 |
|
CLINICIAN'S CORNER
Prophylaxis Strategies for Contrast-Induced Nephropathy
Neesh Pannu, MD;
Natasha Wiebe, MMath, PStat;
Marcello Tonelli, MD, SM; for the Alberta Kidney Disease Network
JAMA. 2006;295:2765-2779.
ABSTRACT
| |
Context Contrast-induced nephropathy is associated with significant economic and clinical consequences, including prolonged hospitalization, the requirement for dialysis, and an increased risk of death.
Objectives To summarize the current state of evidence for prophylaxis of contrast-induced nephropathy, provide evidence-based recommendations regarding management of high-risk patients undergoing angiographic procedures, and identify new avenues for research.
Data Sources Systematic searches of peer-reviewed publications were performed in MEDLINE, EMBASE, and the Cochrane database from 1966 to January 2006. Search terms included radio contrast nephropathy, contrast media, acetylcysteine, theophylline, sodium bicarbonate, HMG Co-A reductase inhibitors, ascorbic acid, kidney diseases, renal insufficiency, kidney failure, nephropathy, fenoldopam, diuretics, and saline or half saline.
Study Selection Observational studies of risk factors and randomized controlled trials of prophylaxis strategies for contrast-induced nephropathy that specified a definition of contrast-induced nephropathy or postprocedure creatinine level as an outcome measure.
Evidence Synthesis Important patient-related risk factors for contrast-induced nephropathy include chronic kidney disease, diabetes mellitus, heart failure, older age, anemia, and left ventricular systolic dysfunction. Non–patient-related risk factors include high-osmolar contrast, ionic contrast, contrast viscosity, and contrast volume. Practice guidelines recommend obtaining preprocedural serum creatinine levels among patients with renal disease, diabetes, proteinuria, hypertension, gout, or congestive heart failure. Available evidence, largely based on small- to medium-sized trials, supports the use of hydration, bicarbonate, and low volumes of iso- or low-osmolar contrast in patients at risk. N-acetylcysteine or ascorbic acid may be of value in very high-risk patients.
Conclusions While several risk factors for contrast-induced nephropathy have been identified, the development of an effective prophylaxis strategy for contrast-induced nephropathy has been limited by our poor understanding of the pathophysiology and the clinical significance of this condition. Future research should focus on correctly identifying higher-risk patients and testing therapies in the setting of large well-powered clinical trials.
INTRODUCTION
Nephropathy induced by contrast media is a recognized complication of diagnostic and therapeutic procedures requiring parenteral administration of contrast and is the third leading cause of hospital-acquired acute renal failure, accounting for 12% of cases.1 Contrast-induced nephropathy is associated with significant consequences, including prolonged hospitalization, the requirement for dialysis, and an increased risk of death.2-3 Clinical outcomes associated with acute renal failure following cardiac catheterization can be catastrophic, with an in-hospital mortality rate of 20% in unselected patients and a 1-year mortality rate of up to 66% in patients with acute myocardial infarction and preexisting renal dysfunction.3-5
This review will critically evaluate current evidence for strategies to prevent contrast-induced nephropathy, present an evidence-based approach to this clinically important problem, and identify key areas for future research.
METHODS
A literature search was performed for English-language journal articles reporting risk factors for contrast nephropathy. We searched MEDLINE and EMBASE to identify publications from 1966 to January 2006, using the key terms radio contrast nephropathy, contrast media, risk, diabetes, nephrotoxicity, creatinine, coronary disease, coronary procedures, dehydration, and hypovolemia. A total of 59 studies were identified by the authors as being potentially relevant.
A separate literature search was performed for English-language clinical trials in contrast nephropathy. We searched MEDLINE and EMBASE to identify trials from 1966 to January 2006, using the key terms radio contrast nephropathy, contrast media, acetylcysteine, theophylline, sodium bicarbonate, HMG Co-A reductase inhibitors, ascorbic acid, kidney diseases, renal insufficiency, kidney failure, nephropathy, fenoldopam, saline, and diuretics. The citations of existing reviews and trials identified were evaluated by 2 reviewers to identify pertinent trials. Any study considered relevant by one or both reviewers was retrieved for further consideration. A total of 331 studies were identified as potentially relevant, of which 63 were randomized controlled trials (RCTs) of prophylaxis strategies for contrast-induced nephropathy. Criteria for study selection were publication in a peer-reviewed journal, controlled study design, and English language. Quality assessment was based on concealment of treatment allocation, double-blind design, intention-to-treat analysis, and rate of loss to follow-up.
