Randomised trial of high-dose isosorbide dinitrate plus low-dose furosemide versus high-dose furosemide plus low-dose isosorbide dinitrate in severe pulmonary oedema 

Authors Cotter G, Metzkor E, Kaluski E, Faigenberg Z, Miller R, Simovitz A, Golik A, et al 
Source Lancet. 351:389-93. February 7, 1998. 
Institutions Multi-institutional in Israel. 
Support None indicated. 


The standard treatment of acute pulmonary edema includes the administration of morphine, oxygen, diuretics and nitrates. Diuretics such as furosemide are thought to acutely lower preload by a veno-dilatory effect, followed later by diuresis. Diuretics can cause a reflex increase in heart rate, however, which can aggravate cardiac ischemia. Nitrates at low doses are predominantly preload reducers; at high doses they also act as general and coronary arterial dilators, which could be beneficial in cardiogenic pulmonary edema. This study was designed to compare the effects of high dose nitrates and low dose furosemide with low dose nitrates and high dose furosemide in pulmonary edema. 



Patients were recruited from the emergency medical services of three cities in Israel. 

Inclusion criteria: pulmonary edema by chest x-ray and O2 saturation (by pulse oximetry) less than 90%. 

Exclusion criteria: treatment with nitrates at more than a low dose; treatment with furosemide at more than 80 mg/day and BP less than 110/70. 


Patients were randomised to either high-dose nitrates and low dose furosemide (Group A) or to high-dose furosemide and low dose nitrates (Group B). 

All patients received oxygen (10 l/min by face-mask), 3 mg of morphine intravenously and 40 mg of furosemide intravenously. 

Heart rate, BP, respiratory rate and O2 saturation were determined at baseline and every three minutes during treatment. 

Patients randomised to the high dose nitrates group received 3 mg of isosorbide dinitrate intravenously every 3 minutes. 

Patients randomised to the high dose furosemide group received 80 mg of furosemide iv every 15 minutes. In addition, they received an infusion of (low dose) isosorbide dinitrate: starting at 1 mg/hour and increased by 1 mg/hour every 10 minutes. 

The nitrates/furosemide treatment was continued until O2 saturation increased to at least 96% or mean blood pressure fell by 30% or to below 90 mmHg. 

Patients were intubated if their O2 saturation fell below or remained below 80%, or if their respiratory distress worsened. 

ECG and CPK levels were performed on admission and 24 hours later; all patients underwent echocardiography during their hospital stay. 


Primary outcome measures were death, need for intubation within 12 hours and development of myocardial infarction. 

Other parameters followed were blood pressure reduction, arrhythmias, heart rate and O2 saturation during the first hour. 




During one year starting July 1, 1996, 446 patients were screened. 217 were excluded because their heart failure was either too severe (requiring immediate intubation) or too mild; 119 were excluded because of treatment with higher doses of furosemide or nitrates, or low BP on presentation; 6 were excluded after randomisation because chest x-ray did not confirm pulmonary congestion. Thus, 104 patients were randomised and studied, 52 in each group. 

Baseline characteristics were well-matched between the groups. Some of these (both groups combined unless otherwised specified) included: 

  • Age: 74; male sex 52%; hypertensive 49%; diabetes 40%; coronary disease 65%. 
  • Mean arterial BP: 132 mmHg group A; 124 mmHg group B. Mean heart rate 115; O2 saturation 78%. 
  • Treatment with nitrates prior to evaluation: 80%; treatment with furosemide: 86%. 
  • EF by echocardiography: 42%. Aortic stenosis was present in 11%; mitral regurgitation in 27% group A; 17% group B. 
Primary endpoints

Mechanical ventilation was required in 7 patients in the high dose nitrates group, vs. 21 in the high dose furosemide group (13% vs. 40%; p=0.004). 

Myocardial infarction occurred in 9 vs. 19 patients (17% vs. 37%; p=0.047). 

