Comparison of differences in cohort (forwards) and case control (backwards) methodological approaches for validation of the Hypotension Prediction Index.

BACKGROUND: The Hypotension Prediction Index (the index) software is a machine learning algorithm that detects physiological changes that may lead to hypotension. The original validation used a case control (backwards) analysis that has been suggested to be biased. We therefore conducted a cohort (forwards) analysis and compared this to the original validation technique. METHODS: We conducted a retrospective analysis of data from previously reported studies. All data were analysed identically with 2 different methodologies and receiver operating characteristic curves (ROC) constructed. Both backwards and forwards analyses were performed to examine differences in area under the ROC for HPI and other haemodynamic variables to predict a MAP < 65mmHg for at least 1 minute 5, 10 and 15 minutes in advance. RESULTS: Two thousand and twenty-two patients were included in the analysis, yielding 4,152,124 measurements taken at 20 second intervals. The area-under-the-curve for the index predicting hypotension analysed by backward and forward methodologies respectively was 0.957 (95% CI, 0.947-0.964) vs 0.923 (95% CI, 0.912-0.933) 5 minutes in advance, 0.933 (95% CI, 0.924-0.942) vs 0.923 (95% CI, 0.911-0.933) 10 minutes in advance , and 0.929 (95% CI, 0.918-0.938) vs. 0.926 (95% CI, 0.914-0.937) 15 minutes in advance. No other variable had an area-under-the-curve > 0.7 except for MAP. Area-under-the-curve using forward analysis for MAP predicting hypotension 5, 10, and 15 minutes in advance was 0.932 (95% CI, 0.920-0.940), 0.929 (95% CI, 0.918-0.938), and 0.932 (95% CI, 0.921-0.940). The R 2 for the variation in the index due to MAP was 0.77. CONCLUSION: Using an updated methodology, we found the utility of the HPI index to predict future hypotensive events is high, with an area under the receiver-operating-characteristics curve similar to that of the original validation method.

Relationship between intraventricular mechanical dyssynchrony and left ventricular systolic and diastolic performance: An in vivo experimental study.

Left ventricular mechanical dyssynchrony (LVMD) refers to the nonuniformity in mechanical contraction and relaxation timing in different ventricular segments. We aimed to determine the relationship between LVMD and LV performance, as assessed by ventriculo-arterial coupling (VAC), LV mechanical efficiency (LVeff ), left ventricular ejection fraction (LVEF), and diastolic function during sequential experimental changes in loading and contractile conditions. Thirteen Yorkshire pigs submitted to three consecutive stages with two opposite interventions each: changes in afterload (phenylephrine/nitroprusside), preload (bleeding/reinfusion and fluid bolus), and contractility (esmolol/dobutamine). LV pressure-volume data were obtained with a conductance catheter. Segmental mechanical dyssynchrony was assessed by global, systolic, and diastolic dyssynchrony (DYS) and internal flow fraction (IFF). Late systolic LVMD was related to an impaired VAC, LVeff , and LVEF, whereas diastolic LVMD was associated with delayed LV relaxation (logistic tau), decreased LV peak filling rate, and increased atrial contribution to LV filling. The hemodynamic factors related to LVMD were contractility, afterload, and heart rate. However, the relationship between these factors differed throughout the cardiac cycle. LVMD plays a significant role in LV systolic and diastolic performance and is associated with hemodynamic factors and intraventricular conduction.

Indicators of haemodynamic instability and left ventricular function in a porcine model of esmolol induced negative inotropy.

To investigate if the Hypotension Prediction Index was an early indicator of haemodynamic instability in a negative inotropy porcine model, and to assess the correlation of commonly measured indicators of left ventricular systolic function. Eight anaesthetised pigs were volume resuscitated and then underwent an incremental infusion of esmolol hydrochloride (0-3000 mg/hr), following which it was then reduced in a stepwise manner. Full haemodynamic measurements were taken at each stage and measurements of left ventricular systolic function including left ventricular stroke work index, ejection fraction and peripheral dP/dT were obtained. At an infusion rate of 500 mg/hr of esmolol there were no significant changes in any measured variables. At 1000 mg/hr MAP was on average 11 mmHg lower (95% CI 1 to 11 mmHg, p = 0.027) with a mean of 78 mmHg, HPI increased by 33 units (95% CI 4 to 62, p = 0.026) with a mean value of 63. No other parameters showed significant change from baseline values. Subsequent increases in esmolol showed changes in all parameters except SVV, SVR and PA mean. Correlation between dP/dt and LVSWI was 0.85 (95% CI 0.77 to 0.90, p < 0.001), between LVEF and dP/dt 0.39 (95% CI 0.18 to 0.57, p < 0.001), and between LSWI and LVEF 0.41 (95% CI 0.20 to 0.59, p < 0.001). In this model haemodynamic instability induced by negative inotropy was detected by the HPI algorithm prior to any clinically significant change in commonly measured variables. In addition, the peripheral measure of left ventricular contractility dP/dt correlates well with more established measurements of LV systolic function.

