Pulse pressure variation using a novel smartphone application (Capstesia) versus invasive pulse contour analysis in patients undergoing cardiac surgery: a secondary analysis focusing on clinical decision making

Hemodynamic Monitoring by Smartphone—Preliminary Report from a Comparative Prospective Observational Study

Background: Advanced hemodynamic monitoring supports making therapeutic decisions in critically ill patients. New technologies, including mobile health, have been introduced into the hemodynamic monitoring armamentarium. However, each monitoring method has potential limitations—content, technical and organizational. The aim of this study was to assess the comparability between measurements obtained with two arterial pressure cardiac output methods: Capstesia™ smartphone hemodynamic software (CS) and LiDCO Rapid™ uncalibrated hemodynamic monitor (LR). Methods: The initial analysis included 16 patients in the period 06–09 2020 without limitations that could make the results obtained unreliable. Eighty pairs of cardiac output measurements were obtained. The comparability of cardiac output results obtained with both methods was assessed using the Spearman’s rank correlation coefficient (R), the intra-class correlation (CCC) and the Bland–Altman curves analysis (B-A). Results: The median (IQR) cardiac output measured with CS and LR were 4.6 (3.9–5.7) and 5.5 (4.6–7.4) L min⁻¹, respectively. In the B-A analysis, CS cardiac output values were on average 1.2 (95% CI −2.1–4.4) L min^-1 lower than LR values. The correlation between cardiac output with CS and LR was moderate (r = 0.5; p = 0.04). After adjusting for the presence of the dicrotic notch on the pulse waveform, in the group of eight patients with a visible dicrotic notch, the CS and LR results differed by only 0.1 (95% CI −0.8–1.1) L min⁻¹, the correlation between CS and LR was close to complete (r = 0.96; p < 0.001), and the percentage error was 40%, with a CCC-CS of 0.98 (95% CI 0.95–0.99). Conclusions: The Capstesia^TM smartphone software can provide an alternative method of cardiac output assessment in patients meeting arterial pressure cardiac output evaluation criteria with a clearly discernible dicrotic notch on the arterial pulse pressure waveform. It is necessary to confirm the obtained observations on a larger group of patients; however, it may potentially make objective hemodynamic measurements ubiquitous in patients with invasive arterial pressure monitoring with a clearly discernible dicrotic notch.

Evaluation of a new smartphone optical blood pressure application (OptiBP™) in the post-anesthesia care unit: a method comparison study against the non-invasive automatic oscillometric brachial cuff as the reference method

We compared blood pressure (BP) values obtained with a new optical smartphone application (OptiBP™) with BP values obtained using a non-invasive automatic oscillometric brachial cuff (reference method) during the first 2 h of surveillance in a post-anesthesia care unit in patients after non-cardiac surgery. Three simultaneous BP measurements of both methods were recorded every 30 min over a 2-h period. The agreement between measurements was investigated using Bland-Altman and error grid analyses. We also evaluated the performance of the OptiBP™ using ISO81060-2:2018 standards which requires the mean of the differences ± standard deviation (SD) between both methods to be less than 5 mmHg ± 8 mmHg. Of 120 patients enrolled, 101 patients were included in the statistical analysis. The Bland-Altman analysis demonstrated a mean of the differences ± SD between the test and reference methods of + 1 mmHg ± 7 mmHg for mean arterial pressure (MAP), + 2 mmHg ± 11 mmHg for systolic arterial pressure (SAP), and + 1 mmHg ± 8 mmHg for diastolic arterial pressure (DAP). Error grid analysis showed that the proportions of measurement pairs in risk zones A to E were 90.3% (no risk), 9.7% (low risk), 0% (moderate risk), 0% (significant risk), 0% (dangerous risk) for MAP and 89.9%, 9.1%, 1%, 0%, 0% for SAP. We observed a good agreement between BP values obtained by the OptiBP™ system and BP values obtained with the reference method. The OptiBP™ system fulfilled the AAMI validation requirements for MAP and DAP and error grid analysis indicated that the vast majority of measurement pairs (≥ 99%) were in risk zones A and B.Trial Registration ClinicalTrials.gov Identifier: NCT04262323.

Estimation of pulse pressure variation and cardiac output in patients having major abdominal surgery: a comparison between a mobile application for snapshot pulse wave analysis and invasive pulse wave analysis

Pulse pressure variation (PPV) and cardiac output (CO) can guide perioperative fluid management. Capstesia (Galenic App, Vitoria-Gasteiz, Spain) is a mobile application for snapshot pulse wave analysis (PWAsnap) and estimates PPV and CO using pulse wave analysis of a snapshot of the arterial blood pressure waveform displayed on any patient monitor. We evaluated the PPV and CO measurement performance of PWAsnap in adults having major abdominal surgery. In a prospective study, we simultaneously measured PPV and CO using PWAsnap installed on a tablet computer (PPV_PWAsnap, CO_PWAsnap) and using invasive internally calibrated pulse wave analysis (ProAQT; Pulsion Medical Systems, Feldkirchen, Germany; PPV_ProAQT, CO_ProAQT). We determined the diagnostic accuracy of PPV_PWAsnap in comparison to PPV_ProAQT according to three predefined PPV categories and by computing Cohen’s kappa coefficient. We compared CO_ProAQT and CO_PWAsnap using Bland-Altman analysis, the percentage error, and four quadrant plot/concordance rate analysis to determine trending ability. We analyzed 190 paired PPV and CO measurements from 38 patients. The overall diagnostic agreement between PPV_PWAsnap and PPV_ProAQT across the three predefined PPV categories was 64.7% with a Cohen’s kappa coefficient of 0.45. The mean (± standard deviation) of the differences between CO_PWAsnap and CO_ProAQT was 0.6 ± 1.3 L min^− 1 (95% limits of agreement 3.1 to − 1.9 L min^− 1) with a percentage error of 48.7% and a concordance rate of 45.1%. In adults having major abdominal surgery, PPV_PWAsnap moderately agrees with PPV_ProAQT. The absolute and trending agreement between CO_PWAsnap with CO_ProAQT is poor. Technical improvements are needed before PWAsnap can be recommended for hemodynamic monitoring.