Disclosures
Dr. Friedewald has received honoraria for consulting from AstraZeneca, Wilmington, Delaware. Dr. Goldfarb has received honoraria for speaking from GE Healthcare. Dr. Vetrovec has received honoraria for speaking from Eli Lilly, Indianapolis, Indiana; Daiichi Sankyo, Parsippany, New Jersey; Gilead, Foster City, California. Dr. Vetrovec has received research funding from Cordis Corporation/Johnson & Johnson, Warren, New Jersey; Schering Plough, Kenilworth, New Jersey; and Corindus, Natick, MA. Dr. Vetrovec has Endpoint and DSMB contracts with Merck (Endpoint); Pfizer, New York, New York (DSMC); and Abbott Laboratories, Abbott Park, Illinois (DSMC). Dr. Vetrovec owns stock in Abbott Laboratories; Merck; Boston Scientific Corporation, Natick, Massachusetts; Johnson & Johnson, New Brunswick, New Jersey; Cardinal Health, Dublin, Ohio; Hospira, Lake Forest, Illinois; Medco Health Solutions, Columbus, Ohio; Schering Plough; and Owens and Minor, Richmond, Virginia. Dr. Roberts has received honoraria for speaking from AstraZeneca;and Merck, Whitehouse Station, New Jersey.
Disclosures
Dr. Friedewald has received honoraria for consulting from AstraZeneca, Wilmington, Delaware. Dr. Goldfarb has received honoraria for speaking from GE Healthcare. Dr. Vetrovec has received honoraria for speaking from Eli Lilly, Indianapolis, Indiana; Daiichi Sankyo, Parsippany, New Jersey; Gilead, Foster City, California. Dr. Vetrovec has received research funding from Cordis Corporation/Johnson & Johnson, Warren, New Jersey; Schering Plough, Kenilworth, New Jersey; and Corindus, Natick, MA. Dr. Vetrovec has Endpoint and DSMB contracts with Merck (Endpoint); Pfizer, New York, New York (DSMC); and Abbott Laboratories, Abbott Park, Illinois (DSMC). Dr. Vetrovec owns stock in Abbott Laboratories; Merck; Boston Scientific Corporation, Natick, Massachusetts; Johnson & Johnson, New Brunswick, New Jersey; Cardinal Health, Dublin, Ohio; Hospira, Lake Forest, Illinois; Medco Health Solutions, Columbus, Ohio; Schering Plough; and Owens and Minor, Richmond, Virginia. Dr. Roberts has received honoraria for speaking from AstraZeneca;and Merck, Whitehouse Station, New Jersey.
Introduction
Contrast-induced nephropathy (CIN) is acute renal dysfunction occurring within days after the administration of an imaging contrast agent and is a relatively common cause of hospital-acquired cases of renal insufficiency. The risk for CIN is likely increasing, given that increasing age and diabetes mellitus, which are both risk factors, are increasingly prevalent in patients who undergo procedures requiring contrast agents. In this Editor’s Roundtable, the participants discuss the role of contrast agents themselves, with particular emphasis on the American College of Cardiology (ACC) and American Heart Association (AHA) update guideline for ST elevation myocardial infarction (STEMI) and other issues related to CIN.
Discussion
Dr. Friedewald: Let’s begin with the faculty’s opinion about the ACC/AHA STEMI update guideline as it relates to CIN in patients with chronic kidney disease (CKD).
Dr. Roberts: How do you define CKD as it pertains to patients in the setting of an acute myocardial infarction (AMI)?
Dr. Goldfarb: That is a major problem in discussing this guideline. This has been a controversial topic. The history goes back about 8 years, when the National Kidney Foundation forwarded the concept that clinicians should widely employ a formula for estimating the glomerular filtration rate (GFR) that was developed by the National Institutes of Health for a large prospective study, the Modification of Diet in Renal Disease (MDRD). This formula had been created several years earlier to get a better measure of kidney function than an isolated serum creatinine. The MDRD formula places great emphasis on patient age as a determinant of estimated GFR (eGFR). Thus, an older person with a serum creatinine in the normal range—with “normal range” defined by the standard ways of measuring the serum creatinine in thousands of people and then establishing the boundaries for normal—can have a normal serum creatinine and yet fall within the category of stage 3 kidney disease, with an eGFR of <60 ml/min. Unfortunately, this number has become a criterion for inclusion in various studies and is the real issue, because an older person with a calculated eGFR of <60 ml/min does not necessarily have kidney disease or any kidney malfunction, since eGFR declines with age in otherwise normal individuals. Furthermore, an eGFR <60 ml/min is not necessarily associated with hypertension, proteinuria, or progression to more advanced kidney disease in most older individuals. The idea that an individual with an eGFR of 59 ml/min is at risk for CIN because they fall into the category of stage 3 CKD is without a sound basis.
