Meta-analysis of frusemide to prevent or treat acute renal failure
BMJ 2006; 333 doi: https://doi.org/10.1136/bmj.38902.605347.7C (Published 24 August 2006) Cite this as: BMJ 2006;333:420All rapid responses
Rapid responses are electronic comments to the editor. They enable our users to debate issues raised in articles published on bmj.com. A rapid response is first posted online. If you need the URL (web address) of an individual response, simply click on the response headline and copy the URL from the browser window. A proportion of responses will, after editing, be published online and in the print journal as letters, which are indexed in PubMed. Rapid responses are not indexed in PubMed and they are not journal articles. The BMJ reserves the right to remove responses which are being wilfully misrepresented as published articles or when it is brought to our attention that a response spreads misinformation.
From March 2022, the word limit for rapid responses will be 600 words not including references and author details. We will no longer post responses that exceed this limit.
The word limit for letters selected from posted responses remains 300 words.
Acute renal failure in the hospital is associated with a high rate of
mortality. As oliguric renal failure is considered a poor prognostic sign,
many clinicians administer diuretics to increase urine output despite a
lack of evidence demonstrating true benefit.(1) In their recent meta-
analysis, Ho and Sheridan conclude that frusemide does not have
significant benefit in prevention or treatment of acute renal failure and
may lead to ototoxicity. (2) These results are based on a number of small
randomized control trials and are in agreement with a larger observational
study which found that diuretics may cause harm. (1) These conclusions
also echo a recent meta-analysis on another unproven pharmacotherapy for
acute renal failure: low dose dopamine. In that study, Friedrich et al.
showed that low dose dopamine increases urine output but has no effect on
overall renal function or mortality.(3) The message appears to be that
while both frusemide and dopamine may increase urine output, neither
confers clinical benefit.
Although diuretics may not provide renal protection or improved
survival, experts have recommended such treatment for fluid management in
oliguric acute tubular necrosis (4) with a presumed goal of shorter time
to discharge. Ho and Sheridan conclude that frusemide actually increases
hospital stay, although this data was based on only two studies, both on
prevention rather than treatment of renal failure. We believe that this
conclusion requires further study in studies of treatment before it is
substantiated by data.
Finally, we have some concern regarding statistical heterogeneity in
the meta-analysis. The authors report significant heterogeneity for
outcomes other than in-hospital mortality and ototoxicity. However, they
report I-squared of 0% both for the number of dialysis sessions required
and for the length of hospital stay. More importantly, the proportion of
patients requiring renal replacement therapy – one of the primary outcomes
measured – and proportion of patients who remain oliguric had I-squared
values of 64% and 91%, respectively, representing significant
heterogeneity. The authors do not address possible causes of this
heterogeneity and proceed with pooling of these potentially very different
trials in meta-analysis. While we laud the author’s attempts at reporting
as much data as possible, such differences should be explored further and
interpreted with caution. (5)
Saul Blecker, MD
Nirav R. Shah, MD, MPH
Department of Internal Medicine, New York University School of Medicine
References
(1) Mehta RL. Pascual MT. Soroko S. Chertow GM. PICARD Study Group.
Diuretics, mortality, and nonrecovery of renal function in acute renal
failure. JAMA. 2002;288:2547-53.
(2) Ho KM. Sheridan DJ. Meta-analysis of frusemide to prevent or
treat acute renal failure. BMJ. 2006;333:420,
doi:10.1136/bmj.38902.605347.7C
(3) Friedrich JO. Adhikari N. Herridge MS. Beyene J. Meta-Analysis:
Low-Dose dopamine increases urine output but does not prevent renal
dysfunction or Death. Ann Inten Med. 2005; 142:510-524.
(4) Rose, BD. Pathogenesis and prevention of postischemic acute
tubular necrosis. In: UpToDate, Rose, BD (Ed), Waltham, MA, 2006.
(5) Higgins JPT, Green S, editors. Cochrane Handbook for Systematic
Reviews of Interventions 4.2.5 [updated May 2005].
http://www.cochrane.org/resources/handbook/hbook.htm. Accessed 5th Sept
2006.
