Low Variability of Blood Pressure Predicts Abnormal Electroencephalogram in Infants with Hypoxic Ischemic Encephalopathy.

OBJECTIVE: This study aimed to evaluate the role of an objective physiologic biomarker, arterial blood pressure variability, for the early identification of adverse short-term electroencephalogram (EEG) outcomes in infants with hypoxic-ischemic encephalopathy (HIE). STUDY DESIGN: In this multicenter observational study, we analyzed blood pressure of infants meeting these criteria: (1) neonatal encephalopathy determined by modified Sarnat exam, (2) continuous mean arterial blood pressure (MABP) data between 18 and 27 hours after birth, and (3) continuous EEG performed for at least 48 hours. Adverse outcome was defined as moderate-severe grade EEG at 48 hours. Standardized signal preprocessing was used; the power spectral density was computed without interpolation. Multivariate binary logistic regression was used to identify which MABP time and frequency domain metrics provided improved predictive power for adverse outcomes compared with standard clinical predictors (5-minute Apgar score and cord pH) using receiver operator characteristic analysis. RESULTS: Ninety-one infants met inclusion criteria. The mean gestational age was 38.4 ± 1.8 weeks, the mean birth weight was 3,260 ± 591 g, 52/91 (57%) of infants were males, the mean cord pH was 6.95 ± 0.21, and 10/91 (11%) of infants died. At 48 hours, 58% of infants had normal or mildly abnormal EEG background and 42% had moderate or severe EEG backgrounds. Clinical predictor variables (10-minute Apgar score, Sarnat stage, and cord pH) were modestly predictive of 48 hours EEG outcome with area under curve (AUC) of 0.66 to 0.68. A composite model of clinical and optimal time- and frequency-domain blood pressure variability had a substantially improved AUC of 0.86. CONCLUSION: Time- and frequency-domain blood pressure variability biomarkers offer a substantial improvement in prediction of later adverse EEG outcomes over perinatal clinical variables in a two-center cohort of infants with HIE. KEY POINTS: · Early outcome prediction in HIE is suboptimal.. · Patterns in blood pressure physiology may be predictive of short-term outcomes.. · Early time- and frequency-domain measures of blood pressure variability predict short-term EEG outcomes in HIE infants better than perinatal factors alone..

A regression-based differential expression detection algorithm for microarray studies with ultra-low sample size.

Global gene expression analysis using microarrays and, more recently, RNA-seq, has allowed investigators to understand biological processes at a system level. However, the identification of differentially expressed genes in experiments with small sample size, high dimensionality, and high variance remains challenging, limiting the usability of these tens of thousands of publicly available, and possibly many more unpublished, gene expression datasets. We propose a novel variable selection algorithm for ultra-low-n microarray studies using generalized linear model-based variable selection with a penalized binomial regression algorithm called penalized Euclidean distance (PED). Our method uses PED to build a classifier on the experimental data to rank genes by importance. In place of cross-validation, which is required by most similar methods but not reliable for experiments with small sample size, we use a simulation-based approach to additively build a list of differentially expressed genes from the rank-ordered list. Our simulation-based approach maintains a low false discovery rate while maximizing the number of differentially expressed genes identified, a feature critical for downstream pathway analysis. We apply our method to microarray data from an experiment perturbing the Notch signaling pathway in Xenopus laevis embryos. This dataset was chosen because it showed very little differential expression according to limma, a powerful and widely-used method for microarray analysis. Our method was able to detect a significant number of differentially expressed genes in this dataset and suggest future directions for investigation. Our method is easily adaptable for analysis of data from RNA-seq and other global expression experiments with low sample size and high dimensionality.