This is a parallel-group, double-blinded, randomised controlled, single-centre trial with a 1:1 allocation ratio. The study has been designed per the ‘Declaration of Helsinki’ and was approved by the institutional review board (IRB) at Xijing Hospital (KY20232194-C-1), followed by registration on ClinicalTrials.gov (NCT05920967. Date of registration: 24 June 2023). This trial will be carried out in the Department of Cardiovascular Surgery at Xijing Hospital. This trial does not involve extra biological specimens beyond clinical practice, and any modifications shall be approved by the IRB and updated online with the registration. The study protocol (V2.0, 2023/7/3) follows the Standard Protocol Items Recommendations for Interventional Trials.39 The schedule of enrolment, allocation and assessments is listed in figure 1. The flow diagram is illustrated in figure 2.

The study is set out to investigate and test whether TPVB with ropivacaine plus D-PAL emulsion has superior analgesic effects in relieving incidence of postoperative chronic pain within 3 months over TPVB with ropivacaine plus DSP, among participants planning to undergo intercostal MICS.

This study aims to enroll patients aged between 18 and 65 years, classified NYHA I–III (New York Heart Association Classfication of Heart Failure), who are scheduled to undergo elective intercostal minimally invasive cardiac surgery. Those who meet the following conditions will be excluded: (1) undergoing urgent surgery; (2) body mass index (BMI) ≥35 kg/m²; (3) had a previous history of cardiac surgeries; (4) known allergy to local anaesthetic; (5) with skin damage or infection at the TPVB site; or (6) with any of liver, kidney or blood coagulation dysfunctions; (7) already experiencing chronic pain preoperatively.

Eligible participants will be randomly assigned at a 1:1 ratio using a web-based randomisation system (Research Electronic Data Capture, REDCap, developed by Vanderbilt University). A trial statistician was responsible for generating the randomisation seeds and allocations. The randomisation strategy employed was permuted-block randomisation, stratified by surgery type (with or without CPB, cardiopulmonary bypass), using random block sizes of 4 or 6. The blinded allocations were then securely uploaded to REDCap, which has strictly encrypted access. No other individuals were involved in this process.

The group allocation will be blinded to all participants, caregivers and investigators responsible for outcome assessment and data analyses. Throughout the whole study process, only one qualified nurse preparing and reconstituting the study drugs will be unblinded. The study drug will be covered by an opaque shelter and kept blind to all other stakeholders of the trial considering the difference between the lipid emulsion in the infiltration solution and the clear solution.

Participants will be randomised to receive either 1 mL (4 mg) D-PAL emulsion (Guangzhou Green Cross Pharmaceutical, China) or 1 mL (5 mg) DSP (XinChen Pharmaceutical, China), combined with 0.5% ropivacaine 20 mL (RuiYang Pharmaceutical, China). 1 mL of D-PAL and DSP contains an equivalent dose of dexamethasone.40 To achieve the target concentration, 10 mL of 1% ropivacaine (RuiYang Pharmaceutical, China) will be diluted with 10 mL of normal saline. The diluted intervention drugs will be administered as a pre-emptive bolus of TPVB infiltration before the procedure with a sterile continuous peripheral nerve block catheter set (B. Braun Melsungen AG). An 18-gauge needle will be inserted into the designated site, and subsequently, a continuous peripheral nerve block catheter will be carefully positioned within the corresponding paravertebral space. Following the surgical procedure, a Patient-Controlled Analgesia (PCA) device will be connected to the catheter, administering a continuous infusion of 250 mL ropivacaine (0.2%) into the paravertebral space. The background infusion will be maintained at a rate of 5 mL/hour, with a PCA volume of 5 mL and a lockout interval set at 30 min. Considering the volume limit of the disposable PCA infusion bag, the PCA will be used for 2 days after surgery.