Dr. Friedewald: What should be measured for CIN risk in the patient undergoing coronary arteriography?
Dr. Goldfarb: First, a serum creatinine. A serum creatinine >1.5 mg/ml indicates that the patient may have underlying CKD, although this level itself is not proof . An elevated serum creatinine is simply better evidence for possible CKD than an eGFR <60 ml/min in an older patient. Studies that use an eGFR of <60 ml/min as a criterion for “high-risk” patients may be studying many patients who actually are at low risk.
Dr. Roberts: When does the age cutoff become significant?
Dr. Goldfarb: Age is a continuous variable in all formulas used to assess eGFR.
Dr. Friedewald: What is the significance of measurement variance with the eGFR?
Dr. Goldfarb: With an eGFR of 30 ml/min in a stable outpatient, the variance is about ±10 ml/min, so a patient may have a real eGFR of 40 ml/min when the report states that it is 30 ml/min. When the eGFR is 60 ml/min, the variance is even greater, as large as ±20 ml/min. In essence, the eGFR is not a precise measurement, but research studies often use it as though it were precise, which gives practitioners incorrect measures of renal function .
Dr. Laskey: The same concerns pertain to serum creatinine, which is quite variable from day to day. Establishing a stable baseline serum creatinine in a particular patient is difficult and is another significant problem in many research trials.
Dr. Goldfarb: I agree. Instability of the serum creatinine is well documented, especially in hospitalized patients. Thus, the issue germane to these new guidelines is how to extrapolate results of studies using eGFR and serum creatinine data such that they apply to different populations of patients for whom the Guidelines are intended.
Dr. Vetrovec: As a coronary interventionist, how should I manage the 76-year-old patient with a creatinine of 1.3 mg/ml and a low eGFR? There are insurance reviewers who look at these data and commonly question our judgment in these types of situations.
Dr. Goldfarb: There are published data to defend the safety of giving contrast agents in a patient with an eGFR of 51 ml/min according to the MDRD equation, such as trials by Dr. Laskey. The question is whether information from trials in which most of the “high-risk” patients actually have minimal renal abnormalities (or even merely the low eGFR of aging) can be extrapolated to patients whose renal function is outside the range of those trials. For example, if a trial uses an eGFR of <60 ml/min as the definition of CKD, the results do not automatically apply to the patient with a serum creatinine of >2 mg/dl or to a patient with diabetes mellitus and proteinuria. The risks for such patients with advanced CKD have not been studied.
Dr. Vetrovec: I am particularly concerned with patients in the “gray zone” of increased risk, that is, what do you tell them about increased risk? And in some cases, do you simply not perform the procedure when the indication for angiography may be equivocal? In other words, what is the risk/benefit of CKD versus interventional treatment of coronary artery disease? There are a lot of confusing data. I also accept that a patient with a serum creatinine of 1.3 mg/dl is quite different from >2 mg/dl, but there remains a great deal of unpredictability and an absence of consensus about how to correctly adjust the risk.
Dr. Laskey: I agree, and this is a large, complex issue. Adding to Dr. Goldfarb’s comments, the current data cannot easily be extrapolated or generalized, as in these new Guidelines. A study in which 50% of the patients were in the eGFR range of 50 to 60 ml/min cannot be applied to the patient with an eGFR of 20 ml/min. Expanding further on Dr. Vetrovec’s point, there are a lot of factors other than the baseline serum creatinine and eGFR that must be taken into account to predict risk. The issue with the new guideline, specifically Table 9, is that although the title of the updated guideline refers to patients with an AMI, the cited evidence has nothing to do with an AMI population .
Dr. Roberts: What is the eGFR variance when >60 ml/min?
Dr. Goldfarb: The eGFR has less variability around 30 ml/min. Above 60 ml/min, however, its variance is about 20 ml/min, so it loses much of its value in serial determinations at higher levels. The original formula was based on 2 serum creatinines measured 6 weeks apart in stable outpatients. In acutely ill patients, whom we are concerned with in this discussion, the evidence that the eGFR is an accurate reflection of CKD is extraordinarily poor. This is an area of great uncertainty, and unfortunately, it is very difficult to use existing data in the larger trials to accurately diagnose kidney function at any 1 point in time.