DEFINITIONS AND EPIDEMIOLOGY
Although there is no universally accepted definition, contrast-induced nephropathy refers to the development of acute renal impairment following the intravascular administration of radiocontrast in the absence of other identifiable causes of renal failure. Most studies have used a 25% elevation in serum creatinine (SCr) or an absolute increase of 0.5 mg/dL (44 µmol/L) 2 to 7 days following contrast administration. Although relatively mild, these changes in kidney function are associated with clinically important adverse short- and long-term outcomes.2-3,6-7 The causal pathway linking the development of contrast-induced nephropathy to adverse cardiovascular outcomes has not been established. Whether contrast-induced nephropathy is directly responsible for the observed increase in mortality, or is simply a marker for illness acuity and/or comorbidity, remains unknown.
Based on these definitions, the incidence of contrast nephropathy in patients undergoing diagnostic interventions requiring contrast is low (1.6%-2.3%).8 Intra-arterial administration of radiocontrast might be more likely to lead to contrast-induced nephropathy than the intravenous route,9 although other causes of acute renal failure (such as atheroemboli) may also be triggered by arteriography. The incidence of contrast-induced nephropathy is dependent on the definition and is considerably lower when cases are defined based on a greater absolute increase in SCr level, or if the postprocedure SCr measurement is obtained earlier.10 Acute renal failure due to contrast-induced nephropathy is generally nonoliguric and reversible. In most cases, the SCr level peaks between 2 and 5 days after contrast exposure and returns to normal within 14 days.11-12 Registry data report the incidence of contrast-induced nephropathy requiring dialysis treatment to be approximately 0.4%.13
RISK FACTORS
The most commonly identified risk factors for contrast-induced nephropathy are listed in the Box. Most have been identified through retrospective analysis of databases cataloging coronary angiographic procedures. Unfortunately, periprocedural hydration and an accurate assessment of comorbidity have rarely been captured in these data sources, so estimates of the risk attributable to individual factors are unreliable.
| Box. Risk Factors for Contrast-Induced Nephropathy
Patient Related
- Chronic kidney disease3, 13-19
- Diabetes mellitus13, 16-17,19
- Urgent/elective procedure16
- Intra-aortic balloon pump16-17,20
- Congestive heart failure13, 15, 17, 20
- Age3, 17
- Hypertension21-22
- Low hematocrit17, 23
- Hypotension17, 22
- Left ventricular ejection fraction <40%22
Not Patient Related
- Contrast properties
- High osmolar contrast7, 24
- Ionic contrast25-28
- Contrast viscosity29-30
- Contrast volume3, 13, 16-19,31-34
RETURN TO TEXT |
|
Patient-Related Factors
Patients with diabetes and chronic kidney disease appear to be at the highest risk for developing contrast-induced nephropathy. These patients have reported rates of contrast-induced nephropathy that are approximately 4-fold higher than those without diabetes or preexisting renal impairment.25, 35
Hypovolemia and/or decreased effective circulating volume are well-recognized risk factors for contrast-induced nephropathy, but have never been directly assessed in clinical trials. Indirect evidence comes from studies that show a benefit of intravenous hydration36-37 and the deleterious effect of diuretics.38 Conditions resulting in a low effective circulating volume such as cardiogenic shock,16 use of an intra-aortic balloon pump,16-17,20 hypotension,17, 22 congestive heart failure (CHF),13, 15, 17, 20 and ejection fraction less than 40%22 are also identified risk factors for contrast-induced nephropathy.
Female sex is a frequently cited14, 21, 39 but somewhat controversial risk factor for contrast-induced nephropathy. A more recent analysis of 1383 patients40 suggests that women may have unfavorable baseline characteristics (older age, more frequent hypertension and diabetes, lower baseline kidney function) that put them at risk for contrast-induced nephropathy. After adjustment for these confounders, female sex did not appear to independently increase risk.
