Dexketoprofen Trometamol Plus Sufentanil for Analgesia After Thoracoscopic Radical Resection of Lung Cancer: A Randomized, Double-Blind, Placebo-Controlled Trial

Abstract

Background

Evidence for adding dexketoprofen trometamol to sufentanil-based patient-controlled intravenous analgesia (PCIA) after thoracoscopic radical resection of lung cancer remains limited.

Objective

To evaluate 48-hour opioid consumption, pain outcomes, and short-term adverse events with dexketoprofen trometamol plus sufentanil versus sufentanil-based PCIA alone.

Methods

In this randomized, double-blind, placebo-controlled trial, 108 patients scheduled for thoracoscopic radical resection of lung cancer were randomized preoperatively (54 per group). After 18 post-randomization exclusions before PCIA initiation, 45 patients per group received the allocated intervention and were analyzed. All underwent unilateral uniportal anatomical segmentectomy with mediastinal lymph node dissection or sampling and received a standardized intercostal nerve block. Group S received sufentanil-based PCIA, and Group R additionally received dexketoprofen trometamol. The primary outcome was total 48-hour postoperative sufentanil consumption, including rescue doses.

Results

Sufentanil consumption was lower in Group R than in Group S (78.03 ± 10.66 vs 96.78 ± 13.69 μg; reduction 18.75 μg, 95% CI 13.61 to 23.89; P<0.001), representing an approximately 19% reduction. Resting pain scores were lower at 12 and 24 hours (both P=0.004), meeting the revised Bonferroni threshold of P<0.005 for 10 time-point comparisons; the coughing-pain difference at 12 hours (P=0.006) did not meet this threshold. Resting-pain area under the curve was lower (P=0.009), but generalized estimating equations showed no significant overall group effect or group-by-time interaction for resting or coughing pain. Nausea occurred in 11.1% versus 28.9% (Pearson χ² P=0.035).

Conclusion

Adding dexketoprofen trometamol reduced 48-hour sufentanil consumption in the analyzed population. Pain differences were time-specific secondary findings and did not establish sustained analgesic superiority. The nausea finding requires confirmation, and the trial was insufficient to establish safety for uncommon adverse events.

Introduction

Video-assisted thoracoscopic surgery (VATS) can cause clinically important postoperative pain from incision trauma, intercostal nerve injury, and pleural irritation. Pain may impair coughing and mobilization and can persist beyond the acute postoperative period.1,2 Effective analgesia is therefore an important component of enhanced recovery after surgery. Procedure-specific guidance recommends paracetamol with an NSAID or cyclooxygenase-2 inhibitor and regional analgesia, with opioids used for rescue.3

Sufentanil is used in PCIA because of its potent analgesic effects,4 but opioid exposure is associated with nausea, vomiting, pruritus, urinary retention, and respiratory depression.5 Non-opioid adjuncts can reduce opioid requirements through complementary analgesic mechanisms.6,7

Dexketoprofen is the active S-enantiomer of ketoprofen; its trometamol formulation has a rapid onset of action and is used for acute postoperative pain.8,9 These properties motivate its evaluation as a systemic adjunct, but do not establish superiority over ketorolac, parecoxib, or other NSAIDs. Nonselective NSAIDs also require attention to gastrointestinal, renal, platelet-related bleeding, and respiratory adverse effects in susceptible patients.9 Evidence specifically evaluating dexketoprofen added to sufentanil-based PCIA after thoracoscopic lung cancer resection remains limited.

This trial evaluated the incremental opioid-sparing effect of dexketoprofen trometamol added to sufentanil-based PCIA after thoracoscopic radical resection of lung cancer. Cumulative sufentanil consumption over 48 hours was prespecified as the primary endpoint for three reasons. First, the PCIA regimen—including the co-administered dexketoprofen—was prescribed for 48 hours, so cumulative sufentanil consumption at 48 hours reflects total opioid exposure over the entire treatment period, the parameter most directly related to opioid-related adverse outcomes and recovery. Second, postoperative pain commonly extends beyond the first 24 hours: in the multicentre OCTOPUS trial, 28% of patients still reported a numerical rating scale score >3 at 48 hours despite morphine PCA, and the morphine-sparing benefit of multimodal non-opioid regimens was sustained through 48 hours.10 Third, after thoracoscopic surgery, intercostal chest drains typically remain in situ for 24–72 hours (median 1–2 days in published series).11,12 Pain on coughing and mobilisation during this period remains a principal source of morbidity, as the chest drain stimulates both the parietal pleura and the intercostal nerves. The current PROSPECT guidelines for video-assisted thoracoscopic surgery therefore recommend continuing multimodal opioid-sparing analgesia throughout this period.3 Pain scores and adverse events were secondary outcomes; longer-term recovery and chronic pain were outside the study follow-up.