Dr. Laskey: When you use postprocedural serum creatinine to assess the incidence of CIN, this becomes even more important if the variability in that determination is substantial. This uncertainty is a real “Achilles’ heel” because there is an absence of rigor in the ascertainment of serum creatinine in numerous studies of CIN.
Dr. Friedewald: What is the mechanism whereby contrast agents damage the kidney?
Dr. Goldfarb: The data are conflicting about the pathophysiology of CIN as it relates to contrast agents themselves. In general, high-osmolar agents, defined as agents with osmolalities 7 to 8 times the osmolality of plasma, definitely produce vasoconstriction in the kidney. The kidney, particularly within the medulla, operates at a very low oxygen tension and a very low blood flow, which allows the normal kidney to function as a concentrating system. Thus, agents that produce intense vasoconstriction in the medulla cause ischemic injury and a reflow phenomenon that generates oxygen free radicals. This pathophysiologic effect has been the rationale for some of the pharmacologic manipulations, such as N-acetyl cysteine, a free radical scavenger, that have been proposed to reduce the incidence of CIN. The same effect probably extends to the so-called low-osmolar agents, which possess 2 to 3 times higher osmolality than plasma. The potential nephrotoxic action of agents with an osmolality up to threefold higher than plasma is the basis for the contention that an iso-osmolar agent is less likely to induce renal injury, which is supported by some data. It also is possible that these agents have direct toxic effects, unrelated to effects on blood flow, but this is a very difficult area to study. Although other mechanisms might be possible most nephrologists believe that vasoconstrictive-induced ischemia is the underlying mechanism of injury. The next question is whether the osmolality of the agent is a driving force in these changes in blood flow.
Dr. Friedewald: Does the etiology of the underlying CKD (i.e., diabetic nephropathy vs hypertensive nephropathy) have any bearing on whether a contrast study is more or less likely to cause CIN?
Dr. Goldfarb: I believe that the etiology of renal disease is important. Diabetic nephropathy seems to have a unique propensity to be associated with CIN. Whether this is due simply to the nephropathy per se, particularly in older persons, or due to unique changes in renal blood flow associated with diabetes is uncertain. Many years ago we showed that patients with diabetes mellitus had about 5 times the risk of getting CIN with low- or high-osmolar agents, compared to patients with similar degrees of CKD who were not diabetic.
Dr. Friedewald: You believe that a person with diabetes mellitus and a serum creatinine of 1.5 mg/dl has a greater risk for CIN than a nondiabetic with the same levels of serum creatinine and systemic arterial pressure?
Dr. Goldfarb: I believe that is true, but my opinion is based on an extrapolation from studies.
Dr. Laskey: Age also is important, because many older patients have significant co-morbidities. If no co-morbidity, such as hypertension, is present, is not the older patient still at increased risk for CIN because the eGFR is lower than in a younger person? In other words, is not a normal older kidney more predisposed to hypoxic injury than a normal younger kidney?
Dr. Goldfarb: We do not have good data to support that notion, and I do not believe it to be true. It is an area, however, being actively investigated. My view is there are many older patients who need to be exposed to contrast because they benefit from the resulting evaluation and treatment. Thus, we should do all we can not to restrict use of these agents, and place more focus on protocols that minimize CIN risk, even if they require use of more expensive agents or more time for volume expansion. We should target older patients who really are at high risk rather than just assume that all older patients are at high risk, because they are not. Busy interventional cardiologists and radiologists too often throw their hands up and say, “We are being flooded by aging patients and they all have a low eGFR. What are we supposed to do?” From my perspective, our answer is to target truly high-risk patients, which does not include patients who only have an age-related reduced eGFR.
Dr. Vetrovec: There are other factors that contribute to CIN. The state of hydration is very important. Sometimes patients are nil per os all day and undergo coronary angiography at 6 o’clock at night. That is a physiologically different person than the individual who drank fluids until 2 am and their procedure was performed at 8 am . Another, often overlooked issue involves heart failure (HF), which may modulate the relative importance of the eGFR in the context of the patient’s age. Many older patients undergoing cardiac catheterization have significant HF, which markedly increases the risk of CIN.
Dr. Friedewald: Let’s discuss contrast agent osmolality.
Dr. Laskey: The classification of contrast agents for many years was divided into ionic and nonionic, but today, in the nonionic era, there are many shades of gray, so the classification is now according to osmolality or osmolarity, but the term molecular weight , as used in the guideline, has never been part of the categorization of these agents, and its use in the guideline has raised a new level of confusion that needs correction.
