Abstract / Summary
Abstract Background Change in the corrected flow time of the common carotid artery (ccFT Δ ) during a preload challenge (PC) is a surrogate for stroke volume change (SV Δ ) to infer preload responsiveness (P R ). Methods This is a retrospective cohort study comprising patients with suspected sepsis in a large, publicly funded, academic emergency department; no patient health information is linked to the Doppler assessments. Carotid Doppler spectrograms from a wireless, wearable critical care Doppler ultrasound system are analyzed for ccFT Δ during a PC (either infusion of 300 mL of crystalloid over 3 min or a Trendelenburg maneuver). The primary objectives of this study are to describe the proportion of P R patients by previously described ccFT Δ thresholds (i.e., both 4 and 7 ms ccFT Δ ), describe time-to-peak (TTP) ccFT Δ , and evaluate the ability of baseline ccFT to detect ccFT Δ during PC, using both the equations of Wodey and Bazett. Comparisons of related outcomes were adjusted for multiple comparisons using the Holm procedure, baseline ccFT discrimination was quantified using area under the receiver operating characteristic curve, and TTP was estimated using patient-level bootstrap resampling. Threshold-agnostic ccFT Δ trajectories were additionally explored using unsupervised k-means clustering of patient-level temporal responses for both correction equations. Results A PC was attempted in 4367 patients and an adequate ccFT Δ was collected from 4,128 patients (95%). In total, 982,937 Doppler cardiac cycles are included. The median ccFT Δ was 7.3 milliseconds (ms). 51.9% of patients had a ccFT Δ ≥ 7 ms, while 68.6% had a ccFT Δ ≥ 4 ms by the equation of Wodey. For Bazett, these fractions were 61.2% and 73.5%, respectively. For patients identified as P R based upon either a 4 ms or 7 ms ccFT Δ threshold, the time-to-peak ccFT Δ was 27.5 s for both equations. Baseline absolute ccFT did not discriminate P R from P UR patients under both correction formulae (AUROC 0.56–0.57). Conclusions To our knowledge this is the largest reported dataset of ED patients with suspected sepsis receiving P R assessments by Doppler ultrasound, comprising nearly 1 million Doppler cardiac cycles obtained from over 4100 ED patients. Both 7 ms and 4 ms ccFT Δ thresholds described clinically-accepted rates of P R . The TTP was roughly 30 s into PC and baseline ccFT did not accurately detect ccFT Δ during PC. Finally, a hypothesis-generating analysis identified four distinct ccFT Δ trajectories during PC which may represent unique preload response patterns during P R testing.