Water, taken in moderation, cannot hurt anybody (Mark Twain 1835–1910)

Few areas of clinical practice arouse such strong feeling as intra-venous fluid therapy. For many years clinicians have hotly debated both the type and dose of intra-venous fluids they should use. In many cases fluid therapy is required simply to meet maintenance requirements. These can be estimated from body mass with acceptable accuracy and there is general agreement regarding the optimal approach. Greater controversy surrounds the treatment of patients with fluid deficits. Trauma, major surgery and severe sepsis are common causes of life threatening reductions in plasma volume. In some cases fluid deficit may be reliably estimated, but this is not usually possible and clinicians must use more subjective criteria. Consequently, such patients are at risk of the adverse effects of either inadequate or excessive fluid administration. Some commentators suggest the existence of two schools of thought, one advocating a ‘wet’ approach to fluid therapy and the other a ‘dry’ approach (Bellamy, 2006). Whilst in reality there is no evidence of such a dichotomy, the concept serves to illustrate just how little we know about determining the optimal dose of intra-venous fluid. It is unlikely that two quite different approaches to fluid therapy are equally effective but sadly plausible that both are equally ineffective. In truth, clinicians all strive to achieve the same objective: adequate plasma volume without tissue oedema.

At the heart of this problem lies the systemic inflammatory response to tissue injury. The aetiology of this phenomenon may vary but in every case, endothelial dysfunction results in a range of cardiovascular abnormalities including myocardial dysfunction, vasodilatation, abnormalities of microvascular flow and increased vascular permeability (Hotchkiss & Karl, 2003; Jhanji et al. 2009). The direct effects of the systemic inflammatory response on organ function are currently unavoidable and the physician must focus on maintaining homeostasis by artificial means such as mechanical ventilation and renal dialysis. Whilst abnormalities of vascular tone and myocardial contractility may be promptly corrected with vasoactive drug therapies, increased capillary permeability may persist for many hours or days, leading to a sustained reduction in plasma volume, cardiac output and hence tissue blood flow. The challenge is to ensure that organ injury is not further compounded by the effects of inadequate plasma volume. Clearly, the early administration of intra-venous fluid is essential. The use of stroke volume measurements as an end-point for the dosing of intra-venous fluid may be the most reliable tool but this approach has yet to gain widespread acceptance (Bellamy, 2006). Regardless of the approach chosen, in the presence of abnormal vascular permeability, adequate fluid resuscitation comes with the inevitable cost of tissue oedema. This in itself will compromise organ function over the days which follow. As a general rule, the patient will not recover before this oedema is corrected.

Whilst clinical academics strive to identify the best way of estimating fluid requirements, the solution perhaps lies elsewhere. Rather than working out how to treat the consequences of the problem, perhaps we should focus on the root cause: abnormal capillary permeability. In this issue of The Journal of Physiology, an interesting paper illustrates the continued importance of basic physiological science to our understanding of the pathophysiology of disease. Curry and colleagues have investigated the effects of atrial natriuretic peptide (ANP) on the transfer of both water and plasma proteins across the endothelial barrier (Curry et al. 2010). The regulation of vascular permeability by ANP appears to play a central role in control of plasma volume. Perhaps a key aspect is that the ANP-induced decrease in plasma volume relates to enhanced clearance of both water and plasma proteins. The increased capillary permeability which is characteristic of systemic inflammation results in significant loss of water and plasma proteins from the circulation. This partly explains low levels of plasma albumin in critical illness, although impaired synthetic function also plays a role. Clearly any attempt to reduce or prevent the loss of plasma volume resulting from the systemic inflammatory response must address the loss of both protein and water.

Sepsis is associated with abnormal plasma concentrations of any number of biomarkers. It is therefore no surprise to learn that increased ANP levels are a feature of sepsis associated with poor survival (Hartemink et al. 2001; Morgenthaler et al. 2005). However, whilst the investigation of natriuretic peptides has focussed mainly on use in diagnosis and prognostication, there has been less investigation of the physiological significance of raised ANP in critical illness. Could this peptide hormone play a key role in the loss of plasma volume so characteristic of trauma, major surgery and severe sepsis? More importantly, could specific ANP receptor antagonists regulate the loss of plasma volume in such patients? Clinical sepsis research has proved a graveyard for novel therapies that showed promise in bench studies but which did not translate into improved survival. These failings may relate to the difficulty in stemming an already advanced inflammatory cascade; modulation of a single step in such a complex pathway is unlikely to be effective. The need to administer specific sepsis therapies at a very early stage greatly limits their clinical efficacy and utility. Importantly, the use of ANP antagonists would not be an attempt to stem the tide of systemic inflammation but to control a downstream consequence. The persistent nature of increased capillary permeability means this remains a treatment target for a longer time frame than other features of the systemic inflammatory response. Further research must establish the role of ANP in the regulation of plasma volume in specific disease models before then evaluating the effects of specific antagonists. If such studies showed promise, there would be a sound rationale for investigating the effects of ANP antagonists in critically ill patients. We must be cautious about the potential for adverse effects of such a treatment approach. For example, ANP may have possible protective effects in ischaemia–reperfusion injury (Witthaut, 2004). However, abnormal vascular permeability is of profound importance to the clinical course of many tens of thousands of critically ill patients each year. A novel solution to this problem is a very worthwhile pursuit.

Conflict of interest statement

There are no conflicts.