Dr. Friedewald: The term molecular weight also was not used in reference 19 of this statement in the guideline, which describes the CARE (Cardiac Angiography in Renally Impaired Patients) study of iopamidol and iodixanol.
Dr. Laskey: Low osmolar , not low molecular weight , is the proper term.
Dr. Friedewald: Is there a difference between iso-osmolar and low-osmolar agents in terms of risk? Furthermore, among the low-osmolar agents, is there a difference, as suggested by the guideline?
Dr. Vetrovec: The data are very conflicting, and a case can be made for almost any point of view.
Dr. Goldfarb: The primary issue relates to the details of the studies addressing these questions. The study that generated the updated guideline was the NEPHRIC (Nephrotoxicity in High-Risk Patients Study of Iso-Osmolar and Low-Osmolar Non-Ionic Contrast Media) study, which was a European study performed several years ago. NEPHRIC showed that there was an important difference between the risk of CIN in patients getting intra-arterial studies with an iso-osmolar agent, iodixanol, compared to a low-osmolar agent, iohexol. Subsequent studies have addressed the question whether there is something about this particular iso-osmolar agent other than osmolality that drove the difference in outcomes. One study, which Dr. Laskey participated in, showed that the differences vanished when comparing another low-osmolar agent, iopamidol, to the iso-osmolar agent, suggesting that perhaps in the original study there was something about iohexol that led to the outcome. But there are differences between the studies, in that while both were well performed, the later study used much more volume expansion and much lower contrast volumes, which do reduce CIN incidence.
My conclusion from these studies was that contrast agents used in small amounts (≤100 ml) in well-hydrated patients provide similar outcomes regardless of whether the agent is either iopamidol or iodixanol. Outcomes for iodixanol also were better in the more recent study of iodixanol versus iopamidol than they were in the original NEPHRIC study. I suspect the better outcomes were due to the use of smaller volumes of contrast agent and more aggressive volume expansion. For patients in whom extracellular volume expansion is minimal or in whom larger doses of contrast agents are required, however, I am uncertain what can be extrapolated from the more recent study. Thus, although the studies are well done, the study populations are different.
Dr. Laskey: I agree, but our concern in this discussion are patients with an AMI, which is the group targeted by these updated guidelines. Neither the original NEPHRIC study nor our recent study involving a diabetic population with renal insufficiency focused on patients undergoing angiographic procedures with an AMI. Rather, the patient populations were a mix of clinically and hemodynamically stable patients undergoing coronary intervention and/or diagnostic studies, usually with a 60:40 predominance in favor of diagnostic studies versus interventional procedures. There has not been a study of AMI patients uniformly undergoing coronary intervention looking at the relative safety of contrast agents. That is 1 reason why I believe there is not a discernible difference between agents. There are many other reasons, but for the “sickest of the sick”—patients with an AMI—we are lacking appropriate studies. Thus, firm statements as in this guideline should not be made; rather, the absence of existing data for the target population should be acknowledged.
Dr. Friedewald: What contrast agent would you use for an 80-year-old woman with no known diabetes or underlying renal disease and a serum creatinine of 1.4 mg/ml having coronary arteriography for acute chest pain and electrocardiographic changes of acute ischemia, but no definite infarction? According to this guideline, you could use either an iso-osmolar agent or low-osmolar agent other than ioxaglate or iohexol.
Dr. Laskey: The guidelines do not truly address this type of patient.
Dr. Friedewald: That is the reason for the question.
Dr. Laskey: The utility of guidelines, and why people do or do not follow them, is a function of whether they apply to specific patients. The issue is primarily about the generalizability of the supporting data . Do the data apply to other patient populations in other settings, and this does not even address “how good are the data?” This is a great example of why you may not be able to, and therefore should not, extrapolate from published studies. As is so often the case in medicine, we must rely on our own integrated personal experiences, because when there are no solid data, one needs to fall back on one’s clinical gestalt. The latter does not necessarily negate or neglect the “evidence base” but incorporates all available information, in an on-line Bayesian sense, to form a decision path.
Dr. Roberts: So what would you do in this case?
Dr. Laskey: My experience goes back to the COURT (Contrast Media Utilization in High Risk PTCA) trial, which is an older study comparing iodixanol to ioxaglate. It was a trial of very sick patients and supported the use of iodixanol in them. That is the agent I would use in this case.