There was one death in the high dose nitrates group vs. 3 in the high dose furosemide group (p=0.61, NS). 

Other parameters

The mean dose of isosorbide dinitrate administered in group A was 11.4 mg vs. 1.4 mg in group B. The corresponding furosemide doses were 56 mg and 200 mg. 

Mean BP dereased from 132 to 107 in group A (a reduction of 19%), and from 124 to 103 in group B (a 15% reduction). The difference between the groups was not significant. An excessive drop in mean BP (greater than 30% or below 90 mmHg) occurred in 10% of patients in group A and 13% in group B, and did not require any specific measures. 

Heart rate after treatment decreased by 15 bpm in group A and by 9 in group B (p=0.024). Respiratory rate after treatment decreased by 11 per minute in group A vs. 5 in group B (p<0.0001). O2 saturation increased to 96% in group A and to 92% in group B (p=0.0063). 

Author's discussion

According to the authors, the treatment protocol was designed to prevent the occurrence of significant hypotension, which can jeopardize the perfusion of ischemic myocardium. They administered both furosemide and isosorbide dinatrate to both groups, at different doses, because they felt that neither medication could be ethically withheld. The mean dose of furosemide given in the high dose nitrates group was higher than specified by protocol (56 mg rather than the specified 40 mg) because of several protocol violations by physicians who felt that the patients "needed" more furosemide. 

Although both groups received both drugs, the authors point out that the dose of furosemide given in the high dose nitrates group was similar to these patients' daily furosemide dose, and the dose of isosorbide given to patients in the high dose furosemide group was also similar to their previous nitrate intake. Finally, and perhaps most importantly, the authors note that nitrates have different effects at low and high doses. At low dose, they act principally as veno-dilators, decreasing preload. At high doses, they also act as arteriodilators for both the systemic and coronary circulations, producing afterload reduction and reducing ischemia. It is the arteriodilatory effect of nitrates at high doses that the authors feel is responsible for the beneficial effect in group A, in this study. They also believe that the effect of nitrates in reducing ischemia may be responsible for the decrease in myocardial infarctions in the high dose nitrate group. 



This study compared high dose isosorbide dinitrate, given as 3 mg boluses every 5 minutes plus standard dose furosemide, to high dose furosemide plus a low dose infusion of isosorbide, for the treatment of acute pulmonary edema. The authors found a significant reduction in the need for intubation and in the rate of myocardial infarction in the high dose nitrates group. 

From a methodological standpoint, the study appears to have been unblinded (or at most single-blinded). This important point is not stated explicitly anywhere in the article but can be inferred from the protocol and should have been addressed in the discussion. Significant bias can be introduced by lack of blinding. For example the decision to intubate a patient could easily be influenced by knowledge of treatment assignment. 

This problem aside, studies that look at different combinations of accepted therapies are often harder to find funding for than those looking at new and often expensive treatments. Comparative studies of conventional approaches are much needed, however, and have the potential to produce significant clinical benefit at comparatively low cost. From this standpoint, I believe this is a valuable trial. 

The direct applicability of this trial to clinical practice in the United States is limited, since bolus administration of intravenous isosorbide dinitrate is not used here. Further studies looking at other modalities of nitrate delivery and at varying doses will be needed and hopefully will be undertaken. 

April 8, 1998 


References related to this article from the NLM's PubMed database. 

Reader Comments

Date: Mon, 25 May 1998
From: "J.M. Pontious MD" <michael-pontious@ouhsc.edu>

With this review, I sense that there are significant selection biases up front in determining which patients to include in the study.  Anytime that I hear selection criteria such as "considered to have CHF that was too severe and would probably require intubation"...gives me some anxiety from the onset.

The other issue (in addition to the fact that I can't give my patient bolus intravenous isosorbide dinitrate) is how do I make the "leap of faith" in treating my patients?  I note that the PubMed references discuss Sublingual NTG, is it equivalent?  IV NTG only, or would acute use of NTG paste work as well?

