1 The burden of fluid overload

Fluid overload in critically ill patients is defined as a 10% increase in cumulative fluid balance (CFB) from baseline body weight and represents a well-known problem which detrimentally affects intensive care unit (ICU) patient’s clinical course and their outcome [1,2,3]. Therefore, timely recognition and correction of fluid overload, or even better, identification of patients at risk for fluid accumulation in an early stage is of great importance. However, this remains challenging in clinical practice, as we miss an accurate and reliable tool for correct assessment of fluid status [4]. Current methods include the calculation of daily and CFB, by recording fluid inputs and outputs, the use of clinical or biochemical signs of fluid overload, monitoring filling pressures, however, none of these methods allow for a close monitoring for fluid balance, and intercompartmental fluid shifts [1, 5].

2 The promise for BIA

Bio-electrical impedance analysis (BIA) has gained increased interest to help physicians to determine volume status and fluid distribution in critically ill patients (see Table 1) [1, 2, 4, 6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21]. Indeed, several data published in the last decade suggest that BIA may provide useful information not only in different well-established patient groups (dialysis, AIDS, malnutrition), but also in critically ill patients with burns, trauma or sepsis undergoing fluid resuscitation. This tool may offer a non-invasive, fast and reliable assessment of volume status and fluid distribution as well as evaluation of dynamic changes in fluid distribution. It measures total body water (TBW), extracellular water (ECW), and intracellular water (ICW). BIA can calculate absolute fluid overload (AFO), the difference between normal, expected ECW and the actual, measured ECW, expressed in liters, as well as relative fluid overload (RFO), AFO/ECW, expressed in percentages. BIA measurements may be performed easily at the bedside, do not require extensive training and have limited inter-observer variations [22]. Thereby, BIA sounds as a promising diagnostic tool in the ICU or operating room to assess fluid status. Recently, new BIA-devices have been introduced allowing not only calculation of TBW, ECW and ICW but also the intra- and extravascular fluids (IVF, EVF) and, in case of dialysis, Kt/V (Fig. 1).

Table 1 Overview of recent studies published on the use of bio-electrical impedance analysis in critically ill patients
Full size table
Fig. 1
figure 1figure 1figure 1

The future of bio-electrical impedance analysis (BIA)

Full size image

3 Shedding new light

We read with great interest the work by Ciumanghel et al. [23]. Their study highlights an important clinical problem, fluid overload assessment, in a field that lacks clinical data. The study population was appropriate as abdominal surgery is among the most common elective surgical procedures [24]. The results confirm the potential role and usefulness of BIA monitoring to quantify body fluid composition and intercompartmental shifts in a major abdominal surgery perioperative setting. In this study 71 adult patients undergoing elective major abdominal surgery were included. Patients were then divided in two subgroups according to the presence of pre- and postoperative cumulative fluid overload (CFO): Normal Hydration (NH) subgroup with CFO < 5%, and Fluid Overload (FO) subgroup with CFO > = 5%. The authors found some differences between the subgroups regarding peroperative and postoperative parameters like median surgery duration and mean fluid infusion rate, amongst others. Positive intraoperative fluid balance (2.4 ± 1.0 L) resulted in a significant increase of TBW (1.4 ± 2.4 L) and of ECW (1.4 ± 1.2 L). Intraoperative fluid balance significantly correlated with TBW change (r = 0.23, p = 0.04) and with AFO change (r = 0.31, p < 0.01). A significant correlation was found between pre- and postoperative AFO and RFO on one hand, and ICU-LOS on the other. In addition, we would like to provide some additional comments as food for further thought.

First of all, before surgery, patients in the FO subgroup had significantly lower haemoglobin values and significantly lower diastolic blood pressure than patients in the NH subgroup. Furthermore, patients in the FO more frequently underwent duodeno-pancreatectomy, esophagectomy and aorto-femoral bypass. This was the case in around 72.7% (16 out of 22 patients) in the FO subgroup. This form of extensive surgery usually concerns patients presenting with a serious illness like cancer or peripheral vessel disease, also having more comorbidities.

