Did you know? Fluid‐and‐electrolyte replacement and the uncertainty principle

Surprises are not always welcome. We have studied total body water, quantified the volume and determined the extracellular space, interstitial, and plasma volumes. Nonetheless, estimating these parameters clinically (at the bedside) is frustratingly imprecise. The author ruminates over 50 years’ clinical experience in intensive care units and grapples with the literature. In September, 1831, Asiatic cholera erupted in England. Faced with a mounting death rate and lacking effective therapy, the medical profession grasped at an array of empirical remedies ranging from cayenne pepper given by mouth to turpentine enemas. In response to the need for rational therapy a young physician named William Brooke O'Shaughnessy was persuaded to study the chemical pathology of the blood in cholera. These studies laid the foundations of modern intravenous fluid therapy. The same man was to become a pioneer in the field of telegraphy and, inadvertently, to make a significant contribution to English history. Our understanding of the vasculature, the idea that the arteries are merely a grandiose extension of the heart (and vice-versa), the heart's vasomotor control, and the regulation of blood pressure play pivotal roles in volume assessment. For instance, 300 years ago de Sénac wrote; “The arteries (and the heart) have the same movements; they alternate with dilations and contractions without pause.”3 The force inherent in the arteries depends on their muscular fibres and their length. Regulation of the cardiovascular tree concerns us to this date and 300 years of research, until the studies of McGee et al,4 prior to the “molecular” era, brought us to where we are. We are endowed with a reasonable knowledge of the circulatory intrinsic, local, neuronal, and central regulation. These advancements must not be forgotten.3 The >300 years of cardiovascular research include the contributions of Bernard, Brown-Sequard, Magendie, Bayliss, Gaskell, Langley, Cannon, Dale, Frank, Starling, Hering and others. These brilliant investigators carried physiology on their (not-yet molecular biology) shoulders, not to mention the clinical relevance of hypotension, hypertension, and central nervous system, cardiovascular, osmolar and renal-volume, and regulation. Years after O'Shaughnessy and many others, clinicians earnestly sought to determine who, if anybody, needs volume-replacement therapy? Not quite 200 years later, McGee et al attempted to answer this exceedingly difficult clinical (those are patients who are conscious and can stand upright) issue.4 They asked the pivotal question: Is this patient hypovolemic? At their disposal were history, physical examination, stethoscopes and fingertips. The authors reviewed the literature and technical possibilities of the time. They found that in patients with increase in heart rate >30 beats/min with standing, altered consciousness, and a history of volume depletion were probably volume depleted. Serum studies on urea nitrogen concentration (BUN) and creatinine levels might also be of assistance so the authors added, but the evidence was uncertain. Can it be that the subject matter influences the outcomes? Aside from Hippocrates and his advice, “It is a bad sign in acute illness when the extremities become cold,” could invasive measurements help us to assess needs for volume replacement? Or could something as simple as passive leg raising along with minimally invasive monitoring help in guiding volume replacement in shock patients? In a preliminary study, investigators assessed passive leg raising manoeuvre as a means to alter biventricular preload, in combination with real-time measurement of cardiac output changes, and suggest that these measurements are an easy-to-use, fast, relatively unbiased, and accurate bedside test to guide fluid management and to avoid fluid overload during early septic shock treatment.5 Ideally, we would like to measure preload but that requires knowing muscle-fibre length. End-diastolic volume determinations would be a good estimate. But clinically, cardiac filling pressures have played a prominent role in evaluating the circulating fluid volume. Central venous pressure has been used for 40 years; however, the relationship between circulating blood volume and central venous pressure is inconsistent and not statistically significant.6 Thus, central venous pressure (although it is invariably monitored) should not be used to make decisions regarding fluid management. Pulmonary artery catheters to assess left-atrial filling pressures could be better but are much more cumbersome. Comparable data as compared to CVP measurements are not available. Nevertheless, a Cochrane Review on “Pulmonary Artery Catheters for Adult Patients in Intensive Care“ found that pulmonary artery catheters are a diagnostic and hemodynamic monitoring tool, but not a therapeutic intervention.7 The review concluded that use of techniques reflecting left-atrial measurements (filling pressures) did