In this single-center retrospective cohort study, we retrieved and analyzed the clinical and follow-up data of patients undergoing thoracoscopic lung nodule surgery at Nanjing Drum Tower Hospital from September 2021 to August 2022, and those who met the inclusion and exclusion criteria were retrospectively followed up with by telephone at 1 year post-surgery. This study was performed in accordance with the Helsinki Declaration and approved by the Ethics Committee of Nanjing Drum Tower Hospital (approval number 2023-296). The ethics committee was aware the 1-year follow-up would be conducted prospectively and waived the requirement for the signing of the informed consent form. When we followed up with patients, we explained the purpose of our study and indicated that the results might be published. The writing of this study adhered to Strengthening the Reporting of Observational studies in Epidemiology (STROBE).

The inclusion criteria of this study included (1) patients who were 18–85 years of age, (2) patients who underwent elective thoracoscopic pulmonary nodule surgery, and (3) patients whose American Society of Anesthesiologists (ASA) classification was 1–3. The exclusion criteria included (1) patients transferred to the intensive care unit (ICU) after surgery; (2) patients who underwent bilateral pulmonary nodule resection or re-do cardiothoracic surgery; (3) patients with a history of malignant tumors; (4) patients whose pathological finding was small-cell lung cancer or who needed postoperative radiotherapy and chemotherapy; (5) patients with severe cardiopulmonary, liver, and kidney function impairment before surgery; (6) patients with chronic pain or severe preoperative anxiety and depression status on long-term medication; (7) patients discharged from hospital with chest drain tubes; and (8) patients who refused follow-up or had incomplete follow-up data.

Patients were divided into a DEX group and a control group according to whether dexmedetomidine was infused intravenously on the night after surgery or not. Patients in the DEX group were injected with 0.2 μg/kg/h dexmedetomidine, initiated from 9:00 p.m. on the day of surgery to 6:00 a.m. on POD1, while patients in the control group received the same postoperative interventions except for dexmedetomidine infusion.

We calculated the sample size based on the study by Li et al. They compared the effect of perioperative application of dexmedetomidine on CPSP after open-heart surgery, where the incidence of CPSP was 7.89% in the dexmedetomidine group compared with 27.03% in the control group [16]. With 90% power and 5% alpha error (two sided), we calculated that a minimum sample size of 91 per group was required. Thus, the sample size was adequate both before and after matching in our retrospective study.

Routine anesthesia protocol was used. Preoperative nerve block was performed at the pre-anesthesia holding area. Anesthesia induction was performed with midazolam, fentanyl, propofol, and vecuronium bromide, and total intravenous anesthesia maintenance was performed with remifentanil, propofol, vecuronium bromide, and dexmedetomidine to maintain the bispectral index (BIS) at 40–60, with 50 μg fentanyl intermittently injected as light anesthesia during surgery. Thoracoscopic surgery consisted of single-port and two-port procedures. Single-port thoracoscopic surgery involved a 3–4 cm incision from the anterior axillary line to the mid-axillary line of the fourth or fifth intercostal space on the operated side, which served as an observation hole as well as an operation hole. In two-port thoracoscopic surgery, a 1–2 cm incision was made in the mid-axillary line of the sixth or seventh intercostal space on the operated side as an observation hole, and a 3 cm incision was made in the anterior axillary line to the mid-axillary line of the third or fourth intercostal space as an operation hole. The extent of resection was determined by the size as well as the type of pathology of the pulmonary nodule.

When the surgery was completed, the patients were given 4 mg ondansetron to prevent postoperative nausea and vomiting (PONV). Every patient was equipped with patient-controlled intravenous analgesia for postoperative analgesia consisting of 2 µg/kg sufentanil (up to 100 µg), 0.1 mg/kg ondansetron, and 10 mg dexamethasone in 100 ml saline before they were transferred from the post-anesthesia care unit (PACU). We routinely evaluated the patients’ grip strength and their ability to sustain a head lift off the bed for over 5 s before extubation in combination with quantitative neuromuscular blockage assessment using a TOF-watch® monitor to avoid residual muscle relaxation. In addition, all patients had a Steward score ≥ 5 before they were discharged from the PACU.