Methods

Study Design and Ethical Approval

This randomized controlled double-blind clinical trial was performed at the Hangzhou First People’s Hospital, affiliated with Westlake University School of Medicine, in China, from October 2024 to April 2025. The study protocol was approved by the hospital ethics committee in July 2024 (Approval No. IIT-20240701-0233-02). The trial was prospectively registered at the Chinese Clinical Trial Registry (https://www.chictr.org.cn) on 19 July 2024 (Registration No. ChiCTR2400087079, https://www.chictr.org.cn/showproj.html?proj=238056). All study participants signed written informed consent. The conduct and reporting of this clinical trial adhered to the CONSORT guidelines.

Participants and Surgical Procedures

Patients scheduled for elective thoracoscopic radical resection of lung cancer and postoperative PCIA were eligible if they had American Society of Anesthesiologists (ASA) physical status I–III and no coagulation dysfunction. Exclusion criteria were allergy or other contraindication to the study medications or PCIA; severe cardiac, hepatic, or renal dysfunction; airway hyperresponsiveness; pregnancy or lactation; concurrent participation in another clinical trial; or other investigator-determined conditions precluding participation. Written informed consent was obtained before enrollment.

All procedures were unilateral uniportal VATS for a single pulmonary lesion and comprised anatomical segmentectomy with systematic or lobe-specific mediastinal lymph node dissection or sampling. Lobectomy and wedge resections were not included. The same fixed thoracic surgical team performed the operations.

Prespecified criteria for withdrawal after randomization but before initiation of the allocated PCIA intervention were withdrawal of consent, cancellation of surgery, and operative duration exceeding 2 hours. The duration criterion was intended to limit heterogeneity in intraoperative anesthetic and analgesic exposure associated with prolonged procedures.

Randomization and Blinding

After preoperative screening, written informed consent, and enrollment, 108 patients were randomized before surgery in a 1:1 ratio to Group S (sufentanil) or Group R (dexketoprofen trometamol plus sufentanil). Simple randomization without blocking or stratification was used; the sequence was generated in advance using SPSS 29.0 (IBM Corp., Armonk, NY, USA), and allocation was concealed in sequentially numbered, opaque, sealed envelopes opened after enrollment. The allocated PCIA intervention began postoperatively. An independent nurse who was not involved in outcome assessment, data collection, or statistical analysis prepared the study solutions. In Group S, 0.9% sodium chloride replaced dexketoprofen trometamol; both groups had a final reservoir volume of 100 mL, with identical PCIA pumps and labeling. Patients and the clinical, outcome-assessment, and analysis teams were blinded to allocation; the drug-preparation nurse was not part of these blinded teams.

Preanesthesia Preparation and Induction

Selected patients were required to fast and abstain from drinking for more than 8 hours before surgery. Upon entering the operating room, intravenous access was established in an upper extremity, and routine monitoring of non-invasive blood pressure, invasive arterial blood pressure, electrocardiogram (ECG), and pulse oxygen saturation (SpO₂) was performed. As part of the standardized protocol in both groups, before induction of anesthesia, atropine 0.01 mg·kg⁻¹ and dexamethasone 0.1 mg·kg⁻¹ were injected intravenously. All patients underwent general anesthesia with endotracheal intubation, and pure oxygen (6 L·min⁻¹) was inhaled via a face mask for more than 3 minutes. The induction of anesthesia regimen consisted of midazolam 0.01 mg·kg⁻¹, propofol 2 mg·kg⁻¹, cisatracurium 0.05 mg·kg⁻¹, and sufentanil 0.5 μg·kg⁻¹. After 5 minutes of assisted ventilation, intubation was performed. Baseline heart rate (HR) and mean arterial pressure (MAP) were measured before induction after patients entered the operating room and rested for several minutes.