Dr. Goldfarb: If you called me as the renal consultant for this case, I would say she is probably going to do fine as long as she does not have a hemodynamic catastrophe associated with the study. If she does, it probably does not matter what agent was used. Another question is what agents are going to be available to the cardiologist to make that choice, as opposed to what choice they would make.
Dr. Friedewald: That is a good question, because this guideline appears to be influencing formulary decisions.
Dr. Goldfarb: Exactly. Clinicians need to have options available to them in the absence of definitive data on the safety of each agent in a wide spectrum of patients.
Dr. Vetrovec: I would not use iodixanol, but others in our laboratory disagree with me. One thing worth mentioning about iodixanol is that it is associated with a slightly higher risk of delayed hypersensitivity skin reactions, but not anaphylactic reactions. This reaction can confuse the management of these patients who also are on other drugs like clopidogrel, and there is the risk that that drug could be discontinued if the rash is attributed to it.
Dr. Laskey: The confusion with clopidogrel is important. The exact incidence of delayed hypersensitivity, however, is unknown but likely quite low, probably in the 1% to 2% range.
Dr. Friedewald: The reference for the specific area of the guideline under discussion is the CARE study, which did not compare ioxaglate to iohexol, but Table 9 in the guideline makes the recommendation of using a low-osmolar agent “other than ioxaglate or iohexol.” To make a recommendation based on a trial that did not study the agents not being recommended is misleading. The referenced article did state in its discussion, however, that “the inescapable conclusion is that iohexol and iopamidol, both low-osmolality contrast agents, have different renal safety profiles.” Is this truly an “inescapable conclusion”?
Dr. Goldfarb: I am unaware of any study that directly compares iohexol and iopamidol, and there probably never will be. Only such a study would allow any “inescapable conclusion.” The flaw in these sorts of extrapolations from various studies comes in the fact that these studies have been conducted on different populations using different protocols.
Dr. Friedewald: Is there a biological reason why the 2 low-osmolar agents, iohexol and iopamidol, might have differing effects on renal function?
Dr. Goldfarb: The osmolality and chemical structure of iohexol and iopamidol are quite similar. Thus, there is no obvious basis to assume a difference based on current information.
Dr. Laskey: The CARE study, reference 19 in the guideline, also has nothing to do with AMI patients. This guideline blurred it all together, which is sloppy and confusing. One final point about the contrast agents themselves is that they clearly have effects other than on the kidney, many of which have yet to be delineated. For the last several years, we have been using high-throughput proteomic methods in patients receiving contrast agents. We have seen dynamic changes in clotting factors, lipoprotein metabolites, and markers of endothelial function. Thus, this focus on effects on the kidney and serum creatinine changes is historically understandable but is misplaced because intravascular contrast agents stir up a fair amount of havoc over the course of 24 to 48 hours, some of which appear deleterious. I do not believe we have paid enough attention to some of the hemorheologic effects of contrast agents.
Dr. Goldfarb: I agree. Both studies were well performed. The problem with the guideline is the extrapolation from studies of populations different from the population targeted by the guideline.
Dr. Roberts: What is the relative cost of these agents?
Dr. Goldfarb: My understanding is that the iso-osmolar agent is some multiple of the cost of the low-osmolar agent, predominantly related to manufacturing cost. While it does not translate into much per individual study—approximately $20 to $40—for a purchasing agent in a busy radiology program or catheterization laboratory, multiplying that by the number of studies ends up being a significant amount of money. This becomes a question of institutional budgets as opposed to the relatively small incremental cost for an individual study.
Dr. Laskey: I agree.
Dr. Friedewald: What do you recommend when monitoring patients for CIN?
Dr. Goldfarb: There are some recent data suggesting that increases in serum creatinine begin very early after administering a contrast agent among patients having a 25% increase in serum creatinine, which generally is the criterion that is used for CIN. Creatinine is a complicated issue, because there are issues of creatinine production and volume of distribution in specific patients, and there are many concerns about various interventions (e.g., volume infusions or diuretic use) which may change the volume of distribution of creatinine and cause the serum creatinine to have different results from 1 time period to another. There also is a variability that has been observed that is specific for cardiac patients, with as many as 25% of patients having spontaneous changes in the serum creatinine, which has been attributed to contrast agents but may simply represent variability.
Dr. Friedewald: When should the serum creatinine be measured after cardiac catheterization when monitoring patients for CIN?
Dr. Goldfarb: This is uncertain, but the first measurement is usually made within 24 to 72 hours after the procedure, generally determined largely by patient convenience more than urgency. From the point of view of identifying patients for clinical trials, I recommend daily measurements for at least the first few days, in which transient creatinine elevations may occur, and then they typically return to normal. Whether a 25% increase in serum creatinine, particularly if it is in the range of 1.4 to 1.7 mg/ml, is clinically important is a complex question.
