DYNAMIC ANALYSIS OF EXTRACELLULAR FLUID EXCHANGES IN THE BURN PATIENT DURING CONTINUOUS ARTERIOVENOUS HAEMOFILTRATION

Gòmez-Cia T. (1,2) Ortega-Martinez J.I.,(1,2) Roa L .(2)

(1) Burn Unit, Department of Plastic and Reconstructive Surgery, Virgen del Roc[o University Hospital, Seville, Spain
(2) Bioengineering Team, University of Seville

SUMMARY. An algorithm is presented for the analysis of effects of continuous arteriovenous haernofiltration (CAVH) on critical burn patients. The objective of CAVH is the controlled elimination of vascular fluids and low molecular weight solutes. Th s method requires rigorous control of fluid intake and elimination. The algorithm, calculated on the basis of easily available daily varia venous haemotocrit and plasma protein concentration), can show us the behaviour in time of variables such as blood volume and fluid shifts between plasmatic and interstitial compartments. These variables are indicative of the real state work also presents the results obtained using the algorithm in a series of burn patients subjected to CAV1-1.

Conditions of fluid overload and/or acute renal failure (ARF) not responding to conventional treatment are not infrequent in the critical bum patient. Continuous arteriovenous haernofiltration (CAVH) has shown its clinical efficiency in the extracorporal clearance of fluids and dissolved molecules in plasma not bound to proteins and inferior to 50,000 daltons molecular weight.
In bum patients the main objectives of CAVH are:

1. Removal of volume overloads, with obvious and immediate benefit for heart and lung function
2. Clearance of substances accumulated in the plasma, mainly urea and creatine, reinstating kidney function
3. Treatment of metabolic disturbations such as acidosis or alkalosis
4. Other less common indications

CAVH is a low-flow, low-pressure mechanism, compared to other conventional methods of dialysis. The patient's own blood pressure and the negative pressure of the fluid obtained by ultrafiltration - known as the siphon effect - operate in conjunction with filtration. Plasma protein oncotic pressure opposes the fluid leaking through the semipermeable membrane into the haemofilter (Fig. 1).
Owing to the need to maintain, on an hourly basis, a fluid balance that prevents the risk of hypovolaemia, it is necessary to monitor patients subjected to CAVE The strategy to be followed will vary according to the pathology of each individual patient. Negative fluid balances are therefore established in volume overload situations by replacing only a greater or lesser part of the fluid lost through ultrafiltration. Fluid balance is however usually compensated in acute renal failure situations, and we remove urea, creatinine and other dissolved molecules in les (fluid balance,olurne, interstitial f the system.

The effects of CAVH on the interstiti~l and plasmatic compartments in burn patients are analysed in this work. With this aim in mind, we have created ]an algorithm to analyse the behaviour in time of several variables that cannot be determined in daily clinical practice. In this way, starting from venous haernotocrit values,~ plasma protein concentration and timetable fluid balance~ we can follow blood volume, plasmatic and interstitial variations more closely, and thus know the dynamic behaviour of fluid shifts between vascular and interstitial~ spaces during CAVH.

The algorithm is based on the equations that control the exchange of fluids at the capillary membrane level in bum patients.' It has a modification to allow for fluid losses through ultrafiltration:

dPV/dt = FI - DI - FS + LF - UF
dIV/dt = FS - LF - LEV

where:

PV = plasma volume (ml)
FI = fluid input (ml/min)
DI = diuresis (ml/min)
FS = fluid shift at capillary level (ml/min)
LF = lymphatic flow (ml/min)
UF = volume obtained by ultrafiltration (ml/min)
IV = interstitial volume (mllmin)
FEV = evaporation fluid output (ml/rnin)

For the calculation of these equations we consider one minute as the integratiOn time (At). We use the following concept of haematocrit:

HQt + At) = BV(t + At) / [PV(t + At) + JIV(t + At)] where:

We can change [1] to:
dPV(t) / dt = FI(t) - Dl(t) - FS(t) + LF(t) - UF(t)

Using Euler's discretization method, the previous equation changes to:
PV(t + At) = PV(t) + At x [FI(t) - DIffi - UF(t) + LF(t) - FS(t)l [21

Likewise:
IV(t + At) = IV(t) + At x [FS(t) - FEV(t) - LF(t) [31

We can then continue the algorithm construction, knowing' that:

IP(t) f (IV(t))
FS(t) f (IP(t))

where:

IP = interstitial compartment pressure
Returning to [21, we can now find the values of LF and FS:

LF(t) - FS(t) = [PV(t + At) - PV(t)j /At - Fl(t) + Dl(t) + UF(t) = NF(t)

where:

NF = net fluid shift between plasmatic and interstitial compartments at the capillary level
Thus, NF(t) = LF(t) - FS(t)