Competing interests:
None declared
Competing interests: No competing interests
Thank you for their interest in our meta-analysis and the questions
raised by Drs. Bagshaw, Delaney and Bellomo.
First, the trials that evaluated frusemide in the prevention or
treatment of acute renal failure were stratified in the meta-analysis with
the results separately reported in the forest plots. Both prevention and
treatment trials were considered in this meta-analysis because there is
experimental evidence to support the use of frusemide before acute renal
insult occurs. The interaction between the results of the two strata of
trials (prevention or treatment) was tested statistically and the results
were not significantly different.
Second, contrary to their interpretation of the meta-analysis, we
have acknowledged the limitations very clearly in the method, result, and
discussion section of the meta-analysis. We specified a priori criteria
for trial inclusion in this meta-analysis and as such, we had included a
study of single dose of frusemide compared with prolonged continuous
infusion (as stated on line 8 in the method section) in the initial
analysis. However, this study was excluded in the sensitivity analysis and
this did not change the magnitude and direction of the results.
Heterogeneity was observed in all outcomes except hospital mortality and
ototoxicity (as stated on the first line of the third paragraph in the
result section) and also the I2 values (which signify the degree of
inconsistency) were reported in the results. Furthermore, we also made it
very clear that the sample size of pooled studies (n=204) in this meta-
analysis might still be too small to exclude a small protective effect
(relative risk <30%) of frusemide on the risk of renal replacement
therapy. This point was discussed when the limitations of the meta-
analysis were discussed in the discussion section of the paper. Meta-
analyses are prone to bias. A separate sensitivity analysis that included
higher quality studies with adequate allocation concealment could not
demonstrate beneficial effects of frusemide in the setting of acute renal
failure.
Third, frusemide had been evaluated to prevent acute renal failure
and treatment of early phase and also in established phase of acute renal
failure in the included studies. Frusemide has been shown to be useful in
preventing renal tubular injuries in isolated perfused kidney. The limited
data available in the literature, however, do not suggest frusemide is
more useful in a specific phase of acute renal failure.
Fourth, it might be argued that a correction should be made to the
overall significance level and confidence interval to account for the
pairwise comparisons between the three treatment arms in one study. Use of
a Bonferroni correction would result in widened confidence intervals and a
rejection of the null hypothesis at smaller p values. In the present
study, which used a nominal 5% level with Bonferroni correction, a test of
the relative risk being different from 1 would be regarded as significant
only if p<0·017 (ie, 0·05 divided by 3). Based on this method, the
significance of all p-values would remain the same for the outcomes that
included this study.
Finally, we agree with Dr. Bagshaw, Delaney, and Bellomo that a large
randomised controlled trial is needed to confirm the results of our meta-
analysis. Until we have the results of a large randomised controlled
trial, clinicians should be cautious when high doses of frusemide (1 to
3.4g per day) are used to treat acute renal failure.
Competing interests:
None declared
Competing interests: No competing interests
Dear Editors:
We read with interest the meta-analysis by Ho et al regarding the use
of frusemide in the prevention or treatment of acute renal failure.1 We
would like to raise several concerns we have with their meta-analysis.
First, the reporting of a pooled analysis of trials that use
frusemide both for the prevention and for the treatment of acute renal
failure may have the consequence of generating confusion and lead to
erroneous conclusions.
Second, the authors have failed to adequately acknowledge the
limitations of the included trials. These trials inconsistently report
both on primary outcomes and adverse effects. Likewise, all patients in
two trials and the majority in another were already receiving renal
replacement therapy at inclusion, thus making pooled estimates on this
outcome invalid.2-4 Furthermore, these trials are generally small and of
poor overall quality. Lastly, most of these trials were published several
decades ago.2 4 5 The approach to, and technical capabilities for,
management of acute renal failure have evolved.
