Machine Learning Approaches for Predicting Acute Respiratory Failure, Ventilator Dependence, and Mortality in Chronic Obstructive Pulmonary Disease

Chronic obstructive pulmonary disease (COPD) is one of the leading causes of mortality and contributes to high morbidity worldwide. Patients with COPD have a higher risk for acute respiratory failure, ventilator dependence, and mortality after hospitalization compared with the general population. Accurate and early risk detection will provide more information for early management and better decision making. This study aimed to build prediction models using patients’ characteristics, laboratory data, and comorbidities for early detection of acute respiratory failure, ventilator dependence, and mortality in patients with COPD after hospitalization. We retrospectively collected the electronic medical records of 5061 patients with COPD in three hospitals of the Chi Mei Medical Group, Taiwan. After data cleaning, we built three prediction models for acute respiratory failure, ventilator dependence, and mortality using seven machine learning algorithms. Based on the AUC value, the best model for mortality was built by the XGBoost algorithm (AUC = 0.817), the best model for acute respiratory failure was built by random forest algorithm (AUC = 0.804), while the best model for ventilator dependence was built by LightGBM algorithm (AUC = 0.809). A web service application was implemented with the best models and integrated into the existing hospital information system for physician’s trials and evaluations. Our machine learning models exhibit excellent predictive quality and can therefore provide physicians with a useful decision-making reference for the adverse prognosis of COPD patients.

The outcomes of patients with diabetes mellitus in The Philippine CORONA Study

Patients diagnosed with diabetes mellitus (DM) who are infected with severe acute respiratory syndrome coronavirus 2 (SARS-COV-2) belong to the most vulnerable patient subgroups. Emerging data has shown increased risks of severe infections, increased in ICU admissions, longer durations of admission, and increased mortality among coronavirus disease 2019 (COVID-19) patients with diabetes. We performed a subgroup analysis comparing the outcomes of patients diagnosed with DM (n = 2191) versus patients without DM (n = 8690) on our data from our study based on a nationwide, comparative, retrospective, cohort study among adult, hospitalized COVID-19 patients involving 37 hospital sites from around the Philippines. We determined distribution differences between two independent samples using Mann–Whitney U and t tests. Data on the time to onset of mortality, respiratory failure, intensive care unit (ICU) admission were used to build Kaplan–Meier curves and to compute for hazard ratios (HR). The odds ratios (OR) for longer ventilator dependence, longer ICU stay, and longer hospital stays were computed via multivariate logistic regression. Adjusted hazard ratios (aHR) and ORs (aOR) with 95% CI were calculated. We included a total of 10,881 patients with confirmed COVID-19 infection (2191 have DM while 8690 did not have DM). The median age of the DM cohort was 61, with a female to male ratio of 1:1.25 and more than 50% of the DM population were above 60 years old. The aOR for mortality was significantly higher among those in the DM group by 1.46 (95% CI 1.28–1.68; p < 0.001) as compared to the non-DM group. Similarly, the aOR for respiratory failure was also significantly higher among those in the DM group by 1.67 (95% CI 1.46–1.90). The aOR for developing severe COVID-19 at nadir was significantly higher among those in the DM group by 1.85 (95% CI 1.65–2.07; p < 0.001). The aOR for ICU admission was significantly higher among those in the DM group by 1.80 (95% CI 1.59–2.05) than those in the non-DM group. DM patients had significantly longer duration of ventilator dependence (aOR 1.33, 95% CI 1.08–1.64; p = 0.008) and longer hospital admission (aOR 1.13, 95% CI 1.01–1.26; p = 0.027). The presence of DM among COVID-19 patients significantly increased the risk of mortality, respiratory failure, duration of ventilator dependence, severe/critical COVID-19, ICU admission, and length of hospital stay.