Likewise, [3] changes to:
IV(t + At) = IV(t) + At x [-NF(t) -FEV(t)]

If we analyse the previous equatios:

1. HQ FI, DI and UF can be measured; and
2. LF and FEV can be calculated.

These can therefore now be solved.
The algorithm determine the dynamic the fluid shift at capillary 'Vel, as w and intestinal volumes, which all incr treated with CAVH. An advanced version is accessible tists interested in the mathematical a rithm.
Between February 1992 and February 1994 eight patients admitted to the Burn Unit in our Hospital required, CAVE The indications for CAVH treatment were fluid overload and/or acute renal failure, with refractory responses to conventional therapy, in critical bum patients. Table I represents the parameters measured and the schedule in patients subjected to CAVH therapy.

The clinical characteristics of the eight patients treated with CAVH are shown in Table IL The age range was from 16 to 65 years, with a burned body surface area between 55 and 90%. All the patients had suffered flame burns.
All the patients were resuscitated with colloids on admission to the Burns Unit (B.E.T. references: Roa & Gòmez-Cfal and G6mez-Cfa & Roal'). Topical treatment was carried out in all cases with 1% silver sulphadiazine (Flammazineo').

Two patients showed volume overload which did not respond to treatment with diuretics. It was decided to include these patients in our CAVH treatment programme with the aim of eliminating the excess of administered liquids (more than 200% of the normal intestinal Volume calculated in both cases) during the first 48 hours ppst-burn.
The six remaining patients were treated with CAVH because they showed acute renal failure in progress.
The time of initiation of treatment, its duration and the time spent between the patient's entry and death (when that occurred) are shown in Table III.
The results obtained in two representative patients in each group for which CAVH is indicated (fluid overload and acute renal failure) are now described.

A 21-year-old male suffered a 90% BSA flame bum. He was resuscitated with B.E.T. therapy, but renal function did not respond either to crystalloid overload or to a mixed solution of crystalloids and colloids. Administration of diuretic drugs also did not achieve the desired effect. The patient, 57 hours post-bum, thus showed a very positive hydric balance, with clinical repercussions on pulmonary and cardiac function, as well as on oxygen transport to the tissues, secondary to tissue oedema.
The excess fluid was eliminated by CAYH in approximately 48 hours. The clinical results were evident: the oedema disappeared almost entirely and pulmonary function improved considerably, to the point that oxygenotherapy was not required at the end of treatment.

Fig. 3 shows fluid infusion during the resuscitation phase, together with the diuresis obtained (an average of 0.55 ml/kg body weight per hour until 57 hours post-bum). During CAVH treatment, volume loss was replaced by lactated Ringer in the necessary quantities to obtain a negative balance of six litres in two days. Diuresis increased at the end of the treatment (over 1 ml/kg body weight per hour).

Fig. 3 - Case 1: realization of negative fluid balance during CAVH, by addition of perspiratory and burn exudate to the diuresis and ultrafiltration obtained by CAVH (Fl = fluid input; DI = diuresis; UF = ultrafiltered liquid). Fig. 4 - Case 1: increase of plasma protein concentration (PPC) after initiation of CAVH.

Our algorithm enables us to obi haviour of variables not available in ( but indicative of the patient's real s volume (Fig. 5) - from parameters measured in this kind of patient, e.g and plasma protein concentration (F can observe how the interstitial volun of the second day post-burn had inc 200% of its initial value; CAVH pi the interstitial volume to more not shows the behaviour in time of the ne plasmatic to the interstitial compartr algorithm.ain the dynamic beaily clinical practice tate - such as blood that are commonly venous haematocrit g. 4). In Fig. 5, we ie variable at the end reased to more than ogressively reduced mal figures. Fig. 6 fluid shift from the ient obtained by the ment. NLS increases immediately after the bi high 48 hours post-burn. After initiation of CA rected. NLS becomes close to zero after a few of treatment (after about 100 hours) NLS beca entering the interstitial compartment is greah plasmatic compartment.

Fig. 5 - Case 1: values obtained by applyi patient. A decrease of the interstitial volume associated to an increase of blood volume (BL the process of elimination. On termination c dency is re-established. Fig. 6 - Case 1: evolution of net fluid shift

Case 2

A 28-year-old woman with 62% BSA flame burns developed acute renal failure, probably secondary to the deficient resuscitation volume she received between the time of the accident and her admission to our Bum Unit (over four hours). The patient's burn shock had to be treated with dopamine because of decreasing systolic and diastolic arterial pressure, in spite of initial aggressive resuscitation on admission.
Fig. 7 shows renal response to treatment during the resuscitation period. Diuresis was initially less than 0.35 ml/kg body weight per hour, despite fluid infusions and diuretic drug administration. In this case - an acute renal failure syndrome without volume overload - the strategy followed consisted in the maintenance of an adjusted fluid balance by making up for the liquid removed through ultrafiltration with CAVH.