Third, the authors failed to specify a priori criteria for trial
inclusion. As a result, the trial by Brown et al 5 has been included in
their pooled analysis. This trial is a comparison of two frusemide
regimens rather than with a placebo control. They have duplicated the
patients allocated to control in a study by Cantarovich et al for the
assessment of hospital mortality.2 Both of these will unduly influence
the event rates and summary effect estimates. Finally, when assessing
need for renal replacement therapy or the proportion of patients remaining
oliguric, the authors have failed to account for or consider the observed
heterogeneity across studies.
From these findings, we can only conclude that it is not possible to
draw any strong conclusions and that a more cautious interpretation is
warranted, in particular for those trials examining whether treatment with
frusemide impacts outcome. These discrepant findings taken with those
from larger observational studies, suggest there exists a genuine
equipoise on the efficacy of frusemide.6 7
Thus, a large and suitably power multi-centre randomized controlled
trial is needed to definitively resolve the question of the role of
frusemide in the management of acute renal failure. Such a trial should
ideally incorporate clinically relevant and patient-centered outcomes as
well as important secondary outcomes addressing issues of harm,
physiologic effects (i.e. urine output) and dose-response.
Sean M Bagshaw, MD, MSc
Department of Intensive Care, Austin Hospital, Heidelberg, Victoria,
Australia
Rinaldo Bellomo, MD
Department of Intensive Care, Austin Hospital, Heidelberg, Victoria,
Australia
Anthony Delaney, MBBS, MSc
Department of Intensive Care, Royal North Shore Hospital, St. Leonards,
New South Wales, Australia
References:
1. Ho KM, Sheridan DJ. Meta-analysis of frusemide to prevent or treat
acute renal failure. BMJ 2006.
2. Cantarovich F, Fernandez JC, Locatelli A, Perez Loredo J.
Frusemide in high doses in the treatment of acute renal failure.
Postgraduate Medical Journal 1971;47:Suppl:13-7.
3. Cantarovich F, Rangoonwala B, Lorenz H, Verho M, Esnault VL. High-
dose furosemide for established ARF: a prospective, randomized, double-
blind, placebo-controlled, multicenter trial. Am J Kidney Dis
2004;44(3):402-9.
4. Kleinknecht D, Ganeval D, Gonzalez-Duque LA, Fermanian J.
Furosemide in acute oliguric renal failure. A controlled trial. Nephron
1976;17(1):51-8.
5. Brown CB, Ogg CS, Cameron JS. High dose frusemide in acute renal
failure: a controlled trial. Clinical Nephrology 1981;15(2):90-6.
6. Mehta RL, Pascual MT, Soroko S, Chertow GM, Group PS. Diuretics,
mortality, and nonrecovery of renal function in acute renal failure. JAMA
2002;288(20):2547-53.
7. Uchino S, Doig GS, Bellomo R, Morimatsu H, Morgera S, Schetz M, et
al. Diuretics and mortality in acute renal failure. Crit Care Med
2004;32(8):1669-77.
Competing interests:
None declared
Competing interests: No competing interests
Thank you for the question raised by Dr. Hayanga regarding the
possibility of contraction alkalosis caused by prolonged frusemide
infusion.
First, acid base status was seldom reported in many trials in which
frusemide was evaluated. It was impossible to pool results that were not
reported in individual clinical trials together in a meta-analysis.
Second, biochemical complications such as alkalosis or hypokalaemia
induced by frusemide, either by a bolus injection or by a prolonged
infusion, are not patient orientated outcomes. They are amendable to
correction by avoiding hypovolaemia or by replacing the potassium
deficiency, respectively. Furthermore, for the trials that evaluated
frusemide as a preventive agent in patients at risk of acute renal
failure, hypovolaemia was prevented by infusing intravenous fluid and
there was no significant difference in the changes of the total body
weight before and after treatment with either the placebo and frusemide.
And as such, contraction alkalosis was unlikely to be a significant
biochemical complication in these patients. On the other hand, ototoxicity
is a patient orientated clinical outcome and the associated symptoms can
be very disturbing to our patients, especially if they are irreversible.
Finally, there was no signficant improvement in the proportion of patients
with persistent oliguria in patients who already had acute renal failure
as shown by our meta-analysis. Therefore, it would be unlikely that
frusemide infusion in the pooled studies could induce significant
alkalosis as the urine output did not increase significantly. Contraction
alkalosis would not have happened after the use of frusemide if
hypovolaemia was not induced in these patients.
