Introduction with a clinical conundrum
Imagine that you admit a patient with septic shock. You resuscitate the patient as best you can with inopressors, fluids, and antibiotics. An adequate blood pressure is achieved. A reasonable amount of fluid is administered. Despite all these measures, the urine output remains minimal. What should you do next?
- Do nothing, sit on your hands and wait for the patient to improve.
- See if a volume challenge will improve urine output.
- Attempt a vasopressor challenge (e.g. raise the MAP, see if this improves urine output).
- Attempt an inotrope challenge (e.g. administer dobutamine, see if this improves urine output).
- Perform a furosemide stress test.
The crux of this problem is difficulty sorting out the following two possibilities:
- Smart Kidney: The patient’s kidneys may have temporarily reduced urine output due to hypoperfusion (“pre-renal azotemia”). The decrease in urine output is an adaptive response, in attempts to conserve volume. In this scenario, hemodynamic manipulation (increased MAP or cardiac output) could improve renal perfusion, increase the urine output, and avoid persistent kidney injury.
- Dumb Kidney: Realistically, you’ve already established pretty decent hemodynamics – and the kidneys haven’t responded. Therefore, it's quite possible that the patient has intrinsic renal failure at this point. Further hemodynamic manipulation won't help this. In fact, giving more fluid may accelerate the development of volume overload and hasten the need for hemodialysis.
Spoiler alert: There is no simple answer to this quandary. The best approach will depend on the specifics of the patient. The purpose of this post is to explore the furosemide stress test as one potential option which might be useful in some patients.
Furosemide Stress Test
Getting started
A furosemide stress test (FST) is used when approaching a patient with oliguria to determine whether underlying renal function is intact. The test consists simply of administering a bolus of 1 mg/kg furosemide (if the patient is furosemide naive) or 1.5 mg/kg furosemide otherwise.1
- If >200 ml urine is produced over the next two hours then the patient has “passed” the furosemide stress test – this indicates that tubular function is intact and short-term renal recovery is likely.
- If <200 ml urine is produced, then the patient has “failed” the furosemide stress test – this predicts ongoing renal dysfunction which may require dialysis.
If the patient does produce urine, this will generally be replaced in the form of IV crystalloid (unless the patient is felt to be volume overloaded). The overall goal is to test the ability of the kidneys to respond to furosemide.
Physiology of the FST
In order for the patient to respond to furosemide, all of the following processes must occur:
- The glomerulus must be filtering fluid into the renal tubule.
- The function of the proximal convoluted tubule must be intact. Furosemide enters the proximal tubule via an active transport process through the human organic anion transporter system.
- The function of the loop of Henle must be intact, because this is the physiologic target of furosemide.
- There must be no obstruction anywhere between the glomerulus and the Foley catheter.
Thus, passing the FST requires integrated function of several components of the nephron:
- Urine output will improve if the nephron is functionally intact (but merely trying to conserve sodium – a “smart” kidney).
- Urine output won't improve if any portion of the nephron is truly dysfunctional (i.e. “dumb” kidney).
Fractional excretion of sodium (FeNa) is a conceptually attractive approach to this problem (differentiation between pre-renal vs. intrinsic renal dysfunction). However, emerging evidence questions the value of FeNa in the ICU.2 It's conceivable that over time the FST might come to replace the FeNa as a diagnostic tool to better understand the patient's physiology.
Evidentiary basis for the FST
Experienced clinicians know intuitively that when a patient doesn't respond to a big slug of furosemide, that's a bad sign. Alternatively, if a small dose of furosemide elicits buckets of urine, that's an optimistic sign. However, this assumption is imprecise – the concept needs to be operationalized and validated.
Chawla et al 2013: Development and standardization of a furosemide stress test to predict the severity of acute kidney injury
This was the initial study that designed the FST.1 77 patients with early AKI were involved (consisting of a 23-person retrospective cohort and a 54-person prospective cohort). Inclusion required evidence of acute tubular necrosis (either granular or epithelial cell casts on urine sediment, or a fractional excretion of sodium >1%).
The volume of urine produced in response to furosemide was strongly predictive of progression to Stage III AKI, with an area under the receiver-operator-curve of 0.87. Using a cutoff of 200 ml over 2 hours, the test had a sensitivity of 87% and a specificity of 84%.
One potential weakness of this study is that Stage-III AKI is a somewhat arbitrary endpoint that isn't patient-centered. In a follow-up analysis of these same patients, the FST was also capable of predicting the need for dialysis (with an area under the receiver-operator curve of 0.86).3
Matsuura R et al. 2018: Response to different furosemide doses predicts AKI progression in ICU patients with elevated plasma NGAL levels [a renal biomarker].