What dose of furosamide are they looking at for the low and high dose wings of the studies?

J.Michael Pontious M.D.
Associate Professor
Department of Family and Preventive Medicine
Program Director
OU/Enid Family Medicine
Enid, Oklahoma

    The exclusion of patients who were considered to require urgent intubation certainly is a subjective one. However, since these patients were excluded before randomization, it should not have introduced bias in one direction or another. 

    The problem of transferring the results to any other form of nitroglycerin is a major one, which greatly limits the applicability of this study.

    The mean furosemide dose administered in the low dose group was 56 mg, in the high dose group it was 200 mg.  -- mj

Date: Wed, 23 Dec 1998
From: "Dr.D.Prabhakar" <prabhud@giasmd01.vsnl.net.in>


I have noticed that the use of nitroglycerin at the maximum tolerated dose titrated to clinical endpoints along with a maximal dose of furosemide is the best combination. I do not find any harm in giving both in large doses. What about the role of morphine ? My experience is it produces dramatic results.

Coronary Care Unit
ChennaiMedical College
Chennai - INDIA - 600 003

I would agree that most of us who treat acute pulmonary edema tend to use high doses of nitrates, furosemide and often morphine, with acceptable results. For a condition that is so common, however, there is little data on the relative merits of these various therapies, which is what makes this study interesting.  -- mj

Date: September 8, 1999
From: R.Kumm [r.kumm@ukf.net]

Great !

Especially in situations with high blood pressure and beginning pulmonary oedema / asthma cardiaque I have always been using GTN first line; in a wet lung simultaneously GTN plus Furosemide 60mg i.v. bolus.

As GTN should work faster, this has priority with me. I have both in my home visiting bag.

Kind regards


Date: November 11, 1999
From: Dr. Bernard Silke [b.silke@qub.ac.uk]

There is a large haemodynamic data base on the effects of nitrates in acute myocardial infarction. As the authors state, the low-dose therapy acts mainly as a venodilator, but in the clinical dose-range, there are usually systemic arteriolar actions. Frusemide also acts as a venodilator to reduce cardiac pre-load. The benefits on cardiac pumping function depend on the degree of initial cardiac stretch and PCW. Lowering the 'wedge' into the physiological range (say 15 - 20 mmHg) from higher levels (20 - 30 mmHg) will usually increase cardiac output, whereas excessive pre-load reduction may, because of the Starling mechanism, lead to a further fall in the cardiac output. Without titration (based on Swan - Ganz measurements), the degree of preload reduction achieved is unpredictable. The hazard in using both loop diuretics and nitrates together are excessive pre-load falls with a lower cardiac output than would be achieved from each therapy used in isolation. In an ideal world, with a knowledge of the PCW, one could dose-titrate : clearly this is unrealistic in the real world of clinical medicine.

There are studies on the GTN patch in AMI, and on most types of nitrates. All types can be shown to be haemodymically effective. The optimal combination to use is a matter of empiricism and clinical experience; however in principle large doses of nitrates and loop diuretics used together may not always be a good plan!

Dr. Bernard Silke, M.D., D.Sc., F.R.C.P.,
F.R.C.P.(I), F.R.C.P.(Edin), F.E.S.C.,
Reader in Pharmacology & Therapeutics,
The Queen's University of Belfast.
Consultant Physician, Belfast City Hospital.


Avoiding excessive preload reduction is definitely important.  In the absence of a Swan-Ganz catheter, closely monitoring BP is necessary.  Also, it should be noted that this was a trial of furosemide and nitrates in pulmonary edema, not in acute MI (although about a third of the patients in this trial did have an infarct).  Patients with acute ischemic syndromes may have stiffer ventricles and may be particularly sensitive to preload.  -- mj


Submit a comment

Journal Club homepage

Site Meter