not alter the outcomes (mortality) or costs for adult patients in intensive care. Were the predictions concerning preload from pulmonary artery catheters more informative than CVP measurements, we might have expected a more positive outcome. Marino excellently reviewed this topic earlier.8 Serum lactate, lactate clearance, acid-base deficit, fluid challenges, passive leg raising and other manoeuvres are also not clearly helpful.9 Blood volume can be measured directly within 1 h. Yu et al performed a randomized controlled trial to determine whether or not measuring blood volume was better than guidance from a pulmonary artery catheter (reflecting left ventricular filling pressures) in patients with hypovolemic shock.10 The control group received fluid management based on pulmonary artery catheter parameters and red blood cell transfusions based on hematocrit values. The blood-volume group received fluid and red blood cell transfusions based on blood-volume analyses in addition to pulmonary artery catheter parameters and enjoyed improved survival. Recently, a method determining direct glomerular filtration rate (GFR) measurements and plasma volume has been presented.11 However, these promising data will have to be confirmed and standardized. Thus, physicians are commonly left with uncertainty and are asked to act without perfect guidance. Physicians need to know precisely what the results of their infusions are likely to be (Figure 1).8, 12 An isotonic 5% of glucose solution will expand the plasma volume very little but should contribute to the total body water (circa 40 L). Thus, selection of this option in shock patients is probably not helpful. With 0.9% of saline, we can reason that bulk of the solution will reside extracellularly (13 L). Attractive are colloid solutions such as those containing albumin or hetastarch. These solutions should substantially increase plasma volume, a common goal in shock patients. A research option is hypertonic saline, by which small volumes should have large effects.8 We now have sound data on most of these strategies. Which strategy is actually superior? Much has already been tested. Intravenous fluid therapy in critically ill adults has recently been reviewed1; we add only a brief reappraisal here. Early goal-directed therapy (EGDT) involves adjustments of cardiac preload, afterload, and contractility to balance oxygen delivery with oxygen demand. Rivers et al randomly assigned patients who arrived at an urban emergency department with severe sepsis or septic shock to receive either 6 hours of early goal-directed therapy or standard therapy (as a control) before admission to the intensive care unit.13 The EGDT patients received a central venous catheter; however, the focus was on central-venous oxygen saturation, not solely pressure. A 500-mL bolus of crystalloid was given every 30 minutes to achieve a central venous pressure of 8-12 mmHg. If the mean arterial pressure (MAP) was less than 65 mmHg, vasopressors were given to maintain a MAP of at least 65 mmHg. If the MAP was greater than 90 mmHg, vasodilators were given until it was 90 mmHg or below. If the central venous oxygen saturation was less than 70%, red cells were transfused to achieve a hematocrit of at least 30%. Vasopressors and vasodilators were administered according to a defined protocol. Less information is provided on the standard care group, but they also received a central venous catheter to guide a pressure 8-12 mmHg, their mean blood pressure goal was >65 mmHg, and their urine output was monitored in an effort to maintain >0.5 ml/kg/h. The in-hospital mortality was 30.5% in the group assigned to early goal-directed therapy, as compared with 46.5% in the group assigned to standard therapy. Ninety-day mortality was not reported, but 60-day mortality was 44% versus 57%. This report generated much excitement and provided needed guidance. Nevertheless, the study was single-centred and not above critique. Three subsequent trials of this issue were less favourable. As a result, a meta-analysis was performed that compared 1857 EGDT patients to 1880 standard-care patients.14 The severity of illness scores and the care-delivery characteristics were similar to the earlier report. Mortality at 90 days was similar for DGDT (25%) and standard care (25%). EGDT was more cumbersome and costly. Fifteen years had passed and how are the discrepancies to be explained. Should EGDT be abandoned? The meta-analysis authors14 could not determine that their patients were less sick than those of Rivers et al13 The authors also argue that the standard-care group in their analysis was also probably not different from those described in the much smaller study of Rivers et al13 Of course, care standards, training, and improvements could occur over 15 years. General advances in the provision of care for sepsis and shock to the benefits of all patients could explain the differences in these reports. A crucial question is: “Why did the colloid-based studies fail?” Basic physiology coupled with molecular