This was a retrospective analysis of 95 patients challenged with furosemide in one medical-surgical ICU.4 At the time of the FST, 44 patients had no AKI, 34 patients had stage-I AKI, and 17 patients had stage-II AKI. Over one week, 18 patients progressed to stage-III AKI. Furosemide responsiveness predicted progression to Stage-III AKI with an area under the receiver-operator curve of 0.87, replicating the Chawla study above. Furosemide responsiveness also outperformed renal biomakers as a predictor of AKI progression:
These authors used low, variable doses of furosemide and reported their results in terms of urine output per mg of furosemide administered.5 Lower furosemide dosing may reflect that these patients were less ill at baseline than in other studies, limiting comparison with other studies. Nonetheless, the study does provide some general support for the validity of FST.
Lumlertgul N, et al. 2018: Early versus standard initiation of renal replacement therapy in furosemide stress test non-responsive acute kidney injury patients (the FST trial)
This was a multicenter RCT involving 162 ICU patients meeting the following criteria:6
- AKI of any stage, attributed to acute tubular necrosis based on any of the following criteria: the presence of granular or epithelial casts, a fractional excretion of sodium >1%, a fractional excretion of urea >50%, or a plasma NGAL > 150 ng/mL.
- Completed resuscitation to a euvolemic state
- No immediate indication for dialysis
The patient population included mostly medical patients (68%), a majority of whom were septic (58%), intubated (82%), and on vasopressors (78%). All patients underwent a FST. Patients who failed to respond to the FST were randomized to early dialysis (within six hours) versus standard care (dialysis when prompted by traditional indications such as hyperkalemia or volume overload).
This study was designed as a feasibility study to plan subsequent trials. Nonetheless, it does provide some useful information about the FST. Among 44 patients who responded to the FST, only six patients (14%) required dialysis. Alternatively, among patients who failed to respond to the FST and were randomized to conventional management, 45 (75%) ultimately required dialysis. Thus the FST alone was a fairly good predictor of the need for dialysis. FST was also a strong predictor of death: mortality was 34% among FST-responders versus 60% among FST-nonresponders. Indeed, the FST was a more accurate predictor of mortality than the APACHE II or SOFA score.
Potential limitations of the furosemide stress test
- Chronic renal insufficiency: Every study of FST has excluded patients with chronic renal insufficiency and baseline GFR <30 ml/min. Performance in this scenario is unknown.
- Severe hypoalbuminemia: Albumin is involved in the binding of furosemide and transport to the proximal convoluted tubule. Some trials of FST have excluded patients with albumin <2 g/dL.6
- Frank hypovolemia: Studies of FST have excluded patients believed to be hypovolemic in the opinion of the treating team.
- Patients without evidence of acute tubular necrosis? Most studies of FST have required some evidence to suggest acute tubular necrosis (e.g. abnormal urinary sediment, elevated FeNa, or elevated renal biomarkers). However, the FST could probably be applied to a general population of ICU patients without need for a priori chemical or cellular analysis for the following reasons. First, the poor performance of FeNa as a diagnostic test suggests that this prerequisite is relatively meaningless.2 Second, in the validation study by Lumlergul only 5/297 patients were excluded due to absence of acute tubular necrosis.6
Returning to our clinical conundrum: Could the FST be used in our septic patient with persistent oliguria?
Let's return to our case of the septic patient who is resuscitated yet continues to have poor urine output.
Potential use of the FST is shown above:
- If the patient is an FST-nonresponder, this supports the presence of intrinsic renal failure. The urine output here is an unreliable reflection of end-organ perfusion. Further attempts to improve the urine output are likely to be futile and potentially harmful. In this situation renal recovery may occur, but it will likely take days. Basic hemodynamic support must be continued, but aggressive efforts to elicit increased urine output should be curtailed.
- If the patient is an FST-responder, this suggests underlying functionality of the kidney. Urine output provoked by the FST should be replaced with IV crystalloid to prevent volume depletion. In this situation, ongoing hemodynamic manipulation to improve renal perfusion and maintain urine output may be beneficial.
In short, the primary benefit of the FST might be allowing us to identify patients with intrinsic renal failure who won't benefit from further fluid loading. Without any definitive test to diagnose this condition, there is often an insurmountable temptation for each sequential provider to perform their own fluid bolus, which over time, adds to worsening fluid overload. Early identification of these patients may also facilitate preparation for the potential need for dialysis.
Please note that the presently outlined use of the FST hasn't been formally validated. However, the FST has been linked to clinically meaningful outcomes (e.g. dialysis), which is more than can be said about most hemodynamic monitoring devices. Overall the FST is simply one piece of the clinical picture, which would be combined with other sources of hemodynamic and renal information.
