Persistent cough is a common postoperative complication following lung resection [1, 3]. With the development of ERAS, more attention is being paid to the quality of life of postoperative patients. Some studies have found that CAP can affect patients' postoperative quality of life, thus hindering their postoperative recovery [23]. Therefore, more and more researchers are actively studying CAP to accelerate the recovery of patients. Several risk factors have been shown to be associated with the occurrence of CAP [24]. To our knowledge, this is the first meta-analysis to investigate the risk factors for CAP. By including independent risk factors for CAP, our meta-analysis confirmed that young age, female sex, preoperative cough, right lobe surgery, lobectomy, subcarinal lymph node dissection, mediastinal lymph node removal, closure of bronchial stump with stapler, peritracheal lymph node resection, and postoperative acid reflux were independent risk factors for CAP. In the future, these identified risk factors can be used to construct predictive models to identify high-risk patients. For those high-risk patients, preoperative communication about CAP should be enhanced and intraoperative measures should be taken to prevent CAP, as well as more aggressive postoperative follow-up.

Our study showed that female sex was an independent risk factor for CAP. Previous studies have shown a preponderance of females with chronic cough, which may be related to the influence of female hormones [25, 26]. Back in 1989, researchers found that women who received angiotensin-converting enzyme inhibitors (ACEI) were more likely to cough than men [27]. Women have also been observed to be more sensitive than men to cough reflexes triggered by the inhalation of citric acid, tartaric acid and capsaicin [28,29,30,31]. C- fibers of the vagus nerve are the most important cough receptors and are mainly distributed in the larynx, trachea, carina and larger bronchi in the lung, which are sensitive to various chemical stimuli [32]. The ability of C fibers to sense chemical substances mainly depend on the expression of transient receptor potential (TRP) V1/A1 channels and other ion channels. Zhu et al. [33] discovered that the level of TRPV1 in patients with acute or chronic cough after lung cancer surgery was higher than that in patients without cough. Several studies have shown that estrogen can affect C fiber activation by affecting TRPV1 activation/sensitization [34]. Therefore, some researchers speculated that women's susceptibility to cough may be related to the influence of estrogen on TRPV1. In the future, drugs blocking the TRPV1 signaling pathway may be created to treat CAP. Interestingly, some studies have revealed that women were also at increased risk for chronic pain compared to men. Thus, some pain physiologists believed that women's greater susceptibility to chronic cough was part of an enhanced or overdeveloped visceral sensitivity that was the result of an evolutionary selection process [35]. Nonetheless, more relevant researches are needed to explore the mechanisms involved. A study indicated that the health-related quality of life (HRQOL) of women was more adversely affected than that of men, the longer a cough lasted [36]. Therefore, for female patients, more attention should be paid to cough and cough-related quality of life after lung cancer surgery. For female CAP patients, more aggressive treatments may be required.

Our study found that age was a risk factor for CAP, and younger people were more likely to develop CAP. This may be related to the relatively sluggish receptors of the cough reflex in the elderly [19]. However, age was not a risk factor for CAP when the surgical method was limited to thoracoscopic surgery, which is consistent with the conclusion of previous studies [3]. Due to the different age thresholds and few included studies, we were unable to determine which age group of patients were more likely to suffer CAP, future studies could further explore this. Our study found that longer anesthesia duration was not a risk factor for CAP, which was not consistent with previous studies [2, 8]. They thought the relationship between longer anesthesia duration and CAP may be due to the fact that the longer the time of tracheal intubation, the stronger the stimulation of the airway, resulting in a stronger inflammatory response of the tracheal tissue. However, in our study, cutoff values for anesthesia time varied across different studies, which may be a source of heterogeneity in our results. More research is warranted in the future to explore the association between duration of anesthesia and CAP. Previous studies have shown that patients with a history of smoking had less CAP after surgery, which was consistent with our research results. The reason may be that long-term smoking can reduce the sensitivity of airway cough receptors and weaken the sensitivity of cough reflex to the stimulation caused by surgery [37, 38]. Only two studies [21, 22] reported the relationship between preoperative cough and CAP, and the overall result indicated that preoperative cough was a risk factor for CAP. Airway hyperreactivity may be present in patients with cough before surgery. Therefore, appropriate drugs can be used preoperative to improve airway hyperreactivity in these patients.

Postoperative acid reflux was also an independent risk factor for CAP in current study, which was consistent with previous studies. This may be due to gastroesophageal reflux activate the vagus nerve from the esophagus to the lungs, as the vagus nerve innervates not only the bronchus but also the esophagus. However, the included studies had different definitions of postoperative acid reflux. Pan [2] et al. used the Reflux Diagnostic Questionnaire (RDQ) to assess the frequency and severity of reflux symptoms in postoperative patients. Postoperative acid reflux was diagnosed when the sum of the two scores was greater than 12. Sawabata [11] et al. diagnosed gastroesophageal reflux by asking patients if they had symptoms such as heartburn, nausea, chest pain and the characteristics of these symptoms. Thus, to further investigate the relationship between CAP and acid reflux, a 24-h esophageal pH monitor may be needed. Sawabata [11] et al. treated 20 CAP patients with acid reflux with proton pomp inhibiter and pro-kinetic agent and found significant improvement in cough in most patients. We therefore recommended that the RDQ could be used preoperatively to evaluate reflux. Proton pump inhibitors and prokinetic agents might be used to treat CAP patients with gastroesophageal reflux. The need for prophylactic use of proton pump inhibitors in patients with preoperative symptoms of reflux still requires further study.