Non–Patient-Related Factors
Radiocontrast media are frequently classified on the basis of osmolality, which is determined by the ratio of iodine atoms to osmotically active particles.41-42 A comparison of commonly used radiocontrast agents is presented in Table 1.
|
|
|
|
Table 1. Properties of Commonly Used Radiocontrast Media
|
|
|
A direct correlation between osmolality and nephrotoxicity is well established in contrast agents with an osmolality greater than 780 mOsm/kg. A 1992 meta-analysis43 pooling data from 25 randomized trials showed that the risk of contrast nephropathy was significantly greater with high-osmolality (>1400 mOsm/kg) radiocontrast agents in patients with preexisting renal disease.
While the previously publicized findings of a study comparing iohexol (a low-osmolar agent [600-800 mOsm/kg]) with iodixanol (an iso-osmolar agent [290 mOsm/kg]) initially suggested additional nephroprotection with a further reduction in radiocontrast osmolality,24 the differing physicochemical properties of radiocontrast agents may in fact be a more important mediator of nephrotoxicity. Pooled analyses evaluating the nephrotoxicity of differing contrast agents used in recent randomized interventional trials29-30 showed that iohexol was associated with a significantly increased risk of developing contrast-induced nephropathy compared with either iopamidol (another agent with similar osmolality) or iodixanol (an iso-osmolar radiocontrast agent) (contrast-induced nephropathy 25% vs 13.5% and 11%, respectively; both P<.05). More importantly, there was no significant difference in the reported rates of contrast nephropathy associated with iopamidol and iodixanol, which can perhaps be explained by the increased viscosity of iodixanol relative to many of the low-osmolar agents.
Whether ionic compounds are more nephrotoxic than nonionic compounds remains somewhat controversial, as previous studies are frequently confounded by differences in osmolality. Several randomized trials of ionic vs nonionic contrast showed no difference in rates of contrast-induced nephropathy.25-28 However, post hoc analysis of one study did demonstrate that patients with preexisting renal dysfunction were less likely to develop contrast-induced nephropathy when nonionic radiocontrast media was used.25
RISK STRATIFICATION
Several attempts have been made to develop a clinical tool for the purposes of risk stratification,13, 15-17,44-45 but none have been validated prospectively or in other databases. In addition, none of the databases captured comorbidity or prophylactic interventions to prevent contrast nephropathy such as the administration of hydration. As all risk scores were derived from patients undergoing coronary angiography, they may not specifically apply to the use of parenterally administered contrast in other settings (ie, intravenous contrast).
Two risk scores16-17 for contrast-induced nephropathy developed from large interventional cardiology databases may be the most generalizable to this patient population. Bartholomew et al16 (n = 20 479) identified 8 variables that were associated with contrast-induced nephropathy (creatinine clearance <60 mL/min [1.0 mL/s], use of an intra-aortic ballon pump, urgent coronary procedure, diabetes, CHF, hypertension, peripheral vascular disease, contrast volume) and created 4 risk categories based on their analysis. They defined contrast-induced nephropathy as a greater than 1 mg/dL [88.4 µmol/L] rise in SCr with no specified time frame for post SCr measurement. By this definition, contrast-induced nephropathy occurred in 2% of patients. Patients in the highest risk group had a 28% risk of developing contrast nephropathy and a 17% risk of death. Mehran and colleagues17 (n = 8357) identified 3 additional characteristics that were associated with increased risk: older age, the presence of hypotension, and anemia. They used a less stringent definition of contrast-induced nephropathy (change in SCr  25% or 0.5 mg/dL [44.2 µmol/L] at 48 hours), which may partially account for the higher reported incidence of contrast-induced nephropathy (13.1%).
PROPHYLAXIS STRATEGIES
The mechanism by which contrast agents produce nephrotoxicity is poorly understood but probably includes a reduction in renal perfusion resulting in regional hypoxia, as well as direct tubular toxicity. Therapies studied to date have targeted renal vasoconstriction and hypoxia-induced oxidative stress with limited success.
