Contrast media viscosity versus osmolality in kidney injury: lessons from animal studies

Abstract

Iodinated contrast media (CM) can induce acute kidney injury (AKI). CM share common iodine-related cytotoxic features but differ considerably with regard to osmolality and viscosity. Meta-analyses of clinical trials generally failed to reveal renal safety differences of modern CM with regard to these physicochemical properties. While most trials' reliance on serum creatinine as outcome measure contributes to this lack of clinical evidence, it largely relies on the nature of prospective clinical trials: effective prophylaxis by ample hydration must be employed. In everyday life, patients are often not well hydrated; here we lack clinical data. However, preclinical studies that directly measured glomerular filtration rate, intrarenal perfusion and oxygenation, and various markers of AKI have shown that the viscosity of CM is of vast importance. In the renal tubules, CM become enriched, as water is reabsorbed, but CM are not. In consequence, tubular fluid viscosity increases exponentially. This hinders glomerular filtration and tubular flow and, thereby, prolongs intrarenal retention of cytotoxic CM. Renal cells become injured, which triggers hypoperfusion and hypoxia, finally leading to AKI. Comparisons between modern CM reveal that moderately elevated osmolality has a renoprotective effect, in particular, in the dehydrated state, because it prevents excessive tubular fluid viscosity.

Figure 1

Simplified scheme summarizing major factors/mechanisms that influence tubular fluid viscosity following CM administration. For detailed explanations see text.

Figure 2

Changes of concentration and viscosity of solutions of six different contrast media (four LOCM depicted in blue: iopromide 300, iohexol 300, ioversol 300, and iomeprol 300; two IOCM depicted in red: iodixanol 320 and iosimenol 350) caused by in vitro dialysis to emulate the renal tubular concentration process. Iodine concentration (a) and viscosity (b) of the respective contrast agent solutions as marketed/formulated for clinical use (marked as CA) and after dialysis with PEG solutions with osmolalities of 290, 400, 500, 700, and 1000 mosm/kg H₂O. Data are mean ± SEM. Please note that the viscosity of both IOCM solutions after the dialysis at 700 and 1000 mosm/kg H₂O was so high that it even exceeded the upper measurement limit of the viscometer. Redrawn from data in [34].

Figure 3

Viscosity of urine samples, urine flow rate, and glomerular filtration rate (GFR; measured by creatinine clearance) in rats before (control) and following contrast media administration (six 10 min sampling periods). Iopromide 370 mg I/per mL or iodixanol 320  mg I/mL was injected into the thoracic aorta as a bolus of 1.5 mL. Rats had access to drinking water prior to the experiment but were not hydrated by infusions. Data are mean ± SEM. *P < 0.05 iopromide versus iodixanol. In all sample periods after contrast media injection, urine viscosities and urine flow rates were significantly higher than in the respective control sample. In rats receiving iodixanol, GFR was significantly lower than control GFR 10 to 40 min after iodixanol injection, whereas GFR remained unchanged in rats receiving iopromide. Note that GFR values for the first period following contrast media injection are not depicted, as high creatinine clearance values obtained for this period do not represent actual increases in GFR but rely on the dead-space effect. Redrawn from data in [39].

Figure 4

Exemplary computed tomographic (CT) scans to assess renal iodine retention and exemplary histological images (hematoxylin-eosin staining) to assess formation of vacuoles in proximal tubular cells, both taken 24 hours after injection (24 h p.i.) of either saline (a), marketed iopromide 300 mg I/mL solution (b), iodixanol 320 mg I/mL solution with mannitol added to elevate the solution's osmolality (c), or marketed iodixanol 320 mg I/mL solution (d). CM were administered intravenously at a dose of 4 g I/kg of body mass. CT scans show predominantly cortical iodine retention 24 h p.i. for the marketed iodixanol solution, less retention following the iodixanol/mannitol solution, and virtually none following iopromide and saline. Formation of vacuoles (arrows) in proximal tubular cells was prominent 24 h p.i. for the marketed iodixanol solution, slightly less following the iodixanol/mannitol solution and sparse after iopromide and saline. Reprinted from [40].

Figure 5

Renal tissue expression (mRNA levels as analysed by real-time PCR) of kidney injury molecule 1 (KIM-1) and of neutrophil gelatinase associated lipocalin (NGAL) as quantified 24 hours after injection and that of plasminogen activator inhibitor-1 (PAI-1) as quantified 2 hours after injection of either saline, marketed iopromide 300 mg I/mL solution, marketed iodixanol 320 mg I/mL solution, or iodixanol 320 mg I/mL solution with mannitol added to elevate the solution's osmolality. CM were administered intravenously at a dose 4 g I/kg of body mass. Data are mean ± SEM. *P < 0.05 versus saline. Data taken from [35, 40, 58].

Figure 6

Elevated proliferation rate of renal cells following CM administration as assessed by immunofluorescence analysis of the incorporation of bromdesoxyuridine (BrdU; red staining) into proximal tubular cells (indicated by green staining of aquaporin-1; scale bar 100 μm). The BrdU incorporation time was 46 hours, starting 2 h after i.v. injection of either saline (a), iopromide 300 mg I/mL (b), or iodixanol 320 mg I/mL solution (c); CM doses were 4 g I/kg of body mass. Semiquantitative analysis of the BrdU-incorporated cells (d). Reprinted from [58].

Figure 7

Time course of alterations in renal oxygenation as estimated by blood oxygen level-dependent (BOLD) MRI for the cortex and the medulla depicted as percentage changes (mean ± SEM) in R ₂* from baseline (100%, dotted lines). Increase in R ₂* above baseline signifies reduced (blood) oxygenation, that is, hypoxia. Intravenous injection of either saline, iopromide 300 mg I/mL, or iodixanol 320 mg I/mL was done at time 0; CM doses were 4 g I/kg of body mass. Data taken from [40].