Maintenance of Anesthesia and Intercostal Nerve Block

After endotracheal intubation, ventilation used a tidal volume of 6–8 mL·kg⁻¹ and respiratory rate of 12–18 breaths·min⁻¹, adjusted to maintain end-tidal carbon dioxide at 35–45 mmHg. Both groups received sevoflurane 1.0%–2.0%, propofol 3–5 mg·kg⁻¹·h⁻¹, and remifentanil initially at 0.1 μg·kg⁻¹·min⁻¹, titrated to a bispectral index of 40–60. Sufentanil 0.1 μg·kg⁻¹ was administered before skin incision and after wound closure, in addition to the induction dose of 0.5 μg·kg⁻¹, giving a protocol-specified intraoperative total of 0.7 μg·kg⁻¹ in both groups. Cumulative remifentanil exposure was not prospectively recorded and could not be reliably retrieved. Sevoflurane was stopped 5 minutes before the end of surgery, and intravenous maintenance infusions were stopped at completion.

At the end of surgery, the same trained surgical team performed the standardized thoracoscopic intercostal nerve block under direct visualization on the operative side. The incision interspace and two interspaces above and below it were blocked, for five levels in total. Each level received 3 mL of 0.5% ropivacaine, for a total of 15 mL (75 mg). The protocol was identical in both groups.

Postoperative PCIA and Rescue Analgesia

After surgery, an anesthesiology nurse not involved in allocation connected the PCIA pump in the postanesthesia care unit (PACU). Group S received sufentanil 2 μg·kg⁻¹ and tropisetron 10 mg diluted with 0.9% sodium chloride to 100 mL. Group R additionally contained dexketoprofen trometamol 300 mg in the same 100-mL final volume. The background rate was 1 mL·h⁻¹, the bolus volume 2 mL, and the lockout interval 30 minutes. PCIA was prescribed for 48 hours.

The 300-mg amount denotes reservoir content, rather than a fixed administered dose. At the stated concentration of 3 mg/mL, the programmed background infusion corresponds to 3 mg/h (144 mg over an uninterrupted 48 hours), with 6 mg per successfully delivered bolus. Cumulative exposure should be determined as the pump-recorded delivered volume multiplied by 3 mg/mL. Parenteral prescribing information permits up to 150 mg of dexketoprofen per day for no more than 2 days;9 a 300-mg reservoir over 48 hours does not itself establish compliance with a daily maximum because bolus delivery varies. Study-specific physicochemical compatibility and stability testing of the dexketoprofen–sufentanil–tropisetron mixture was not performed.

The same rescue protocol applied throughout the 48-hour observation period in the PACU and thoracic surgery ward. Intravenous sufentanil 0.1 μg·kg⁻¹ was administered upon patient request when pain remained ≥4 despite PCIA activation, or when resting VAS was ≥7. Ward staff did not routinely administer additional systemic analgesics, including acetaminophen, flurbiprofen axetil, or other NSAIDs. All rescue sufentanil was included in the total 48-hour postoperative consumption.

In the PACU, blood pressure, HR, and SpO₂ were monitored continuously, and awakening time, agitation, and emergence pain were recorded. Propofol 50 mg was administered intravenously for agitation. Patients were discharged from the PACU when the Aldrete score reached 9.

Outcome Measurements

The primary outcome was cumulative postoperative sufentanil consumption over 48 hours, including the background PCIA infusion, delivered boluses, and rescue doses. Secondary outcomes were VAS pain scores at rest and during coughing at 1, 6, 12, 24, and 48 hours; total PCA attempts; effective PCA attempts; and the effective attempt ratio. The ratio was calculated separately for each patient as effective attempts divided by total attempts and was then summarized across patients; it was not calculated from the group medians. Cumulative pain burden was summarized by the area under the VAS–time curve (AUC) from 1 to 48 hours, using the trapezoidal rule.

Safety assessments within 48 hours included nausea, vomiting, pruritus, urinary retention, clinically apparent gastrointestinal symptoms or bleeding, postoperative bleeding, and other suspected drug-related events; thoracic drainage volume was recorded. The recorded oxygen-desaturation episodes (defined as a single SpO₂ reading <92% on ward pulse oximetry) are reported descriptively rather than classified as confirmed respiratory depression. Postoperative serum creatinine, blood urea nitrogen, and estimated glomerular filtration rate were not systematically reassessed. Trained research coordinators blinded to allocation recorded outcomes on standardized case report forms.

