Fluid resuscitation for the burns patient

Question 21 from the first paper of 2014 and Question 26 from the second paper of 2016 present the candidates with a scenario of a haemodynamically unstable patient with burns to 45% of his body surface. Question 14 from the first paper of 2019 was very similar. Specific attention was paid to the type of fluid, the rationale for that specific fluid, and how the fluid requirements would be estimated. The examiners showed a preference for a balanced isotonic crystalloid, eschewing saline for fear of hyperchloraemic acidosis. The Parkland or modified Brooke formulae were mentioned, the latter being potentially better. According to the italicized comments in the paper, the examiners were unhappy with the lack of rationale for the choice of fluid. Sure, the candidates chose the correct fluid, but they didn't explain why. Some of this chapter hopefully addresses this concern.

The LITFL Burns page from the CCC is an excellent starting point for reading about fluid resuscitation in burns.

Urine output is the goal of resuscitation in burns

Traditionally, urine output is  used to guide fluid management in burns. All the formulae seem to use it as their endpoint. The goal is 0.5-1.0ml/kg/hr. It has been well validated in the literaure.  Paratz et al (2014) performed a thorough systematic review of burns resuscitation endpoints, and found no survival advantage of haemodynamic monitoring over hourly urine output, at least among well-designed studies.

Comparison of formulae for estimation of fluid resuscitation requirements in burns

For convenience, here is a quick reference table of the formulae which will be discussed below.
Only the adult formulae are listed.

The Parkland formula

• The "modified" Parkland formula is quoted below:
• First 24 hrs:
  • Hartmanns (or Ringer's Lactate)
  • 4 ml per kg per percent BSA (counting full-thickness burns only)
  • First half of this volume is given over the first 8 hours
  • No colloid in the first 24 hours
• Next 24 hours
  • No more crystalloid
  • Albumin infusion: .0.3-1.0ml.kg/%BSA/16/hr
    (no specific albumin concentration is being targeted)
  • Maintain a urinary output of 0.5–1 ml/hour

Decades after its introduction, the Parkland formula is still the most widely quoted resuscitation protocol for burns.  The original Parkland formula was introduced by Baxter and Shires in 1968. It has since undergone various modifications ( eg. the volume is now 4ml/kg/%, whereas the original called for 3.7-4.3ml/kg/%). If you use urine output to guide your therapy,  the Parkland formula seems to underestimate the fluid requirements (Blumetti et al, 2008).

The Brooke formula

• The "modified" Brooke formula:
• First 24 hours:
  • Hartmanns (or Ringer's lactate)
  • No colloids.
  • 2 ml/kg/% BSA
• Next 24 hours:
  • Albumin infusion at 0.3–0.5 ml/kg/% burn
  • No more crystalloid

The Evans formula

• First 24 hours:
  • Crystalloid 1 ml/kg/% BSA
  • PLUS albumin at 1 ml/kg/% burn
  • PLUS 2000 ml of 5% dextrose
• Next 24 hours:
  • Crystalloid at 0.5 ml/kg/% burn,
  • PLUS albumin at 0.5 ml/kg/% burn
  • PLUS 2000 ml of 5% dextrose

The Monafo formula

• First 24 hours:
  • Weird sodium-rich solution:
  • 250mmol/L Na⁺
  • 150 mmol/L lactate
  • 100mmol/L chloride
  • No specific volume prescription
• Next 24 hours
  • Same weird solution
  • Titrate with one third saline (i.e. NaCl 0.3%) to maintain urine output

Shriner's formula (paediatric)

• For older children:
  • Lactated Ringer’s solution 4 ml/kg/% BSA
  • PLUS 1500 ml/m²
  • Half of the total volume over 8 hours, rest of the total volume during the following 16 hours
• For younger children:
  • Lactated Ringer’s solution mixed with 50 mmol NaHCO_3, 4 ml/kg/% burn +1500 ml/m² total, in the first 8 hours
  • Lactated Ringer’s solution without the bicarbonate in the second 8 hours
  • Lactated Ringer’s solution with isooncotic albumin in the third 8 hours

Galveston's formula (paediatric)

• First 24 hours:
  • Lactated Ringer’s solution  5000 ml/m² burn
  • PLUS 2000 ml/m²
  • Half of total volume over 8 hours, rest of the total volume in 16 hours

The influence of inhalation injury on fluid resuscitation requirements

• A famous study by Naver et al (1985) demonstrated that patients with smoke inhalation injury and airway burns require a larger volume of fluid resuscitation.
• The volume is increased from 35% up to 65%. Saffle (2007) discusses the evidence for this; it is compelling, in spite of the fact that most of the studies are retrospective and paediatric.

