Introduction

The release of proinflammatory cytokines and catecholamines by tissues damaged by a significant burn injury may induce a hypermetabolic state, raising the patient’s basal temperature above the normal level [1]. This, together with episodes of sepsis, may lead to periods of hyperpyrexia during which the patient’s temperature may reach 40 °C or above. Sustained hyperpyrexia leads to protein denaturation, disruption of enzyme systems, cellular injury and death.

Elevated temperature has been shown to be independently associated with longer Intensive Care Unit (ICU) and hospital stay and increased ICU mortality [2]. Strategies for temperature regulation include pharmacological agents, such as paracetamol and non-steroidal anti-inflammatory drugs (NSAIDs), external cooling by ambient temperature regulation, external cooling devices or gel pads and invasive techniques [3]. These include intravascular heat-exchange systems (Coolgard®, Zoll USA) and extracorporeal renal replacement techniques.

Several different extracorporeal techniques exist for renal replacement therapy including haemodialysis, haemofiltration and haemodiafiltration. Haemodialysis relies primarily on the diffusion of molecules from the blood through a semipermeable membrane into the diasylate, which is travelling in a counter current direction. Haemofiltration works by convection of molecules through a highly permeable membrane. In this technique, the waste material is removed as an ultrafiltrate and the fluid can be replaced with a sterile reinfusion if needed. Haemodifiltration combines the techniques of convection and diffusion by using a highly permeable membrane and a counter current diasylate. The renal replacement circuit can run with either an arterial input and venous output (arteriovenous) or a venous input and output (veno-venous) and be intermittent or continuous.

As described above, heamofiltration and haemodiafiltration (HDF) are procedures primarily used as renal replacement therapy [4, 5], but recent research has demonstrated its role in severe sepsis [6, 7] multiorgan failure [8, 9] and hyperpyrexia [10]. These techniques are believed to be effective by removal of the proinflammatory mediators, cytokines and catecholamines [10, 11]. Through the same mechanism, continuous veno-venous heamofiltration has been used successfully for temperature regulation in heat stroke patients with the added benefit of removing circulating myoglobin and preventing rhabdomyolysis related kidney disease [12]. Haemodiafiltration (HDF) combines convection and diffusion and therefore optimises the clearance of molecules of differing molecular weights [13].

The use of HDF in the burns unit is not new and its role in managing fluid overload, hypernatraemia and acute kidney injury is well documented [14,15,16,17,18,19]. The benefits of HDF include a reduced vasopressor requirement, improved lung function and reduced mortality rate [20].

The process of continuous veno-venous heamodiafiltration (CVVHDF) involves the removal of blood from the patient through a double-lumen central venous line. This blood then passes through the haemodialyser which contains a semipermeable membrane that allows diffusion of small molecules from the blood into the diasylate. The treated blood is then returned to the patient. CVVHDF circuits have a limited lifespan and cease to function when they become clotted or obstructed. On each occasion that this occurs in our practice, the requirement for continued treatment is reassessed. Treatment is discontinued when either renal failure was seen to be resolving, temperature fell below 40 °C or further treatment was considered futile.

CVVHDF protocol for hyperyrexia

We developed a protocol for the use of CVVHDF for hyperpyrexia in patients managed at our burns centre in 2001, when CVVHDF was being used routinely in the General ICU to manage renal failure. We consider using CVVHDF if a patient’s temperature is elevated to 40 °C or higher for 6 h consecutively or to more than 41 °C for 2 h consecutively when conventional non-invasive methods of thermoregulation have failed to reduce the temperature (including external cooling, paracetamol and NSAIDs). CVVHDF is also considered when there is refractory acidosis (pH < 7.2) or incipient renal failure, with persistent hyperkalaemia >7 mmol/L or anuria with fluid overload.

The aim of this study was to investigate: (1) the number of patients treated with CVVHDF over a set year period, (2) the indications for treatment in these patients, (3) the impact of CVVHDF on hyperpyrexia and (4) the survival of patients treated with CVVHDF.

Materials and method

A retrospective analysis was carried out of all patients admitted to the burns ICU between 2005 and 2012. Patients were identified using electronic burns and renal databases. Data was retrieved retrospectively from patient’s medical records, detailing their demographic information, indication for CVVHDF, observations, temperature changes and eventual outcome. We focused on the temperature change within 3 h as we felt this most closely reflected the impact of CVVHDF. Data from the group who received CVVHDF was compared with that from other patients treated on the burns ICU who did not receive CVVHDF. Statistical analysis was carried out using Graphpad® software.

Standard burn care

All patients received standard burn care including immediate assessment in theatre including burns assessment, essential surgery if needed (escharotomies) and bronchoscopy for inhalational injury. If inhalational injury is found then treatment including bronchial lavage is implemented. The patients are admitted to a burns intensive care unit (ITU) and once stable (within 24 h) undergo early burns excision and grafting with autograft or allograft. All patients receive early nutritional support with NG or NJ feeding and elemental supplementation (Zinc, selenium etc.). We do not routinely use beta-blockers to control the hypermetabolic response.

