Reports

This report is based on medical evidence presented at sanctioned medical congress, from peer reviewed literature or opinion provided by a qualified healthcare practitioner. The consumption of the information contained within this report is intended for qualified Canadian healthcare practitioners only.

Transcatheter Cardiovascular Therapeutics 2006

Washington, DC / October 22-27, 2006

Reported by: James A. Tumlin, MD

Professor of Medicine Renal Division Emory University School of Medicine Atlanta, Georgia

Epidemiologic studies of contrast-induced nephropathy (CIN) have yielded somewhat contradictory results. Some evidence suggests the frequency of the complication has decreased by more than 50% in the past decade, from about 15% to about 7%, owing to increased awareness of CIN, better preventive measures and improved contrast agents (Bartholomew et al. Am J Cardiol 2004;93(12):1515-9). However, the growing volume of contrast-requiring procedures may prevent the incidence of CIN from decreasing any further. Two different studies conducted several years apart found evidence that CIN may occur in as many as 15% of unselected cardiology patients undergoing imaging studies requiring contrast (McCullough et al. Am J Med 1997;103(5):368-75, Iakovou et al. J Invasive Cardiol 2003;15(1):18-22).

CIN confers a substantial risk of morbidity and mortality. Multiple studies have demonstrated that CIN significantly increases the risk of complications and death (McCullough et al. Am J Cardiol 2006;98(6A):5K-13K). In one investigation involving >16,000 patients, in-hospital mortality was 34% in patients who developed CIN compared to 7% in patients who did not (Levy et al. JAMA 1996;275(19):1489-94).

Some studies have suggested that CIN leading to dialysis occurs in <1% of patients. Other studies have indicated the dialysis rate might be as high as 3%, particularly among patients undergoing percutaneous coronary revascularization procedures. Patients with underlying renal impairment have a threefold greater risk of CIN-associated dialysis compared to patients with normal baseline renal function (McCullough et al. Am J Cardiol 2006;98(6A):5K-13K). The need for dialysis has dire prognostic implications. In a study of patients undergoing percutaneous coronary intervention (PCI), researchers reported an in-hospital mortality of 27.5% in patients requiring dialysis after PCI compared to 1% in those who did not undergo dialysis (P<0.0001) (Gruberg et al. Catheter Cardiovasc Interv 2001;52(4):409-16). One-year mortality was 54.5% in the dialysis patients vs. 6.4% in patients who did not require dialysis (P<0.0001). Significantly more dialyzed patients had myocardial infarctions and event-free survival was about half that of the non-dialyzed patients.

In addition to its physical toll, CIN adds substantially to the cost of health care. A study of 600 hospitalized diabetic patients showed that the average duration of hospital stay was 17 days in those who developed CIN vs. 10 days for those who did not. Hospital charges were threefold higher in the CIN group (ASN 2002, Philadelphia, PA, Abstract 447A). Measures currently thought to reduce the risk of CIN include judicious use of contrast media, optimal hydration, prophylaxis with N-acetylcysteine (NAC) and use of an iso-osmolar contrast agent, such as iodixanol, which is considered to be less nephrotoxic.

A study reported at the ASN 2005 meeting suggested that use of iodixanol might reduce health care costs despite the higher upfront cost of the contrast agent itself. A decision analysis model based on published literature was used to compare the cost-effectiveness of routine use of iodixanol and a low-osmolar contrast agent. The analysis revealed a net savings of $410 per patient and a net increase in quality-adjusted life-years with iodixanol in patients who had a moderate to high baseline risk of CIN.

Various properties of iodinated contrast media may contribute to the risk of CIN. These properties include osmolality, viscosity and ionic vs. nonionic. As a consequence, the choice of contrast medium affects the risk of CIN. A considerable amount of evidence shows that low-osmolar contrast media are less nephrotoxic than high-osmolar agents. Current evidence suggests that nonionic iso-osmolar contrast poses the lowest risk of CIN in patients with chronic kidney disease, particularly those with diabetes. Within-class comparisons have shown that higher volumes of iodinated contrast medium are associated with an increased risk of CIN (Davidson et al. Am J Cardiol 2006; 98(6A):42K-58K).

Small, prospective clinical studies, typically involving about 100 patients, have compared different contrast agents in patient populations with a high risk for CIN. Some of them demonstrated significant differences in CIN risk among contrast agents, but the results were not consistent. A recent meta-analysis has shed more light on safety differences among iodinated contrast agents, specifically the risk of CIN (McCullough et al. J Am Coll Cardiol 2006;48(4):692-9). This analysis comprised 16 randomized, double-blind, controlled clinical trials involving 2727 patients undergoing diagnostic and interventional procedures requiring contrast.