The phenomenon of "fluid creep"

Significantly more fluid is given to burns patients  then is predicted by any formula (Mitra et al, 2006). When using adequate urine output as the marker of a "well resuscitated" patient, one finds that most end up receiving around 6ml/kg/hr - in total, up to 250ml/kg , or  25% of their body weight in fluid (Blumetti et al, 2008) An excellent article by Jeffrey Saffle (2007) discusses this phenomenon in detail.

In short, it seems people have arrived at the conclusion that fluid is good, and therefore more fluid is more good. Or something. This is not without consequences. Complications of excessive fluid resuscitation are predictable, and include facial swelling, abdominal compartment syndrome, and compartment syndrome of the extremities. The way to overcome this is to use less volume to meet the same haemodynamic end-points; colloid is recommended in the formulae for this very reason.

The whole concept of torrential resuscitation has been questioned. Arlati et al (2007) have suggested that this dogma is without foundation; their "permissive hypovolemia" protocol (only 3ml/kg/%) ended up producing better organ system function. In recent articles reviewing burns resuscitation (eg. Bittner et al, 2015), the formulae are now being described as "helpful guidelines", to be acknowledged but not followed prescriptively. Sensibly, modern authors call for an individually tailored approach, guided by the monitoring of physiological responses.

Choice of replacement fluid

Colloid resuscitation in burns

As discussed above, the use of colloid in burns resuscitation is a reaction to the adverse effects of massive fluid resuscitation. It is recommended as a "crystalloid-sparing agent"; the objective is to reduce the total fluid requirement while still achieving a urine output of 0.5-1.0ml/hr, or whatever endpoint you decided on. Most of the abovementioned formulae do not specify which colloid you should use, except the modified Parkland and Shriner's formula which call for 5% albumin specifically. In the spirit of controlling resuscitation volume, one could think that 20% concentrated albumin would be the better choice.

After their acute resuscitation, burns patients go on to develop hypoalbuminaemia - likely as a part of their stress response. Some centres make albumin replacement a part of their routine care. For instance, the Alfred protocol calls for albumin infusion (4%, at 100ml/hr) to keep the serum albumin level around 25g/L for the first 4-5 days after a burn. There is probably no harm in it. We can recall that the SAFE study demonstrated an equivalence for albumin and saline in terms of adverse effects and mortality, although notably burns patients were excluded. . However, there does not seem to be very much evidence to support the practice of routine albumin replacement (i.e. non-reuscitative pursuit of specific serum albumin levels). For instance, Melinyshyn et al (2013) were unable to find any difference in outcomes or organ dysfunction scores between retrospectively audited groups of patients whose albumin levels were 20g/L and 25g/L in the first 30 days after the burn. Similarly, a 2006 RCT by Cooper et al (ALBUR) could not find any improvement in organ system dysfunction scores when 5% albumin was added to Ringer's lactate in the first 14 days of resuscitation.

Crystalloid resuscitation in burns

Most of the abovementioned formulae suggest Ringer's lactate, which is very similar (but not identical) to Hartmanns solution found in Australian hospitals. In brief, Ringer's has the same ingredients by in a slightly lower concentration: the osmolality of Ringers is 273 mOsm/kg, whereas Hartmanns is 279 mOsm/kg. The difference is trivial and the fluids can be used interchangeably. 

The use of balanced crystalloid seems sensible in this context. The fluid lost by burns patients is similar in its concentration to the extracellular fluid, and it would make sense to replace it with something like Plasmalyte or Hartmanns. Generally speaking, the authors all recommend against saline, given the adverse effects of hyperchloremic acidosis. A retrospective case control study by Walker et al (2001) were able to demonstrate a significant difference in acid-base balance, strongly favouring the balanced solutions.

Hypertonic resuscitation fluid

The theory behind this is the reduction in total fluid requirements and improvement in tissue oedema. William Monafo (1971) was able to achieve haemodynamic goals with a smaller volume, using a solution with 250mmol/L of sodium. However, subsequently studies had revealed that there was no significant difference in total  positive fluid balance at 48 hours, with or without hypertonic sodium lactate. Furthermore, Huang et al (1995) found the hypertonic solutions were associated with a fourfold increase in the risk of renal failure and a twofold increase in the risk of death. In their answer to Question 21 from the first paper of 2014, the college examiners remarked that "use of hypertonic saline does not provide better outcomes than isotonic saline".

On the other hand, burns patients do tend to become hyponatremic. Do we treat this? This is still a topic of debate, and many authorities maintain an attachment to hypertonic resuscitation. For instance, the Alfred protocol calls for 3% saline to be given at 200-300ml/hr, with the goal of maintaining a serum sodium of 145-150mmol/L. That document is not referenced, but one might surmise that the Alfredians may be using saline for its immunomodulatory effect. Hypertonic saline is said to be anti-inflammatory; however, at this stage it has not been demonstrated in the severely burned rat model (Ye-Xiang et al, 2013).