Mehta LM, Pascual MT, Soroko S, Chertow GM. Diuretics, mortality, and nonrecovery of renal function in acute renal failure. JAMA 2002; 288:2547-2553.
* PRACTICE RECOMMENDATIONS
Although widely used to treat acute renal failure, diuretics may actually be harmful.
The results of this observational study demonstrated a higher risk of death and nonrecovery of renal function when diuretics were initiated during the first week of hospitalization. It didn't matter whether a single or combination diuretic was used.
A randomized controlled trial would better answer this question by minimizing the inherent flaws in an observational study. Although this study doesn't conclusively prove that diuretics cause poorer outcomes, it certainly raises the possibility and should prompt us to think twice before initiating diuretic therapy for acute renal failure.
* BACKGROUND
Diuretics continue to be widely used for treating acute renal failure despite the lack of supporting evidence. The ability to promote renal salt and water excretion with diuretics and extracellular volume overload in patients with acute renal failure influences the decision of many practitioners to use these medications.
This study evaluated the effects of diuretics on mortality, renal function, and length of hospital stay in hospitalized patients with acute renal failure.
* POPULATION STUDIED
The authors studied 552 critically ill patients with acute renal failure in 4 California academic medical center ICUs. In patients without previous kidney disease, acute renal failure was defined as blood urea nitrogen >40 mg/dL or creatinine >2 mg/dL. In others, acute renal failure was defined as creatinine levels rising at >1 mg/dL compared with baseline. Patients were excluded if they had previous dialysis, urinary tract obstruction, or hypovolemia.
* STUDY DESIGN AND VALIDITY
Patients in this prospective cohort study were placed into groups according to which day diuretics were initiated during the first week following consultation. Patients were also categorized as "ever" or "never" having received diuretics. Patients received 1 or more of the following: furosemide, bumetanide, metolazone, and hydrochlorothiazide.
The researchers monitored vital signs, urine output, blood urea nitrogen, and serum creatinine levels each day until hospital discharge. They calculated disease-specific severity-of-illness scores daily in the ICU based on the number of organ systems in failure.
Because these patient groups were given therapies not randomly assigned, the researchers adjusted for confounding variables with regression methods based on propensity scores of illness severity.
The study design used in this research limits our ability to draw conclusions regarding any true causal relationship between diuretic use and poorer outcomes. A randomized controlled trial is needed to definitively establish cause and effect.
Also, the results from this study of critically ill patients cannot be generalized to patients with less severe forms of acute renal failure. The results may not apply to patients in other medical institutions where management of acute renal failure and availability of dialysis differs.
* OUTCOMES MEASURED
The primary outcomes were mortality, nonrecovery of renal function, and length of hospital stay.
* RESULTS
Of the 552 patients included in the final sample, 294 (53%) died in the hospital. 0f the 258 patients who survived, 17 required dialysis following discharge. Diuretics were used in 326 patients (59%).
Based on adjusted models, the use of diuretics was associated with a 68% increase in mortality (odds ratio [OR]=1.68: 95% confidence interval [CI], 1.06-2.64) and a 79% increase in the nonrecovery of renal function (OR=1.79; 95% CI, 1.19-2.68). Length of stay was not affected if diuretics were started on the first day of consultation (median 21.5 vs 22.5 days). However, diuretics initiated any other day during the first week prolonged hospital stays by a median of 4 to 10 days.
Patients who received diuretics at any time during that week had higher risk of death or nonrecovered renal function compared to patients who never received a diuretic (OR=3.12; 95% CI, 1.73-5.62). Patients with low urine output despite higher-dose diuretics died or needed dialysis sooner than patients who became nonoliguric with lower-dose diuretics. No significant differences in mortality, nonrecovery of renal function, and length of hospital stay occurred when comparing patients receiving combination diuretics vs single diuretics.
Lawrence Dybedock, MD, and Kevin Kane, MD, MSPH, Department of Family & Community Medicine, University of Missouri, Columbia. E-mail: dybedockl@health.missouri.edu.
COPYRIGHT 2003 Dowden Health Media, Inc.
COPYRIGHT 2003 Gale Group