Sepsis is defined as a life-threatening organ dysfunction due to dysregulated host response to an infection. In the pathophysiology of sepsis, endothelial dysfunction plays a most significant role. It leads to generalized vasodilation due to the increased NO production and increased capillary permeability leading to the loss of intravascular fluid into the interstitial space. The net result is severe intravascular hypovolemia and hypotension. Fluid resuscitation and vasopressors are the cornerstones in treating septic patients. According to Surviving Sepsis Campaign (SSC) international guidelines for management of sepsis and septic shock from 2021 for patients with sepsis, induced hypoperfusion or septic shock at least 30 mL/kg of IV crystalloid fluid should be given within the first three hours of resuscitation. This recommendation is downgraded from strong (2016 guidelines) to weak with low quality of evidence after a serious criticism and the lack of patient benefit from goal-directed therapy.^{[1, 2]} The exact amount, type, or rate of fluid administration in septic patients admitted to the ICU is still unclear and fluid therapy recommendations are based mainly on expert opinions. Although intravascular volume expansion is the main treatment option in septic patients, if applied inadequately it can lead to major complications and can be associated with increased ICU mortality.^[3] The main reason for the ambiguity is the lack of universal monitoring technique for fluid status assessment in sepsis. The bioelectrical impedance analysis (BIA) is a non-invasive, easy-to-use, inexpensive, portable, and fast tool for body composition and body fluids evaluation. In critically ill patients, it is used to assess the nutrition and hydration status and is used also in perioperative risk assessment and prognostication.

The aim of this study was to assess the usefulness of bioimpedance analysis in fluid status evaluation in patients with sepsis and septic shock admitted to the adult ICU.

We conducted a prospective, observational, clinician-blind study from April 2023 to January 2024 for patients admitted to the adult ICU. The inclusion criteria were a diagnosis of sepsis at admission to ICU, a stay in ICU of at least 72 hours, and the first BIA measurement to be conducted within the first three hours of ICU admission. We performed whole-body BIA measurements upon admission and every 24 hours thereafter for at least four consecutive measurements. We used BodyStat MultiScan 5000 bioimpedance analyzer with self-adhesive disposable electrodes provided by the manufacturer. Four electrodes were attached to the right hand and right foot - behind the knuckles, on the wrist, behind the toes, and on the ankle as recommended by the manufacturer. All enteral and parenteral fluids administered to the patients were recorded, as well as the loses through drains, tubes, aspiration, and urine. The cumulative fluid balance (CFB) was calculated every 24 hours for every patient without the insensible losses. The treating physicians were unaware of the bioimpedance measurements results. Determination of the volume of fluid therapy was based on constant monitoring of heart rate and invasive or non-invasive pressure measurements, CFB, standard ICU daily lab tests, but was also based on individual judgement. Dynamic parameters of fluid responsiveness were unavailable to use.

A total of 51 patients (30 men and 21 women) with a mean age of 62 years were included in the final analysis. Their mean ICU stay was 12 days, and the mortality rate was 48%. Baseline patient characteristics on admission are shown in Table 1. Biochemical data and hydration evaluation on admission are shown in Table 2.

CFB gradually increased during the first 72 hours of ICU stay from 2003±133 mL at 24 hours to 3680±2368 mL at 48 hours, and 5217±2642 mL at 72 hours. The bioimpedance measured parameters also showed a gradual increase. Tables 3–7 show the total body water (TBW), extracellular water (ECW), the ECW/TBW ratio, intracellular water (ICW), and overhydration (OHY) and their changes during the first 72 hours. There was a statistically significant positive correlation between the increase in CFB and the increase in TBW, ECW and OHY. Table 8 displays Pearson’s correlation coefficient and the corresponding p-value for the calculated CFB and BIA measured hydration parameters. Changes in the measured ICW did not correlate with CFB. The scatter plots displayed in Figs 1–3 show the correlation between the calculated total CFB and the overall changes in the BIA measured TBW, ECW, and OHY.