Sample Size Calculation

The study was planned using G*Power 3.1.9 with an a priori assumed 30% reduction in opioid consumption, a two-sided α of 0.05, and 80% power; the stated target was 45 participants per group. The 48-hour primary endpoint was selected to cover the complete PCIA period, but directly comparable 48-hour planning data were unavailable. The external reference, Yoon et al,13 reports 24-hour fentanyl consumption of 450 ± 245 μg, rather than direct sufentanil data. Applying these data to cumulative 48-hour sufentanil consumption entails cross-drug and cross-timepoint extrapolation, which limits the planning rationale. The assumed 30% effect is not an established minimum clinically important difference. Interpretation of the trial is based on the observed effect estimate and its confidence interval.

Statistical Analysis

Analyses included the 90 participants who received their allocated PCIA intervention and completed follow-up. Because 18 randomized participants were excluded before intervention, this was not an intention-to-treat analysis of all 108 randomized participants. Statistical analysis used R version 4.2.2. Normality was assessed with the Shapiro–Wilk test. Continuous variables were summarized as mean ± SD or median (interquartile range) and compared using independent-samples t tests or Mann–Whitney U-tests, respectively. Categorical variables were summarized as n (%) and compared by Pearson χ² or Fisher’s exact test according to expected cell counts. Mean differences and Hodges–Lehmann location-shift estimates were reported with 95% confidence intervals.

Longitudinal resting and coughing VAS scores were analyzed separately using generalized estimating equations (GEE) with an exchangeable working correlation structure to account for within-patient correlations across repeated postoperative measurements. Robust sandwich standard errors were used. The models included group, time, and the group-by-time interaction. The assessment times and use of Bonferroni correction were prespecified. The original approach treated resting and coughing scores as separate five-comparison families (threshold P<0.01). During revision, all 10 time-point comparisons were treated as one family with a more conservative threshold of P<0.005 (0.05/10). AUC comparisons were analyzed separately at P<0.05; their relationship to the time-point comparison family was not explicitly prespecified. AUC and individual time-point findings are interpreted as secondary observations.

For the adverse-event analysis, binary effects are risk ratios (RRs), defined as risk in Group R divided by risk in Group S, with 95% confidence intervals calculated on the log scale. Pearson χ² was used for nausea (expected event count 9 per group); two-sided Fisher’s exact tests were used for the other binary outcomes. The drainage-volume effect is the mean difference (Group R minus Group S). These safety analyses were not adjusted for multiplicity. All tests were two-sided.

Results

Participant Flow and Characteristics

Of 130 patients assessed for eligibility, 22 were excluded before randomization: 6 did not meet the inclusion criteria, 15 declined participation, and 1 was excluded for another reason. The remaining 108 patients were randomized preoperatively, 54 to each group. Before PCIA initiation, 5 patients in Group R withdrew consent and 4 had surgery lasting more than 2 hours; in Group S, 8 withdrew consent and 1 had surgery canceled. Thus, 45 patients in each group received the allocated intervention and were included in the analysis. There was no loss to follow-up among these 90 patients and no further exclusion from analysis (Figure 1).

Figure 1.

CONSORT participant flow. Of 130 patients assessed, 108 were randomized (54 per group); 45 per group received the allocated intervention and were analyzed. Group S, sufentanil-based PCIA; Group R, dexketoprofen trometamol plus sufentanil-based PCIA.

Abbreviations: CONSORT, Consolidated Standards of Reporting Trials; PCIA, patient-controlled intravenous analgesia.

The available demographic and perioperative characteristics are shown in Table 1. Although ASA I–III was permitted by the eligibility criteria, the analyzed cohort included only ASA I and II patients. HR and MAP refer to measurements before anesthesia induction.

Table 1.