In the DEX group, 0.2 μg/kg/h dexmedetomidine was pumped intravenously from 9:00 p.m. on the night of the operation to 6:00 a.m. the next day, and if clinically significant hypotension (systolic blood pressure < 90 mmHg or 30% reduction of baseline) or bradycardia (heart rate < 50 beats per minute) occurred during the procedure, dexmedetomidine was discontinued. On POD1, patients were encouraged to cough and move out of bed to promote postoperative pulmonary rehabilitation. During hospitalization, multimodal analgesia including patient-controlled intravenous analgesia and non-steroidal anti-inflammatory drugs or weak opioids were administrated to maintain pain visual analogue scale < 4. Oral analgesics were given to the patients for 7 days after discharge.

An independent researcher reviewed the patients’ electronic medical record system and anesthesia system to extract the following data: sex, age, ASA classification, body mass index (BMI), education level, comorbidities, nodule size, duration of operation, thoracoscopic port, extent of excision, types of nerve block, intraoperative use of vasoactive medications, fentanyl dosage, remifentanil dosage, fluid intake and output, postoperative discontinuation of dexmedetomidine, postoperative days of drainage, and postoperative rescue analgesia. Another independent researcher followed up with patients who met the study criteria by telephone at 1 year after the operation, asking whether the patient was suffering from CPSP. If the patient answered that they were suffering from CPSP, the patient’s pain status was further evaluated by using the 11-point numeric rating scale (NRS) to assess the patient’s pain at rest and on movement in the last 24 h, where an NRS score of 0 represented “no pain” and a score of 10 represented the “worst pain imaginable.” Then the Douleur Neuropathique 4 questions (DN4), containing 10 questions related to neuropathic pain, would be used to assess neuropathic pain. If four of these items were met, the patient was diagnosed with neuropathic pain [17]. Furthermore, we also assessed the quality of patients’ recovery using the 12-Item Short Form Health Survey (SF-12), which consisted of 12 entries from eight dimensions, and the final physical and mental scores, as well as the total score, were calculated [18].

The primary outcome was the incidence of CPSP at 1 year after the operation. The secondary outcomes included NRS pain score at rest and on movement, the incidence of postoperative neuropathic pain, the incidence of moderate-to-severe pain (NRS ≥ 4), quality of recovery according to SF-12 scores, and postoperative rescue analgesia. The discontinuation of dexmedetomidine was presented as descriptive information. We then analyzed the differences between the CPSP and non-CPSP groups to study the risk factors for the development of CPSP.

Data were analyzed using SPSS26.0 and R4.3.1. All statistical tests were two-sided tests, and P < 0.05 was considered statistically significant. The Kolmogorov–Smirnov method was used to test the data distribution. Measurements data obeying normal distribution were described by X(± S), and the t- or t′-test was used for comparison between groups; measurements data not obeying normal distribution was described by M[P25, P75], and the Mann–Whitney U test was used for comparison between groups. Count data were described by frequency (ratio), and the Chi-square test was used for comparison between groups. For missing perioperative data, the multiple interpolation method was used to insert the missing data.

Propensity score matching was used, and matching data included age, sex, BMI, ASA classification, education level, comorbidities (hypertension, diabetes, coronary artery disease, history of cerebral infarction), nodule size, duration of operation, thoracoscopic port, extent of excision, types of nerve block, intraoperative use of vasoactive medications, fentanyl dosage, remifentanil dosage, fluid intake and output (bleeding, urine, blood transfusion, and fluid transfusion), and postoperative days of drainage. Patients were matched in a 1:1 ratio using the nearest neighbor matching method, with caliper set to 0.01 without replacement. Baseline data before and after matching was assessed using standardized mean difference (SMD) to estimate the balance between the two groups, with a SMD less than 0.1 indicating balance.

For the analysis of CPSP risk factors, multifactorial logistic regression was used, and the indicators with P < 0.1 in the univariate analysis were included. To avoid multicollinearity among multiple indicators, logistic regression was performed using a stepwise forward approach. Finally, the indicators with significance in multivariate logistic regression were shown in the form of a nomogram to reveal the predictive relationship of each indicator on CPSP, which was evaluated using concordance index (C-index).