Factors that Predict Survival From Mesenteric Ischemia in the Geriatric Patient: Not a Death Sentence

Background The mortality rate from mesenteric ischemia is reported to be as high as 80%. The goal of our study was to identify demographic and clinical predictors of post-operative mortality and discharge disposition among elderly patients with mesenteric ischemia. Methods All patients 65 years and older who underwent emergency surgery (ES) for the management of mesenteric ischemia in the American College of Surgeons–National Surgical Quality Improvement Program database from 2007 to 2017 were included. Univariate analyses and logistic regressions were used to identify independent predictors of mortality and discharge disposition. Results A total of 2438 patients met inclusion criteria, with a median age of 77 years and 60.8% being female. The 30-day mortality of the overall cohort was 31.5% and the 30-day morbidity was 65.3%. The following were the major predictors of 30-day mortality: pre-operative diagnosis of septic shock [OR: 2.46, (95% CI: 1.94-3.13)], dialysis dependence [OR: 2.05, (95% CI: 1.45-2.90)], recent weight loss [OR: 1.80, (95% CI: 1.16-2.79)], age ≥80 years [OR: 1.67, (95% CI: 1.25-2.23)], and ventilator dependence [OR: 1.65, (95% CI: 1.23-2.23)]. In the absence of these predictors, survival rate was 84%. The major predictors of discharge to post-acute care (PAC) included age ≥80 years [OR: 3.70, (95% CI: 2.50-5.47)] and pre-operative septic shock [OR: 2.20, (95% CI: 1.42-3.41)]. Conclusion In the geriatric patient, a diagnosis of mesenteric ischemia does not equate to an automatic death sentence. The presence of certain pre-operative risk factors confers a high risk of mortality, whereas their absence is associated with a high chance of survival.

Update: Gender differences in CABG outcomes—Have we bridged the gap?

Background Appreciation of unique presentation, patterns and underlying pathophysiology of coronary artery disease in women has driven gender based risk stratification and risk reduction efforts over the last decade. Data regarding whether these advances have resulted in unequivocal improvements in outcomes of CABG in women is conflicting. The objective of our study was to assess gender differences in post-operative outcomes following CABG. Methods Retrospective analyses of institutional data housed in the Society of Thoracic Surgeons (STS) database for patients undergoing CABG between 2002 and 2020 were conducted. Multivariable regression analysis was conducted to investigate gender differences in post-operative outcomes. P-values were adjusted using Bonferroni correction to reduce type-I errors. Results Our final cohort of 6,250 patients had fewer women than men (1,339 vs. 4,911). more women were diabetic (52.0% vs. 41.2%, p<0.001) and hypertensive (89.1% vs. 84.0%, p<0.001). Women had higher adjusted odds of developing ventilator dependence >48 hours (OR: 1.65 [1.21, 2.45], p = 0.002) and cardiac readmissions (OR: 1.56 [1.27, 2.30], p = 0.003). After adjustment for comorbidity burden, mortality rates in women were comparable to those of age-matched men. Conclusion The findings of our study indicate that despite apparent reduction of differences in mortality, the burden of postoperative morbidity is still high among women.

Delayed positive COVID19 nasopharyngeal test, a case study with clinical and pathological correlation

Background There are various reasons for delayed positive nasopharyngeal PCR tests for coronavirus disease 2019 (COVID19) in not only asymptomatic but also severely diseased patients. The pathophysiological attributes are not known. We explore this possibility through a case report. Case presentation A 64-year-old male with history of pulmonary fungal infection, asthma and chronic pulmonary obstructive disease (COPD), diabetes, coronary artery disease presented with shortness of breath, fever and chest image of ground opacity, reticular interstitial thickening, highly suspicious for COVID19. However, nasopharyngeal swab tests were discordantly negative for four times in two weeks, and IgG antibody for COVID19 was also negative. However, serum IgE level was elevated. No other pathogens are identified. His symptoms deteriorated despite corticosteroid, antibiotics and bronchodilator treatment. Bronchoalveolar lavage (BAL) and open lung wedge biopsy were performed for etiology diagnosis. They demonstrated COVID19 viral RNA positive fibrosing organizing pneumonia with respiratory tract damage characterized by suspicious viral cytopathic effect, mixed neutrophilic, lymphoplasmacytic, histiocytic and eosinophilic inflammation and fibrosis besides expected asthma and COPD change. One week later, repeated COVID19 nasopharyngeal tests on day 40 and day 49 became positive. Conclusion Our case and literature review indicate that allergic asthma and associated high IgE level together with corticosteroid inhalation might contribute to the delayed positive nasopharyngeal swab in upper airway; COPD related chronic airways obstruction and the addition of fibrosis induced ventilator dependence and poor prognosis in COVID19 pneumonia, and should be therapeutically targeted besides antiviral therapy.