Performance of the Hypotension Prediction Index with non-invasive arterial pressure waveforms in non-cardiac surgical patients.

An algorithm derived from machine learning uses the arterial waveform to predict intraoperative hypotension some minutes before episodes, possibly giving clinician's time to intervene and prevent hypotension. Whether the Hypotension Prediction Index works well with noninvasive arterial pressure waveforms remains unknown. We therefore evaluated sensitivity, specificity, and positive predictive value of the Index based on non-invasive arterial waveform estimates. We used continuous hemodynamic data measured from ClearSight (formerly Nexfin) noninvasive finger blood pressure monitors in surgical patients. We re-evaluated data from a trial that included 320 adults ≥ 45 years old designated ASA physical status 3 or 4 who had moderate-to-high-risk non-cardiac surgery with general anesthesia. We calculated sensitivity and specificity for predicting hypotension, defined as mean arterial pressure ≤ 65 mmHg for at least 1 min, and characterized the relationship with receiver operating characteristics curves. We also evaluated the number of hypotensive events at various ranges of the Hypotension Prediction Index. And finally, we calculated the positive predictive value for hypotension episodes when the Prediction Index threshold was 85. The algorithm predicted hypotension 5 min in advance, with a sensitivity of 0.86 [95% confidence interval 0.82, 0.89] and specificity 0.86 [0.82, 0.89]. At 10 min, the sensitivity was 0.83 [0.79, 0.86] and the specificity was 0.83 [0.79, 0.86]. And at 15 min, the sensitivity was 0.75 [0.71, 0.80] and the specificity was 0.75 [0.71, 0.80]. The positive predictive value of the algorithm prediction at an Index threshold of 85 was 0.83 [0.79, 0.87]. A Hypotension Prediction Index of 80-89 provided a median of 6.0 [95% confidence interval 5.3, 6.7] minutes warning before mean arterial pressure decreased to < 65 mmHg. The Hypotension Prediction Index, which was developed and validated with invasive arterial waveforms, predicts intraoperative hypotension reasonably well from non-invasive estimates of the arterial waveform. Hypotension prediction, along with appropriate management, can potentially reduce intraoperative hypotension. Being able to use the non-invasive pressure waveform will widen the range of patients who might benefit.Clinical Trial Number: ClinicalTrials.gov NCT02872896.

Dynamic Arterial Elastance as a Ventriculo-Arterial Coupling Index: An Experimental Animal Study.

Dynamic arterial elastance (Eadyn), the ratio between arterial pulse pressure and stroke volume changes during respiration, has been postulated as an index of the coupling between the left ventricle (LV) and the arterial system. We aimed to confirm this hypothesis using the gold-standard for defining LV contractility, afterload, and evaluating ventricular-arterial (VA) coupling and LV efficiency during different loading and contractile experimental conditions. Twelve Yorkshire healthy female pigs submitted to three consecutive stages with two opposite interventions each: changes in afterload (phenylephrine/nitroprusside), preload (bleeding/fluid bolus), and contractility (esmolol/dobutamine). LV pressure-volume data was obtained with a conductance catheter, and arterial pressures were measured via a fluid-filled catheter in the proximal aorta and the radial artery. End-systolic elastance (Ees), a load-independent index of myocardial contractility, was calculated during an inferior vena cava occlusion. Effective arterial elastance (Ea, an index of LV afterload) was calculated as LV end-systolic pressure/stroke volume. VA coupling was defined as the ratio Ea/Ees. LV efficiency (LVeff) was defined as the ratio between stroke work and the LV pressure-volume area. Eadyn was calculated as the ratio between the aortic pulse pressure variation (PPV) and conductance-derived stroke volume variation (SVV). A linear mixed model was used for evaluating the relationship between Ees, Ea, VA coupling, LVeff with Eadyn. Eadyn was inversely related to VA coupling and directly to LVeff. The higher the Eadyn, the higher the LVeff and the lower the VA coupling. Thus, Eadyn, an easily measured parameter at the bedside, may be of clinical relevance for hemodynamic assessment of the unstable patient.

