Predictors of dyspnea in patients with advanced cancer

Background: More than 70% of patients with advanced cancer experience dyspnea. Dyspnea is predictive of shorter survival and interferes with quality of life (QOL). The present study aimed to identify predictors of the presence and severity of dyspnea in advanced cancer patients. Methods: A prospective database collected from patients attending a palliative radiotherapy clinic was analyzed for patient demographics, Edmonton Symptom Assessment System (ESAS) scores, Patient-Reported Functional Status (PRFS), history of smoking and respiratory conditions, pulse oximetry readings, and primary cancer site. Using the ESAS shortness of breath item, dyspnea was classified as mild [1-3], moderate [4-6] or severe [7-10]. Logistic regression analysis and generalized estimating equations (GEEs) were used to identify predictors of the severity of dyspnea and presence of moderate/severe dyspnea (ESAS >= 4) at patients' first visit and over time, respectively. Results: A total of 252 patients with dyspnea data were included (median age 71.3 years, 61.5% male, 44.4% had dyspnea) in a demographic analysis. Multivariable analysis showed liver metastases (P=0.01, OR =2.04), a history of respiratory conditions (P=0.03, OR =2.09) and PRFS >= 3 (P=0.03, OR =1.75) were predictive of the severity of dyspnea at the first visit. Analyzed over time, liver metastases (P=0.02, OR =1.80), lymph node metastases (P=0.02, OR =1.79), a history of respiratory conditions (P=0.006, OR =2.50) and pulse oximetry <90 (P=0.003, OR =3.32) were predictive of greater severity of dyspnea symptoms. Patients with multiple radiation treatments in the thorax region were less likely to have severe dyspnea symptoms over time (P=0.01, OR =0.32). Lung metastases (P=0.04, OR =2.03), a history of respiratory conditions (P=0.01, OR =2.60) and PRFS >= 3 (P=0.009, OR =2.30) were predictive of moderate/severe dyspnea at the first visit. Over time, lymph node metastases (P=0.003, OR =2.51), a history of respiratory conditions (P=0.04, OR =2.37) and pulse oximetry <90 (P=0.0004, OR =5.15) were predictive of moderate/severe dyspnea. Conclusions: Liver, lung and lymph node metastases, a history of respiratory conditions, pulse oximetry <90 and PRFS >= 3 were predictive of the severity of dyspnea and moderate/severe dyspnea. Physicians should be aware of predictive factors that could lead to dyspnea to promote early intervention for improved patient care and the creation of screening tools for clinical practice.