaration of Helsinki and CONSORT reporting principles. Written informed consent was obtained from all participants before enrollment, and patient-identifiable data were anonymized.