Anticoagulation of the dialysis circuit

This chapter is related to the aims of Section H3(i) from the 2017 CICM Primary Syllabus, which expects the exam candidate to "describe the principles of dialysis and renal replacement fluid". Extending this expectation ever so slightly, one could find the matter of  circuit anticoagulation to be contained somewhere within that statement. Anticoagulation is actually a feature almost unique to the ICU, where the dialysis circuit pumps blood very slowly, in contrast to the comparatively violent rush produced during  intermittent haemodialysis.

Though the college has never asked about any of this in any of the previous primary exam papers, it has  come up a few times in the Part II. Multiple questions ask either about circuit anticoagulation in generic terms (Question 17 from the first paper of 2007) or in patients who cannot have heparin (Question 29 from the first paper of 2014 and Question 4 from the second paper of 2010). This chapter does not adequately serve the needs of the fellowship exam candidate, who already know all this and require merely a brief table of reminders. This table can be found in the  Required Reading chapter on the strategies used to prolong the life of the CRRT circuit. Additionally, as external reading, one can recommend this excellent article by Notohamiprodjo et al (1986). 

Why do we need to anticoagulate the dialysis circuit?

The answer is that blood in its usual course is accustomed to encountering only itself and the inside of blood vessels, and to meet any other substance tends to provoke clotting. Of the dialysis circuit, we have plastic tubing and we have filter. The filter generally has about 20 times the surface area as compared to the circuit, and therefore the filter is the most thrombogenic component. 

Mechanism of anticoagulation by citrate

Citrate is a calcium chelator, and by robbing the clotting cascade of its ionised calcium it disables the steps of the cascade in which calcium plays a role (many people dont realise that calcium used to be Factor IV). This has been used to preserve the fluidity of blood samples for many decades, being first discovered around 1890.

The following are clotting cascade proteins which require calcium to function:

• Factor VIIa
• Factor IXa
• Factor X
• Factors Xa
• Factor II (prothrombin)

So, 2, 7 9 and 10. Same as the Vitamin K-dependent factors.

The net effect is the inhibition of thrombin formation.

In order to work, the citrate dose must be adjusted to achieve an ionised calcium concentration less than 0.4mmol/L within the filter.

Metabolic fate of citrate

A good overview of this is available from an article by Bellomo and Kellum. Of the infused citrate, some proportion is actualy removed by the CRRT circuit (as it is a small molecule), and the rest is rapidly metabolized in the Krebs cycle reactions - particularly in the liver, muscle and renal cortex. The chelated calcium is released by this process.

As a metabolic substrate, citrate has a measurable nutritional value: about 3kcal/g, or 0.59kcal/mmol (in case you are wondering, 1g citrate = 5mmol). And because it is consumed in Krebs cycle, it has the advantage of increasing the strong ion difference by being a strong anion which disappears without a trace. For each molecule of citrate burned in the Krebs furnace, three sodium ions are left behind, which has the equivalent effect of adding 3 HCO₃^- molecules.

A non-Stewartian explanation is also often given for this phenomenon. It describes the process of metabolism as something which consumes 3 H⁺ ions, and therefore has the equivalent effect of adding 3 HCO₃^- molecules. Even more crudely, it is generally said that "citrate generates three bicarbonate molecules". It is true - its metabolism is the equivalent of buffering, and in excess citrate can cause a metabolic alkalosis.

Evidence for the safety and efficacy of citrate as a regional anticoagulant in dialysis

Is it any good, you ask.

The answer is, probably yes.

