Utilizing machine learning to improve clinical trial design for acute respiratory distress syndrome.

Heterogeneous patient populations, complex pharmacology and low recruitment rates in the Intensive Care Unit (ICU) have led to the failure of many clinical trials. Recently, machine learning (ML) emerged as a new technology to process and identify big data relationships, enabling a new era in clinical trial design. In this study, we designed a ML model for predictively stratifying acute respiratory distress syndrome (ARDS) patients, ultimately reducing the required number of patients by increasing statistical power through cohort homogeneity. From the Philips eICU Research Institute (eRI) database, no less than 51,555 ARDS patients were extracted. We defined three subpopulations by outcome: (1) rapid death, (2) spontaneous recovery, and (3) long-stay patients. A retrospective univariate analysis identified highly predictive variables for each outcome. All 220 variables were used to determine the most accurate and generalizable model to predict long-stay patients. Multiclass gradient boosting was identified as the best-performing ML model. Whereas alterations in pH, bicarbonate or lactate proved to be strong predictors for rapid death in the univariate analysis, only the multivariate ML model was able to reliably differentiate the disease course of the long-stay outcome population (AUC of 0.77). We demonstrate the feasibility of prospective patient stratification using ML algorithms in the by far largest ARDS cohort reported to date. Our algorithm can identify patients with sufficiently long ARDS episodes to allow time for patients to respond to therapy, increasing statistical power. Further, early enrollment alerts may increase recruitment rate.

Tracheal bronchus: a cause of prolonged atelectasis in intubated children.

Tracheal bronchus is a common anomaly that occurs in approximately 2% of people. Two children with multiple medical problems which led to endotracheal intubation are described. The hospital course for each child was complicated by persistent right upper lobe atelectasis. The presence of a tracheal bronchus was not recognized in either case initially; identification of this anatomic variant allowed appropriate changes in airway management. The potential for tracheal bronchus to cause, or be associated with, localized pulmonary problems is reviewed. The diagnosis of tracheal bronchus should be considered early in the course of intubated patients with right upper lobe complications.

Arrhythmias in children prenatally exposed to cocaine.

OBJECTIVES: To describe the time of detection, electrophysiologic mechanism, and severity of hemodynamic sequelae of arrhythmias in infants and children prenatally exposed to cocaine and to determine whether the incidence of severe neonatal arrhythmia is related to prenatal exposure to cocaine. DESIGN: Characteristics of arrhythmias are described for all significant arrhythmias detected at Boston (Mass) City Hospital in infants and children with known cocaine exposure, as well as a convenience sample of children from Children's Hospital, Boston. A historical cohort was used to calculate the rates of cardiac consultation for arrhythmia among children prenatally exposed to cocaine and among children with no known cocaine exposure. STUDY POPULATION: Characterization of the arrhythmias is based on case studies of 18 children. The rate of arrhythmia consultation was calculated from 554 infants who had urine toxic screens for cocaine and from 13 arrhythmias detected between 1988 and 1991. OUTCOME MEASURES: Prenatal cocaine exposure; the time of detection, electrophysiologic mechanism, and severity of hemodynamic sequelae of arrhythmias; and the incidence of cardiology consultation for arrhythmia in infants and children prenatally exposed to cocaine. RESULTS: Fetal arrhythmia persisted into the neonatal period in three cocaine-exposed infants; two were delivered via emergency cesarean sections for presumed fetal bradycardia and were subsequently found to have asymptomatic frequent and blocked atrial premature beats. Including these three infants, arrhythmia was observed in 13 cocaine-exposed neonates; 12 had a variety of supraventricular arrhythmias and four had low-grade ventricular ectopy. Arrhythmia resulted in congestive heart failure in five (38%) of 13 neonates. Six occurrences of arrhythmia were observed beyond the neonatal period among five cocaine-exposed infants. Late arrhythmias included high-grade ventricular arrhythmias and resulted in two cardiorespiratory arrests. In addition, neonates with known exposure to cocaine were more likely to have a consultation for arrhythmia than neonates without known exposure. CONCLUSIONS: Sustained arrhythmias may result from an increased number of potential initiating premature beats in children prenatally exposed to cocaine. These effects persist beyond the period of exposure and are associated in some children with congestive heart failure, cardiorespiratory arrest, and death. Prenatal cocaine exposure increases the incidence of consultation for atrial and ventricular arrhythmias.

Cardiovascular abnormalities in infants prenatally exposed to cocaine.

This study utilized a historical cohort to examine the relationship between maternal cocaine use during pregnancy and the occurrence of congenital cardiovascular abnormalities. All neonatal drug screens performed at Boston City Hospital during an 18-month period were reviewed (n = 554); for 214 (39%) screened high-risk neonates, results of toxicologic screens were positive for cocaine, and 340 (61%) neonates had no detectable cocaine. We compared the occurrence of cardiovascular malformations and electrocardiographic abnormalities in these two groups. Matches were sought between these 554 infants and our pediatric cardiology data base, which consisted of inpatient consultation, outpatient consultation, and electrocardiography. Forty-nine patients had drug screens and were also entered into our cardiology data base: 25 had both consultations and electrocardiograms, and 24 had electrocardiograms only. The rate of cardiac anomalies among the cocaine-positive infants was significantly higher (relative risk = 3.7; 95% confidence interval: (1.4, 9.4)) than the rate of these anomalies among the cocaine-negative comparison group (65/100 vs 18/1000); the rate for cocaine-positive infants was also significantly higher than published rates for general populations of infants. Several electrocardiographic abnormalities, high-grade ventricular ectopy, and cardiorespiratory arrests were also noted in our study population. We conclude that cocaine exposure during prenatal life appears to predispose infants to structural cardiovascular malformations, electrocardiographic abnormalities, and, possibly, cardiopulmonary autonomic dysfunction.