- The furosemide stress test (FST) is a protocolized furosemide challenge, which has been shown to predict whether patients are likely to progress towards advanced AKI and dialysis.
- Critically ill patients commonly develop oliguria due to severe intrinsic AKI. In this situation, ongoing efforts to elicit urine output through repeated fluid boluses or other hemodynamic manipulations is generally futile and potentially harmful.
- FST could conceivably be utilized among patients with persistent oliguria despite initial resuscitation, to facilitate early identification of patients with severe intrinsic renal failure in whom urine output is an unreliable measurement of tissue perfusion.
Expert Commentary from Dr. Richard Solomon, Chief of Nephrology at the University of Vermont Medical Center:
Very nice job and balanced discussion of the FST. As you correctly point out, there are not a lot of patients studied in these trials. That said, the predictive value of the FST is better than any other biomarkers of renal injury that are available, including the Nephrocheck (TIMP-2 & IGFBP7, two renal biomarkers). In addition to being a diagnostic study, it may have therapeutic importance. For example, if a patient is acidotic after resuscitation and responds to furosemide in the FST, you can infuse bicarbonate and continue to use furosemide to remove the sodium, keeping the patient in sodium balance. Finally, the use of furosemide to prevent AKI is very controversial. In humans, it is often difficult to prevent volume depletion, even in the ICU setting with hourly urine outputs, and the lack of benefit of furosemide for prevention of AKI may be confounded by the associated volume depletion. The one exception is when you use a device that automatically infused an equal volume of crystalloid for every drop of urine that the kidney puts out (RenalGuard). Then even small doses of furosemide (0.25 mg/kg) are protective. The mechanism probably has to do with furosemide decreasing oxygen consumption in the TAL and realigning consumption and delivery to prevent ischemia in the medulla.
Acknowledgement: Thanks to Dr. Gilman Allen for thoughtful comments on this post.
- Pulmcrit wee: The cutoff razor - April 15, 2024
- PulmCrit Blogitorial – Use of ECGs for management of (sub)massive PE - March 24, 2024
- PulmCrit Wee: Propofol induced eyelid opening apraxia – the struggle is real - March 20, 2024
It very good post. My question is how to deal with a septic, oliguric pt with underlying CKD 3 or 4, and what is the importance of albumin level?
I don’t know. Hopefully more data will emerge about FST in patients with severe hypoalbuminemia or advanced underlying CKD.
Oliguria, in absence of intrinsic renal failure, is not always produce by pre-renal azoemia. In fact, in sepsis, renal perfusion is not decreased. I think is important to stress the fact that the inflammatory kidney response could be responsible for oliguria, where the treatment is the interventions in the primary cause. Wouldn’t be adecuate I think to give the idea that oliguria always needs interventions directly in the kidney, when many times less is more and only time is needed. The same with acidosis and bicarbonate. Specially with fluids. More fluids = more renal failure in septic shock. (the problem… Read more »
I agree, the lasix challenge does appear to predict whose renal fxn will get worse and whose won’t, but whether giving fluids or hco3 will have a positive impact on renal fxn or more importantly an overall patient centered positive impact, in either group, is really unknown. Especially since as you said 1) septic aki is it’s own thing and generally with high renal a blood flow and 2) in the medical icu these pts generally have already received a fair amount of fluid by the time this test is being considered and are already considered euvolemic/resuscitated (at least those… Read more »
Why not use FeUrea instead of a lasix challenge?
Excellent post that both summarizes the data supporting our imprecise intuitions and outlines areas for future study design.
Excellent post What is your opinion about the use of furosemide in infusion in AKI In my UTI we have managed to reduce the use of renal replacement therapy between 70 and 80% and the progression from AKI to stage 3 in only 9% of patients In severe Sepsis good response in 60% of patients with a progression to stage 3 of AKI in 11 to 14% we give as an infusion pump combined with: furosemide 400 mg with aminophylline 200 mg with 18% mannitol 6 mg in 500 cc of[spoiler title=” ” solution to be passed in fixed doses… Read more »
The beauty of the FST is the easy reproducibility and simplicity in performing the test with minimal cost and really not much risk for the patient. Something to debate is to add or not albumin during furosemide injection.
Hi Josh, great piece. I stumbled across this random in teaching a student who found this on google. As a nephrologist I use the FST for deciding about starting RRT or CRRT in the next 24 to 48 hours. My only question is in the final comment it says Then even small doses of furosemide (0.25 mg/kg) are protective. The mechanism probably has to do with furosemide decreasing oxygen consumption in the TAL and realigning consumption and delivery to prevent ischemia in the medulla. I’ve seen this in animal studies but there has been no proof of this in human… Read more »