Ability of an Arterial Waveform Analysis-Derived Hypotension Prediction Index to Predict Future Hypotensive Events in Surgical Patients.

BACKGROUND: Intraoperative hypotension is associated with worse perioperative outcomes for patients undergoing major noncardiac surgery. The Hypotension Prediction Index is a unitless number that is derived from an arterial pressure waveform trace, and as the number increases, the risk of hypotension occurring in the near future increases. We investigated the diagnostic ability of the Hypotension Prediction Index in predicting impending intraoperative hypotension in comparison to other commonly collected perioperative hemodynamic variables. METHODS: This is a 2-center retrospective analysis of patients undergoing major surgery. Data were downloaded and analyzed from the Edwards Lifesciences EV1000 platform. Receiver operating characteristic curves were constructed for the Hypotension Prediction Index and other hemodynamic variables as well as event rates and time to event. RESULTS: Two hundred fifty-five patients undergoing major surgery were included in the analysis yielding 292,025 data points. The Hypotension Prediction Index predicted hypotension with a sensitivity and specificity of 85.8% (95% CI, 85.8%-85.9%) and 85.8% (95% CI, 85.8%-85.9%) 5 minutes before a hypotensive event (area under the curve, 0.926 [95% CI, 0.925-0.926]); 81.7% (95% CI, 81.6%-81.8%) and 81.7% (95% CI, 81.6%-81.8%) 10 minutes before a hypotensive event (area under the curve, 0.895 [95% CI, 0.894-0.895]); and 80.6% (95% CI, 80.5%-80.7%) and 80.6% (95% CI, 80.5%-80.7%) 15 minutes before a hypotensive event (area under the curve, 0.879 [95% CI, 0.879-0.880]). The Hypotension Prediction Index performed superior to all other measured hemodynamic variables including mean arterial pressure and change in mean arterial pressure over a 3-minute window. CONCLUSIONS: The Hypotension Prediction Index provides an accurate real time and continuous prediction of impending intraoperative hypotension before its occurrence and has superior predictive ability than the commonly measured perioperative hemodynamic variables.

Determinants of left ventricular ejection fraction and a novel method to improve its assessment of myocardial contractility.

BACKGROUND: The aim of this study was to quantify the impact of different cardiovascular factors on left ventricular ejection fraction (LVEF) and test a novel LVEF calculation considering these factors. RESULTS: 10 pigs were studied. The experimental protocol consisted of sequentially changing afterload, preload and contractility. LV pressure-volume (PV) loops and peripheral arterial pressure were obtained before and after each intervention. LVEF was calculated as stroke volume (SV)/end-diastolic volume (EDV). We studied global cardiac function variables: LV end-systolic elastance (Ees), effective arterial elastance (Ea), end-diastolic volume and heart rate. Diastolic function was evaluated by means of the ventricular relaxation time (τ) and ventricular stiffness constant (ß) obtained from the end-diastolic PV relationship. Ventriculo-arterial coupling (VAC), an index of cardiovascular performance, was calculated as Ea/Ees. LV mechanical efficiency (LVeff) was calculated as the ratio of stroke work to LV pressure-volume area. A linear mixed model was used to determine the impact of cardiac factors (Ees, Ea, EDV and heart rate), VAC and LVeff on LVEF during all experimental conditions. LVEF was mainly related to Ees and Ea. There was a strong relationship between LVEF and both VAC and LVeff (r2 = 0.69 and r2 = 0.94, respectively). The relationship between LVEF and Ees was good (r2 = 0.43). Adjusting LVEF to afterload ([Formula: see text]) performed better for estimating Ees (r2 = 0.75) and improved the tracking of LV contractility changes, even when a peripheral Ea was used as surrogate (Ea = radial MAP/SV; r2 = 0.73). CONCLUSIONS: LVEF was mainly affected by contractility and afterload changes and was strongly related to VAC and LVeff. An adjustment to LVEF that considers the impact of afterload provided a better assessment of LV contractility.