To make an adequate evaluation of pain intensity, quantitative sensory testing (QST) was used to evaluate the magnitude of pain sensation in surgical patients. QST was performed using an Electronic von Frey device (EvF device, IITC Life Science, Woodland Hills, California, USA) by one trained investigator blinded to the group assignment. The EvF device has an 800 g probe with a rigid tip of 0.8 mm in diameter. The probe was pressed to the skin surface at a right angle with increasing force at a controlled rate. When the patient perceived pain, the result was stored and displayed in grams, and the probe was then removed. Before TPVB, the basic mechanical pain threshold will be defined as measured by EvF device at eight markers (T2, T4, T6, T8 at midclavicular line on both left and right sides). The pain intensity will also be measured by EvF device at the same markers after the TPVB operation and 1 day after to confirm the effect of TPVB. The pain intensity will be measured on postoperative days 2 and 3 as needed.

PCA regimen will be discontinued on cessation of necessity. Each button press will be meticulously documented by an electronic memory system, including valid and invalid presses. For rescue analgesia, patients will receive intravenous tramadol at a single dose of 50–100 mg, administered via an infusion pump at a rate of 20–40 mg/hour, if the NRS score exceeds four after four PCA boluses. Oxycodone will be administered intravenously if tramadol cannot achieve satisfactory analgesia. A single dose ranging from 1 to 10 mg of oxycodone will be administered, and additional doses will be administered at an interval of no less than 4 hours. The cumulative doses of rescue analgesics will be recorded.

Prior to study initiation, all investigators will be trained. All interventions will strictly adhere to established clinical practice guidelines. In the preoperative visit, eligible patients will be introduced to the pain Numerical Rating Scale (NRS), which ranges from 0 (indicating no pain) to 10 (representing the worst possible pain). Comprehensive guidance will be provided to eligible patients for proper utilisation of a PCA device.

This is a single-centre study. All the perioperative management will follow the institutional routine, which has been previously described in detail.41 All patients will be treated according to the best available anaesthesia management in our centre. Briefly, all participants will receive perioperative intensive treatment according to the institutional practice, including general anaesthesia, mechanical ventilation, inotropic drugs, pacing, blood and fluid management, postoperative sedation and analgesia, diuretics, antibiotics and invasive monitoring. The monitoring includes but is not limited to, invasive arterial pressure, ECG, central venous pressure, cardiac output, pulse oximetry, temperature, urine output, arterial blood gases, coagulation test and routine laboratory examinations.

Anaesthesia is induced with a bolus injection of 0.02–0.04 mg/kg midazolam, 0.1–0.2 mg/kg etomidate and 0.5–1 µg/kg sufentanil. When the consciousness of patients is lost, rocuronium is administered at 0.6–0.9 mg/kg to facilitate endotracheal intubation. Anaesthesia is maintained by continuously infusing 4 mg/kg/hour propofol or continuously inhaling 1–2% sevoflurane, 1–2 µg/kg/hour sufentanil and 4–12 µg/kg/min rocuronium. During surgery, the brain bispectral index was maintained at 40–60. The doses of maintenance anaesthetics are adjusted based on the haemodynamics, bispectral index and the experience of the anaesthetist. During general anaesthesia, patients are mechanically ventilated in a volume-controlled mode with a tidal volume of 6–8 mL/kg, 12–14 breaths per minute at FiO₂ 0.5 with PEEP of 5 cmH₂O for double-lung ventilation. In the case of one-lung ventilation, volume-controlled ventilation with a tidal volume of 4–6 mL/kg, a respiratory rate of 14–18 breaths per minute, and the lowest possible FiO₂ to maintain SPO₂ between 95% and 100% will be employed. When clinically indicated, inotropes or vasopressors were administrated to maintain stable systemic haemodynamics. If urine output was reduced during CPB or in the ICU (<0.5 mL/kg/hour) and hypovolaemia and dehydration were excluded, diuretics, such as furosemide, would be administered intravenously.

