(This question follows from Question 5.1 from the first paper of 2020; the original stem is as follows:)

A 51-year-old male with a history of cirrhosis secondary to Hepatitis C is admitted for the first time with haematemesis. His gastroscopy is complicated by aspiration. He is admitted to ICU ventilated.

The patient becomes anuric and 6 hours after commencing continuous veno-venous haemodiafiltration (CWHDF) with citrate anticoagulation has the biochemistry results shown below:


Patient Value

Adult Normal Range







89 mmHg (11.9 kPa)


31.0 mmHg (4.1 kPa)*

35.0 - 45.0 (4.6 - 6.0)




9.0 mmol/L*


Base Excess

-18.0 mmol/L*



2.1 mmol/L*



142 mmol/L



4.4 mmol/L

3.5 -5.0


107 mmol/L*



8.0 mmol/L*

3.5 -6.0

Ionised calcium

0.69 mmol/L*


Calcium corrected

3.70 mmol/L"'


Give the likely cause of the biochemical abnormality with your reasoning.

What adjustments  may  be made to the CWHDF?
(40% marks)

[Click here to toggle visibility of the answers]

College answer

Citrate accumulation is likely. Features suggesting citrate toxicity are the high anion gap metabolic acidosis, history of liver disease, reduced ionized calcium and increased Ca gap (Ca total minus iCa). The dose of citrate should be reduced (e.g. by reducing the citrate-containing filtration replacement rate or increasing the dialysate rate (which will remove citrate, there are numerous protocols prescribing rate changes). Consider changing the CVVHDF circuit to one without citrate.


Again, let's go through this systematically, remembering that these abnormalities are all observed in a shocked liver disease patient, whom somebody put on citrate haemodialysis for some reason (one might suppose because bleeding makes a heparinised circuit sound even more silly).

  • The A-a gradient remains increased. (713 × 0.4) - (31 / 0.8) - 89 = 157.5 mmHg. Whatever these people are doing about his ventilation, it is not helping.
  • There is acidaemia, which has worsened since the commencement of dialysis (pH was 7.16 in the last question). 
  • There is a severe metabolic acidosis, which is essentially unchanged from the previous gas result (the SBE has decreased to -18.0)
  • The CO2 is unhelpful - the expected CO2 would be (40-18) = 22 by the SBE method, or (9 ×1.5)+8 = 21.5 mmHg by the old Boston rules. Either way, there's clearly a respiratory acidosis. Given that the patient is ventilated, this must be attributed to unhelpful ventilator settings.
  • The anion gap is (142-107-9)= 26, whereas previously it was only 20. 
  • The delta ratio has therefore changed from 0.54 (previous question) to  (26-12)/(24-9) = 0.93, i.e. the high anion gap component is now winning (whereas previously this was more of a NAGMA).
  • The ionised calcium is desperately low. The total calcium is disturbingly high. What could possibly have given rise to this unmistakably characteristic pattern

Without further analysis, one could conclude that this high anion gap metabolic acidosis with a widened ionised-total calcium gap could only be due to citrate toxicity.

What adjustments could be made to the CVVHDF?

  • Decrease the citrate dose. Or, decrease the blood flow rate (if you can). Usually the machine runs a default citrate dose indexed to the blood flow rate, at 3.0 mmol/L of blood. With this doubtlessly coagulopathic cirrhosis patient, one might make the argument that the blood is already thin enough, and that a lower citrate dose per litre of blood would still achieve satisfactory anticoagulation
  • Alternatively, one could slow the rate of citrate delivery by decreasing the blood flow rate. The patient is anuric, that is true - but looking at their previous blood results, there is no imperative to clear uraemic toxins from the blood. A slower blood flow rate would be less efficient, but super-efficient dialysis is not what you're looking for here. 
  • Run the circuit with no anticoagulation. That's an option if the blood flow rate is fast and the replacement fluid is being given pre-blood-pump 
  • Stop the dialysis altogether. How essential is it to continue? This patient does not have life-threatening biochemical abnormalities, except for those you yourself have caused with the dialysis. Surely he would be better off without your cack-handed management.
  • Increase the patient's resistance to citrate toxicity. One can do nothing about the damaged liver, but one can perfuse it with better blood flow. With fluid resuscitation and inotropes/vasopressors, hepatic blood flow will improve, and citrate dialysis may be better tolerated.
  • Serious, think about heparin as an alternative. Sure, the patient presented with uncontrolled haematemesis and may not have two platelets to rub together, but really, how badly do you want to dialyse him? If the need is significant, and the bleeding is under control, some conservative in-circuit heparin dose could be safe.
  • Think about SLED as an alternative. How essential is it to persist with a continuous RRT modality? SLED could be performed with minimal or no anticoagulation, if pushed. If the patient can be chemically compelled to tolerate the process haemodynamically, this is a valid option.


Wonnacott, Rob, Brandi Josephs, and Jill Jamieson. "CRRT regional anticoagulation using citrate in the liver failure and liver transplant population." Critical care nursing quarterly 39.3 (2016): 241-251.