Reliability of effective arterial elastance using peripheral arterial pressure as surrogate for left ventricular end-systolic pressure.

To compare the effective arterial elastance (Ea) obtained from the arterial pressure with Ea calculated from left-ventricular (LV) pressure-volume analysis. Experimental study. LV pressure-volume data was obtained with a conductance catheter and arterial pressures were measured via a fluid-filled catheter placed in the proximal aorta, femoral and radial arteries. Ea was calculated as LV end-systolic pressure (ESP)/stroke volume (SV). Experimental protocol consisted sequentially changing afterload (phenylephrine/nitroprusside), preload (bleeding/fluid), and contractility (esmolol/dobutamine). 90% of systolic pressure (Eaao_SYS, Eafem_SYS, Earad_SYS), mean arterial pressure (Eaao_MAP, Eafem_MAP, Earad_MAP), and dicrotic notch pressure (Eaao_DIC, Eafem_DIC, Earad_DIC) were used as surrogates for LV ESP. SV was calculated from the LV pressure-volume data. When Ea was compared with estimations based on 90% SAP, the relationship was r2 = 0.95, 0.94 and 0.92; and the bias and limits of agreement (LOA): - 0.01 ± 0.12, - 0.09 ± 0.12, - 0.05 ± 0.15 mmHg ml-1, for Eaao_SYS, Eafem_SYS and Earad_SYS, respectively. For estimates using dicrotic notch, the relationship was r2 = 0.94, 0.95 and 0.94 for Eaao_DIC, Eafem_DIC and Earad_DIC, respectively; with a bias and LOA: 0.05 ± 0.11, 0.06 ± 0.12, 0.10 ± 0.12 mmHg ml-1, respectively. When Ea was compared with estimates using MAP, the relationship was r2 = 0.95, 0.96 and 0.95 for Eaao_MAP, Eafem_MAP and Earad_MAP, respectively; with a bias and LOA: 0.05 ± 0.11, 0.06 ± 0.11, 0.06 ± 0.11 mmHg ml-1, respectively. LV ESP can be estimated from the arterial pressure. Provided that the SV measurement is reliable, the ratio MAP/SV provides a robust Ea surrogate over a wide range of hemodynamic conditions and is interchangeably in any peripheral artery, so it should be recommended as an arterial estimate of Ea in further research.

Performance comparison of ventricular and arterial dP/dtmax for assessing left ventricular systolic function during different experimental loading and contractile conditions.

BACKGROUND: Maximal left ventricular (LV) pressure rise (LV dP/dtmax), a classical marker of LV systolic function, requires LV catheterization, thus surrogate arterial pressure waveform measures have been proposed. We compared LV and arterial (femoral and radial) dP/dtmax to the slope of the LV end-systolic pressure-volume relationship (Ees), a load-independent measure of LV contractility, to determine the interactions between dP/dtmax and Ees as loading and LV contractility varied. METHODS: We measured LV pressure-volume data using a conductance catheter and femoral and radial arterial pressures using a fluid-filled catheter in 10 anesthetized pigs. Ees was calculated as the slope of the end-systolic pressure-volume relationship during a transient inferior vena cava occlusion. Afterload was assessed by the effective arterial elastance. The experimental protocol consisted of sequentially changing afterload (phenylephrine/nitroprusside), preload (bleeding/fluid bolus), and contractility (esmolol/dobutamine). A linear-mixed analysis was used to assess the contribution of cardiac (Ees, end-diastolic volume, effective arterial elastance, heart rate, preload-dependency) and arterial factors (total vascular resistance and arterial compliance) to LV and arterial dP/dtmax. RESULTS: Both LV and arterial dP/dtmax allowed the tracking of Ees changes, especially during afterload and contractility changes, although arterial dP/dtmax was lower compared to LV dP/dtmax (bias 732 ± 539 mmHg⋅s- 1 for femoral dP/dtmax, and 625 ± 501 mmHg⋅s- 1 for radial dP/dtmax). Changes in cardiac contractility (Ees) were the main determinant of LV and arterial dP/dtmax changes. CONCLUSION: Although arterial dP/dtmax is a complex function of central and peripheral arterial factors, radial and particularly femoral dP/dtmax allowed reasonably good tracking of LV contractility changes as loading and inotropic conditions varied.