Table 2, Table 3, and Table 4 summarize all identified RCTs investigating prophylaxis strategies for contrast-induced nephropathy. Most studies were small and were consequently underpowered to detect a clinically significant benefit. Few studies present sample size or power calculations, and most were not double-blinded (Table 5). Many were not analyzed as intention to treat, and loss to follow-up was rarely reported (Table 5).
|
|
|
|
Table 2. Study Demographics and Clinical Characteristics: Hydration and Diuretics, Dopamine, Fenoldopam, and Theophylline/Aminophylline
|
|
|
|
|
|
|
Table 3. Study Demographics and Clinical Characteristics: Calcium Channel Blockers and N-Acetylcysteine (NAC)
|
|
|
|
|
|
|
Table 4. Study Demographics and Clinical Characteristics: Hemodialysis, Hemofiltration, and Other
|
|
|
|
|
|
|
Table 5. Quality Assessment of the Included Studies
|
|
|
Hydration
Early studies evaluating the renal effects of radiocontrast administration in dogs demonstrated a reduction in renal perfusion lasting up to 20 hours after radiocontrast administration.108 While no RCT has studied the benefits of hydration alone, it seems plausible that adequate hydration may counteract some of the putative hemodynamic effects that may lead to contrast-induced nephropathy.
We identified 10 studies that evaluated the effects of various hydration protocols and diuretics in the incidence of contrast-induced nephropathy. Four studies38, 49, 51, 54 compared forced diuresis (furosemide and/or mannitol) with hydration, of which 3 showed a significant increase in the rate of contrast-induced nephropathy in the groups receiving diuretics. Two studies46-47 evaluated bolus intravenous infusions of 0.9% saline (250-300 mL) immediately before or during cardiac catheterization vs slow intravenous hydration 12 hours prior to the procedure. Neither found a significant difference between treatment groups; however, both studies were small (n = 39 and n = 37) and the event rates were low. Two additional studies36, 50 compared oral hydration with prolonged intravenous hydration (12 hours before and after) and found contradictory results. Taylor et al50 (n = 36) found no difference between treatment groups, although the oral hydration group in this study received 6 hours of intravenous hydration in addition to their oral intake. Trivedi et al36 (n = 53) found that oral hydration alone appeared to be inferior to intravenous hydration with respect to the development of contrast-induced nephropathy (34.6% vs 3.7%; P = .005) in patients with normal renal function undergoing cardiac catheterization. Interestingly, the incidence of contrast-induced nephropathy in the oral hydration group was much higher than expected in this patient population.
In addition to timing and route of hydration, other factors, such as fluid tonicity and fluid composition, may also play a role. Single studies supporting the use of isotonic vs half isotonic saline37 and sodium bicarbonate48 suggest that isotonic fluids may be superior to hypotonic fluids, likely because of their enhanced ability to expand intravascular volume. Sodium bicarbonate may provide additional renoprotection by alkalinizing renal tubular fluid and thereby minimizing tubular damage.
In summary, there is suggestive, but incomplete, evidence supporting the use of hydration as a prophylaxis measure for contrast-induced nephropathy. Questions still remain about whether all patients benefit equally from this treatment, as well as the optimal type, route, volume, and timing of hydration administration. Although hydration is a generally benign therapy, it is logistically challenging to implement, as most angiographic procedures are performed on outpatients (in whom 6 to 12 hours of preprocedural hydration may not be possible). These issues deserve further study.
Vasodilators
Dopamine/Fenoldopam. The benefits of vasodilation and increased renal blood flow associated with "renal dose" dopamine were studied in 4 RCTs,52-54,60 evaluating a variety of patients with both normal and impaired baseline renal function. While not directly comparable (all used different definitions of contrast-induced nephropathy), none showed a benefit of dopamine administration. Currently, there is no basis for the use of dopamine in preventing contrast-induced nephropathy.
Fenoldopam, a dopamine-1 receptor agonist with vasodilatory properties, has also been extensively studied. Three randomized trials55, 58-59 (416 participants) have compared fenoldopam with placebo. The relative risk of contrast-induced nephropathy in one trial favored fenoldopam55 but was not significant. Continued interest in fenoldopam has persisted, despite the lack of benefit shown in the CONTRAST trial.58 Additional randomized studies comparing fenoldopam with N-acetylcysteine56-57 found no benefit associated with the use of fenoldopam.
Critics argue that the doses of dopamine and fenoldopam used in these trials may have been insufficient to produce renal vasodilation.109 Despite this, we do not feel that further studies are warranted, given the significant adverse effect profile of these drugs (ie, dopamine: arrhythmias; fenoldopam: systemic hypotension) and the difficulties associated with intravenous administration.