insights at a very local level may provide an answer. Starling's law of the capillaries gave insight into the forces inherent in movement across this barrier. Starling's forces remain largely the same. The colloid solutions may add to the colloid oncotic pressure. Now, we are aware of mechanisms involving intracellular movements and the extracellular endothelial surface, including the glycocalyx (a site of sodium storage). Our data strongly suggest the skin is a source of sodium storage (5 kg of skin filled with it). We know that a host of other surfaces are filled with a glycocalyx, including obviously the endothelium.15 The glycocalyx, a pericellular matrix of glycose-aminoglycans and glycolipids, is located on the apical surface of vascular endothelial cells, which line the lumen. The layer is strongly influenced by sodium.16 At this level, membrane-bound angiotensin converting enzymes, antithrombin-III, lipoprotein lipase, apolipoproteins, growth factors, and chemokines exercise their effects.17 Endothelial damage includes the glycocalyx and profoundly alters microcirculatory fluid shift as described by Starling. The Starling forces remain the same. However, the permeability and reflection coefficient may change. There are already large differences in permeability: the lung allows much albumin to pass the endothelium (a requirement, because hydrostatic pressure is so low), whereas most other tissues hardly let any albumin pass. Thanks to years of careful, randomized, clinical trials, much has been done. Volume presentation remains highly variable, inflexible and sensible.1 We learnt that crystalloid fluids should generally be given, Colloid-administration (although highly appealing physiologically) showed no benefit over crystalloid administration, an exception could be persons with severe liver impairment and bacterial peritonitis,18 the (very attractive) idea that hydroxyethyl starch (HES) could substitute for albumin (should a substitute be necessary) failed in numerous, clearly documented, prospective clinical trials.1 This failure remains a vexing puzzle to intensive-care physicians and requires further research. The most recent report involved post-operative patients receiving 6% of HES, compared to 0.9% of saline.19 The solutions were given in 250 mL boluses according to a hemodynamic algorithm for up to 24 h after their procedures. Outcomes were death and complications at 2, 4 and 12 weeks. The use of HES for volume replacement therapy compared with 0.9% of saline resulted in no significant difference in a composite outcome of death or major post-operative complications within 14 days after surgery. These findings do not support the use of HES for volume replacement therapy in such patients. The key-points emphasized by Finfer et al1 are revealing and can be summarized as follows: (a) Crystalloids (not colloids) should be used in first-line therapy. (b) Albumin-infusions may provide some advantage in certain patients with cirrhosis. (c) Synthetic colloids should not be used. (d) Avoiding volume expansion (causing pulmonary oedema) is surely sensible, although guidelines are elusive. The most recent report supports this conclusion.19 The results of these trials are surprising. We rely on Starling's principle of the capillaries and the importance of oncotic pressure to sustain the vasculature. Could it be that under these circumstances our expected laws of the capillaries are less reliable than we believed? The reasons that the colloid solutions failed have not yet been defined. Why did they; all the experts predicted otherwise? An 80-year-old patient is admitted with generalized peritonitis and a systolic BP of 75 mmHg. I give 1 L of 0.9% of NaCl rapidly and add noradrenaline (NE) to correct hypotension (salvage phase). I do an echocardiogram and insert a central venous catheter to be able to give vasopressors safely and measure CVP, changes in which help to interpret tolerance to fluid (optimization). After a couple of hours, the patient's hemodynamic status stabilizes on a small dose of NE (stabilization), but oedema is clearly present. I reduce fluids, discontinue the NE infusion, and consider a diuretic (de-escalation). The author of this review would do it like that on findings such as history, physical including the and but also on diagnostic The is so that clinicians can in their These will the For instance, with in is a or bedside to assess in and with modern coupled with training, this will The was recently by et assigned adults to receive 0.9% of saline or solution according to the of the to which they were The outcome was major within 30 the composite of replacement therapy, or creatinine of at the of or 30 In those the use of for intravenous fluid to reduce the composite outcome of in-hospital replacement therapy, and compared with the use of and differences are do we have a reason to 0.9% of A similar study was performed in ill patients. In that study, no significant of saline versus crystalloid on days could be However, in the saline group to compared