Machine-learning Algorithm to Predict Hypotension Based on High-fidelity Arterial Pressure Waveform Analysis.

WHAT WE ALREADY KNOW ABOUT THIS TOPIC: WHAT THIS ARTICLE TELLS US THAT IS NEW: BACKGROUND:: With appropriate algorithms, computers can learn to detect patterns and associations in large data sets. The authors' goal was to apply machine learning to arterial pressure waveforms and create an algorithm to predict hypotension. The algorithm detects early alteration in waveforms that can herald the weakening of cardiovascular compensatory mechanisms affecting preload, afterload, and contractility. METHODS: The algorithm was developed with two different data sources: (1) a retrospective cohort, used for training, consisting of 1,334 patients' records with 545,959 min of arterial waveform recording and 25,461 episodes of hypotension; and (2) a prospective, local hospital cohort used for external validation, consisting of 204 patients' records with 33,236 min of arterial waveform recording and 1,923 episodes of hypotension. The algorithm relates a large set of features calculated from the high-fidelity arterial pressure waveform to the prediction of an upcoming hypotensive event (mean arterial pressure < 65 mmHg). Receiver-operating characteristic curve analysis evaluated the algorithm's success in predicting hypotension, defined as mean arterial pressure less than 65 mmHg. RESULTS: Using 3,022 individual features per cardiac cycle, the algorithm predicted arterial hypotension with a sensitivity and specificity of 88% (85 to 90%) and 87% (85 to 90%) 15 min before a hypotensive event (area under the curve, 0.95 [0.94 to 0.95]); 89% (87 to 91%) and 90% (87 to 92%) 10 min before (area under the curve, 0.95 [0.95 to 0.96]); 92% (90 to 94%) and 92% (90 to 94%) 5 min before (area under the curve, 0.97 [0.97 to 0.98]). CONCLUSIONS: The results demonstrate that a machine-learning algorithm can be trained, with large data sets of high-fidelity arterial waveforms, to predict hypotension in surgical patients' records.

Effects of phenylephrine on cardiac output and venous return depend on the position of the heart on the Frank-Starling relationship.

INTRODUCTION: phenylephrine is used daily during anesthesia for treating hypotension. However, the effects of phenylephrine on cardiac output (CO) are not clear. We hypothesized that the impact of phenylephrine on cardiac output is related to preload dependency. METHODS: eight pigs were studied at a preload independent stage (after CO augmentation) and at a preload dependent stage (after a 21 ml/kg hemorrhage). At each stage, phenylephrine boluses (0.5, 1.0, 2.0, and 4.0 µg/kg) were given randomly while mean arterial pressure (MAP), CO, inferior vena cava flow (IVCf) (both measured using ultrasonic flow probes), and pulse pressure variation were measured. RESULTS: at the preload independent stage, phenylephrine boluses induced significant increases in MAP (from 72 ± 6 to 100 ± 6 mmHg; P < 0.05) and decreases in CO and IVCf (from 7.0 ± 0.8 to 6.0 ± 1.1 l/min and from 4.6 ± 0.5 to 3.8 ± 0.6 l/min, respectively). At the preload-dependent stage, phenylephrine boluses induced significant increases in MAP (from 40 ± 7 to 65 ± 9 mmHg), CO (from 4.1 ± 0.6 to 4.9 ± 0.7 l/min), and IVCf (from 3.0 ± 0.4 to 3.5 ± 0.6 l/min; all data presented are for 4 µg/kg). Incremental doses of phenylephrine induced incremental changes in cardiac output. A pulse pressure variation >16.4% before phenylephrine predicted an increase in stroke volume with a 93% sensitivity and a 100% specificity. CONCLUSION: impact of phenylephrine on cardiac output is related to preload dependency. When the heart is preload independent, phenylephrine boluses induce on average a decrease in cardiac output. When the heart is preload dependent, phenylephrine boluses induce on average an increase in cardiac output.

Predicting fluid responsiveness with stroke volume variation despite multiple extrasystoles.