Baseline and Perioperative Characteristics

Characteristic	S Group N = 45	R Group N = 45	SMD (95% CI)
Age, years	56 (53, 61)	55 (48, 61)	0.23 (−0.19, 0.64)
Gender, male/female	24/21	23/22	0.04 (−0.37, 0.46)
Body mass index, kg/m2	23.89 ± 1.88	23.66 ± 2.21	0.11 (−0.30, 0.53)
ASA classification			0.04 (−0.37, 0.46)
I	20 (44.4%)	21 (46.7%)
II	25 (55.6%)	24 (53.3%)
Duration of surgery, min	83 ± 18	81 ± 18	0.12 (−0.29, 0.54)
Preinduction HR, beats/min	78 ± 9	78 ± 10	0.03 (−0.39, 0.44)
Preinduction MAP, mmHg	96 ± 7	95 ± 8	0.16 (−0.25, 0.57)

Notes: Data are mean ± SD, median (IQR), or n (%). Group S: sufentanil; Group R: dexketoprofen trometamol plus sufentanil. HR and MAP were measured before anesthesia induction after several minutes of rest. No analyzed participant was ASA III. SMDs and their CIs are retained from the original statistical output.

Abbreviations: SMD, standardized mean difference; CI, confidence interval; ASA, American Society of Anesthesiologists; HR, heart rate; MAP, mean arterial pressure.

Sufentanil Consumption and PCA Use

Total 48-hour postoperative sufentanil consumption, including rescue doses, was 96.78 ± 13.69 μg in Group S and 78.03 ± 10.66 μg in Group R. The reduction was 18.75 μg (Group S minus Group R; 95% CI 13.61 to 23.89; P<0.001), or approximately 19% relative to Group S (Table 2).

Table 2.

Sufentanil Consumption and PCA Use Within 48 Hours

Variables	S Group (n=45)	R Group (n=45)	Mean Difference or Hodges–Lehmann Estimate (95% CI)	P value
Sufentanil consumption, μg	96.78 ± 13.69	78.03 ± 10.66	18.75 (13.61 to 23.89)	<0.001
Total PCA attempts	17.00 (15.00–18.00)	15.00 (13.00–16.00)	2.00 (1.00 to 3.00)	0.003
Effective PCA attempts	13.00 (12.00–14.00)	15.00 (13.00–16.00)	1.00 (0.00 to 2.00)	0.012
Effective attempt ratio	0.80±0.10	0.81±0.08	−0.02 (−0.05 to 0.02)	0.426

Notes: Data are mean ± SD or median (IQR). Differences are Group S minus Group R. The primary outcome includes PCIA infusion, delivered boluses, and rescue sufentanil. Effective attempt ratio is calculated for each patient as effective attempts/total attempts and then summarized; it is not the ratio of group medians.

Abbreviations: PCA, patient-controlled analgesia; CI, confidence interval.

Total PCA attempts were lower in Group R than in Group S: median 15.00 (IQR 13.00–16.00) versus 17.00 (15.00–18.00), with a Hodges–Lehmann estimate of 2.00 (Group S minus Group R; 95% CI 1.00 to 3.00; P=0.003). Effective attempts and the effective attempt ratio are reported in Table 2. The effective attempt ratio was 0.80 ± 0.10 in Group S and 0.81 ± 0.08 in Group R (P=0.426); group medians of attempt counts cannot reconstruct this patient-level ratio.

Two patients in Group S (2/45, 4.4%) each received one rescue dose of sufentanil 0.1 μg·kg⁻¹; no patient in Group R received rescue analgesia. Rescue medication contributed to the reported total 48-hour consumption. The reservoir amount of dexketoprofen should not be interpreted as the actual cumulative dose delivered. Using the recorded effective PCA attempts and assuming an uninterrupted 48-hour background infusion, the calculated cumulative dexketoprofen dose in Group R was 215.6 ± 8.1 mg over 48 hours (median 216.0 mg, interquartile range 210.0–222.0 mg; range 198–240 mg). The calculation was 144 mg from background infusion plus 6 mg per effective attempt. These 48-hour totals do not specify the dose delivered during each 24-hour period.

Postoperative VAS Scores

Resting-pain AUC from 1 to 48 hours was lower in Group R than in Group S (84.64 ± 22.75 vs 97.09 ± 21.68; reported mean difference, Group S minus Group R, 12.44; 95% CI 3.13 to 21.75; P=0.009). Resting VAS scores at 12 and 24 hours were lower in Group R (both unadjusted P=0.004), meeting the revised Bonferroni threshold of P<0.005; other resting time-point comparisons did not meet this threshold (Figure 2a).

Figure 2.

