Frusemide in critical care

Frusemide has been the topic of multiple primary exam questions, but in the Part Two the first serious mention of this drug came in the form of Question 7 from the first paper of 2019. The college asked about "the role of frusemide" in Intensive Care, including "potential indications, proposed benefits, adverse effects and a summary statement of the available evidence". Though the "potential indications" for frusemide include multiple non-renal applications, the bulk of Question 7 focused on the evidence behind its use in renal failure, specifically as a means of delaying or preventing dialysis in renal failure patients with fluid overload.  That is the rationale for putting this undifferentiated question into the "renal failure and dialysis" section.

The free article by Ho & Power (2010)  is probably enough as a single reference to answer Question 7.  Joannidis et al (2019) have published an excellent (free) editorial in Intensive Care Medicine which answers some of the misconceptions regarding the uses of frusemide, but it is probably not enough of a review by itself.  

Uses of frusemide in Intensive Care patients are numerous, and it would be difficult to group them under "rationale" or "indications", particularly as not everybody would agree to every indication. Instead, it would probably be better to divide the physiological effects of frusemide as "useful" or "not useful", to discuss the various ways in which it has been applied in critical care, and to evaluate the evidence for these applications.

Useful effects of frusemide 

• Diuresis. As a loop diuretic, the main role of frusemide is to produce water loss, classically in the form of torrential dilute urine. This is generally grouped among its useful effects because intensivists are chronically disappointed with the fluid balance of their patients, usually because the fluid balance is too positive.  The association between frusemide use and a negative fluid balance is quite well defined. The relationship between this negative balance and improved outcomes is less well defined, though studies (eg. Shen et al, 2019) were able to demonstrate some relationship between loop diuretic use and decreased hospital mortality. More broadly, an increase in the urine volume and flow rate is viewed as an end in itself if the urine is somehow toxic,eg. if it is full of cyclophosphamide one may use forced diuresis to prevent haemorrhagic cystitis (Droller et al, 1982)
• Venodilation and therefore preload reduction is one of the consequences of an intravenous dose of frusemide. Dikshit et al (1973) found marked changes in venous capacitance and LV filling pressure which occurred within 5 minutes of an IV frusemide bolus, long before the diuretic effects became apparent. The mechanism for this is unknown, but is probably some sort of indirect effect,
• Hypokalemia and hypocalcemia can be desirable effects if the patient is hyperkalemic or hypercalcemic.  Frusemide produces a rather unpredictable amount of potassium wasting in the urine. According to Morgan et al (1980), in outpatients this effect is less with frusemide than with thiazide diuretics: after several days of therapy, the potassium tends to fall only by 0.3 mmol/L. In patients with hyperkalemia, loop diuretics are one of the potential therapies to decrease the serum potassium (Mushiyakh et al, 2012), particularly where other methods (eg. cation exchange resins) are for whatever reason contraindicated.
• Alkalosis can be a desirable effect of loop diuretic therapy if the patient has acidosis. In the context of Type 4 renal tubular acidosis, for example, frusemide can be a useful therapy to increase urinary acidification (Rastogi et al, 1985).
• Urinary acidification itself may be a goal of therapy in toxicology, where acidic urine favours the ionization of alkaline compounds and causes "ion trapping" of weak alkali in urine. Experience with this appears to be mainly in the case of amphetamine or PCP toxicity, though it is unclear what advantage this treatment has over simple supportive care.
• Increased excretion of toxic cations may be promoted by frusemide. Toxicological uses include increased elimination of thallium  (Lameijer et al, 1979) and amelioration of mercury-induced nephrotoxicity by inhibiting the absorption of  Hg-conjugated organic anions (Hazelhoff et al, 2015)