Dr. Friedewald: Would you say that if the creatinine has returned to normal by 1 week, after a transient rise, that CIN will not occur?
Dr. Goldfarb: Yes, in general. The issue is whether such transient changes in serum creatinine produce other clinical outcomes, which is an interesting and uncertain area. When serum creatinine rises, outcomes tend to be worse for long-term cardiovascular morbidity and mortality, with some cardiac events occurring quite late in time. Whether this is only a reflection of sicker patients or a direct effect of the kidney dysfunction in some unidentified way is unknown. I suspect that it reflects the general condition of the patient rather than a specific injury to the kidney, although there are nonhuman animal studies suggesting that transient renal ischemia is associated with distant effects on other organs. Whether those studies relate to patients is unknown, and extrapolation from those studies may be unwarranted.
Dr. Vetrovec: Another question about transient rises in serum creatinine after the use of contrast agents is how that may relate to inflammation. Perhaps serum creatinine rise is a marker of an inflammatory response or causes an inflammatory response that could later mediate vascular risk. I too have been struck by the fact that a small serum creatinine rise portends greater risk for cardiac events at a later date, and not necessarily coronary restenosis, just a greater risk of adverse cardiac events. Furthermore, it is possible that rises in serum creatinine may be even more sensitive markers of later cardiac risk than troponin or creatine kinase markers, but there is an absence of data on this.
Dr. Goldfarb: There have been serious attempts to look at this issue, but most studies are retrospective. Prospectively, it has been more difficult to show impact, but there are ongoing studies. There are nonhuman animal studies suggesting that this is biologically plausible. For example, placing a clip on the rat’s renal artery and releasing it after 20 minutes causes ischemic renal injury, as well as increased inflammation in the brain, some myocardial cell dropout, and changes in lung permeability. Thus, there presumably is a release of renal factors that enter the circulation and produce vascular injury elsewhere in the body, but there are insufficient data about this potential effect in humans.
Dr. Friedewald: I am concerned that the new plethora of meta-analytic studies, such as cited in the guidelines for contrast agents, may replace good scientific prospective research. There is an important role for meta-analyses, but their data can also be misleading, mainly in that different studies look at different populations, and meta-analyses can lead to wrong conclusions because of this “apples and oranges” mixing that occurs.
Dr. Goldfarb: Meta-analyses are best used as hypothesis-generating studies, and when used in other ways, especially for specific guidelines, we are on very thin ice. The problem is that it is a lot easier to look at different populations in this way, and the limitations of this approach need to be appreciated. These kinds of analyses are always based on dichotomous variables because such variables are easily coded for database creation. For example, if a patient becomes hypotensive for a brief period (i.e., for 20 minutes during a study), such a transient event may never be recorded. Moreover, its severity, duration, and comparison to the patient’s baseline blood pressure is rarely captured. Thus, transient hypotension is an important variable that is not addressed by these types of retrospective meta-analytic studies. All of those very fancy statistics try to circumvent these uncaptured clinical events, but in the end, only an adequately powered prospective trial can allow rejection or acceptance of a hypothesis.
Dr. Laskey: This illustrates the important cut point between whether a patient gets put into an “unconfounded” CIN category or a CIN category which is confounded by other events. We had this problem in the most recent study with iodixanol and iopamidol. We tried to substantiate CIN in the absence of any other confounding features during the study by blindly reviewing very detailed records. Thus, patients with short durations of hypotension did not go into the unconfounded CIN box; rather, they went into the confounded CIN box, which means we can never be certain whether CIN was secondary to the agent or the hypotension or both. This is so important, and it is never even discussed in these sorts of über-meta-analytic studies because the studies themselves do not take the pains to find this out. I believe this is one of the sources of anxiety that Vince expressed that clinicians may not weigh these considerations. Although there are a lot of smart clinical people who do meta-analytic work, too, the point that Dr. Goldfarb just raised is critical to this topic because it determines whether patients go into the “yes” box or the “no” box. These determinations are carried forward, and must be accounted for when selecting studies upon which guidelines are based.
Dr. Goldfarb: The definition of CIN includes acute renal failure without any other explanation . Thus, other explanations are highly relevant.
Dr. Laskey: The main problem is the loss of specificity when all this “noise” is present, and what follows is a loss of confidence in what number of patients have true CIN. The studies themselves are often limited, with weak methodology, and the problem is compounded, not clarified by a guideline that lumps them all together.