Postoperative VAS pain scores at rest (a) and during coughing (b). Group S, sufentanil-based PCIA; Group R, dexketoprofen trometamol plus sufentanil-based PCIA. The revised significance threshold is P<0.005 for the 10 time-point comparisons; coughing VAS at 12 hours (P=0.006) does not meet that threshold.

Abbreviations: VAS, visual analogue scale; AUC, area under the curve; PCA, patient-controlled analgesia.

Coughing-pain AUC did not differ significantly (134.38 ± 26.37 in Group R vs 145.16 ± 28.36 in Group S; mean difference, Group S minus Group R, 10.78; 95% CI −0.70 to 22.25; P=0.065). The unadjusted P value for coughing VAS at 12 hours was 0.006 and did not meet the revised P<0.005 threshold. No coughing time-point comparison met that threshold (Table 3 and Figure 2b).

Table 3.

Postoperative VAS Scores and Cumulative Pain AUC

Variables	S Group (n=45)	R Group (n=45)	Effect Estimate (Reported 95% CI)	Unadjusted P value
Cumulative AUC at rest
AUC1h-48h	97.09±21.68	84.64±22.75	12.44 (3.13 to 21.75)	0.009
VAS at rest
1 h	2.00 (1.00–2.00)	2.00 (1.00–2.00)	0 (0–0)	0.830
6 h	3.00 (2.50–3.00)	3.00 (2.00–3.00)	0 (0–0)	0.387
12 h	3.00 (2.00–3.00)	2.00 (2.00–3.00)	1(0–1)	0.004
24 h	2.00 (2.00–3.00)	2.00 (1.50–2.00)	0(0–1)	0.004
48 h	1.00 (1.00–1.50)	1.00 (1.00–1.50)	0 (0–0)	0.763
Cumulative AUC with cough
AUC1h-48h	145.16±28.36	134.38±26.37	10.78 (−0.70 to 22.25)	0.065
VAS during coughing
1 h	3.00 (2.00–3.00)	3.00 (2.00–3.00)	0 (0–0)	0.483
6 h	4.00 (4.00–5.00)	4.00 (3.00–4.00)	0 (0–0)	0.054
12 h	4.00 (3.00–4.00)	3.00 (3.00–4.00)	1(0–1)	0.006
24 h	3.00 (3.00–3.50)	3.00 (2.50–3.00)	0 (0–0)	0.078
48 h	2.00 (2.00–3.00)	2.00 (1.50–2.00)	0 (0–0)	0.955

Notes: Data are mean ± SD for AUC or median (IQR) for individual VAS scores. Effect estimates are Group S minus Group R: mean differences for AUC and Hodges–Lehmann estimates for time-point VAS scores. P values are unadjusted. During revision, Bonferroni correction treated all five resting and five coughing comparisons as one family (P<0.005); only resting scores at 12 and 24 hours met this threshold. AUC comparisons were separate secondary analyses at P<0.05.

Abbreviations: VAS, visual analogue scale; AUC, area under the curve from 1 to 48 hours, in VAS score·hours; CI, confidence interval.

The results of the generalized estimating equation analysis showed that in the resting state, the time effect was statistically significant (Wald χ² = 47.59, P = 0.003), while the group effect (Wald χ² = 0.02, P = 0.897) and the time-by-group interaction effect (Wald χ² = 1.82, P = 0.393) were not significant. In the coughing state, the time effect was also significant (Wald χ² = 51.75, P = 0.003), but the group effect (Wald χ² = 0.72, P = 0.396) and the time-by-group interaction effect (Wald χ² = 7.13, P = 0.120) did not reach statistical significance (Table S1).

Postoperative Adverse Events

Nausea occurred in 5/45 patients (11.1%) in Group R and 13/45 (28.9%) in Group S (RR 0.38, 95% CI 0.15 to 0.99; Pearson χ² P=0.035). A two-sided Fisher’s exact sensitivity calculation gave P=0.063, indicating that this secondary finding is sensitive to the test used. Other recorded binary outcomes showed no statistically significant between-group differences (Table 4). Drainage volume at 48 hours was 59.00 ± 17.89 mL in Group S and 60.67 ± 17.98 mL in Group R (mean difference, Group R minus Group S, 1.67 mL; 95% CI −5.84 to 9.18; P=0.660).

Table 4.

