Identifying critical DO₂ with compensatory reserve during simulated hemorrhage in humans

Abstract

Background Based on previous experiments in nonhuman primates, we hypothesized that DO2crit in humans is 5–6 ml O2·kg−1 min−1. Study Design and Methods We measured the compensatory reserve (CRM) and calculated oxygen delivery (DO2) in 166 healthy, normotensive, nonsmoking subjects (97 males, 69 females) during progressive central hypovolemia induced by lower body negative pressure as a model of ongoing hemorrhage. Subjects were classified as having either high tolerance (HT; N = 111) or low tolerance (LT; N = 55) to central hypovolemia. Results HT and LT groups were matched for age, weight, BMI, and vital signs, DO2 and CRM at baseline. The CRM‐DO2 relationship was best fitted to a logarithmic model in HT subjects (amalgamated R2 = 0.971) and a second‐order polynomial model in the LT group (amalgamated R2 = 0.991). Average DO2crit for the entire subject cohort was estimated at 5.3 ml O2·kg−1 min−1, but was ~14% lower in HT compared with LT subjects. The reduction in DO2 from 40% CRM to 20% CRM was 2‐fold greater in the LT compared with the HT group. Conclusions Average DO2crit in humans is 5.3 ml O2·kg−1 min−1, but is ~14% lower in HT compared with LT subjects. The CRM‐DO2 relationship is curvilinear in humans, and different when comparing HT and LT individuals. The threshold for an emergent monitoring signal should be recalibrated from 30% to 40% CRM given that the decline in DO2 from 40% CRM to 20% CRM for LT subjects is located on the steepest part of the CRM‐DO2 relationship.

Noninvasive Monitoring of Simulated Hemorrhage and Whole Blood Resuscitation

“…Although LBNP ramp protocols have been used for investigation of the hemodynamic responses to progressive central hypovolemia [ 38 ], this is the first study in which the compensatory reserve has been analyzed with a constant ramp LBNP protocol that simulates both hemorrhage and resuscitation. Most previous LBNP protocols were designed to depressurize the vacuum chamber using a step profile in which negative pressure was maintained at a certain value [ 15 , 16 , 23 ] for a fixed period of time between depressurization steps. The ramped protocol applies a constant rate for both depressurization and repressurization, more accurately simulating both scenarios of whole blood hemorrhage or resuscitation, which are continuous processes in reality.…”

“…Furthermore, using this model they showed that a target CRM of 35% during whole blood resuscitation required only 40% of total blood volume lost to sustain adequate DO 2 to maintain hypotensive resuscitation [ 3 ]. Their most recent work establishes a CRM of 40% as a key threshold to ensure a patient at greatest risk for early onset of shock (i.e., individual with low tolerance to central hypovolemia) is at an adequate DO 2 level [ 23 ]. Unique to the current study is the varying rates of simulated whole blood hemorrhage and resuscitation used to develop the HRDN model.…”

“…This work supports the sound physiological basis for the concept of a compensatory reserve metric. Previous work by Convertino et al has established the efficacy of estimating compensatory reserve in a variety of human LBNP models using a step protocol [ 15 , 16 , 20 , 21 , 23 , 24 , 38 ]. This model has been shown to be an accurate simulation of real blood loss and resuscitation in humans and primates [ 3 , 10 , 16 , 29 , 39 ].…”

“…The effect of variable rates of simulated hemorrhage and resuscitation were also evaluated by averaging RMSE and R 2 per ramp speed: −9, −6, −3, +3, +6, and +9 mmHg/min, where the negative rates indicate the depressurization (simulated hemorrhage) and the positive rates indicate repressurization (simulated resuscitation). For AUC, ROCs were computed for binary classification at the clinically relevant CRM thresholds of 40% and 70% based on the work of Koons et al [ 23 ]. A CRM of 40% corresponds to the critical threshold at which reduction in DO 2 places individuals with low tolerance to reductions in central blood volume at greatest risk for imminent hemodynamic decompensation.…”

“…Convertino et al reported high correlations between CRM and percent blood lost and resuscitated in nonhuman primates. They calculated a clinically meaningful threshold of systemic oxygen delivery (DO 2 ) of 5.3 mL O 2 ·kg −1 ·min −1 and accordingly proposed an endpoint for resuscitation of CRM ~40% [ 3 , 23 ]. Li et al demonstrated the ability to track resuscitation using a ML model with feature sets extracted from electrocardiogram (ECG) and photoplethysmogram (PPG) waveforms using a porcine model of induced hemorrhagic shock and fluid resuscitation [ 14 ].…”