Does Safe Dosing of Iodinated Contrast Prevent Contrast-Induced Acute Kidney Injury?

Abstract

Background

Previous work on contrast-induced acute kidney injury (CI-AKI) has identified contrast volume as a risk factor and suggested there is a maximum allowable contrast dose (MACD) above which the risk of CI-AKI is markedly increased. We hypothesized there is a relationship between contrast volume and CI-AKI and there might be reason to track incremental contrast volumes above and below the MACD limit.

Methods and Results

Consecutive patients undergoing percutaneous coronary intervention (PCI) were prospectively enrolled from 2000 – 2008 (N=10,065). Patients on dialysis prior to PCI were excluded (N=155). MACD was defined as (5 mL×body weight (kg))/baseline serum creatinine (mg/dL)) and divided into categories where 1.0 reflects the MACD limit: ≤MACD ratios (< 0.5, 0.5-0.75, 0.75-1.0) and exceeding MACD (1.0-1.5, 1.5-2.0, and > 2.0). CI-AKI was defined as a ≥0.3 (mg/dL) or ≥50% increase in serum creatinine from baseline or new dialysis. Multivariable regression was conducted to evaluate the effect of exceeding the MACD on CI-AKI. 20% of patients exceeded the MACD. Risk-adjusted CI-AKI increased by an average of 45% for each category exceeding the MACD (OR: 1.45; 95%CI: 1.29-1.62) Adjusted odds ratios (OR) for each category exceeding the MACD were 1.60 (95%CI: 1.29-1.97), 2.02 (95%CI: 1.45-2.81), and 2.94 (95%CI: 1.93-4.48). CI-AKI for contrast dose <MACD showed no statistical difference (p=0.5).

Conclusions

Contrast volume is a key risk factor for CI-AKI and matters the most in the highest risk patient. The incremental use of contrast beyond the MACD is associated with an increased risk of CI-AKI.

Short Commentary

Contrast volume has been one of the many causal factors for patients developing contrast-induced acute kidney injury (CI-AKI). We investigated whether calculating a tailored dose of contrast based on the size of the patient (body weight, kg) and baseline renal function (serum creatinine, mg/dL). This tailored contrast dose is referred to as the maximum allowable contrast dose (MACD): 5mL × body weight (kg)/serum creatinine (mg/dL). We discovered that when the MACD dose was exceeded during a percutaneous coronary intervention, the risk of CI-AKI increased linearly, supporting a dose-dependent effect. This work can assist operators in approximating the increased risk of CI-AKI when the MACD is exceeded by 0 – 50% (60% increased risk of CI-AKI), 50 – 100% (2-fold increased risk of CI-AKI, and >100% (3-fold increased risk of CI-AKI). The MACD may be used as a tool to guide safe dosing of contrast in the catheterization laboratory and incorporated as part of the team’s “time out” prior to initiating a case.

Over 1.2 million cardiac catheterizations are conducted in the United States each year,¹ precipitating approximately 120,000 new cases of contrast-induced acute kidney injury (CI-AKI). CI-AKI results from acute tubular necrosis from nephrotoxic iodinated contrast dye administered intra-arterially during the procedure and from other homeostatic factors^{2, 3} and is a serious complication resulting in increased short- and long-term mortality and morbidity.⁴

Recently, there has been more focus on preventing CI-AKI with hydration and pharmaceutical interventions than on the older concept of a maximal acceptable contrast dose (MACD).^{5, 6} While hydration and pharmaceutical interventions to protect against kidney injury are important, the additional issue of volume of contrast and risk of kidney injury has been lacking. Therefore, we sought to identify the rate of CI-AKI and other clinical outcomes among patients that did and did not exceed the MACD of iodinated contrast. The purpose of this investigation is to identify the relationship between the volume of current contrast agents and patient body weight and baseline renal function and determine if tailoring the dose of contrast could place patients at higher or lower risk of CI-AKI during a percutaneous coronary intervention (PCI).

Methods

The Dartmouth Dynamic Registry (DDR) is a large prospective, clinical, consecutive patient registry of patients undergoing diagnostic or interventional cardiovascular catheterization procedures at the Dartmouth-Hitchcock Medical Center.⁴ In the present study, consecutive patients undergoing PCI were prospectively enrolled in the Dartmouth Dynamic Registry from 2000–2008 (N=10,065). Patients with pre-existing dialysis-dependent renal failure were excluded (N=155). The registry was approved by the Dartmouth Center for the Protection of Human Subjects. The authors had full access to the data and take responsibility for its integrity. All authors have read and agree to the manuscript as written.