The parameters for assessment of the hydration status measured by bioimpedance vector analysis (BIVA) – the vector length (VL) and impedance ratio (IR) were also analyzed. The daily changes of VL and IR as well as the overall 72-hour changes showed statistically significant correlation with the CFB. Table 9 shows Pearson’s correlation coefficient and the corresponding p-value for the calculated CFB and the IR and the BIVA-measured VL. The scatter plots displayed in Figs 4, 5 show the correlation between the calculated total CFB and the overall changes in the BIVA-measured VL and IR.

Fluid status assessment in critically ill septic patients is extremely challenging. Dynamic tests for predicting fluid responsiveness are not always applicable. This may leave the static fluid status assessment parameters as the only option. Due to pathogenetic changes, the intravascular hypovolemia and extravascular fluid overload are common in septic patients. Balanced crystalloid solutions as a first-line treatment for intravascular hypovolemia show a relatively low plasma-expanding effect and worsen the interstitial tissue edema.^[4] The results in our study show a clear trend towards increase in the measured TBW and ECW during the resuscitation phase of critically ill septic patients. The increased ECW/TBW ratio is associated with higher mortality.^[5] Using BIA during initial resuscitation, we can monitor the extent of extravascular fluid overload occurring during large-volume resuscitation with balanced crystalloid solutions. BodyStat MultiScan 5000 uses spectroscopy to calculate the volume of fluid overload in liters and is displayed as OHY. This is of particular importance when assessing or monitoring dry or target weight for individual dialysis patients. OHY is a parameter showing extracellular fluid overload. In our study, OHY in critically ill septic patients showed good correlation with the CFB.

When patients’ body weight is not known, the IR can still be used. In critically ill septic patients, body weight and composition change significantly during their ICU stay. Monitoring body weight is also not available in all ICU beds. IR is the ratio between the impedance measured at 200 kHz and 5 kHz. In a healthy individual, at 5 kHz the resistance to the flow of the current will be higher because the current cannot penetrate the cell membrane (so it can only measure the ECW). At 200 kHz, the current can penetrate the cell membrane wall (the impedance is lower and measures TBW). IR, known also as prediction marker, can be used as fluid status assessment and as a prognostic marker. The maximum value of IR is 1.00 and higher values are interpreted as less cellular health or excess amount of ECW. In our study, IR showed good correlation with the increase in ECW and CFB.

BIVA (or ‘Rxc Graph’) uses no complex mathematical equations. It uses only the Resistance (R) and Reactance (Xc) at 50 kHz and is standardized to the subject’s height (weight is not required). The results are illustrated as dots on the vector. The line from the beginning to the dot indicates the VL. The shorter line indicates more tissue edema (as water and electrolytes have good conductivity, so the resistance to flow is less). In our study population, the increase in CFB showed a good correlation with the shortening of VL.

Our study has two major limitations. Firstly, daily CFB can be affected by a variety of factors due to the increase in insensible losses. Some examples include fever, open wounds, mechanical ventilation, etc., all of which are relatively common in septic patients. Adjusting the CFB with the insensible losses (those that we cannot measure) is merely impossible at the bedside. Nevertheless, CFB was used to compare the BIA-measured hydration parameters. The reason is that most clinicians adjust their fluid therapy empirically in cases of increased insensible losses. The aim of our study was to see if the BIA-measured parameters can track changes in body fluids, not to validate the method. In septic patients with generalized endothelial dysfunction, it is expected from physiology that with fluid therapy ECW will increase to a greater extent than ICW which is fairly supported by the results of our study.

The second limitation is that BIA provides static volume status parameters. Changes of these parameters over time showed moderate correlation with CFB. Like the central venous pressure (CVP), guiding fluid therapy based on static parameters would be too optimistic. Dynamic measures of fluid responsiveness (stroke volume variation [SVV], pulse pressure variation [PPV], echocardiography, etc.) should be used in critically ill septic patients. Although noninvasive and minimally invasive, access to modern devices and the availability of trained ICU personnel are still not universal, including in our department.

Guiding fluid therapy in septic patients is based on complex assessment of a variety of parameters. BIA is a non-invasive, easy-to-use, inexpensive, portable, and fast tool for fluid status assessment. In critically ill septic patients, it can be a useful tool in fluid therapy management.