Discussion
This study is the first one done to determine whether or not continuous infusion of 25% albumin enhances furosemide-induced diuresis in critically ill patients. We found that the effect of coadministration of furosemide and albumin was no greater than that of continuous infusion of furosemide alone.
Various beneficial mechanisms of the action of albumin beyond simple volume expansion have been described. For example, an early evaluation of furo - semide mixed with an equimolar solution of albumin in analbuminemic rats suggested that albumin might play an important role in delivering furo - semide to its site of action in the kidneys, thereby enhancing diuresis. In contrast, in a study in rats with nephrotic syndrome, the response to furosemide was compromised when the drug was given with albumin, suggesting that furosemide binding to albumin reduced the availability of the active compound. When medications were added that displaced furosemide from its albumin binding site, an improved diuretic response occurred. These conclusions suggest that the relationship between albumin and furosemide is not well understood.
The method of administration may influence the efficacy of albumin for diuresis. For example, results in normal and analbuminemic rats suggest that albumin and furosemide administered together form a complex that carries the furosemide to the kidney for uptake by renal tubular cells. Our study is the first to investigate the diuretic effects of a continuous infusion of albumin in medical ICU patients. In previous studies focused on specific populations of patients to whom albumin was administered viaintermittent boluses or short-term infusions, the results were conflicting. In a trial of 8 patients with nephrotic syndrome, Akcicek et al administered furosemide as a 60-mg intravenous bolus followed by a 4-hour intravenous infusion of furosemide with and without 20% albumin. Maximal diuretic and natriuretic responses occurred during the drug infusions for all groups. The addition of albumin had no diuretic benefit. In contrast, Fliser et al conducted a randomized, double- blind, placebo, controlled trial in 9 patients with nephrotic syndrome. Patients received a 60-mg furosemide bolus plus placebo, a 60-mg furo - semide bolus plus 200 mL of 20% albumin, or placebo given as a bolus plus 200 mL of 20% albumin. The increase in urine output between the 3 groups during the 8-hour monitoring period were significant. Fliser et al noted that the same result could possibly be achieved by optimizing the dose of furosemide alone rather than by adding albumin.
The combination of furosemide plus albumin for the management of ascites in patients with cirrhosis has also had conflicting results. Gentilini et al performed a randomized, controlled trial of 126 patients with cirrhosis and ascites. Patients were randomized to receive escalating doses of diuretics with or without 25% albumin. End points of the inpatient phase of the trial included disappearance of ascites and duration of hospital stay. The benefits of furosemide plus albumin were significantly better than the benefits of furosemide alone for both end points. In contrast to these findings, Chalasani et al found a lack of benefit with furosemide plus albumin. They performed a randomized crossover study in 13 patients with cirrhosis and ascites to evaluate the effects of albumin on the response to furosemide as indicated by urinary excretion of sodium and urine volume. Patients received each of the following intravenously over 30 minutes: 40 mg of furosemide alone, 25 g of albumin alone, 40 mg of furosemide mixed with 25 g of albumin, and 40 mg of furosemide and 25 g of albumin infused simultaneously in different arms. Urine output for furosemide alone was similar to that for either combination. The diuretic and natriuretic responses returned to normal within 6 hours of drug administration in all arms of the study. Chalasani et al concluded that coadministration of furosemide and albumin does not improve diuresis in patients with cirrhosis with ascites and most likely does not enhance diuresis in other populations of patients.
Martin et al studied furosemide plus albumin in patients with hypoproteinemia and acute lung injury treated with mechanical ventilation. In this randomized, double-blind, placebo-controlled, multicentered trial, change in oxygenation over 24 hours was the primary end point. Net fluid loss was a secondary outcome. A total of 40 patients received furosemide with either a placebo or albumin. The control group received a furosemide bolus of 20 mg followed by a titrated infusion of up to 10 mg/h of furosemide for 72 hours with a normal saline placebo substituted for an equivalent volume of albumin. The treatment group received 25 g of 25% albumin immediately before the furosemide bolus and every 8 hours thereafter for the same duration. The improvement in oxygenation and net negative fluid balance in patients who received furosemide plus albumin was significantly greater than the changes in patients given furosemide plus placebo. These data are the most compelling to date that suggest albumin may enhance the effectiveness of furosemide in a specific population of patients.
Although a large percentage of our patients had a ratio of PaO2 to fraction of inspired oxygen less than 300, we did not observe a beneficial diuretic effect with the addition of albumin. One difference in design between our study and that of Martin et al is that we used continuous infusions of albumin rather than boluses. Also, we cannot exclude a relationship between increasing albumin concentrations and increased diuretic response. This possibility might account for the net fluid loss noted at 72 hours in the study by Martin et al.
In our regression analysis, albumin concentration at 48 hours was associated with urine output, but this finding was based on data from only 5 patients. Furthermore, by 48 hours, significantly more furosemide had been administered to patients receiving furosemide plus albumin than to patients given furosemide alone, a situation that could also explain the increased urine output at this time. In our study, the one consistent factor significantly related to increased urine output was increased fluid intake.
Although hyperoncotic albumin may have potential benefits in specific populations of patients, we conclude that the addition of a 25% albumin infusion to continuous infusion with furosemide does not improve diuresis in critically ill patients with hypoalbuminemia. In agreement with previous studies, we recommend optimizing the furosemide dosing before considering the addition of colloid. Furthermore, if diuretic resistance occurs, the addition of a thiazide diuretic may produce a better result and be more cost-effective in achieving diuresis in patients with volume overload whose hemodynamic status is stable.
Our study had some limitations. Because the study was retrospective, incorrectly recorded information and uncontrolled confounders are possible. For example, details of patients' characteristics that might have influenced the response to furosemide or the combination regimen may have been missed. Because of limited enrollment, we did not perform subgroup analyses to evaluate whether albumin is beneficial in specific populations of patients in the ICU. Furthermore, the limited number of patients at the final 48-hour end point prohibits firm conclusions about an effect or lack of effect of albumin in enhancing furosemide diuresis with more prolonged administration. Finally, we evaluated the use of continuous infusions of albumin and furosemide at only a single institution, so our results may not be generalizable to other ICUs.