The chest is opened through the intercostal incision. For procedures under CPB, CPB is initiated and surgery is performed with mild hypothermia (core temperature range, 28–30°C). Cardiac arrest is achieved by delivery of 15 mL/kg cardioplegic solution and repeated every 30 min as needed (7.5 mL/kg). Standard non-pulsatile CPB is instituted. Blood flow during CPB is adjusted to achieve a mean arterial blood pressure of 50–70 mm Hg, and a mixed venous oxygen saturation, measured at the inflow of the CPB circuit, of greater than 65%. Acid-base balance is performed with α-stat blood gas principles. Haematocrit is adjusted between 25% and 30%, and blood glucose levels are maintained at less than 200 mg/dL. The patient is maintained fully anticoagulated during CPB with an ACT (activated clotting time) above 480 s. Protamine was used to neutralise the effect of heparin by intravenous injection of 1 mg per 100 units of heparin given over 10 min. For off-pump coronary artery bypass grafting (CABG), the anaesthesia management strategy remains consistent.

After surgery, all participants are transferred to the ICU for further treatment.

Intensive care treatment follows standardised protocols established at the Department of Cardiovascular Surgery, ICU division of Xijing Hospital. In the ICU, vital signs and haemodynamics of patients are continuously monitored. When clinically indicated, supportive strategies or treatments are initiated to support organ function.

Data will be collected after enrolment, at the end of surgical procedures, at ICU discharge, at hospital discharge, and 3, 6 and 12 months after surgery. We will record the baseline characteristics such as demographics, comorbidities, medications, previous medical history and preoperative laboratory tests. Data about dosage, timing and mode of administration of all drugs during surgery and other surgical characteristics, including CPB and cross-clamping duration, will also be collected. With regards to the thoracic paravertebral nerve block intervention, we will collect data on practice levels of the operator, baseline nociceptive sensations, body position of patients, block side, time of administration, post-procedure sensation blockade level, etc. Postoperative management data include postoperative medication and fluid balance within 24 hours, the consumption of analgesic drugs and the use of rescue analgesics. Follow-ups at 3, 6 and 12 months will be carried out face to face or by telephone and will focus on the incidence, intensity, and phenotype of CPSP, quality of life and adverse events that compromise prognosis. Data collection and management will be performed via the electronic data collection system (REDCap V.14.0.1 - 2024, Vanderbilt University).

All data will be entered and stored in the secure EDC (electronic data capture) system (REDCap application), in accordance with the guidelines set by our research team as previously reported.42 All identifiable information will be kept strictly confidential, with access restricted by the principal investigator (PI). All accesses shall be granted by the PI for research purposes only. Data will be validated weekly for quality and integrity by study investigators for logic and feasibility (eg, reference range, valid values, units). Identified errors will be traced back and corrected according to the original data source (eg, hard copies of CRFs, original electronic system records). Summaries of these corrections shall be included in the data quality reports.

In this trial, experienced anaesthesiologists and nurses will safely administer the intervention drugs. Continuous monitoring of adverse events will be recorded throughout the follow-ups for 12 months. Incidences shall be reported to IRBs as soon as possible and will be treated accordingly. The trial will be terminated immediately in case of serious life-threatening AEs (adverse events) leading to prolonged hospital stay or death.

This protocol strictly adheres to the Declaration of Helsinki, ICH-GCP guidelines, regulations from the China National Drug Administration, and the institutional standard operating procedures as previously published.41–43 All data collected are important for the evaluation of drug efficacy and safety and will be kept confidentially in compliance with laws and regulations on the protection of subject privacy. All private information will be de-identified before data analyses. Written informed consent (Patient Consent Form) is obtained on recruitment. Patients may decline consent and withdraw at any time point, with all relevant information removed.

The primary endpoint is the incidence of chronic postoperative pain at 3 months following surgery defined as per the updated International Classification of Diseases.2 CPSP will be described as neuropathic pain in nature if it meets the neuropathic pain questionnaire criteria.44 Briefly, CPSP is defined as a new development of pain or an increase in the intensity of pain at the surgical area or projected onto the innervation area of a nerve in this area after a surgical procedure that persists for at least 3 months. Participants will be followed up and asked if they are experiencing persistent pain or any discomfort at the surgical site, anterior or posterior thorax, axilla, or ipsilateral upper limb. The nature and intensity of pain will be evaluated with a Brief Pain Inventory Short-Form (BPI-SF) questionnaire. If the BPI ‘average’ pain intensity score is ≥1, the patient is considered to experience CPSP.

