The use of spirometry testing prior to cardiac surgery may impact the Society of Thoracic Surgeons risk prediction score: A prospective study in a cohort of patients at high risk for chronic lung disease

ObjectivesChronic lung disease is a significant comorbidity in patients undergoing cardiac surgery. Chronic lung disease is currently being classified and reported to the Society of Thoracic Surgeons database by using either clinical interview or spirometric testing. We sought to compare the chronic lung disease classification captured by the 2 methods.MethodsWe performed a prospectively designed study in which patients presenting for cardiac surgery, excluding emergent patients, were screened for a history of asthma, a history of 10 or more pack-years of smoking, a persistent cough, and the use of oxygen. Each selected patient underwent spirometry. The presence and severity of chronic lung disease was coded per Society of Thoracic Surgeons guidelines by using the 2 methods of clinical report and spirometric results. The chronic lung disease classifications were compared, and differences were determined by using concordance and discordance rates. The results were then used to construct Society of Thoracic Surgeons–predicted risk models.ResultsThe discordant rate was 39.1%, with underestimation of the severity of chronic lung disease in 94% of misclassified patients. This affected the Society of Thoracic Surgeons–predicted risk models for prolonged ventilation, morbidity/mortality, and mortality by increasing the predicted risk when spirometry was used for morbidity/mortality by an average of 1.5 ± 1.2 percentage points (P < .001) and prolonged ventilation time by an average of 1.3 ± 1.4 percentage points (P < .001).ConclusionThe use of patient history for symptoms, medication, and/or oxygen use as the only method to determine chronic lung disease for this subgroup of patients led to underreporting of chronic lung disease and underestimation of the risk for adverse outcomes. Therefore data submission to the Society of Thoracic Surgeons database should be designed to capture and correct for potential bias in the definition of chronic lung disease because the rate of spirometry in different centers in defining chronic lung disease is not regulated