The results of this meta-analysis showed that right lobe surgery, lobectomy, subcarinal lymph node dissection, mediastinal lymph node removal and peritracheal lymph node resection were risk factors for postoperative cough after pulmonary resection. Compared with segmentectomy or wedge resection, lobectomy contributes to a larger residual cavity in the thoracic cavity after surgery, which can lead to changes in the anatomical structure in the thoracic cavity, bronchial distortion, residual lung deformity, etc., thus increasing airway sensitivity and causing chronic cough [18]. Additionally, it was also associated with transecting a major bronchus while peforming lobectomy. According to this theory, patients undergoing pneumonectomy, which not only involves major bronchial transection but extensive dissection in the region of the carina, are more likely to develop CAP compared with patients undergoing lobectomy or sublobectomy, but this warrants further study due to the lack of relevant studies. The removal of lymph nodes may damage the vagus nerve, thus increasing the sensitivity of cough receptors and causing CAP. Of the studies we included, some evaluated the association between subcarinal lymph node dissection and CAP, some explored the association between paratracheal lymph node dissection and CAP, others investigated the relationship between mediastinal lymph node dissection and CAP. Therefore, we combined these three groups of data separately when performing the meta-analysis. Since mediastinal lymph nodes included several groups, further studies are needed to determine which group of lymph nodes dissection is more likely to cause CAP. Clarifying which group lymph node dissection is a risk factor for CAP may help us to understand the underlying mechanism. Our study suggested that compared with full-thickness interrupted suture, closure of bronchial stumps with stapler was a risk factor for CAP. Dong et al. speculated that closure of bronchial stump with stapler was not conducive to the discharge of airway secretions, and was prone to airway torsion or chronic inflammatory reaction of the airway stump, which may be the reason why closure of bronchial stump with stapler was an independent risk factor for CAP [13]. However, due to the inclusion of only two studies and the small percentage of bronchial stumps were closed with full-thickness interrupted suture, the result may be unstable and need to be confirmed by more high-quality studies.

Some researchers have conducted studies to explore how to prevent CAP from occurring. Dong et al. [13] found that preoperative lung training was a protective factor for CAP. Filling post-lymphadenectomy residual cavities with fatty tissue autografts has been shown to reduce the incidence of CAP while improving the quality of life of patients [39]. Gu et al. [8] discovered that intraoperative protection of pulmonary vagus nerve branches by sampling around the lymph nodes on the side of surgery reduced the incidence of CAP. Xie et al. [12] treated 41 CAP patients with acupuncture on the 8th week after surgery and found that the Leicester Cough Questionnaire in Mandarin Chinese (LCQ-MC) score was higher in the acupuncture treatment group compared to the no treatment group. Several studies have confirmed the efficacy of inhaled corticosteroid plus β2 agonist and the compound methoxyphenamine capsule for the treatment of CAP [6, 22]. Although some progress has been made, it is still worthwhile to further investigate who with CAP needs to be treated, and when and how to go about it. Our study identified several independent risk factors for CAP, which will provide some theoretical basis for the identification of patients at high risk for CAP, and the prevention and treatment of CAP in the future.

Wu et al. [19] observed no difference in the incidence of CAP between robot-assisted thoracic surgery (RATS) and video-assisted thoracic surgery (VATS). However, in the absence of relevant studies, we were unable to explore the differences in the incidence of CAP among thoracotomy, VATS, and RATS. Similarly, it is worthwhile to investigate whether there is any difference in the incidence of CAP between single-port and multi-ports VATS. Previously, we prospectively followed 88 post-thoracoscopic lung cancer patients and found that the incidence of CAP remained at 66% after 90 days after surgery, while the severity of the cough was gradually reduced [1]. Similar to our results, Lin et al. [23] found that the postoperative cough-related quality of life in lung cancer patients who underwent VATS returned to preoperative levels at approximately 3 months postoperative. However, there were still patients with cough symptoms at 6 months after surgery, who may require more attention. In this study, only one study included patients with cough that lasted longer than 90 days, they found that younger age (< 57 years), preoperative cough, lobectomy, and longer duration of endotracheal intubation (≥ 172 min) were risk factors for CAP [19]. In the future, more studies are needed to explore the characteristics of patients with prolonged postoperative cough.

There were several limitations in our study. First, all of the included studies were conducted in Asia, which may limit the applicability of the conclusions to other areas. Unfortunately, we were unable to find papers and data on CAP in regions other than Asia, and therefore, we were unable to provide specific information on CAP in other regions. Second, some proven risk factors, such as right upper lobectomy and difficult airway, could not be subjected to meta-analysis as they were reported in only one study. Hence, it is hoped that future research will give greater emphasis to the study of these risk factors. Third, some risk factors, such as age and anesthesia time, may lead to unstable conclusions due to different cutoff values adopted in different studies. We suggest that future studies could standardize these metrics. Fourth, CAP was inconsistently defined in different studies. Some studies consider CAP as postoperative cough lasting for at least two weeks, while others define it as lasting for a minimum of eight weeks. Moreover, certain studies did not provide a precise definition for CAP; rather, it was assessed by clinical physicians to determine whether patients experienced CAP. Therefore, there is a need to harmonize the definition of CAP in the future. Fifth, owing to the lack of relevant studies, we were unable to explain some of the underlying mechanisms. More in-depth studies are needed for the underlying mechanism of CAP. Given the number and quality of the included studies, more high-quality studies should be conducted to explore the risk factors for CAP, to better improve the quality of life of patients after lung resection.

In conclusion, young age, female sex, preoperative cough, right lobe surgery, lobectomy, subcarinal lymph node dissection, mediastinal lymph node removal, closure of bronchial stump with stapler, peritracheal lymph node resection, and postoperative acid reflux were independent risk factors for CAP. Patients with these risk factors may need more active intervention and postoperative follow-up to help them recover quickly and return to normal life.