Adverse effects of frusemide

• Hypotension: though one's intentions may be to decrease the patients positive fluid balance, the uncontrolled use of diuretics may lead to an unexpectedly large fall in the circulating volume, and therefore hypotension.
• Worsening renal function is listed in the college answer to Question 7 from the first paper of 2019,even though it is also viewed as one of the myths of frusemide use by Joannidis et al (2019). Certainly, it is logical that if one is suffering from AKI due to dehydration, the use of diuretics will make that worse. Equally, it is illogical to include such patients in trials looking at the risk of AKI associated with diuretics (Levi et al, 2012) . 
• Electrolyte derangement: The process of desirable diuresis may give rise to undesirable electrolyte depletion, hypernatremia and metabolic alkalosis.
• Urate accumulation: Frusemide increases the absorption of urate, and decreases ts active secretion (Kahn, 1988). Worsening gout symptoms may ensue.
• Lithium toxicity: Frusemide is implicated in increasing the risk of lithium toxicity  (Oh et al, 1977) but this is probably somewhat overhyped by panicky case reports. Saffer et al (1983) followed numerous lithiated patients for weeks and found no evidence of increased levels.
• Non-renal toxicity of the drug:  frusemide is ototoxic in high doses. Wigand et al (1971) found that (occasionally reversible) hearing loss occurred in 50% of patients receiving 1000mg over 40 minutes. Vertigo and nystagmus are also listed among its side effects, which appear to be direct neurotoxic effects (Baldwin et al, 2008).  This is apparently potentiated by hypoalbuminaemia: frusemide is 98% protein-bound and any loss of protein may give rise to a significant increase in the free fraction of the drug.
• Allergic reaction:  like everything, frusemide may cause an allergic reaction. The risk is apparently quite small, to the effect that it belongs in case reports such as Hansbrough et al (1987). The reaction is thought to be a case of cross-reactivity to the sulfonamide nucleus shared by thiazides, frusemide, sulfonylureas, and numerous other drugs. 
• High doses of frusemide can induce systemic vasoconstriction. Jhund et al (2000) report the results of several studies where frusemide boluses were found to have an acute arterial vasoconstrictor response. Again, this is thought to be some sort of indirect effect. 
• Paralysis of mucociliary clearance: frusemide affects the viscosity of respiratory mucus, making it thicker and therefore more difficult to transport. This was demonstrated in a gross experiment by Kondo et al (2001) involving frog mucus and a mechanical coughing machine. 
• Drug interactions can occur because frusemide is far from chemically inert (Ho & Power, 2010):
  • Frusemide reduces clearance of  the following drugs:
    • theophylline
    • gentamicin
    • benzylpenicillin
    • cephalosporins
    • active metabolite of oseltamivir
    • bumetanide, not that you'd use the both simutaneously
  • Frusemide reduces the therapeutic effect of warfarin, and warfarin reduces the diuretic effect of frusemide
  • Frusemide increases the anti-epileptic effect of valproate
  • Frusemide increases the hypotensive effect of ACE-inhibitors