Secondary outcomes include the total oral consumption of morphine equivalent at 24, 48 and 72 hours postoperatively, the incidence of acute pain at rest and during activity via the NRS at the same time points, and examining the incidence of chronic postoperative pain defined by a BPI score greater than one at 6 and 12 months postsurgery. Additional assessments involve patient-reported postoperative pain severity in the past 7 days (Patient-Reported Outcomes Measurement Information System; PROMIS Scale V.2.0 - Pain Intensity 3a), postoperative pain intensity in the past 24 hours (BPI), interference of chronic pain with daily activities in the past 7 days and in the past 24 hours (PROMIS Short Form V.1.1 - Pain Interference 8a and BPI), neuropathic pain (PROMIS Scale V.2.0 - Neuropathic Pain Quality 5a), the health-related quality of life (12-item Short-Form Health Survey V.2, SF-12 V.2.0), sleep quality (Pittsburgh Sleep Quality Index), emotional well-being (Hospital Anxiety and Depression Scale) and self-efficacy (Pain Self-Efficacy Questionnaire at 3, 6 and 12 months postoperatively.

Exploratory outcomes include (1) any pneumothorax, blood pneumothorax, haematoma and related complications caused by the indexed surgical procedure postoperatively; (2) postoperative poor wound healing and (3) postoperative adverse reactions such as hyperglycaemia, hypoxaemia, atelectasis, hypotension, nausea and vomiting, bradycardia, vertigo, agitation, pruritus, chills and delirium. Other exploratory outcomes include intubation duration, postoperative length in bed and postoperative length of stay in the hospital, stress response indexes (blood glucose, C-reactive protein, white blood cells and neutrophils at 24 hours, 48 hours and 72 hours after surgery), the information on first insulin administration within 24 hours (eg, time, dose and blood glucose level) and the cumulative dose of insulin within 48 hours postoperatively.

The sample size was determined based on the group sequential design to maintain a power of 80% at a 5% significance level. The pilot observation of 50 patients who underwent total thoracoscopic assisted cardiac surgery, intercostal small-incision coronary bypass surgery and intercostal small-incision heart valve replacement in Xijing Hospital indicated that CPSP at 3 months following surgery was around 50%. We thus hypothesised that the use of D-PAL can reduce the incidence rate of CPSP by 20%; that is, the hypothetical incidence rate in D-PAL and DSP groups are 40% and 50%, respectively. One interim analysis at a 50% information rate with futility and efficacy bounds approximated using the O’Brien and Fleming method by the Data and Safety Monitoring Board (DSMB). The rpact package was used for calculating the sample size. A total of 902 (451 in each group) participants are required considering 10% drop-out rates.

Demographic and baseline information will be succinctly described using descriptive statistics. Categorical variables will be represented as counts (percentages) and assessed with χ² or Fisher’s exact test, as appropriate. Unadjusted univariable analyses for between-group incidence differences will employ relative risks and 95% CIs, derived from contingency table and distribution asymptotic approximations. Continuous variables will be presented as mean (±SD) or median (IQR). Normality will be checked using visual histogram evaluation and Shapiro-Wilk tests. Between-group differences in continuous variables will be examined using t-tests or Wilcoxon rank-sum tests, depending on distribution normality.

