Abstract
Background: Distinguishing expected postoperative inflammation from early infection remains challenging after thoracic surgery; serial C-reactive protein (CRP) is widely used to aid this differentiation. Methods: We conducted a single-centre retrospective cohort study of adults undergoing thoracic surgery (1 January 2022-31 December 2024). CRP was measured preoperatively and on postoperative days (POD) 1-5; trajectories were compared by surgical approach and extent of resection using repeated-measures ANOVA with Greenhouse-Geisser correction (α = 0.05). Results: Among 144 patients (VATS n = 79; open thoracotomy n = 65; extent: segmentectomy n = 25, lobectomy n = 96, bilobectomy n = 9, pneumonectomy n = 14), overall CRP rose from 26.6 ± 45.0 mg/L preoperatively to a POD2 peak of 200.9 ± 72.7 mg/L, then declined to 118.1 ± 70.7 mg/L by POD5. Thoracotomy showed higher peaks than VATS (POD2 216.1 ± 76.0 vs. 152.3 ± 29.9 mg/L; POD3 206.7 ± 88.7 vs. 159.8 ± 72.4 mg/L), but time × approach was not statistically significant (F = 1.042, p = 0.381; partial η(2) = 0.115). The extent analysis showed the highest peaks with pneumonectomy (POD2 273.7 ± 46.3 mg/L) compared with bilobectomy (155.7 ± 11.0 mg/L) and lobectomy (VATS 132.1 ± 3.7, open 196.8 ± 85.3 mg/L); time × extent was not significant (F = 1.136, p = 0.384; partial η(2) = 0.299). The overall effect of time did not reach significance (F = 1.127, p = 0.352; partial η(2) = 0.124), reflecting variability. Patients with clinically diagnosed infections exhibited more prolonged CRP elevation, often >100 mg/L beyond POD4, whereas uncomplicated cases declined after the POD2 peak; these trends did not achieve statistical significance in this cohort. Conclusions: Early postoperative CRP in thoracic surgery typically peaks at 48-72 h and then falls. Higher peaks with open surgery and more extensive resection were observed but not statistically confirmed; persistence > 100 mg/L after POD3-4 may flag complications. Prospective studies are needed to validate thresholds and refine CRP-based surveillance pathways.