Evidence for the applications of frusemide in Intensive Care

• In cardiac failure, frusemide has an established role in symptom control. Though most people would logically associate diuretic therapy with a decrease in preload, Wilson et al (1981) found that the end-diastolic dimensions of the left ventricle remain largely unchanged, and that afterload reduction is probably the main mechanism by which these symptom effects are achieved. . The 2013 AHA statement recommends loop diuretics specifically with Level 1 grade of recommendation (albeit on the basis of Grade C evidence).  "The ultimate goal of diuretic treatment is to eliminate clinical evidence of fluid retention", they say. A mortality benefit is occasionally demonstrated (Faris et al, 2012). The college answer to Question 7 from the first paper of 2019 mentions the DOSE trial (Felker et al, 2011) which demonstrated some improvement in secondary outcomes and self-reported symptoms when the decompensated CCF patients' regular frusemide dose was increased by 250%.
• Cardiorenal syndrome,  ”a complex pathophysiological disorder of the heart and the kidneys whereby acute or chronic dysfunction in one organ may induce acute or chronic dysfunction in the other organ”, is an indication for the use of frusemide by virtue of the same mechanism as its use in heart failure (Cruz, 2013). The interdependence of the two organs means that the correction of one organ failure (ie. cardiac) should lead to an improvement of the other. Loop diuretics are inevitably mentioned wherever cardiorenal syndrome is discussed, but the evidence to support their use is relatively scant (Ronco et al, 2008)
• Altitude sickness is mentioned in the college answer to Question 7 from the first paper of 2019. The question asked about "the role of frusemide in patients in the ICU"  for whom tiredness and trouble sleeping are less likely to be an issue. More likely, frusemide may find its application in the treatment of HAPE (HAPO?) and HACE (HACO), high altitude pulmonary oedema and cerebral oedema associated with rapid ascent.  The books talk about the use of frusemide and mannitol as second-line therapies for high altitude cerebral oedema, suggesting some confusion regarding the mechanism of their action. 
• To decrease fluid balance for no organ-specific reason, because this is viewed as having some sort of therapeutic benefit in the critically ill population, enough to become a goal of treatment in and of itself. Certainly, there is evidence to support the idea that a positive fluid balance is associated with increased mortality. Most famously, this has been demonstrated among sepsis patients (in the SOAP study as well as in analysis of the VASST data ), but it appears to be common to all patients admitted to ICU, particularly if they have CCF or renal failure (Lee et al, 2015). Logically, from this one might conclude that the active pursuit of a negative fluid balance should decrease mortality. Shen et al (2017)  explored this notion and found that patients with an extremely negative balance (-60ml/kg/48 hrs, or about 2000ml negative per day) had a substantially decreased risk of mortality in the ICU. In 2019 the same authors (Shen et al) found a decreased mortality in shocked patients who received diuretics, again in a non-randomised case matched cohort.  Overall, it seems like a negative fluid balance is a worthy goal, though RCT-level evidence remains elusive.
• Portal hypertension management:  frusemide together with other agents is probably more effective than frusemide alone; or rather, other agents are probably much more effective and the addition of frusemide is only occasionally seen to produce any additional clinical improvement (Garcia-Pagan et al ,1999).
• Acute hypercalcemia: together with fluid resuscitation, frusemide can help drop calcium levels fairly rapidly. This is a very old-school technique and requires insane doses of frusemide to be effective.  Suki et al (1970) reported success with 100mg given every 2 hours.
• Transfusion-related circulatory overload may be managed reactively with loop diuretics, but they apparently do not prevent it when administered prophylactically. In the college answer to Question 7 from the first paper of 2019, the examiners mention "Cochrane review 2015"  which concluded "no benefit from routine use of frusemide in transfusions to avoid TACO" - that was probably Sarai et al (2015), who made those conclusions on the basis of only four studies with a total of 100 patients. 

Evidence for the use of frusemide in renal failure

Though Question 7 from the first paper of 2019 ostensibly asked about the role of frusemide in all intensive care patients, of its 580-word college answer 323 words were dedicated to exploring the controversial use of loop diuretics in renal failure. Proportionally to this allocation of examiner attention, the topic of frusemide in renal failure deserves a thorough dissection. Fortunately, Ho & Power (2010) have an excellent series of tables at the end of their article which has been pillaged for the content of what follows:

Rationale for the use of frusemide in renal failure patients

• Renal failure is typically associated with fluid overload due to oliguria, hyperkalemia and metabolic acidosis
• These derangements are also common indications for dialysis
• Frusemide may mobilise fluid, increase elimination of potassium and promote alkalinisation of the extracellular fluid
• Ergo, the use of frusemide may reduce the need for dialysis

Proposed beneficial effects of frusemide in renal failure

• Decreased oedema and lower venous pressure decreases congestion of the kidneys, improving renal perfusion and therefore renal function 
• Inhibition of tubule function should decrease tubular metabolic rate and therefore should have a nephroprotective effect
• The demonstration of diuresis with frusemide reveals functioning tubules and therefore may be useful to identify those patients who will go on to require dialysis and those who will not. 

