Contrast induced nephropathy: what's the true risk?
Iodinated IV contrast has long been considered a significant contributor to acute kidney injury in hospitalized patients. But so-called contrast induced nephropathy is hard to accurately identify in real clinical circumstances. Acute kidney injury (AKI) can happen from a variety of causes, or their combination, during acute illness. And no randomized trial has established the risk of contrast induced nephropathy. Further, most of the literature reporting AKI after IV contrast was done in an earlier era, when IV contrast agents were hyperosmolar. Newer contrast agents are generally low- or iso-osmolar, and friendlier to the kidneys. If the risk for contrast induced nephropathy is overblown, it could be detrimental to patient care. Tests and treatments requiring intravenous contrast are often deferred in patients with elevated creatinine who need a diagnosis for suspected pulmonary embolism, arterial thrombus, or intra-abdominal catastrophes. The diagnostic gap leaves their doctors with less than the best information to make treatment decisions. An increasing body of research is calling into question the risks to kidneys from IV iodinated contrast, including a study in the Annals of Emergency Medicine.
Authors looked back at 16,801 patients at a single center who were seen in the emergency department, most of whom were hospitalized afterward. They compared creatinine values at presentation and 48 hours afterward, dividing patients into three groups who had received either:
A CT scan with contrast (~40%)
A CT scan without contrast (~30%)
No CT scan at all (30%)
AKI rates (7-9%) were actually higher in patients who got CT scans without contrast -- as would be expected from selection bias and sound clinical decision making. Doctors ordered non-contrast studies in more patients with increased risk for AKI, and contrast studies in patients at lower risk.
To overcome this bias, authors used propensity scoring to match patients with similar comorbidities, creatinine, age, gender, etc. to try to compare similar groups of patients, trying to isolate the rate of AKI according to whether they had received contrast or not.
The conclusion was that no evidence showed contrast increased the risk of AKI -- even among the patients with chronic kidney disease at baseline. For patients who could be evaluated at six months, there was no increased risk of chronic kidney disease, renal replacement need, or kidney transplant, among those who got IV contrast (compared to their propensity-scored counterparts who did not).
Propensity scoring is not a substitute for a randomized controlled trial. It's impossible to eliminate all bias and confounding in any scoring model. Physicians could be selecting patients for contrast studies based on unmeasured variables. It's hard enough to power a randomized trial with any confidence to predict an outcome; it's even more difficult to assess power in a propensity matching study. The study was criticized for including in the propensity scoring chronic comorbidities coded later in the hospital admission -- information presumably unavailable to the ED physicians.
Contrast Induced Nephropathy: Are the Risks Overblown?
Other studies have similarly shown low risk for contrast induced nephropathy from newer lower-osmolar IV contrast agents. A Mayo clinic retrospective study in 53,439 patients who got 157,140 CT scans between 2001 and 2010 had no detectable difference in AKI rates between patients who had received contrast and those who had not. An analysis of 13 non-randomized studies including ~26,000 patients also found no association between contrast and bad clinical outcomes. Rates of AKI, mortality and need for renal replacement were not significantly different among those who did or did not receive IV contrast. Patients with diabetes or chronic kidney disease were not at increased risk in these observational, retrospective cohorts. Another retrospective study among 20,242 patients at University of Michigan, also using propensity scoring, found that IV low-osmolality iodinated contrast material did seem to bring an increased risk for AKI after CT scanning, but only among patients with creatinine levels of 1.6 mg/dL or greater. This is roughly in line with widely used cutoffs for safety by radiologists in performing contrast studies. Other observers have proposed a cutoff of a GFR of 30 mL/min as the point of significantly increased risk for contrast nephropathy.
A significant percentage of hospitalized patients experience AKI, and an even larger percentage receive IV contrast during their hospital stay. There will thus always be the basis for anecdotal observed instances of "contrast induced nephropathy." Whether or not contrast-induced acute kidney injury still exists, or has been eradicated by the use of lower-osmolarity contrast agents, would require a large randomized trial to sort out. Although a RCT enrolling patients with Cr > 1.5 mg/dL could be considered ethical in patients with strong clinical indications for IV contrast, enrollment to achieve the needed power would be difficult and slow, and such a trial is unlikely to ever occur. Patients with normal or near-normal renal function are at very low risk for contrast induced nephropathy using today's available agents in the U.S. The benefits of contrast studies in some patients with elevated creatinine levels may outweigh the risks if the study is vital for optimal care. Read more:
Risk of Acute Kidney Injury After Intravenous Contrast Media Administration. JS Hinson et al. Ann Emerg Med. 2017 Jan 19.
Radiology Studies Shed New Light on CT Contrast Risks. August 01, 2013. Intravenous Contrast Material–induced Nephropathy: Causal or Coincident Phenomenon? Radiology. 2013 Apr;267(1):106-18. Contrast Material–induced Nephrotoxicity and Intravenous Low-Osmolality Iodinated Contrast Material: Risk Stratification by Using Estimated Glomerular Filtration Rate. Radiology. 2013 Sep;268(3):719-28
Intravenous Contrast Medium–induced Nephrotoxicity: Is the Medical Risk Really as Great as We Have Come to Believe? Radiology. 2010 Jul;256(1):21-8