Theophylline. Small studies of theophylline as a potential prophylaxis agent for contrast nephropathy have yielded conflicting results. To date there have been 9 trials60-68 (601 participants) comparing theophylline or aminophylline with no active treatment. Relative risks for contrast-induced nephropathy ranged from 0.07 to 1.7 (median, 0.25). Three of 5 favored theophylline (one was statistically significant). Differences in mean change from baseline in SCr level between treatment groups ranged from –0.29 to 0 mg/dL (median, –0.14 mg/dL). Six of 8 trials reporting outcomes of contrast-induced nephropathy or a change in SCr favored theophylline/aminophylline therapy (2 were statistically significant).
Meta-analysis has identified considerable heterogeneity among the studies.110 There was variability in the inclusion criteria, the method and schedule of theophylline/aminophylline administration, hydration protocols, and the type of contrast administered. Few of the trials compared the incidence of adverse events between treatment groups. We are unable to recommend theophylline as a prophylaxis agent based on the currently available data.
Calcium Channel Blockers/Other Agents. Three small RCTs of calcium channel blockers vs placebo69-71 (240 participants in total) in patients with normal renal function who received contrast media showed no difference between treatment groups. However, these trials lacked sufficient statistical power to detect clinically significant outcomes. Small underpowered trials of other agents with vasodilating properties such as atrial natriuretic peptide,103 an endothelin antagonist,107 prostaglandin E1,102 angiotensin-converting enzyme inhibitors,101, 106 and L-arginine105 have shown no benefit, and in some cases, potential harm107 with the use of these agents.
Antioxidants
N-Acetylcysteine. N-acetylcysteine has been the most widely studied of all prophylaxis strategies, although the mechanism for its purported nephroprotective action is unclear. N-acetylcysteine might act by scavenging oxygen free radicals111-112 or by enhancing the vasodilatory effects of nitric oxide.113
We identified 22 trials72-93 (2918 participants) comparing N-acetylcysteine with placebo. Relative risks for contrast-induced nephropathy ranged from 0.11 to 1.5 (median, 0.72). Eleven of 20 trials that reported contrast-induced nephropathy and 13 of 20 trials that reported a change in SCr level as an outcome favored N-acetylcysteine prophylaxis (5 were statistically significant). Differences in treatment means in change from baseline SCr ranged from –0.6 to 0.1 mg/dL (median, –0.03 mg/dL). Negative values confer lesser reductions or larger improvements in renal function in the N-acetylcysteine groups.
Twelve meta-analyses have been published on this topic to date.114-125 Nine have presented pooled risk estimates suggesting benefit, but most have found significant unexplained heterogeneity in the analyses leading to inconclusive results. Differences in contrast media, definitions of contrast-induced nephropathy, patient selection (undocumented differences in comorbidity), cointerventions, N-acetylcysteine dose and route of administration, as well as the timing of the procedure (urgent vs elective) may have contributed to the heterogeneity observed in the pooled analyses. New trials that address these issues are required before a final conclusion can be reached. We believe that the available data do not allow definitive conclusions about the efficacy of N-acetylcysteine for prevention of contrast nephropathy.
Ascorbic Acid. The results of a double-blind RCT evaluating the use of the antioxidant ascorbic acid10 to prevent contrast-induced nephropathy in 231 patients undergoing coronary angiography are encouraging and deserve further study. The authors defined contrast-induced nephropathy by a 25% or higher rise in SCr level 2 to 5 days postprocedure and found that ascorbic acid significantly reduced the risk of this outcome (odds ratio, 0.38; 95% confidence interval, 0.17-0.85).
Extracorporeal Removal of Contrast
Hemodialysis effectively removes radiocontrast126-128 and has been proposed as a prophylaxis treatment for contrast-induced nephropathy. Four small randomized trials have considered this question in patients with impaired renal function.94-97 Two found no benefit and the largest (n = 113) suggested that hemodialysis was harmful, since more patients in the treatment group required ongoing dialytic support.97
On the other hand, one group has reported that hemofiltration dramatically reduces the risk of contrast-induced nephropathy compared with hydration alone in 2 randomized trials of patients with impaired renal function.98-99 While encouraging, the provision of hemofiltration is invasive and impractical except in those at very high risk of contrast-induced nephropathy. A recent economic analysis129 suggests that hemofiltration may be cost-effective in patients with a baseline SCr measurement greater than 265 µmol/L if the magnitude of risk reduction seen in these studies is reproducible. The apparent benefits of hemofiltration need to be replicated in another center before it can be widely recommended.