Caveats to the use of frusemide in renal failure

• Larger doses are usually required, which increases toxicity
• Diuresis gives rise to decreased circulating volume and therefore may lead to the activation of the renin-angiotensin-aldosterone axis, increasing salt retention
• Increased urine output may be misinterpreted as an improvement in renal function, whereas in fact, no such improvement has taken place (and all sorts of useful diagnostic and management steps for acute renal failure would be delayed)
• With frusemide, the concentration of urinary sodium becomes an unreliable discriminator between pre-renal and intra-renal causes of renal failure (not that it's particularly reliable anyway).
• Acidification of the urine may reduce the solubility of myoglobin and exacerbate free radical damage due to ionised contrast

Evidence for the use of frusemide in renal failure

• Frusemide does not prevent acute renal failure perioperatively. This was the conclusion of Zacharias et al (2013), which is probably the "Cochrane review 2013" which concluded "no clear benefit for any AKI sparing strategies including frusemide" as mentioned by the examiners in the college answer to Question 7. 
• Frusemide does not reduce mortality in established renal failure, according to Bove et al (2018). However, reduced duration of CRRT dependence or improved fluid balance are easier to demonstrate, and these may be interpreted as patient-centred outcomes (i.e. the patients are probably happier because they are free from the circuit for longer periods). 
• Frusemide does not prevent progression to dialysis in acute renal failure. The SPARK study (Bagshaw et al, 2017)  enrolled AKI patients and gave them frusemide, to no apparent effect beyond some undesirable electrolyte disturbances. 
• Frusemide does not improve recovery from anuric renal failure. Those kidneys will stay dead, as was confirmed by van der Voort et al (2009). The investigators infused their CRRT-dependent patients with 0.5mg/kg/hr of frusemide (that's about 40mg per hour or 960mg per day). Though urine output increased, there was no difference in the need for RRT, nor any change in survival chances.
• Frusemide may decrease mortality when used to prevent renal failure. "A trend towards a beneficial effect of intermittent furosemide administration was found when analyzing the subgroup of studies in which furosemide was administered to prevent AKI", found Bove et al (2018). However, ESICM recommend against the use of diuretics as solo agents to prevent AKI (Joannides et al, 2017).
  What do they mean, "prevent"? How does frusemide prevent AKI? Well. Looking at Table 1 from Bove et al (2018), there is an overwhelming number of trials looking at the prevention of acute kidney injury. Usually, the frusemide ends up being administered prophylactically to prevent a rise in creatinine in the context of some nephrotoxic insults such as intravenous contrast. 
• 
  
  In summary, of the patients receiving frusemide boluses, 13.4% ended up with worse renal function, as opposed to 18.2% patients in the control group. That would sound more heroic if one was not aware that the control group in the majority of these studies was either a frusemide infusion, mannitol or torsemide. In short, it is remarkable that any difference in outcome was generated by this meta-analysis. 

"Own practice"

To discuss one's own approach to a contentious problem is usually advisable in a "critically evaluate" question, but "own practice" is not always called for. "Many candidates chose to describe their own practice, not answering the question asked", complained the college examiners in their comments to Question 7 from the first paper of 2019, who clearly just wanted "a summary statement of the available evidence". However, there may come a time when a description of one's own practice may be called for.

Jones et al (2015) surveyed the locals and published their results in Critical care and Resuscitation, which would have been relatively straightforward as the chief editor was one of the co-authors. It turns out that we Australian intensivists do not tend to use frusemide for acute kidney injury. The most common dose was a 40mg IV bolus, the most common indications were acute pulmonary oedema and a positive fluid balance. The exam candidate who confesses to some sort of similar practice will therefore be in good company.