to crystalloid 0.9% of saline solution has an and is by no to the studies that infusions of 0.9% of saline infusions in The 0.9% of saline solutions no but results in In the of 0.9% of saline could be to a However, the difference is the most informative sodium and are is not in the Nevertheless, a randomized trial is patients were randomized to receive 0.9% of saline or lactate as replacement They L of The 0.9% of saline group a of plasma from to a of the results however, the clinical might not be The of could water and the the plasma sodium concentration within the The intracellular is the same in of effective of the to within this cells to or hypertonic with subsequent or The topic has been and to the that the plasma sodium concentration should the of sodium and in total body They this by measurements to determine in the body of the sodium in the body is to and largely in and these tissues could as sodium up or sodium in response to the needs of the This idea has been by and as In persons who sodium can in the without a in the plasma sodium body or extracellular fluid The is a and from the have been by et The has clinical If we reduce the to a clinically we are left sodium body The is a that we all learnt in where is by plasma the has a of and the (b) has a of that plasma is The not pass because not all sodium is in solution sodium Were a to the urine to occur of for an increase in the water must (Figure The of is The pulmonary central nervous and other such as or exercise in The that plasma sodium is a of and total body water and skin storage must Were or other but the extracellular space, water would from the intracellular to the extracellular (Figure With clinical the sodium concentration concentration Were the of sodium to for we would a movement of water from to and sodium concentration would (Figure The would be a in and an increase in If were to diuretic of sodium (Figure the sodium (and would might expand at the of We can (at a that if sodium or intracellular cell are the plasma sodium concentration must be Thus, the plasma sodium concentration is not only by the body but also by the body This body A clinical method to determine the of in the plasma sodium concentration is to rely on In the physiology of the water is the volume of blood plasma that is of water time. An of use is in the of an of However, as we know the determine water clearance, urea other major is Thus, clinically we care If we to know what will to any given patient catheters are we can this of provided that is relatively However, by merely knowing plasma sodium and urine sodium and a we can whether or not the patient's plasma sodium concentration will increase or if the total is greater than the plasma sodium the plasma sodium concentration must If the urine sodium concentration the concentration are less than the plasma sodium the plasma sodium concentration must If we to the effective would 1 urine but for our clinical 1 would it is that if the of is the same as plasma plasma because If the is is positive and is in the and plasma must If the is water is in the and plasma must The exercise the that clinicians must measure sodium and in the urine to effective water urine sodium is not and clinicians have not this Not reviewed is a further analysis of the The physicians must these patients into those who are or As this is not always If the can the patients into those who in either or these authors also a of and that could help in of replacement solutions and their all based on the and adjustments for to the infusions based on data are a strategy could acute the solution is the could be as a recent study For over a we have a to determine where sodium sodium are subject to these include and The of in the body the of sodium so the more This to the sodium that is Patients with are commonly to measure the body sodium are we have a report. This patient with with a plasma sodium concentration of The patient was given and water, which brought his sodium level to We reviewed the patient's and that his body was kg with and kg after of his sodium and to our standards, the changes were with a in sodium concentration in control We have no data on other patients. These patients are commonly the of with the are but the sodium storage an in the The that in of most in water and is a in water balance and body a method to such The techniques to these research are and subsequent data should be We have a We could to The topic is very much but difficult to could measure the but this is no available. has an of and is subject to and we are in the of to But these can also be We clinicians need to what is it we There are is but where is the is not so sodium storage be but the clinical relevance remains uncertain. The is a positive could (a) clinical of at the bedside or if (b) determinations of vascular diagnostic on the cell in addition to (c) therapeutic should be but more in the of randomized (d) up to us to the if any The and to have of this has not yet been by is solely could help us here. The author has no and has to