COVID-19 outcomes of 10,881 patients: retrospective study of neurological symptoms and associated manifestations (Philippine CORONA Study)

Our study aimed to determine the effects of new-onset neurological symptoms (NNS) on clinically relevant outcomes in hospitalized patients with COVID-19 infection. We conducted a nationwide, comparative, retrospective, cohort study among adult, hospitalized COVID-19 patients involving 37 hospital sites from various regions in the Philippines. We included a total of 10,881 patients with confirmed COVID-19 infection (2008 had NNS while 8873 did not have NNS). The adjusted hazard ratios (aHRs) for mortality among the mild and severe cases were significantly higher by 1.660 (95% CI 1.132–2.435) and by 1.352 (95% CI 1.042–1.752), respectively, in the NNS group compared to those in the non-NNS group. The aHRs for respiratory failure in the NNS group were significantly increased by 1.914 (95% CI 1.346–2.722), by 1.614 (95% CI 1.260–2.068), and by 1.234 (95% CI 1.089–1.398) among the mild, severe, and critical cases, respectively. The aHRs for ICU admission in the NNS group were still significantly higher by 1.973 (95% CI 1.457–2.673) and by 1.831 (95% CI 1.506–2.226) among the mild and severe cases, respectively. Patients who had NNS were not significantly associated with a longer duration of ventilator dependence (adjusted odds ratio (aOR) 0.954, 95% CI 0.772–1.179), longer ICU stay (aOR 0.983, 95% CI 0.772–1.252) and longer hospital admission (aOR 1.045, 95% CI 0.947–1.153). The presence of NNS significantly increases the risk of mortality, respiratory failure and ICU admission among COVID-19 patients. Registration and associated protocol publication: ClinicalTrials.gov website (NCT04386083); Espiritu AI, Sy MCC, Anlacan VMM, Jamora RDG. The Philippine COVID-19 Outcomes: a Retrospective study Of Neurological manifestations and Associated symptoms (The Philippine CORONA study): a protocol study. BMJ Open. 2020;10:e040944.

The Combination of SOFA Score and Urinary NGAL May Be an Effective Predictor for Ventilator Dependence among Critically Ill Surgical Patients: A Pilot Study

Background: Ventilator dependence (VD) has been considered as a serious complication in critically ill patients in the intensive care unit (ICU). Acute kidney injury (AKI) is associated with VD as a result of lung–kidney interaction. The aim of our study was to investigate novel biomarkers in predicting ventilator dependence in critically ill surgical patients. Methods: Patients who were admitted to surgical ICU were enrolled and their serum and urine samples were collected. Novel biomarkers including gelatinase-associated lipocalin (NGAL), calprotectin, kidney injury molecule-1 (KIM-1), cystatin C, and growth differentiation factor 15 (GDF-15) were analyzed and correlated with clinical outcome. Results: A total of 33 patients were enrolled and analyzed. The majority of them received abdominal surgery prior to ICU admission. Thirteen patients were classified into the VD group, while the remaining 20 were in a non-ventilator dependence group (nVD). Statistical analysis demonstrated that the following were significantly higher in the VD group than in the nVD group: serum NGAL (420.25 ± 45.18 ng/mL vs. 314.68 ± 38.12 ng/mL, p-value 0.036), urinary NGAL (420.87 ± 41.08 ng/mL vs. 250.84 ± 39.45 ng/mL, p-value 0.002), SOFA score (11.3 ± 1.5 vs. 5.6 ± 0.7, p-value 0.001), and APACHE II score (23.2 ± 2.6 vs. 13.6 ± 0.8, p-value 0.001). The area under the ROC curve (AUROC) of urinary NGAL for VD was 0.808. The combination of urinary NGAL and SOFA score could further increase AUROC for VD to 0.835. Conclusions: The current study demonstrated the predictive capability of urinary NGAL for ventilator dependence among critically ill surgical patients. When combined with SOFA score, the predictive ability was further augmented. Further large-scale studies are warranted to validate our findings.