Competing interests:
None declared
Competing interests: No competing interests
Meta analysis is considered the highest order of evidence and the
definitive source of conclusive data. Ho and Sheridan [1] failed to
report on a more practical and common side effect of prolonged
intravenous frusemide use, namely the acid –base abnormality that was
previously considered under a Henderson-Hasselbach understanding as a
“contraction alkalosis.” This alkalosis is maintained by volume
contraction, decreased glomerular filtration rate and hypokalemia and can
be corrected by the use of a carbonic anhydrase inhibitor.[2] This
alkalosis has, however, more recently been elucidated by the Strong
Ion Difference theories and the Stewart Model. [3,4] A preoccupation
with the lesser observed and even lesser described side effect of
ototoxicity appears to have diverted the author’s attention from what
would have been a more practical and clinically relevant discussion
about this side effect that would have been of greater clinical value
not just within Critical Care but clinical practice at large.
REFERENCES
1 Ho KM, Sheridan DJ. Meta-analysis of frusemide to prevent or treat
acute renal failure. BMJ. 2006 Jul 21; [Epub ahead of print]
2 Gattinoni L, Carlesso E, Cadringher P, Caironi P. Strong ion
difference in urine: new perspectives in acid-base assessment. Crit Care.
2006 Apr 7;10(2):137 [Epub ahead of print]
3 Derksen R, Scheffer GJ, van der Hoeven JG. Quantitative acid-
base physiology using the Stewart model. Does it improve our understanding
of what is really wrong?Eur J Intern Med. 2006 Aug;17(5):330-3
Competing interests:
None declared
Competing interests: No competing interests
Frusemide to prevent or treat acute renal failure
Dear Editors,
we read with interest the meta-analysis by Ho et al regarding the
role of frusemide in the setting of acute renal failure.1 Although we do
not disagree with the conclusion that frusemide does not cure or prevent
acute renal failure, we have some concerns about how this conclusion was
reached and the overall message it sends to medical practioners.
Firstly, none of the studies used the same criteria for the
definition of acute renal failure. As a result, patients with different
degrees of renal failure were included, ranging from patients with a serum
creatinine < 177 micromol/L to patients already on renal replacement
therapy. Previous studies, including our own, have shown that outcome of
acute renal failure clearly depends on how it is defined and the timing of
its onset.2-4
Secondly, it is important to acknowledge that frusemide still has a role
in the treatment of patients with acute renal failure who are fluid
overloaded. In this situation, treatment with a diuretic will not cure
acute renal failure and may also not avoid renal replacement therapy
completely but it may induce diuresis, prevent life-threatening pulmonary
oedema and hyperkalaemia and overall make acute renal failure more
manageable. This is particularly important in situations when renal
replacement therapy is not immediately available. Being able to induce
diuresis can buy enough time to transfer a patient safely to a setting
(occasionally to another hospital) where renal replacement therapy can be
provided. Finally, the meta-analysis is based on the results of 9 studies
of which 4 were conducted more than 25 years ago and may not be relevant
to modern clinical practice any longer.
We feel strongly that these points should be remembered before
frusemide gets abandoned from the management of patients with acute renal
failure.
References
1. Ho KM, Sheridan DJ. Meta-analysis of frusemide to prevent or treat
acute renal failure. BMJ 2006:333;420-425
2. Novis BK, Roizen MF, Aronson S, Thisted RA. Association of
Preoperative Risk Factors with Postoperative Acute Renal Failure. Anesth
Analg 1994;78: 143-149
3. Ostermann ME, Chang RW. Prognosis of acute renal failure: an
evaluation of proposed consensus criteria. Intensive Care Med 2005;31:250-
256
4. Guerin C, Girard R, Selli JM, Perdrix JP, Ayzac L. Initial versus
Delayed Acute Renal Failure in the Intensive Care Unit. Am J Respir Crit
Care Med 2000;161: 872-879
Competing interests:
None declared
Competing interests: No competing interests