SUGGESTED MANAGEMENT STRATEGY
Given the relatively low incidence of contrast-induced nephropathy, it would be impractical and costly to routinely measure SCr levels for all patients scheduled for diagnostic procedures requiring parenteral contrast administration. Guidelines published by the European Society of Urogenital Radiology suggest that all patients referred for contrast-enhanced diagnostic examinations should be asked about a history of renal disease, diabetes, proteinuria, renal surgery, hypertension, and/or gout.130 Given the identification of left ventricular function as a predictor of contrast-induced nephropathy in several retrospective analyses,16-17 we believe that a history of CHF should also be sought. Patients with any of these conditions and all patients undergoing angiographic procedures should have an SCr measurement within 7 days in advance of the scheduled examination. The detection of abnormal renal function with or without diabetes and CHF constitute high risk for contrast-induced nephropathy. In our opinion, those patients requiring urgent procedures before renal function can be measured should also be considered high risk. Our recommended management strategy for patients with risk factors for contrast-induced nephropathy is presented in the Figure.
|
|
|
|
Figure. Strategy for Management of Patients With Risk Factors for Contrast-Induced Nephropathy
*See Box for listing of risk factors for contrast-induced nephropathy.
|
|
|
FUTURE DIRECTIONS
The discovery of novel therapies for the prevention of contrast-induced nephropathy has been hampered by an incomplete understanding of its pathophysiology. Experimental and preclinical studies should remain a priority for investigators working in this field. However, existing trials evaluating currently available therapies also have significant limitations. Our review of the literature has identified 4 main problems with the randomized studies to date: (1) variation in the definition of contrast-induced nephropathy; (2) inability to accurately identify high-risk patients; (3) inconsistency in the administration of cotherapies like hydration; and (4) small sample sizes with suboptimal study designs.
Perhaps the most problematic issue is the wide variation in the definition of contrast-induced nephropathy. No standard definition has been used in clinical trials, and the clinical relevance of commonly used definitions (such as a 25% or 0.5 mg/dL [44.2 µmol/L] increase in SCr) remains unknown. While even modest increases in SCr have been associated with an increase in in-hospital and long-term mortality, it is hard to imagine a pathophysiological link between the two, unless the development of contrast-induced nephropathy is simply a marker for underlying comorbidity. Furthermore, no convincing data demonstrate that reducing the incidence of contrast nephropathy reduces the incidence of adverse outcomes. Therefore, we recommend development and adoption of a consensus definition of contrast-induced nephropathy based on more clinically relevant criteria. One example of such a definition might be the composite of doubling in SCr or the requirement for dialysis.
Second, study investigators have been unable to accurately identify high-risk patients for enrollment in contrast-induced nephropathy prophylaxis studies. Even among patients with preexisting renal impairment, the reported incidence of contrast-induced nephropathy varies between 20% and 50%. Further studies are required to identify additional risk factors for contrast-induced nephropathy and validate and refine existing risk scores. This information can be used to select high-risk patients for inclusion in future studies, thus increasing statistical power and the likelihood of obtaining useful information.
Third, clinical trials have been inconsistent in the administration of co-interventions such as hydration. In addition, the risk imposed by choice of radiocontrast agent and the route of administration has been largely ignored in clinical trials to date. We recommend that future trials ensure equal use of cointerventions between study groups, and that the type of contrast media be standardized.
Fourth, most studies have been small, often poor in quality (Table 5), and usually underpowered to detect a clinically significant difference in outcome. Future studies should be powered to detect differences in clinically meaningful outcomes such as mortality, the need for dialysis, or resource use. A large-scale multicenter study would likely be needed to accrue a sufficient number of patients. We discourage further attempts at small randomized studies, as they use valuable limited research resources yet generally do not yield definitive results. Unfortunately, there is little incentive for pharmaceutical companies to sponsor large clinical studies, as the interventions that have been proposed are unlikely to generate significant revenue. Further investigation in this area will probably need to be funded by national funding bodies.
CONCLUSIONS
|