Adverse Events Within 48 Hours

Adverse Events	S Group (n=45)	R Group (n=45)	Effect Estimate (95% CI)	P value
Nausea, n (%)	13 (28.9)	5 (11.1)	RR 0.38 (0.15 to 0.99)	0.035
Vomiting, n (%)	3 (6.7)	2 (4.4)	RR 0.67 (0.12 to 3.80)	1.000
Pruritus, n (%)	4 (8.9)	4 (8.9)	RR 1.00 (0.27 to 3.75)	1.000
Urinary retention, n (%)	4 (8.9)	2 (4.4)	RR 0.50 (0.10 to 2.59)	0.677
Transient oxygen desaturation, n (%)	5 (11.1)	3 (6.7)	RR 0.60 (0.15 to 2.36)	0.714
Postoperative thoracic drainage volume at 48 h, mL	59.00±17.89	60.67±17.98	MD 1.67 (−5.84 to 9.18)	0.660

Notes: Binary outcomes are n (%); drainage volume is mean ± SD. Risk ratios (RR) compare Group R with Group S; 95% CIs use the log-RR Wald method. Nausea uses Pearson χ² (expected event counts ≥5); other binary outcomes use two-sided Fisher’s exact tests. For nausea, Fisher’s exact sensitivity P=0.063. The drainage estimate is the mean difference (MD), Group R minus Group S, with the original CI reoriented accordingly; the P value is from the independent-samples t test. No adjustment was made across adverse-event outcomes. Oxygen-desaturation events are not confirmed respiratory depression.

Abbreviation: CI, confidence interval.

Eight transient postoperative oxygen-desaturation episodes were recorded (5 patients in Group S and 3 in Group R), all within 6 hours of transfer to the ward. Recorded nadir SpO₂ values were 90%, 90%, 89%, 91%, 90%, 90%, 88%, and 91%. These episodes resolved after increased supplemental oxygen. No patient received naloxone, required non-invasive ventilation or ICU transfer, or discontinued PCIA. These observations do not establish confirmed opioid-induced respiratory depression or comparative pulmonary safety.

Discussion

Adding dexketoprofen trometamol to sufentanil-based PCIA was associated with an approximately 19% reduction in total 48-hour sufentanil consumption in the analyzed population. This finding is consistent with the opioid-sparing rationale for non-opioid adjuncts supported by a network meta-analysis of major-surgery trials.14 In a randomized trial of major orthopaedic surgery, Hanna et al found that intramuscular dexketoprofen reduced 24-hour morphine PCA consumption compared with placebo.15 Earlier thoracic research evaluated pre-emptive intravenous dexketoprofen together with epidural analgesia after open thoracotomy.16 More directly relevant to VATS, Yang et al reported lower 48-hour sufentanil consumption with paravertebral block combined with parecoxib.17 Differences in surgical approach, NSAID, administration route, regional technique, and comparator preclude direct comparison of effect sizes. Our trial therefore adds evidence for dexketoprofen as an adjunct to sufentanil-based PCIA and standardized intercostal block after uniportal anatomical segmentectomy; it does not establish superiority over other multimodal regimens. The observed reduction was smaller than the 30% effect assumed during planning, and statistical significance does not establish a clinically important opioid reduction.

Lower resting VAS scores at 12 and 24 hours and a lower resting-pain AUC were secondary findings. The median resting VAS difference was 1 point at 12 hours and 0 points at 24 hours, underscoring the distinction between distributional significance and the difference between medians. Myles et al estimated a minimal clinically important change of approximately 10 mm on a 100-mm postoperative pain VAS.18 That within-patient change threshold is not automatically a validated threshold for a between-group median difference in this trial. Clinical importance was not established against a prespecified validated threshold. Coughing VAS at 12 hours did not meet the revised multiplicity threshold. GEE showed no significant group effect or group-by-time interaction for either pain condition; the results therefore do not demonstrate sustained improvement in the overall pain trajectory.