OBJECTIVE: To investigate the ability of a new stroke volume variation algorithm to predict fluid responsiveness during general anesthesia and mechanical ventilation in animals with multiple extrasystoles. DESIGN: Prospective laboratory animal experiment. SETTING: Investigational laboratory. SUBJECTS: Eight instrumented pigs. INTERVENTIONS: Eight anesthetized and mechanically ventilated pigs were monitored with an arterial line and a pulmonary artery catheter. Multiple extrasystoles were induced by right ventricular pacing (25% of heart beats). Arterial pressure waveforms were recorded and stroke volume variation was computed from the new and from the standard algorithm. The new stroke volume variation algorithm is designed to restore the respiratory component of the arterial pressure waveform despite multiple ectopic heart beats. Cardiac output was measured before and after 56 fluid boluses (7 mL/kg of 6% hydroxy ethyl starch) performed at different volemic states. MEASUREMENTS AND MAIN RESULTS: A positive response to fluid boluses (>15% increase in cardiac output) was observed in 21 of 56 boluses. The new stroke volume variation was higher in responders than in nonresponders (19% ± 5% vs. 12% ± 3%, p < .05), whereas the standard stroke volume variation was similar in the two groups (29% ± 8% vs. 26% ± 11%, p = .4). Receiver operating characteristic curve analysis showed that the new stroke volume variation was an accurate predictor of fluid responsiveness (sensitivity = 86%, specificity = 85%, best cutoff value = 14%, area under the curve = 0.892 ±, whereas the standard stroke volume variation was not (area under the curve = 0.596 ± 0.077). CONCLUSIONS: In contrast to the standard stroke volume variation, the new stroke volume variation algorithm was able to predict fluid responsiveness in animals with multiple ventricular extrasystoles.

Peripheral vascular decoupling in porcine endotoxic shock.

Cardiac output measurement from arterial pressure waveforms presumes a defined relationship between the arterial pulse pressure (PP), vascular compliance (C), and resistance (R). Cardiac output estimates degrade if these assumptions are incorrect. We hypothesized that sepsis would differentially alter central and peripheral vasomotor tone, decoupling the usual pressure wave propagation from central to peripheral sites. We assessed arterial input impedance (Z), C, and R from central and peripheral arterial pressures, and aortic blood flow in an anesthetized porcine model (n = 19) of fluid resuscitated endotoxic shock induced by endotoxin infusion (7 µg·kg⁻¹·h⁻¹ increased to 14 and 20 µg·kg⁻¹·h⁻¹ every 10 min and stopped when mean arterial pressure <40 mmHg or Sv(O2) < 45%). Aortic, femoral, and radial artery pressures and aortic and radial artery flows were measured. Z was calculated by FFT of flow and pressure data. R and C were derived using a two-element Windkessel model. Arterial PP increased from aortic to femoral and radial sites. During stable endotoxemia with fluid resuscitation, aortic and radial blood flows returned to or exceeded baseline while mean arterial pressure remained similarly decreased at all three sites. However, aortic PP exceeded both femoral and radial arterial PP. Although Z, R, and C derived from aortic and radial pressure and aortic flow were similar during baseline, Z increases and C decreases when derived from aortic pressure whereas Z decreases and C increases when derived from radial pressure, while R decreased similarly with both pressure signals. This central-to-peripheral vascular tone decoupling, as quantified by the difference in calculated Z and C from aortic and radial artery pressure, may explain the decreasing precision of peripheral arterial pressure profile algorithms in assessing cardiac output in septic shock patients and suggests that different algorithms taking this vascular decoupling into account may be necessary to improve their precision in this patient population.

Calculating arterial pressure-based cardiac output using a novel measurement and analysis method.

Work on applying physical and physiological principles for determining cardiac output by analysis of pressure measurements has been pursued for decades. Reference measurements for this kind of cardiac output analysis rely on the pulmonary artery catheter (PAC), considered the clinical gold standard for cardiac output monitoring. Recent advances in signal processing, as well as applied information on the relationships that enable arterial pulse pressure to be used to determine stroke volume, have led to the development of a novel system that can continuously measure cardiac output from an arterial pressure waveform that does not require an external calibration reference method. There are significant challenges in applying statistical- and signal-processing practices to the analysis of complex physiological waveforms. This paper reviews the historical basis for measuring flow from the analysis of pressure in a vessel, establishes the physiological and mathematical basis for this new system and describes its performance under various physiological conditions.