Dr. Roberts: What is the time interval from the procedure to making the diagnosis of CIN? At Baylor University Medical Center, a huge percent of the patients having cardiac catheterization go home the next day. How should they be followed for CIN?
Dr. Goldfarb: This directly relates to identifying high-risk patients. The risk is very low in low-risk patients, <5%, whom we need not closely monitor for CIN. High-risk patients, however—whom I define as patients with preprocedure elevated serum creatinine, particularly diabetic patients—ought to have their serum creatinine measured sometime within 5 days after the study to be certain they are not going into severe renal failure. If it were slightly elevated, it should be measured again at intervals of a few days until it regresses back to its baseline and it is certain that it is not progressively rising. This is a practical approach that suffices in monitoring patients at risk for CIN.
Dr. Vetrovec: Another consideration is that the follow-up serum creatinine may be performed in a different laboratory, which introduces yet another variable.
Dr. Friedewald: Are there any special instructions to the patient for reducing the risk of CIN after discharge?
Dr. Vetrovec: All I tell patients is to call me if they stop making urine.
Dr. Friedewald: Do you tell them to drink a lot of water?
Dr. Vetrovec: No.
Dr. Friedewald: What do you do before the procedure to reduce CIN risk?
Dr. Goldfarb: Volume expansion is most important in normal persons. Volume expansion in a normal person increases the extracellular fluid volume and greatly enhances renal blood flow.
Dr. Friedewald: What about volume expansion in patients with HF?
Dr. Goldfarb: Whether volume expansion reduces CIN risk depends on their hydration starting point. For example, patients on diuretics and a low-salt diet may be intravascularly volume depleted all the time. With severe HF, they may have very modest renal perfusion to begin with, and giving them a small volume of fluid intravenously might raise renal blood flow by 10% to 20%. Management of HF patients is an example of when we should treat to protocol as opposed to outcome. The outcome ought to be a rise in renal blood flow, and there have been some data suggesting that raising the urine output, which is not a perfect index of volume expansion, helps in determining whether volume expansion is effective in a particular patient.
Dr. Friedewald: Does the type of solution used for volume expansion matter?
Dr. Goldfarb: This is controversial. Bicarbonate expansion may be better than with saline, but the data are not consistent.
Dr. Vetrovec: What about saline versus half-normal saline?
Dr. Goldfarb: The assumption has been that the goal is to raise renal blood flow as opposed to the volume of urine, and increased urine volume is a surrogate for improved renal blood flow. That assumption, however, is based on nonhuman animal studies in which ischemia is always the underlying acute kidney injury in experimental models. Volume expansion with isotonic saline raises the extracellular fluid volume and the plasma volume more than the same volumes of half-normal saline, as the former increases the extracellular fluid volume more effectively. The argument has been to use saline because it has a desired physiologic effect: to raise renal blood flow more than half-normal saline and even much more so than by giving patients only free water. There have been a few studies showing that just drinking water does not seem to do much, whereas giving patients salt tablets plus drinking water may be as protective as acute volume expansion. The key is to induce volume expansion, rather than mere hydration. Volume expansion means giving a saline-containing solution, but whether this should be in the form of sodium chloride or sodium bicarbonate is unclear.
Dr. Friedewald: Volume expansion also has added importance in patients with AMI because they often have extensive anorexia and vomiting before presentation. Also, they may have been anorexic due to infections such as influenza or other co-morbid conditions that precipitate AMI.
Dr. Laskey: This is probably the only area that has not changed over the past 40 years. The 1 certainty is that the volume state is key , and volume expansion is probably the only reliable measure that is effective in decreasing the risk of CIN in the current era.
Dr. Goldfarb: Fluid volume, in any form, is a crucial factor. I also agree with Dr. Vetrovec that management of patients with HF is a huge issue in reducing the risk of CIN. HF patients are underperfused due to a functional volume depletion. In every nonhuman animal model of acute kidney injury, the way to be certain the model works is to feed animals a low-salt diet. This potentiates any maneuver to induce acute kidney injury by reducing extracellular fluid volume.
Dr. Friedewald: What is the relative importance of the volume of contrast used in the procedure?
Dr. Laskey: The importance of constraining volume of contrast under most circumstances is a vital aspect of reducing CIN risk.
Dr. Vetrovec: I agree.
Dr. Laskey: The volume of contrast agent has been found in every risk stratification algorithm, of which there are 1 dozen. Whether it is another marker for a more complex procedure in a more complex, sick individual is another issue, but it is definitely a modifiable risk factor.
