Systemic vascular resistance (SVR) classification is useful for the diagnosis and prognosis of critical pathophysiological conditions, with the ability to identify patients with abnormally high or low SVR of immense clinical value. In this study, a supervised classifier, based on Bayes' rule, is employed to classify a heterogeneous group of intensive care unit patients (N = 48) as being below (SVR < 900 dyn s cm(-5)), within (900 ⩽ SVR ⩽ 1200 dyn s cm(-5)) or above (SVR > 1200 dyn s cm(-5)) the clinically accepted range for normal SVR. Features derived from the finger photoplethysmogram (PPG) waveform and other routine cardiovascular measurements (heart rate and mean arterial pressure) were used as inputs to the classifier. In the construction of the classifier model, two techniques were used to approximate the class conditional probability densities--a single Gaussian distribution model (also known as discriminant analysis) and a non-parametric model using the Parzen window kernel density estimation method. An exhaustive feature search was performed to select a feature subset that maximized the performance indicator, Cohen's kappa coefficient (κ). The Gaussian model with multiple features achieved the best overall kappa coefficient (κ = 0.57), although the results from the non-parametric model were comparable (κ = 0.51). The optimum subset in the Gaussian model consisted of PPG waveform variability features, including the low-frequency to high-frequency ratio (LF/HF) and the normalized mid-frequency power (MF(NU)), in addition to the PPG pulse wave features, such as pulse width, peak-to-notch time, reflection index, and notch time ratio. The classifier performed particularly well in discriminating low SVR, with a sensitivity of 85%, specificity of 86%, positive predictive value of 88% and a negative predictive value of 82%. The results highlight the feasibility of deploying a multivariate statistical approach of SVR classification in the clinical setting, simply using a non-invasive and easy-to-measure PPG waveform signal.

BackgroundCardiac output (CO) and systemic vascular resistance (SVR) are two important parameters of the cardiovascular system. The ability to measure these parameters continuously and noninvasively may assist in diagnosing and monitoring patients with suspected cardiovascular diseases, or other critical illnesses. In this study, a method is proposed to estimate both the CO and SVR of a heterogeneous cohort of intensive care unit patients (N=48).MethodsSpectral and morphological features were extracted from the finger photoplethysmogram, and added to heart rate and mean arterial pressure as input features to a multivariate regression model to estimate CO and SVR. A stepwise feature search algorithm was employed to select statistically significant features. Leave-one-out cross validation was used to assess the generalized model performance. The degree of agreement between the estimation method and the gold standard was assessed using Bland-Altman analysis.ResultsThe Bland-Altman bias ±precision (1.96 times standard deviation) for CO was -0.01 ±2.70 L min-1 when only photoplethysmogram (PPG) features were used, and for SVR was -0.87 ±412 dyn.s.cm-5 when only one PPG variability feature was used.ConclusionsThese promising results indicate the feasibility of using the method described as a non-invasive preliminary diagnostic tool in supervised or unsupervised clinical settings.

Frequency spectrum analysis of circulatory signals has been proposed as a potential method for clinical risk assessment of preterm infants by previous studies. In this study, we examined the relationships between various spectral measures derived from systemic and cerebral cardiovascular variabilities and the clinical risk index for babies (CRIB II). Physiological data collected from 17 early low birth weight infants within 1-3 h after birth were analysed. Spectral and cross-spectral analyses were performed on heart rate variability, blood pressure variability and cerebral near-infrared spectroscopy measures such as oxygenated and deoxygenated haemoglobins (HbO(2) and HHb) and tissue oxygenation index (TOI). In addition, indices related to cardiac baroreflex sensitivity and cerebral autoregulation were derived from the very low, low- and mid-frequency ranges (VLF, LF and MF). Moderate correlations with CRIB II were identified from mean arterial pressure (MAP) normalized MF power (r = 0.61, P = 0.009), LF MAP-HHb coherence (r = 0.64, P = 0.006), TOI VLF percentage power (r = 0.55, P = 0.023) and LF baroreflex gain (r = -0.61, P = 0.01 after logarithmic transformation), with the latter two parameters also highly correlated with gestational age (r = -0.75, P = 0.0005 and r = 0.70, P = 0.002, respectively). The relationships between CRIB II and various spectral measures of arterial baroreflex and cerebral autoregulation functions have provided further justification for these measures as possible markers of clinical risks and predictors of adverse outcome in preterm infants.

Near-infrared spectroscopy (NIRS) for cerebral circulation monitoring has gained popularity in the neonatal intensive care setting, with studies showing the possibility of identifying preterm infants with intraventricular hemorrhage (IVH) by transfer function analysis of arterial blood pressure (BP) and NIRS measures. In this study, we examined a number of NIRS-derived measures in a cohort of preterm infants with IVH (n = 5) and without IVH (n = 12) within 1-3 hours after birth. The IVH infants were found to have significantly higher tissue oxygenation index (TOI), lower fractional tissue oxygen extraction (FTOE) and lower coherence between arterial BP and deoxygenated hemoglobin (HHb) in the very low frequency range (VLF, 0.02-0.04 Hz). Further studies with larger sample size are warranted for a more complete understanding of the clinical utility of these NIRS measures for early identification of IVH infants.

Low systemic vascular resistance (SVR) can be a useful indicator for early diagnosis of critical pathophysiological conditions such as sepsis, and the ability to identify low SVR from simple and noninvasive physiological signals is of immense clinical value. In this study, an SVR classification system is presented to recognize the occurrence of low SVR, among a heterogenous group of patients (N = 48), based on the use of routine cardiovascular measurements and features extracted from the finger photoplethysmogram (PPG) as inputs to a quadratic discriminant classifier. An exhaustive feature search was performed to identify a near optimum feature subset. Cohen's kappa coefficient (κ) was used as a performance measure to compare candidate feature sets. The classifier using the following combination of features performed best (κ = 0.56, sensitivity = 96.30%, positive predictivity = 92.31%): normalized low-frequency power (LFNU) derived from PPG, ratio of low-frequency power to high-frequency power (LF/HF) of the PPG variability signal, and the ratio of mean arterial pressure to heart rate (MAP/HR). Classifiers that used either LF(NU) (κ = 0.43), LF/HF (κ = 0.37) or MAP/HR (κ = 0.43) alone showed inferior performance. Discrimination of patients with and without low SVR can be achieved with reasonable accuracy using multiple features derived from the PPG combined with routine cardiovascular measurements.