The patients were randomized to intra-arterial iodixanol or a low-osmolar contrast agent. Patients who received iodixanol had a significantly lower mean maximum increase in serum creatinine (SCr) compared to those who received low-osmolar contrast (P<0.001). Differences in SCr were particularly striking in patients with chronic kidney disease (CKD) and those with CKD and diabetes. The incidence of CIN (defined as an increase in SCr ³0.5 mg/dL within three days after administration of contrast) was significantly lower in all patients treated with iodixanol (1.4% vs. 3.5%, P<0.001), patients with CKD and patients with CKD and diabetes. Use of low-osmolar contrast was an independent predictor of an increased risk of CIN.

CIN cannot be reversed, but various prophylactic strategies can reduce the risk. A recent review of the origin, evolution and treatment of CIN showed that the risk of the complication is directly related to the degree of pre-existing renal insufficiency (Rudnick et al. Cleve Clin J Med 2006;73(1):75-80, 83-7). Of the numerous prophylactic strategies that have been evaluated, hydration with normal saline is the most widely accepted. Authors of the review also found that conflicting results have arisen with NAC; sodium bicarbonate may have some value, but requires validation in large clinical trials; the logistical effort and high cost of hemofiltration mandate thorough assessment of the prophylactic strategy in large, prospective clinical trials; and theophylline cannot yet be recommended for prophylaxis of CIN.

The authors of the review further concluded that newer nonionic, low-osmolar contrast agents pose less of a risk of CIN compared to older agents but are not risk-free. In addition, the authors stated that currently available evidence suggests that iso-osmolar contrast agents may offer advantages over other types of contrast media.

Presentations here at the TCT meeting continued the examination and discussion of iodinated contrast media and the risk of CIN. The results added to the knowledge base surrounding the issue, but the studies should not be construed as bringing a definitive resolution to questions regarding the relative safety of contrast agents.

New Data

Investigators in two different clinical studies compared iodixanol and low-osmolar contrast agents in patients undergoing coronary angiography or PCI. Both studies showed no significant differences in the primary clinical end point of CIN in patients who received low- or iso-osmolar contrast media.

The CARE (Cardiac Angiography in Renally Impaired Patients) study involved 482 patients considered at high risk for CIN because of an estimated glomerular filtration rate (eGFR) £60 mL/min. The patients were randomized to coronary angiography with iodixanol or iopamidol. All patients received prophylaxis against CIN in the form of a standard protocol of volume expansion with sodium bicarbonate and selected patients also received NAC. The study was the largest-ever direct comparison of the two contrast agents, noted Dr. Richard Solomon, University of Vermont, Burlington.

Patients with class IV heart failure were excluded from the study. No significant differences in baseline characteristics existed between the two treatment groups. The primary end point was the proportion of patients who developed CIN, as defined by three different measures: ³0.5 mg/dL increase in SCr, ³25% increase in SCr, or ³25% decrease in eGFR.

Dr. Solomon reported similar rates of CIN in the treatment groups, regardless of the definition employed. For example, when an increase in SCr ³0.5 mg/dL was used to define CIN, the rates were 4.4% in the iopamidol group and 6.7% in the iodixanol group (P=0.39). A planned subgroup analysis of diabetic patients (about 40% of the entire study population) also showed no significant differences in the rate of CIN. Patients in the iopamidol group did have a significantly lower mean rise in SCr.

The ICON (Ionic Versus Nonionic Contrast to Obviate Worsening of Nephropathy after Angioplasty in Chronic Renal Failure Patients) trial involved 145 patients, all of whom had renal insufficiency associated with SCr values between 1.5 and 3.0 mg/dL, reported Dr. Roxanna Mehran, Columbia University, New York. Almost half the patients (46%) were diabetic and 88% were hypertensive. The mean creatinine clearance of the 145 patients was 45 mL/min.

The patients were randomized to coronary angiography with iodixanol or ioxaglate, which was performed with hydration involving 3.7 L of fluid. Individual investigators at the seven participating centres in Canada and the US could administer NAC at their discretion and 72% of patients received prophylaxis with NAC.

Development of CIN was defined as new-onset or exacerbation of renal dysfunction following contrast administration, associated with a relative 25% increase in SCr or an absolute increase of 0.5 mg/dL. The primary end point of the trial was the peak increase in SCr between baseline and day 3.

Similar to CARE, ICON demonstrated no significant differences in peak SCr increases in the two groups. The peak rise in SCr did not differ at any point in time during follow-up or when alternative definitions of CIN were used. In-hospital and 30-day outcomes were similar in the two groups. CIN occurred in 18 to 20% of patients in each group.

The ICON results contrast with those of a larger, recently published study (Jo et al. J Am Coll Cardiol 2006; 48(5):924-30). Among 300 patients undergoing coronary angiography, those randomized to iodixanol had a CIN incidence less than half that of patients who received ioxaglate (7.9% vs. 17%, P=0.021).

Results of the previously reported IMPACT (Impaired Patients Undergoing Computed Tomography) trial also were discussed at the TCT meeting. The study involved 166 patients with stable moderate-to-severe chronic kidney disease associated with a SCr ³1.5 mg/dL or creatinine clearance £60 mL/min. Patients were randomized to contrast-enhanced multi-detector CT scans with iopamidol or iodixanol. (Barrett et al. Invest Radiol 2006;41(11):815-21).

