Haemoperfusion: mechanisms, indications, and complications

Haemoperfusion as a stand-alone topic has not been asked about in any previous fellowship SAQ papers.  However, it had come up in the toxocology-related Question 19.1 from the first paper of 2009, which asked the candiates to "list the relevant physical features of hemodialysis and hemoperfusion filters"  with specific reference to "intoxications". It is not inconceivable that this topic might come up again in a question which might go something like "List 10 drugs which are susceptible to clearance by haemoperfusion therapy", or "Decribe the mechanisms of haemoperfusion therapy, and discuss the complications of its use". Or something much worse.

 The following brief summaries are based largely on the relevant chapters from the 2nd edition of Critical Care Nephrology by Ronco, Bellomo and Kellum.

Mechanisms of haemoperfusion

Haemoperfusion is defined as "the extracorporeal procedure in which the anticoagulated patient’s blood passes through a volume of adsorbent material"

Adsorption is the deposition of molecules on the surface of a medium, rather than within it (which would be aBsorption).

The adsoprtion medium needs to have several favourable properies:

  • High selectivity/ affinity for noxious solutes
  • Rapid adsorption
  • High capacity for adsorption (i.e. large surface area)
    • Modern sorbents have a surface area of around 1000m2 per gram of sorbent
  • Chemical and thermal stability, including low water solubility
  • Good structural integrity (i.e. non-crumbly)
  • Good biocompatibility (to prevent inflammatory response or anaphylaxis)

Practical administration of haemoperfusion

Relevant features to make note of :

  • A normal CRRT machine can be used
  • Blood flow rate is about 250-300ml/min
  • Anticoagulation requirements are similar to normal CRRT
  • A session usually lasts about 4 hours
  • At the end of the 4 hours the cartidge is depleted and must be discarded.

Complications of haemoperfusion

The table presented here is modelled on the table representing the complications of renal replacement therapy in general. It was felt that haemoperfusion is sufficiently unique to merit its own table of complications; however many complications are shared by all RRT types. The better resource for this is a chapter by James Winchester from the textbook "Complications of Dialysis" which is available for free on Google Books.

Complications of Haemoperfusion Therapy
Generic complications, common to all RRT

Access complications

All RRT requires access of some sort. Be it fistula or vas cath, there are risks:

  • Bleeding
  • Vessel damage
  • Bloodstream or localised infection
  • Air embolism

Haemolytic complications

All RRT filters tend to eat red cells, but with the haemoperfusion filters this issue is exaggerated. Haemolysis of some degree is to be expected.

Thrombocytopenia

The consumption of cellular blood components in the cartridge is significant, but the platelets are the most affected.

Inflammatory reponse

The haemoperfusion membrane is a proinflammatory surface. Modern membranes are a massive improvement, but some inflammatory reaction (particularly complement activation) is to be expected. In addition to the proinflammatory effect of broken red blood cells, there is a risk of widespread inflammation due to cartridge embolism.

Blood loss due to circuit loss

If a perfusion cartridge clots, the whole thing is discarded, just like any other dialysis circuit. The perfusion cartridge may be slightly larger, and so the blood loss may be slightly greater.

Hypothermia

The drop in the core body temperature due to heat exchange via the circuit occurs via similar mechanisms. Just as in CVVHDF, one can be cooled by the haemoperfusion circuit.

Hypoxia

Activation of complement and the inflammatory mechanisms leads to an increase in the activity of nitric oxide synthase, which countracts the normal mechanisms of hypoxic pulmonary vasoconstriction. Increased shunt develops; therefore hypoxia ensues.

Electrolyte disturbance

Charcoal does not tend to cause any sort of electrolyte changes, but the macroporous resins can remove calcium phosphate and potassium from the blood stream.

Malnutrition due to adsoprtion of useful molecules

Adsorption of all lipophilic molecules occurs, and thus one ends up missing out on the fatty acids from TPN, fat-soluble vitamins A, D, E and K, or dietary cholesterol.

Over the initial hour or so, glucose and calcium levels can drop (even with charcoal hemoperfusion).

Unique complications, specific to haemoperfusion

Haemodynamic instability

The hemodynamic instability due to haemoperfusion is wholely due to the generation of an inflammatory response due to an incompatible blood/adsorbent interface. The chances of this have been greatly reduced by the use of modern immunoneutral coatings. In comparison, much of the early interest in this technique was lost due to major haemodynamic complications.

Particle embolisation

Again, this is mainly a complication of older, less "evolved" cartridges, where bits of the adsorbent would break off and embolise downstream. In modern cartridges this is almost unheard of.

Toxin elution

Carbon in the cartridge, though a highly purified form, is still an organic product, and therefore prone to the usual peculiarities of natural materials. Weird hydrocarbons and potentially even toxic heavy metals may elute out of the cartridge and into the patient. Repeated treatments will therefore result in accumulation toxicity. This is largely a theoretical complication; as far as I am aware, such heavy metal elution has only ever been observed in vitro.

Drugs susceptible to removal by haemoperfusion

Harbord et al. have a chapter on drug removal by extracorporeal techniques. That chapter within it contains several massive tables, listing all the possible and impossible drugs which can be removed by an extracorporeal circuit. The haemoperfusion-specific information resides in Table 174-2 on page 922.

The list is massive, and impossible to memorise. Instead, it is easier to remember basic features which discriminate between those drugs that are easily removed by dialysis and those that are easily removed by haemoperfusion.

Drug features which favour haemoperfusion rather than dialytic removal

  • High lipid solubility
  • Large volume of distribution
  • High protein binding

Drugs which are easily extracted by haemoperfusion:

  • Paraquat
  • Parathion
  • Theophylline
  • Carbamazepine
  • Phenytoin
  • Paracetamol
  • Barbiturates
  • Digoxin (maybe)
  • Diltiazem
  • Metoprolol
  • Colchicine
  • Promethazine
  • Amanita phalloides mushroom toxin (phalloidin)

Non-toxicological indications for haemoperfusion

  • Lipopolysaccharide endotoxin: The cell wall component of gram-negative bacteria, which is responsible for much of the nastiness you see in septic shock
  • Superantigen: The secreted exotoxin of gram-positive bacteria, which directly activates T cells by binding to the MHC class II molecules.
  • Various cytokines: Both proinflammatory and antiinflammatory ones are cleared by hemoperfusion
  • Hepatic failure: accumulated toxins can be susceptible to haemoperfusion
  • End stage renal failure with aluminium intoxication: where it is used along with a chelating agent

References

Rafael Ponikvar, "Hemoperfusion" in: Critical Care Nephrology (2009) p.1535

Nikolas Harbord, Steven J. Gruber, Donald A. Feinfeld, and James Frank Winchester "Hemodialysis, Hemofiltration, and Hemoperfusion in Acute Intoxication and Poisoning" in: Critical Care Nephrology (2009) p.919

Gil, H-W., et al. "Clinical outcome of hemoperfusion in poisoned patients."Blood purification 30.2 (2010): 84-88.

Winchester, James F. "Complications of Hemoperfusion." In: Complications of Dialysis (2000): p.127.

Fennimore, J., J. C. Kolthammen, and S. M. Lang. "Evaluation of hemoperfusion systems: in-vitro methods related to performance and safety."Artificial Organs (1977). - this article is not available anywhere, even as an abstract!

Cruz, Dinna N., et al. "Early use of polymyxin B hemoperfusion in abdominal septic shock: the EUPHAS randomized controlled trial." Jama 301.23 (2009): 2445-2452.