Dr. Roberts: How much contrast agent do you give?
Dr. Vetrovec: When <100 ml of contrast agent is used for a study serious renal impairment almost never occurs, even in high-risk patients. With more CT (computed tomographic) coronary angiograms being performed, this is important in high-risk patients because routine diagnostic coronary angiograms require much less contrast agent than the CT coronary angiograms.
Dr. Goldfarb: This is very important to CIN reduction. The updated NEPHRIC study employed smaller doses of contrast agent, and that may have been a very important factor in the lower incidence of CIN in both test groups. There are reports that patients with very advanced CKD receiving <40 ml of contrast agent have little risk of further renal deterioration. It is quite uncommon now for me to be consulted on a patient with CIN, regardless of the agent. I believe that increasing skill and awareness of CIN and an increase in better technology and smaller volumes of contrast agent has led to these improved outcomes.
Dr. Laskey: The incidence of CIN was greatly reduced when ionic agents were stopped.
Dr. Friedewald: What is the role of N-acetyl cysteine in preventing CIN?
Dr. Goldfarb: Unclear. I believe that it is acceptable to employ N-acetyl cysteine, but follow-up studies do not show a clear benefit. It definitely does not replace attempts at volume expansion and limiting the contrast agent volume.
Dr. Friedewald: What does the role of glucose control play in preventing CIN in the patient with AMI and diabetes mellitus?
Dr. Goldfarb: That is another one of these dichotomous variables that is never studied, so there is no answer. The patient is coded as either having or not having diabetes. The level of blood glucose at the time of the study is rarely known.
Dr. Laskey: I agree. There are no data about hyperglycemia and CIN risk in such patients. This relates to the complexities of metabolic derangements. When a patient is exposed to a contrast agent, a lot happens beyond an increased serum creatinine, such as increases in inflammatory markers and insulin resistance. There is, however, nothing that has been quantified in these areas.
Dr. Friedewald: What future research is needed?
Dr. Goldfarb: From the nephrology perspective, the first question is whether the osmolality of the contrast agent is important, which would have to be performed in an adequately randomized, controlled trial in patients with CKD, with eGFRs in the 20 to 30 ml/min range, including a substantial number of patients with diabetes mellitus.
Dr. Laskey: Such a study need not be huge (a “body count” study), but should have several arms and end points beyond capturing CIN incidence. The fascination and obsession with CIN is misplaced because we need to know whether CIN is just a surrogate marker or is truly in the causal pathway.
Dr. Vetrovec: What about prostaglandin research for reducing CIN risk?
Dr. Goldfarb: Prostaglandins are proposed to increase renal blood flow, and if they do, they might be protective. There have been several attempts at using such agents that increase renal blood flow. The results have been mixed. In some instances, the effects have even been deleterious.
Dr. Friedewald: I would like to ask each of you for a final commentary on the guideline. Should it be corrected? ACC/AHA guidelines are critically important to our specialty, and anything that might be incorrect or even misleading should be formally addressed.
Dr. Goldfarb: This guideline gives the clinician a false sense of safety about this issue by its implication that 1 particular agent of a class is safer than other agents in that class. In this instance, we do not know if the conclusion is true. I believe there should be great concern about this guideline, because its evidence base may be flawed. Singling out 2 agents as problematic and stating that other agents in the class are safe cannot be based on extrapolated data as found in this guideline.
Dr. Laskey: I agree with Dr. Goldfarb. I want to make 3 points. First, this is an inappropriate and incorrect way to use evidence and is not evidence-based medicine. The latter cannot be accomplished by extrapolation. Second, I object to the way this was folded into a document, based on its title, on the management of patients with AMI. The selected references do not support the titular guideline. Third, the absence of evidence does not imply that there is no evidence; rather, the studies were not powered to detect it. That sort of consideration does not appear anywhere in this guideline. The result is a very uncritical, ingenuous review of the data. What is presented appears to be based only on the most recent article on the topic. The patient population studied does not reflect the target population of the guideline.
Dr. Vetrovec: In general, I agree. The problem with this guideline update is it probably would have been better entitled “Strategic Issues That Have Developed Since the Last Review of Several Areas.” In this particular section, there were not a lot of new data that should have compelled the writers to take a stand on CIN. In some other areas addressed in the guideline, there were new data and there were issues that clinicians were dealing with that were new and important. This area, with contrast agents, however, did not fit that approach.
Dr. Friedewald: Thank you.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