Second, postoperative serum albumin levels dropped dramatically in FO patients, while in NH patients it decreased to a much lesser extent. A drop of serum albumin levels in the early postoperative period is known to be multifactorial: altered metabolism, blood loss/dilution and capillary leakage-related redistribution into the third space [25,26,27,28,29]. The latter being probably the most important mechanism [25], which may account for > 75% of albumin decrease in the early postoperative phase. This significant parameter also appears to be related to the magnitude of postsurgical systemic inflammatory response [26, 27], which, in turn, is directly related to the extent of surgery [25] and is believed to contribute to the risk of developing postoperative complications [28, 29]. Labgaa et al. [25] identified a serum albumin decrease ≥ 10 g/L on postoperative day 1 after elective abdominal surgery to be independently associated with a threefold increased risk for postoperative complications. Therefore, we believe that serum albumin drop in FO subgroup not only mirrors a higher intraoperative fluid regimen but may also indicates a more pronounced postsurgical systemic inflammatory response in those patients.

Altogether, those points led us to the feeling that preoperatively, FO subgroup patients may have been already more ill and therefore may have required increased fluid infusion during surgery. Indeed, even if blood loss was the same in both subgroups, FO patients received significantly more colloid infusion. Data on the inotropic and vasoactive medication use were collected but are not presented in the paper. One could expect more vasopressors in the FO subgroup. However, in order to maintain patient’s blood pressure peroperatively, one can give intravenous fluid infusion or choose vasopressor drugs. Both strategies being widely accepted in clinical practice. Unfortunately, little is known about the fluid management protocol, or whether it recommends early or late use of vasopressors? Even if fluid overload is known to impair the patient's clinical course in view of aforementioned comments, the greater rate of postoperative respiratory dysfunction and longer ICU stay in FO patients [30,31,32] seen in the work of Ciumanghel et al., may not be explained only by the presence of fluid overload [23].

Third, concerning BIA parameters, a comparison of pre- and postoperative BIA parameters was done for the whole group, and then in FO and NH subgroups separately. All these analyses consistently showed a significant increase in TBW, ECW, ECW/ICW ratio, AFO and RFO. We would also suggest a comparative analysis between FO and NH subgroup, as we believe that the increase in BIA-parameters in the FO subgroup would be more pronounced, due both to longer surgery and more severe illness, as partially discussed above. Moreover, the preoperative AFO and RFO values (more refined BIA parameters not related to body weight or height) were significantly higher in FO group. This could, therefore, additionally indicate that NH and FO patients did not present with the same baseline parameters before surgery.

Fourth, regarding AFO and RFO differences between FO and NH patients, we were surprised to observe that FO patients had lower TBW and ECW values preoperatively compared to NH patients. This might be explained by significantly higher BMI values in NH patient’s subgroup. Hence, it would seem appropriate to present BIA data as litre/kg body weight, as it would give a more appropriate estimation of patient’s true fluid composition.

Fifth, as the authors stated themselves, having only one postoperative BIA measurement is a limitation for a broader understanding of dynamic changes in body fluid composition. More prolonged changes in the fluid balance and or distribution could probably have been revealed if later measurements would have been performed. For instance, a BIA measurement 12 and 24 h after ICU admission would have been of great interest. Nevertheless, we recognise that this was probably challenging in the particular setting of this study, as the patients in the NH subgroup (i.e., 2/3 of included patients) had a median ICU stay of 5 h. This could be an interesting topic for further studies in the field.

4 Take home message

In conclusion, the study of Ciumanghel et al. addresses an important clinical problem and proposes a non-invasive, feasible, easy to perform bedside BIA-technique to assess and monitor fluid status and fluid distribution in the perioperative period, which may be of great interest to help physicians to improve management, care and outcome in critically ill patients. In the future, newer techniques may become available that allow not only calculation of TBW, ICW, ECW but also IVF and EVF. This could be of interest to assess performance of dialysis (Kt/V) but also to assess pharmacokinetics and pharmacodynamics of drugs and the fluids they are diluted in. The use of BIA in critically ill patients sounds promising but is probably not ready yet for prime time.