Background: Contrast-induced nephropathy is common in patients with coronary angiography. Mechanistically, forced euvolemic diuresis with mannitol and furosemide ought to prevent contrast-induced nephropathy. Our objectives are to: (1) undertake a randomized trial testing this hypothesis, and (2) conduct a meta-analysis of our findings with 2 earlier studies.
Study design: (1) Randomized allocation-concealed controlled trial with blinded ascertainment of outcomes, and (2) random-effects meta-analysis of 3 trials.
Setting & participants: Single-center study of consenting adults with serum creatinine level greater than 1.7 mg/dL undergoing coronary angiography; patients unable to tolerate fluid challenge or receiving dialysis were excluded. Two previous trials had randomly assigned 159 patients.
Intervention: Forced euvolemic diuresis with saline, mannitol, and furosemide compared with saline hydration controls. All patients were pretreated with at least 500 mL of half-normal saline before angiography; during and 8 hours after, urine output was replaced milliliter per milliliter with half-normal saline.
Outcomes & measurements: The primary outcome was contrast-induced nephropathy within 48 hours of the procedure, defined as a 0.5-mg/dL absolute or 25% relative increase in creatinine level.
Results: Overall, 92 patients were allocated to intervention (n = 46) or control (n = 46). Mean age was 64 +/- 14 (SD) years, 23% were women, 37% had diabetes, 47% used oral furosemide, mean creatinine level was 2.8 +/- 1.6 mg/dL, and most patients (72%) underwent diagnostic catheterization. Patients had a net positive fluid balance (389 +/- 958 mL for intervention versus 655 +/- 982 mL for controls; P = 0.2). Contrast-induced nephropathy occurred in 23 (50%) intervention patients versus 13 (28%) controls (relative risk, 1.77; 95% confidence interval, 1.03 to 3.05; P = 0.03; adjusted odds ratio, 3.73; P = 0.03). Within 48 hours, creatinine level had increased by 0.8 +/- 1.1 mg/dL with intervention versus 0.2 +/- 0.6 mg/dL for controls (P = 0.002). Overall, 11 (12%) patients died or required dialysis, with no difference according to allocation status (P = 0.5). Random-effects meta-analysis of published data (3 trials; 251 patients) suggests furosemide-based interventions lead to significant harm compared with hydration: pooled relative risk, 2.15; 95% confidence interval, 1.37 to 3.37; I(2) = 0%.
Limitations: Small single-center study that cannot determine whether harms were related to furosemide, mannitol, or a combination.
Conclusions: Forced euvolemic diuresis led to a significantly increased risk of contrast-induced nephropathy. This strategy should be abandoned, and our results suggest that oral furosemide therapy perhaps should be held before angiography.
Trial registration: ClinicalTrials.gov NCT00175227.