As reviewed by Dr. Richard Katzberg, University of California, Davis Medical Center, Sacramento, the primary end point of IMPACT was the proportion of patients in each group that developed CIN. Pre-contrast SCr averaged 1.6 mg/dL and the mean creatinine clearance was 44 to 45 mL/min in each group. Consistent with CARE and ICON, IMPACT investigators defined CIN as an absolute increase in SCr >0.5 mg/dL or a relative increase ³25%.

The rate of CIN was low in both groups and did not differ between the groups. An increase in SCr ³0.5 mg/dL occurred in 2.6% of patients that received iodixanol vs. no patient in the iopamidol group. When CIN was defined as a relative increase in SCr ³25%, the CIN rate was 4% in each group.

Discussion

Though consistent among themselves, these three studies are inconsistent with results of the widely disseminated and cited NEPHRIC (Nephrotoxicity in High-Risk Patients Study of Iso-Osmolar and Low-Osmolar Non-ionic Contrast Media) trial, which compared iodixanol and iohexol in patients undergoing coronary or aortofemoral angiography (Aspelin et al. N Engl J Med 2003;348(6):491-9). NEPHRIC involved a total of 166 patients with an increased risk of CIN because of renal insufficiency and diabetes. The results showed that use of iodixanol resulted in a significantly lower peak rise in SCr at 72 hours compared to iohexol.

The lack of consistent results among trials comparing different types of iodinated contrast media is at odds with the basic science literature. Good evidence from experimental settings supports the view that the risk of CIN is lower with iso-osmolar agents compared to low-osmolar contrast media. Several factors may contribute to the disconnection between the experimental and clinical data. First, patients in the various trials have presented differing degrees of baseline risk, complicating attempts to compare results from different studies. A second factor relates to the definitions used to define baseline renal function. In the ICON trial, for example, patients had a mean SCr of 1.8 mg/dL, but that value provides an incomplete picture in the absence of eGFR values.

Another factor that complicates efforts to compare results of different clinical trials involves the changing scientific environment surrounding concepts of renal insufficiency. Nephrologists have long accepted a GFR £60 mL/min as the cut-off point for defining acute renal failure following contrast administration. However, recent data have begun to suggest that a GFR £40 mL/min might be a more appropriate cut-off. In the RECOVER trial comparing iodixanol and ioxaglate, the isotonic contrast agent clearly demonstrated evidence of less nephrotoxicity compared to the low-osmolar agent at GFR <30 mL/min, in addition to the lower risk of CIN observed in the overall study population involving patients with GFR <60 mL/min (Jo et al. J Am Coll Cardiol 2006;48(5):924-30).

As pointed out here during a round table discussion of CARE and ICON, both trials lacked the statistical power to demonstrate a significant difference in CIN between the agents compared. In CARE, for example, an estimated 3300 patients would have been required to ensure the statistical capability to detect a significant difference between iodixanol and iopamidol.

Related to the study size is the potential for a type 2 statistical error (accepting the null hypothesis when in fact a difference exists). Given the lack of statistical power in the design of some studies conducted to date, the possibility of a type 2 error cannot be rejected out of hand.

The impact of prophylaxis also cannot be overlooked as a contributing factor to the results of various trials. In some instances, the total injected volume used to hydrate patients differed between treatment groups, as occurred in the IMPACT study, for example. As the degree of hydration influences the risk of CIN, the impact of an imbalance in injected volume on subsequent rates of CIN is unclear but has the potential to influence the results between randomized groups.

Another protocol-related issue centres on the timing of measurements central to the primary end point. In the IMPACT study, SCr was measured at a single point in time between 48 and 72 hours after administration of contrast. Serial measurements would have offered more assurance that the peak rise in SCr was accurately captured.

The preceding issues illustrate the difficulties in trying to compare results of different trials of iodinated contrast media. Whether comparative trials of adequate size and appropriate design will be conducted remains to be seen. In their absence, clinicians must rely on their best judgment based on assessment of available data, which should include the cumulative evidence from both clinical and experimental settings.

Summary

CIN remains a concern for many patients who undergo imaging studies involving iodinated contrast media. CIN confers a substantial risk of morbidity and mortality and very clearly has an adverse impact on patient outcomes. The comparative clinical trials of iodinated contrast media conducted to date have yet to create a consistent scientific or clinical picture of CIN risk associated with various agents. Low-osmolar, nonionic agents have demonstrated a reduced risk of CIN compared to higher-osmolar and ionic agents. Iso-osmolar contrast media, such as iodixanol, have a strong base of laboratory evidence to support a claim of superior safety compared to low-osmolar agents and this evidence should be weighed along with the clinical evidence when assessing an individual patient’s risk of CIN.