Data collection and definitions

Patient and disease characteristics, including history of comorbidities, procedural characteristics and clinical outcomes are prospectively collected in the DDR Registry and have been previously described.⁴ The prospective cohort did not capture complete pre- and post-procedure prophylaxis for NaCl, NaHCO₃, N-acetylcysteine. Rehman et al., recently demonstrated that N-acetylcysteine does not falsely lower serum creatinine.⁷ The last pre-procedure serum creatinine and a routine post-operative serum creatinine were obtained on all patients within 24 hours of the procedure prior to discharge. Subsequent serum creatinines were available only to the extent that they were thought to be clinically relevant by the patient’s physician and obtained by linking to the in-hospital laboratory system. The last pre-procedure and peak post-procedure creatinine were used to define CI-AKI. CI-AKI was defined as a ≥0.3 (mg/dL) or ≥50% increase in peak serum creatinine from baseline or new onset of dialysis-dependent renal failure according to the AKIN criteria during the index catheterization admission.⁸ Maximum acceptable contrast dose (MACD) was defined as (5 mL × body weight (kg))/baseline serum creatinine (mg/dL)) as defined by Cigarroa and colleagues.⁶ Ratios of contrast volume: MACD were calculated. Ratios ≤1 represent contrast volumes that did not exceed the MACD. Ratios > 1 represent contrast volumes administered that exceeded the MACD threshold.

Statistical Methods

Baseline patient disease and characteristics, procedural characteristics, and clinical outcomes were compared between patients exceeding versus not exceeding the MACD exposure by χ² tests and student’s t-tests. To help control for confounding when comparing the incidence of CI-AKI between these groups, we developed a multivariate logistic regression model for this endpoint. Univariate logistic regression was used to identify risk factors for CI-AKI. Those with a p-value < 0.10 were incorporated into a multivariate logistic regression made parsimonious by comparing nested models using the likelihood ratio chi-square test.⁹ Variables in the final model included age, sex, body surface area, diabetes, hypertension, congestive heart failure, chronic obstructive pulmonary disease, priority of PCI, number of stents inserted, and baseline eGFR. Adjusted rates of CI-AKI were then calculated and compared using direct standardization.

To assess the robustness of our findings, we used propensity score matching as an alternative method to control for differences in the cohorts of patients exceeding versus not-exceeding the MACD. In this approach, logistic regression was used to calculate for each patient the probability they would exceed the MACD based on the previously developed Mehran risk score (congestive heart failure, hypotension, intra-aortic balloon pump, age > 75 years, anemia, diabetes mellitus, and eGFR).¹⁰ Patients were then stratified into 13 blocks based on their predicted probability of exceeding the MACD using the “pscore” algorithm in Stata, such that patients within a given block had similar predicted probabilities. Within each block, the algorithm checked to insure a reasonable balance of patients exceeding versus not-exceeding the MACD, the exposure of interest.

Since contrast volume was an aspect of our exposure (MACD), we did not include contrast volume as part of our propensity matching.¹⁰ We then calculated the average treatment effect associated with exceeding versus not-exceeding the MACD using conditional logistic.

In a subgroup analysis, we evaluated additional categories above and below the threshold (1.0) according to the ratio of contrast volume : MACD; the categories were as follows: < 0.5, 0.5–0.75, 0.75–1.0, 1.0–1.5, 1.5–2.0, and > 2.0. These categories were plotted with crude and direct adjusted rates of CI-AKI as described above. Multivariable logistic regression was used to test for differences across categories above (< 1.0, 1.0–1.5, 1.5–2.0, and > 2.0) and below (< 0.5, 0.5–0.75, 0.75–1.0). All analyses were conducted using Stata 9.0 statistical software (College Station, TX).

Results

Twenty percent of patients undergoing PCI procedures exceeded the MACD (Table 1). Patients differed with regard to baseline (Table 1) and procedural characteristics (Table 2). Patients exceeding the MACD threshold were more likely to be in shock, have two or three vessel disease and left main stenosis, and more stents (Table 2).

Table 1.

Patient and disease characteristics

Variables	≤MACD	>MACD	P-value
Number (N=9,910)	7,952	1,958
Patient and disease characteristics
Age (years)	63.2±12.2	70.2±11.6	<0.001
Sex (female, %)	29.5	28.8	0.529
BSA (%)
<1.8	21.9	35.3	<0.001
1.8–2.0	27.8	32.8
2.0–2.2	31.0	23.0
≥2.2	19.4	8.9
Morbid obesity (%)	9.9	5.1	<0.001
Diabetes (%)	27.2	28.5	0.254
Hypertension (%)	62.4	72.1	<0.001
Peripheral vascular disease (%)	8.5	16.8	<0.001
COPD (%)	8.5	12.5	<0.001
Hypercholesterolemia (%)	63.0	66.0	0.012
Cardiac disease
Congestive heart failure (%)	9.1	18.2	<0.001
Family History CAD (%)	42.3	39.6	0.032
Coronary artery disease (%)	40.1	49.7	<0.001
Prior myocardial infarction (%)	24.8	32.3	<0.001
Prior therapy and priority
Prior coronary intervention (%)	26.8	25.6	0.263
Prior CABG (%)	12.5	20.7	<0.001
Priority (%)
Elective	49.5	47.7	<0.001
Urgent	32.2	39.6
Emergent	18.3	12.7
Baseline eGFR (mL/min/1.73 m2)
≥90	31.8	7.3	<0.001
60–89	51.0	40.0
30–59	16.6	45.6
15–29	0.7	6.1
<15	0.0	1.0
Mehran risk score	6.2±3.8	10.4±4.5	<0.001