Results

Comparing patients who received CVVHDF and those who did not

Five hundred and seventy patients were admitted to the burns ICU during the 7-year period between 2005 and 2012. Of these, 61 were treated with CVVHDF (11%). Demographic data from these two groups are compared in Table 1. The patients treated with CVVHDF had significantly larger burns, were significantly more likely to have had a smoke inhalational injury and had a significantly higher modified Baux score than those who did not receive CVVHDF. They also spent significantly longer receiving ventilator support and spent longer in burns ICU. The groups were compared using either Chi-squared or unpaired t tests, and a p value of <0.05 was accepted as significant. The mortality among patients who received CVVHDF was higher than those who did not.

Table 1 Comparison of patients who received CVVHDF versus those who did not
Full size table

Indication for CVVHDF

The majority of patients received CVVHDF for hyperpyrexia alone (54%), 16% received CVVHDF for renal failure or acidosis and the remaining 30% was for multiple factors (acidosis, renal failure, fluid overload and hyperpyrexia).

Length of CVVHDF

The average length of time CVVHDF was required was 7.5 days for renal failure, acidosis or multiple reasons (range 1–55 days) and 8.5 days (range <24 h–57 days) for hyperpyrexia.

The effect of CVVHDF on temperature in those patients (n = 32) who received CVVHDF for hyperpyrexia control alone

Analysis of the mean temperature in the 3 h before and after the initiation of CVVHDF demonstrated a significant reduction in temperature (paired t test, p < 0.0001, Fig. 1a). Figure 1b shows the individual reduction in temperature in the 3 h following the commencement of CVVHDF. The average (mean) temperature reduction was 0.8 °C as indicated by the horizontal line.

Fig. 1
figure 1

The effect of CVVHDF on temperature. a Temperature changes for the 3 h before and after the initiation of CVVHDF and the temperature at the end of treatment. b Average reduction in temperature in the 3 h following the initiation of CVVHDF

Full size image

The effect of CVVHDF for hyperpyrexia on survival

The cumulative predicted mortality, using modified Baux score, for the group of patients who received CVVHDF for hyperpyrexia was seven patients, however, only two patients died (Fig. 2). This equates to a standardised mortality ratio of 0.29.

Fig. 2
figure 2

Predicted and actual mortality for patients who received CVVHDF for hyperpyrexia

Full size image

Discussion

Continuous veno-venous haemodiafiltration is a technique that has proven to be of value during the treatment of the most severely burn-injured patients in our burns unit. The mortality of this group is understandably high; however, without CVVHDF their mortality is predicted to be higher. In this study, we have focused on the role of CVVHDF for hyperpyrexia (temp >40 °C) alone. We appreciate that ideally we would compare hyperpyrexial patients who received CVVHDF to those who did not however CVVHDF was the standard treatment for all patients with hyperpyrexia.

CVVHDF is an effective method for controlling temperature even when conventional techniques have failed. We have demonstrated that within 3 h of commencement of CVVHDF temperature is significantly reduced, with all patients experiencing a drop to less than 40 °C. The underlying mechanism for this temperature reduction is probably explained by the extracorporal removal of proinflammatory cytokines and catecholamines [21]. The use of haemofiltration for SIRS and sepsis has increased during the past two decades in the general ITU population. Clinical studies have demonstrated that extracorporeal elimination of circulating cytokines, including tumour necrosis factor alpha, interleukin 1β and interleukin 6, downregulates the excessive inflammatory response that leads to multiorgan failure in sepsis [11]. Our technique of CVVHDF focuses on dialysis and relies more on diffusion than convection. In reality, filtration contributes very little because we do not remove or replace large volumes of fluid. It is probable that a high proportion of the cytokine removal takes place by adherence to the filter membranes rather than passage through them. There is also likely to be a mechanical cooling benefit associated with the blood passing through the diafiltration machine.

We have postulated a survival benefit related to the use of CVVHDF for hyperpyrexia, however in reality the benefit may stem from multiple factors. The predicted mortality we have used was based on the Baux score and the survival reflects their entire burns care (including early surgery, antibiotics and organ support). Progress in burns care has demonstrated that the Baux score may no longer be the most accurate method of predicting outcome, however, no better methods exist as yet [22].

There are risks associated with using CVVHDF. Patients require central venous access with its associated risks of deep vein thrombosis, line infection, line failure and bleeding; however, there were no significant complications in our group. The patient also requires anticoagulation, which increases their risk of bleeding and can complicate surgical treatment of their burns. The risks of this treatment have to be weighed against its benefits.

To further investigate the potential benefit of CVVHDF on hyperpyrexia, a randomised controlled trial, possibly as a multicentre trial to increase numbers and feasibility, should be undertaken.