Study population
We retrospectively analyzed the medical records of 342 patients admitted to Korea University Guro Hospital from January 2011 to May 2017. We searched our electronic medical records database with the keywords “COPD” and “Acute exacerbation.” This study was approved by the Institutional Review Board of Korea University Guro Hospital. (KUGH16131-002). The requirement for informed consent from the patients was waived due to the retrospective nature of this study by the institutional review committee.
All patients included only patients who were followed up for more than one year in our hospital under the diagnosis of COPD. COPD and airflow limitation were diagnosed by synthesizing patient-reported respiratory symptoms, PFT (the ratio of FEV1 to forced vital capacity (FVC) was less than 70% in post-bronchodilator spirometry), chest image, and patient`s history (smokers with at least ten pack-years of tobacco exposure, etc.) by an experienced pulmonologist [6]. AECOPD was defined as worsening of the patient’s respiratory symptoms beyond normal day-to-day variation. Severe AECOPD was defined as ‘if the patient needs hospitalization due to AECOPD.’ The spirometry data used in the analysis was previously performed in the outpatient clinic during the stable period. Spirometry value that was measured within one year from the hospitalization day were used. Patients were excluded with the following criteria: 1) the cause of admission was not AECOPD; for example, acute heart failure, acute pulmonary edema, acute pulmonary embolism, pneumothorax, and arrhythmia (These diseases were excluded through cardiac enzyme, electrocardiogram, echocardiogram and chest image.), 2) the patient had cancer, 3) the patient received a major operation within 3 months, 4) the patient had an acute coronary syndrome, brain hemorrhage, or brain infarction within 3 months, 5) the patient had previously been diagnosed with asthma, and 6) the patient had no DLCO results. All patients were 40 years old or older. We retrospectively analyzed the charts by two experienced pulmonologists to exclude various exclusion factors.
We classified 342 severe AECOPD events using DLCO and FEV1. When the DLCO value is more than 50 (% of predicted value), it is defined as the 'DLCO normal group' and when it is 50 (% of predicted value) or less, it is defined as the 'DLCO impaired group' [11]. Likewise, when the FEV1 value is more than 50 (% of predicted value), it is defined as the ‘FEV1 normal group' and when it is 50 (% of predicted value) or less, it is defined as the ‘FEV1 impaired group' (Figure 1).
Data collection
We defined the prognostic factors of severe AECOPD as the length of hospital stay, mortality in hospital, experience of mechanical ventilation, and experience of intensive care unit (ICU) care in the hospital. When the patient was hospitalized more than once, only the first hospitalized events were included, and the others were excluded. The following medical data were analyzed: age, sex, smoking history, comorbidities, baseline spirometry, inhaler and oral medication before admission, length of hospital stay, hospital mortality, experience of mechanical ventilation, and experience of ICU care in hospital.
Statistical analysis
Data were analyzed using SPSS 20 software (SPSS for Windows, SPSS Inc., Chicago, IL, USA). Data are presented as average ± standard deviation or number (percentage). We performed a statistical analysis in two directions. First, two groups were classified using DLCO and FEV1 and analyzed statistically. Continuous variables were compared using the independent t-test, and categorical variables were compared using the chi-squared test. We analyzed the prognostic factors (except length of hospital stay) by multivariate analysis through logistic regression. Multivariate analysis was conducted for variables with a P-value of less than 0.05 in the univariate analysis, except for baseline spirometry (DLCO and FEV1). Multivariate analysis was conducted using a backward elimination procedure and was assessed by the Hosmer-Lemeshow test.
Second, the linear correlation between spirometry factors (DLCO and FEV1) and length of hospital stay were analyzed. In univariate analysis, the correlation coefficients between spirometry factors and length of hospital stay were analyzed using the Pearson correlation analysis. In addition, we performed a multivariate linear regression analysis that included variables with a P-value of less than 0.05 in the univariate analysis, except baseline spirometry. In addition, multivariate linear regression analysis was conducted using a backward elimination procedure. In the multivariate analysis, B was the regression coefficient, and a negative sign of the regression coefficient meant that the variables were negatively associated.
Third, we used receiver operating characteristic (ROC) curve analysis to predict the sensitivity and specificity of DLCO and FEV1 as prognostic markers in severe AECOPD. A P-value of less than 0.05 was considered statistically significant.