In the present study, a nomogram was developed and internally validated to predict the risk of PPCs in children undergoing thoracic surgery requiring OLV. The nomogram constructed in this study exhibited good performance in both predictive model development and internal validation. The incidence of pediatric PPCs could most reliably be predicted by the model including preoperative NLR, maximum Ppeak during OLV, minimum OI during OLV, and intraoperative ventilation mode. These indicators are easily obtainable in clinical practice, rendering the predictive model easy to extend and apply. Therefore, this nomogram is recommended for clinical practice in the pediatric population.
The observed incidence of PPCs in this study was 35.7%. It was similar to those found in some previous articles22,23, but higher than those in other studies12,24 The variation across studies may be attributed to the broad definition of PPCs in the guidelines (e.g., pure pleural effusion not requiring treatment is also considered as PPCs). In addition, among the published studies related to PPCs during OLV, the majority of the study population are adults. However, considering the respiratory physiological characteristics of children1, they are prone to PPCs when undergoing OLV10,12.
OI is utilized to assess pulmonary function and oxygenation status, providing valuable information on the severity of lung disease and predicting outcomes following lung injury. Due to its important guiding significance in OLV, the OI is commonly employed as an indicator of lung injury severity in children25. When children undergo thoracic surgery that requires OLV, the lower OI during OLV reflects a more severe intrapulmonary shunt and poorer lung function, making them more prone to lung injury and PPCs.
In addition to minimum OI during OLV, the ventilation mode applied was a strong predictor associated with the occurrence of PPCs. Consistent with some articles comparing the use of three ventilation modes in adults26–28, we also found in the present study that the incidence of PPCs in children using the pressure control ventilation-volume guaranteed (PCV-VG) mode was significantly lower than that of the other two ventilation modes. Therefore, we suggest using the PCV-VG mode that combines the advantages of the other two modes in children with OLV to reduce the incidence of PPCs.
The NLR reflects the balance of neutrophils and lymphocytes in the body and is more accurate and stable than neutrophils and lymphocytes29. In recent years, it has been considered an important inflammatory response marker for diagnosing and evaluating patient prognosis30–33. In this study, we found that the NLR within 48 hours before surgery is also an important predictor of PPCs in children undergoing thoracic surgery requiring OLV, which is consistent with previous reports34,35.
Ppeak reflects the dynamic compliance of the respiratory system, affected by inspiratory flow and tracheal catheter diameter, which is the main source of resistance during mechanical ventilation36,37. With the initiation of OLV in the lateral decubitus position, the dependent lung solely ventilates and is accompanied by gravitational compression of the mediastinum. Meanwhile, the nondependent lung collapses, leading to atelectasis and decreased compliance. Therefore, elevated Ppeak during OLV is commonly encountered. However, a higher Ppeak will not only affect ventilation/perfusion matching, leading to significant hypoxemia during OLV, but may also cause pulmonary barotrauma, thus increasing the incidence of PPCs1,36,38,39.
The study has some notable strengths. First, it established a prediction model for PPCs in the pediatric population for the first time. Second, the data for model development were derived from one of the largest children’s hospitals in China and the sole tertiary referral pediatric hospital in Hunan Province, which could potentially reduce the impact of geographic and demographic variability. However, several limitations need to be considered. This study was retrospective; therefore, some variables that may have an impact on PPCs were not included due to missing records. Another potential limitation concerning generalizability is that it was a single-center study with a relatively small sample size. Thus, large-sample, multicenter, and prospective studies are warranted in the future to validate and enhance our prediction model.