Here are a few papers:

• Monchi et al (2004): 49 circuits; single-centre RCT looking at circuit lifespan.
  • Conclusion: "Regional citrate anticoagulation seems superior to heparin for the filter lifetime and transfusion requirements". The haemofilters lasted 70 hours on average with citrate, and 40 hours on average with heparin.
• Oudemans-van Straaten et al (2009):215 patients, naldoparin vs. citrate: single centre RCT.
  • Unexpectedly, 3-month mortality was found to be decreased by 15% with citrate.
  • Authors hypothesised that some sort of interference with inflammation improved the SIRS response. Unfortunately, the population was elderly (average age was 73), and their all-cause mortality was high (43-68%).
• Hetzelet al (2010): 174 patients, heparin vs. citrate: multicentre RCT.
  • "Use of citrate resulted in less systemic anticoagulation, a lower risk of bleeding and a longer haemofilter patency. Episodes of hypercalcaemia, hypocalcaemia and the need for additional bicarbonate infusions occurred more often under citrate."
  • No influence on mortality. Neither any safer nor more effective than heparin, all things considered. Conclusion: "neither citrate nor heparin anticoagulation should be regarded as a therapeutic standard, since there is no advantage of one of these substances with regard to patient mortality."
• Bellomo and Kellum scraped the literature for opinions in 2011:
  • Generally the tone of the paper suggests that heparin has been elevated unfairly to a high pedestal, and that perhaps it should not be viewed as a gold standard. Much is made of various little-known idiosyncrasies of heparin action in the critically ill, for instance proinflammatry effects and the resistance to heparin induced by critical illness. Intelligent use of citrate is cautiously urged as a valid alternative.
• Schilder et al (2013): 139 patients, heparin vs citrate: multicenter RCT.
  • Conclusion: "Renal outcome and patient mortality were similar for citrate and heparin anticoagulation during CVVH in the critically ill patient with AKI. However, citrate was superior in terms of safety, efficacy and costs"
  • However, the number was small; the study was powered for 360 patients but enrolment was slow and the authors got bored of it after 6 years, terminating the trial. With less than ghalf of the expected patient numbers, the study was not powered to detect subtle differences in mortality.
• Fiaccadori et al (2015): An Italian position statement.
  • A strongly pro-citrate statement:  "regional citrate anticoagulation (RCA) could represent the most promising strategy" provided one is cautiously protecting their patient from complications.

The college love citrate toxicity, and have asked about it in the past; one can find it in Question 3.3 from the second paper of 2013.

Citrate as a cause of high anion gap metabolic acidosis, followed by metabolic alkalosis

Citrate is an anion. In solution, it is a weak acid, a concept familiar to those of us who have ever rubbed their eyes with a lemon. The presence of citrate in solution may give rise to a degree of acidosis. Even though it is co-administered with lots of cations (trisodium citrate is the most common formulation), the total strong ion difference (SID) of the solution is 0, and so - like all 0-SID fluids (eg. normal saline) - the net effect on the whole-body acid base balance is to produce a degree of acidosis. Furthermore, over the course of a CRRT session the rate of sodium removal may outstrip the rate of citrate removal by the circuit, resulting in the widening of the anion gap.

Citrate is then metabolised (mainly in the liver, though all tissues could theoretically make use of it). The result is that now you have an excess of cations and a deficit of anions, giving rise to an increased strong ion difference. Thus, alkalosis ensues. Intensivists from Berlin reported that in their single-centre experience over 50% of the patients developed a metabolic alkalosis (and none of them had any metabolic acidosis).

Influence of citrate on ionised and total calcium

The ionised calcium in acidosis normally increases. Well, in respiratory acidosis it probably increases more than in lactic acidosis (because lactate forms calcium-lactate complexes), but still - it should be high, not low.

If one is acidotic with a low ionised calcium, one should look to citrate. Its influence is confirmed by the presence of a high total to ionised calcium ratio (i.e. the total calcium is normal, but the ionised fraction is low) - this is because measurement instruments which detect calcium will also measure citrate-calcium complexes in the serum, but the ion-selective electrode which measures ionised calcium will only measure the free fraction, which decreases with citrate toxicity.

Furthermore, as the citrate is metabolised, the chelated calcium is released, resulting in a rebound hypercalcemia.