Comparison of non-invasive cardiac output measurement and pulse-indicated continuous cardiac output monitoring for determining hemodynamic parameters in patients with critical septic shock: a prospective study

Objective To compare non-invasive cardiac output measurement (NICOM) and pulse-indicated continuous cardiac output (PiCCO) monitoring for determining hemodynamic parameters in patients with critical septic shock and to analyze the correlation between the two techniques. Methods Patients with critical septic shock admitted to the from April to December 2015 who required hemodynamic monitoring were enrolled prospectively. Cardiac output (CO) and stroke volume variation (SVV) were measured by NICOM and PiCCO in all patients and compared by Spearman’s correlation and Bland–Altman analyses. registered. Results Thirty-one patients were included in the study (19 males and 12 females, mean age ± standard deviation, 55.5 ± 18.1 years), with a mean Acute Physiology and Chronic Health Evaluation II (APACHE II) score of 22.7±6.1. There was no significant difference in CO measured by the NICOM and PiCCO methods (5.10 4.35, 6.50 L/min vs. 4.89 4.34, 6.23 L/min; P > 0.05). However, SVV measured by NICOM was significantly higher than that measured by PiCCO (13.00 11.00, 16.00 vs. 12.00 9.00, 15.00; P = 0.009). CO and SVV determined by NICOM and PiCCO were significantly correlated according to Spearman’s correlation analysis (CO: R = 0.904, P < 0.001, 95% confidence interval 0.932–1.135; SVV: R = 0.841, P < 0.001, 95% confidence interval 0.601–0.786). Bland–Altman analysis revealed a bias in mean CO of 0.21 L/min (P = 0.0032) and limits of agreement of −1.12 to 1.54 L/min; and a bias in mean SVV of 1.56 (P < 0.0001) and limits of agreement of −2.56 to 5.68. Conclusions Hemodynamic parameters monitored by NICOM and PiCCO differed in patients with critical septic

Keywords: Septic shock; Non-invasive cardiac output measurement; Cardiac output; stroke volume variation Background Severe infection and septic shock are the major causes of death in critically ill patients.
Monitoring the hemodynamics and guiding treatment are crucial measures for improving the prognosis of these patients [1,2]. Cardiac output (CO) is an index reflecting cardiac function in the clinic, and its continuous measurement and corresponding treatment adjustment are important measures in the treatment of severely ill patients. The variation in stroke output volume (SVV) reflects the volume responsiveness of patients, and monitoring this index is also necessary to improve patient mortality [3]. Pulse-indicated continuous cardiac output monitoring (PiCCO) provides continuous CO monitoring and volume responsiveness data by monitoring SVV, and has been widely used in clinical settings [4,5]. However, PiCCO requires invasive arteriovenous catheterization and is associated with certain clinical complications. According to Belda [6], 86.4% of 544 PiCCO catheters obtained arterial access in the first attempt, 3.3% exuded after insertion, 3.5% exuded after extraction, 4.5% inserted catheter, and 1.2% removed catheter . Local hematoma occurred , while site inflammation occurred in 2% of patients, catheter-related infection in 0.78%, ischemia in 0.4%, disappearance of pulse in 0.4%, and femoral artery thrombosis in 0.2%. However, novel, non-invasive hemodynamic monitoring technologies have developed rapidly in recent years, and non-invasive cardiac output monitoring (NICOM) based on bioelectrical impedance has attracted increasing attention [7,8].
The current study aimed to compare the accuracies of NICOM and PiCCO for monitoring CO and SVV in patients with septic shock, and to determine the correlation between the two techniques. The results support the use of NICOM for non-invasive hemodynamic monitoring, to reduce the risk of invasive procedures in patients with septic shock.

Hemodynamic monitoring
All patients received an intra-cervical or subclavian venous catheter connected to a pressure-measuring device, which was in turn connected to a monitor (Philips, the Netherlands). A 4F arterial catheter was placed in the femoral artery (Pulsion Medical Systems, Germany). Hemodynamics were monitored by PiCCO using the thermal dilution method and by NICOM immediately (0 h) and at 24 and 48 h. All measurements were made and recorded by the same investigator.
PiCCO monitoring: Using the thermal dilution method, 15 mL ice-cold normal saline was injected through the central venous catheter. CO and SVV values were measured three times continuously and the mean values were calculated and recorded.
NICOM measurement: Measurements were made using the NICOM monitoring system (Cheetah, USA) with the patient in the decubitus position. Four electrodes were attached to the patient's chest, abdomen, or back, symmetrical upper and lower, around the heart, and connected to the instrument. Data were acquired three times at 5 minute intervals, and the average values were calculated.

Statistical methods
The data were analyzed using SPSS 20 (SPSS Inc., Chicago, USA) and Medcalc software (MedCalc Software, Ostend,Belgium). Normally distributed variables were expressed as mean ± standard deviation and non-normally distributed variables as median (25th, 75th percentile). The results for the two measurement systems were compared by paired ttests. The correlation between the methods was analyzed by Spearman's correlation, and the consistency was evaluated by Bland-Altman analysis. A value of P < 0.05 was considered significant.

Patient characteristics
A total of 31 patients with septic shock were included in the study. There were 19 males and 12 females, aged 19-84 years (mean age 55.5 ± 18.1 years). The median of body mass index was 21.3 (20.9, 21.4) and the mean Acute Physiology and Chronic Health Score II (APACHE II) score was 22.7 (± 6.1). In terms of the type of infection, 19 patients had pneumonia, 11 had abdominal infection, and one had a bloodstream infection. Eleven patients had a previous history of cardiovascular disease, 12 had chronic obstructive pulmonary disease, and eight had chronic renal insufficiency. All 31 patients required continuous ventilator-assisted ventilation. Twenty-five patients received norepinephrine (median 0.5 (0.2, 0.8)) µg/kg/min to maintain blood pressure, and 12 patients needed continuous renal replacement therapy.
Comparison of hemodynamic parameters measured by two methods A total of 93 hemodynamic monitoring data values were obtained in 31 patients. The median of CO values measured by PiCCO and NICOM were 4.89 (4.34, 6.23) L/min and 5.10 (4.35, 6.50), respectively, with no significant difference between the two groups. However, the SVV measured by NICOM (13.00 [11.00, 16.00]) was significantly higher than that measured by PiCCO (12.00 [9.00, 15.00]) (P = 0.009) ( Table 1).  study found a significant correlation between NICOM and PICCO measurements of CO and SVV in patients with septic shock, and conformity analysis also showed that the two measurements were highly consistent. These observations suggest that non-invasive hemodynamic monitoring could replace invasive monitoring to a certain extent.
The human chest is a conductive cylinder, and the change in blood flow in the chest cavity when the heart pumps blood is the main factor affecting conductivity. When a fixed frequency of alternating current passes through the thorax, the impedance is proportional to the voltage at both ends of the thorax, causing a displacement between the released current and the induced voltage.  Bland-Altman analysis of consistency between PiCCO and NICOM for measuring stroke volume variability (SVV) in patients with septic shock