Fig. 7 - Case 2: patient with acute renal failure. Diuresis (DI) is presented in relation to the phase of CAVIL Two different phases can be observed: oliguric and polyuric. CAVII treatment was over when the polyuric phase began. Removal of liquid was not necessary as there was no oedema. A zero hydric balance was achieved in such a way that fluid input (F1) equalled total output: ultrafiltrate (UF), diuresis (D1) and negligible losses. Fig. 8 - Case 2: CAVH stabilized and then decreased urea and creatinine blood levels in the first days post-burn. It was not necessary to use also haemodiafiltration. The increase of urea and creatinine after interruption of CAVH was countered by the polyuric phase of acute renal failure.

In the initial period, urea and blood creatine rose to values of 88 and 3.3 mg/dl respectively (Fig. 8). With CAVH, we managed to maintain an upward evolution and to decrease both parameters to pratically normal figures (67 and 2 rfig/dl respectively). The end of the first period of haernofiltration (54 hours) coincided with the beginning of the polyuric phase of acute renal failure (Fig. 9). For this reason we decided to postpone CAVH treatment, after clotting of a haernofilter. The great quantities of urine removed from the patient in the polyuric phase of acute renal failure (more than 300 ml/h in several determinations) gradually led to normal urea and plasma creatinine values.

Fig. 9 - Case 2: Acute renal failure phases. In the oligoanuric phase, CAVH maintained urea and creatinine at normal levels. Treatment was suspended when the polyuric phase began. Fig. 10 - Case 2: the algorithm shows the small rise in interstitial volume (INT VOL) and the limited influence of CAVH on these values. In this patient, the strategy consisted in the elimination of harmful solutes and the maintenance of fluid volume in the organism.

In Fig. 10, thanks to the algorithm, we can analyse the limited influence of CAVH on the interstitial volume in this case. This variable was maintained at practically normal values throughout the CAVH period.

The initial idea of continuous arteriovenous haernofiltration is closely related to the search for a treatment for patients who are haemodynamically unstable with volume overloads that do not respond to diuretic drugs.' This requirement has been confirmed by many authors with regard to critical patients.
Application of CAVH to burns patients with acute renal failure was described by Lehinkoster et al." Paradiso` described CAV14 treatment in bum patients with volume overload. A dynamic analysis of fluid exchanges between plasmatic and interstitial behaviour during the CAV11 was not carried out in any of these papers.
On the other hand, the differences between CAVH as an extrarenal cleansing procedure and other conventional methods - haemodialysis and peritoneal dialysis - are well established.`,` The main advantage of CAVI1 over conventional haemodialysis is the possibility of its application to patients with haemodynamic instability, a frequent circumstance in critical bum patients; in a slow, continuous manner (even up to 8 or 12 litres, per day), without causing sharp alterations in the patient's circulating volume. Compared to peritoneal dialysis, CAVH avoids the appearance of respiratory insufficiency due to diaphragm excursion restriction, apart from other secondary effects. Table III indicates the disadvantages of these two methods of dialysis which do not occur with CAVE.
In our series, CAVI1 treatment of patients with volume overload produced an obvious improvement in the clinical situation. In Case 1, the removal of the extra volume

RESUME. Les auteurs présente un algorithme pour analyser les effets improved pulmonary function, gra flow and oxygen concentration nece mal arterial oxygen saturation. The ried out before and after CAVH sho intralung water volume (Fig. 10). Acute renal failure patients were with CAVE They maintained an a while the renaffunction recuperate tion associated to CAVH increased t technique in the clearance of used the three patients who required it. The application of our algorithm the behaviour of indicative variables iour in the critical bum patient. On the basis of accessible var offers us a dynamic vision of varia volume, interstitial volume. and fl plasmatic and interstitial comp e tionship was described in 1986~mith of fluid and protein shift in the bu shock phase.' It is in this line of worl rithm is inserted.
The clinical findings, in dditiually decreasing the sary to maintain norhest radiography cars the decrease in the adequately stabilized equate internal state The haemodiafiltrae effectiveness of the etabolic products, in enables us to analyse f extracellular behaviables, the algorithm es such as plasmatic . shift between the ts. This kind of relaegard to the analysis ed patient during the that the present algo a lion the data obtained by means of the algorithm, show the efiliveness of CAV14 in critical burned patients with volume 0 failure refractory to conventional CAVH itself did not appear to reduc patients, although the series analysed to draw conclusions in this sense. analysed only one survived the_heat quences. We have not observed any related to the treatment.