Current VATS recommendations favor paracetamol plus an NSAID or COX-2 inhibitor alongside regional analgesia, with opioids reserved for rescue.3 ERAS/ESTS lung-surgery guidance similarly places opioid-sparing analgesia within a pathway that includes early mobilization and chest-drain management.19 Our comparator included sufentanil-based PCIA and a standardized intercostal block but no routine additional systemic non-opioid analgesic. Consequently, the trial does not compare dexketoprofen with a complete contemporary basic-analgesia regimen, another NSAID, or another regional technique. In the multicenter OPtriAL trial, intercostal block met the prespecified pain noninferiority criterion relative to thoracic epidural analgesia after thoracoscopic anatomical lung resection, whereas paravertebral block did not; quality of recovery was similar across groups.20 This provides context for our use of intercostal block without establishing equivalence between our five-level protocol and the OPtriAL regimen. Dexketoprofen should be interpreted as an adjunct to regional analgesia. Whether its opioid-sparing effect improves mobilization, pulmonary recovery, or hospital stay remains unknown because these outcomes were not assessed. COX inhibition provides a pharmacological rationale; biological synergy was not directly investigated.

The lower observed nausea frequency was a secondary finding with few events, no across-outcome multiplicity correction, and sensitivity to the choice of statistical test. Reduced opioid exposure is a possible explanation, consistent with consensus guidance that recommends minimizing perioperative opioids to reduce baseline postoperative nausea and vomiting risk.21 However, a causal mediation mechanism was not tested in this trial. Dexketoprofen is a nonselective NSAID with potential renal, gastrointestinal, bleeding, and respiratory adverse effects.9 A Cochrane review found uncertainty about the effect of perioperative NSAIDs on acute kidney injury even in adults with normal preoperative kidney function.22 The absence of statistically significant differences in our other recorded events cannot establish safety equivalence. Short-term clinical observation and thoracic drainage volume do not substitute for systematic renal testing or comprehensive assessment of bleeding and pulmonary complications.

Limitations

This single-center trial followed patients for only 48 hours. Functional recovery, time to mobilization, hospital length of stay, chronic pain, and quality of life were not systematically assessed. The study was not designed to detect uncommon serious adverse events; renal laboratory measurements and postoperative pulmonary complications were not systematically evaluated. Exclusion of patients with major organ dysfunction, airway hyperresponsiveness, or coagulation abnormalities limits applicability to higher-risk populations.

Post-randomization exclusions may introduce selection or attrition bias and compromise the comparability created by randomization. The analyzed population excludes prolonged procedures and patients who did not receive PCIA, and the findings do not represent an intention-to-treat estimate for all randomized patients. The sample-size rationale relies on external data for another opioid and another accumulation period; the original planning assumptions require cautious interpretation. The observed effect was also smaller than the assumed 30% reduction.

Unmeasured baseline factors may influence pain and opioid use; comorbidity profiles (including hypertension, diabetes mellitus, and pulmonary comorbidity such as COPD), history of ipsilateral thoracic surgery, smoking status, chronic pain history, preoperative analgesic use, anxiety/depression, and preoperative pain scores were not systematically collected as baseline variables. Cumulative remifentanil exposure was unavailable. Although the intercostal block was standardized, it may have lowered pain and opioid requirements in both groups and attenuated the measurable incremental effect of dexketoprofen. VAS outcomes are subjective. Study-specific admixture stability was not tested, and the nominal reservoir content cannot establish individual delivered exposure or daily dose compliance.

Conclusion

In patients receiving the allocated treatment after thoracoscopic radical resection of lung cancer, adding dexketoprofen trometamol to sufentanil-based PCIA reduced 48-hour sufentanil consumption by approximately 19%. Lower resting pain at selected time points and the resting-pain AUC were secondary findings; the overall longitudinal analyses did not establish sustained analgesic superiority. The nausea finding requires confirmation. Larger studies with complete randomized follow-up, patient-centered recovery outcomes, and systematic safety assessment are needed.

Funding Statement

This work was supported by the Young Scientists Fund of the National Natural Science Foundation of China (No. 82402582).

Data Sharing Statement

The datasets used and/or analyzed during the current study are available from Jianliang Sun and Xinlei Lu upon reasonable request.

Ethics Approval and Consent to Participate

This study was performed in line with the principles of the Declaration of Helsinki and approved by the Medical Ethics Committee of the Affiliated Hangzhou First People’s Hospital, School of Medicine, Westlake University (IIT-20240701-0233-02). The study was registered at the China Clinical Trials Registry (registration number: ChiCTR2400087079; Date of Registration: 2024-07-19). Written informed consent was obtained from all participants, or, when applicable, from their legal representatives.

Disclosure

The authors declare no competing interests.