The primary analyses will be conducted on an intention-to-treat basis with all participants randomised included. In addition, complementary analyses will be performed on the per-protocol data set. The primary outcome will crudely be compared using a χ² test and reported with risk difference and 95% CI. The estimated risk ratio with a 95% CI will also be obtained using Wald’s likelihood ratio approximation test. A generalised linear regression of the binomial family with either logit or log link will be used depending on the incidence rate and will also be applied to estimate adjusted ORs or relative risks with 95% corresponding CIs. Any pre-randomisation data will be adjusted if they are determined to be strong or important prognostic factors.45 We will assess the balance of randomised groups on the baseline and procedural characteristics using absolute standardised difference, defined as the absolute difference in means, mean ranks or proportions divided by the pooled SD. Baseline variables with standardised differences>0.131 (1.96×sqrt (1/451+1/451)) will be defined as imbalanced and be adjusted.

The secondary outcomes will follow the same procedures as described above: crude comparisons will be conducted using the χ² test; risk differences and relative risks will be provided together with 95% CIs; binomial regressions with either log or logit links will be employed for binary data. For continuous responses, linear regressions with appropriate transformations (eg, log, square root, inverse) where necessary will be conducted. For circumstances where response variables fit better with the log-Gaussian assumption, generalised linear regression will be employed. Cox proportional hazard models will be used for time-to-event (eg, when the pain was first reported) where proportionality will be inspected via Kaplan-Meier curves, log-log plots and Schoenfeld residuals. Estimated HRs will be reported with 95% CIs, and non-parametric comparisons will be done using log-rank tests. All adverse reactions will be reported and analysed, where comparisons will be conducted as described above.

The following sensitivity analyses are planned for the primary outcome. First, the same analysing procedures will be conducted on the per-protocol set. Second, the BPI-SF score will be treated as a continuous variable and compared using t-tests. Adjusted inference will be carried out based on appropriate regression methods depending on data distribution. The between-group differences will be estimated and reported with 95% CIs. Third, the interaction between the group and timing for BPI-SF score at 3, 6 and 12 months will be analysed using repeated measure analysis of variance. We anticipate a low missing rate of the primary outcome. If more than 5% of data are missing for the primary outcome, multiple imputations will be carried out as a sensitivity analysis.

Preplanned subgroup analysis considers age (median of the whole cohort), gender (male, female), BMI (≤24, >24), incision type (video-assisted thoracic ports incision that involves several intercostal spaces, single long intercostal incision), CPB (yes, no) and block side (left, right). The interaction effect will be assessed via the inclusion of interaction terms between stratified variables and intervention groups in the corresponding models. Exploratory analysis of the interaction between the randomisation group and different time points will be done on repetitively collected secondary outcomes with risk ratio and 95% CIs estimated. Length of intubation, postoperative length of stay in bed and in hospital will be described and analysed using survival methods. The stress response indices, information on the first insulin administration within 24 hours, and the cumulative dose of insulin within 48 hours will be described and compared without adjustment for multiple comparisons.

Independent auditors will be responsible for ensuring compliance with the clinical trial protocol and guaranteeing that the data collection follows the Good Clinical Practice guidelines. Monitoring and follow-ups shall be conducted according to the protocols established during the initial setup, and all reporting will adhere to current national and international requirements. DSMB, formed by an independent group of experts (one statistician, one anaesthesiologist, one cardiac surgeon and one pharmacist), will review the trial halfway and assess accumulated study data to ensure the safety of participants and the scientific validity of the trial.

No imputation for missing primary outcomes will be applied for the primary analysis. If covariates are missing during the adjusted analysis, multiple imputation will be conducted.

No patients or public were involved in the design, development, analyses or presentation part of the protocol.

Final trial results will be submitted to a peer-reviewed journal or presented at relevant academic conferences.

After the IRB approval and ClinicalTrials.gov registration, participants will continuously be recruited after confirming their eligibility and providing written informed consent. The first participant was enrolled and randomised on 11 July 2023, and the study is estimated to be completed on 30 June 2026.

The PANDORA trial will take place at a teaching tertiary hospital, with its cardiovascular surgery centre recognised as one of the leading centres in China. The hospital handles over 4000 cardiovascular surgeries annually, including approximately 1000 minimally invasive cardiac surgeries via intercostal incision. Given the planned sample size and follow-up duration, the trial is expected to be completed in about 3 years.

Not applicable.