MACD: maximum acceptable contrast dose; P-value: chi-square test or t-test; BSA: body surface area; COPD: chronic obstructive pulmonary disease; CAD: coronary artery disease; CABG: coronary artery bypass graft surgery; eGFR: estimated glomerular filtration rate using MDRD equation.

Table 2.

Procedural characteristics

Variables	≤MACD	>MACD	P-value
PCI Indications
Stable angina (%)	23.9	29.2	<0.001
Unstable angina (%)	31.3	27.0	<0.001
Non-ST elevation MI (%)	14.3	20.9	<0.001
ST elevation MI (%)	10.9	7.4	<0.001
Post-infarct angina (%)	11.5	11.7	0.766
Cardiogenic shock (%)	1.3	2.6	<0.001
Coronary disease (%)
Single vessel	38.8	21.6	<0.001
Double vessel	34.9	35.5
Triple vessel	26.4	43.0
Critical left main	1.6	4.4	<0.001
Significant left main	4.6	9.7	<0.001
Contrast Media
MACD (mL)	469±148	321±105	<0.001
Contrast (mL)	258±102	431±139	<0.001
Low-osmolar (%)	25.2	18.4	<0.001
Iso-osmolar (%)	74.8	81.6
Case time (min)	86±43	128±53	<0.001
Stentcount (n)	1.5±1.0	2.3±1.5	<0.001
Percent of successful interventions (%)	96±19	93±22	<0.001

P-value for chi-square test or t-test; MACD: maximum acceptable contrast dose: ((5mL*body weight kg)/baseline creatinine).

Patients receiving contrast volumes in excess of the MACD were more likely to develop CI-AKI, new onset of dialysis-dependent renal failure, cardiac events, bleeding complications, receive transfusion, have a longer length of stay following PCI, and more likely to die during the PCI admission (Tables 3). Adjusted and propensity-matched odds ratios for the clinical outcomes are reported in Table 4. The propensity-matched results – using the Mehran risk score without contrast volume for matching – reported consistent findings with the adjusted modeling for CI-AKI, dialysis, and in-hospital mortality.

Table 3.

Index Admission Outcomes

Variables	≤MACD	>MACD	P-value
CI-AKI (%)	5.9	15.3	<0.001
Dialysis (%)	0.2	1.5	<0.001
Cardiac events* (%)	10.2	13.4	<0.001
Bleeding (%)	4.1	8.4	<0.001
Transfusion (%)	3.8	8.3	<0.001
In-hospital mortality (%)	1.7	3.4	<0.001
Length of stay, post-procedure (days)	2.2±4.1	2.8±4.6	<0.001

^*Cardiac events (new MI, cardiac arrest, stent thrombosis, not including recurrent angina or new congestive heart failure).

Table 4.

Risk-adjusted and Propensity-Matched Outcomes For Exceeding MACD

Variables	OR	95%CI	P-value
Risk-adjusted
CI-AKI	2.00	(1.67–2.40)	<0.001
Dialysis	6.41	(3.33–12.3)	<0.001
Cardiac events	1.35	(1.14–1.60)	<0.001
Bleeding	1.51	(1.20–1.90)	<0.001
Transfusion	1.59	(1.26–2.00)	<0.001
In-hospital mortality	1.38	(0.97–1.97)	0.073
Propensity-matched
CI-AKI	1.75	(1.49–2.07)	<0.001
Dialysis	3.13	(1.73–5.65)	<0.001
Cardiac events	1.13	(0.97–1.32)	0.125
Bleeding	1.26	(1.02–1.55)	0.032
Transfusion	1.30	(1.05–1.60)	0.016
In-hospital mortality	0.98	(0.71–1.35)	0.904

Adjusted odds ratios (OR) for age, sex, body surface area, diabetes, hypertension, congestive heart failure, chronic obstructive pulmonary disease, priority, baseline eGFR, stent count; propensity-matched for Mehran risk score using conditional logistic regression.

Ratios of contrast volumes that exceeded or did not exceed the MACD threshold are plotted in Figure 1. Risk-adjusted CI-AKI increased by an average of 45% for each category exceeding the MACD (OR: 1.45; 95%CI: 1.29-1.62). Adjusted odds ratios for each category exceeding the MACD were 1.60 (95%CI: 1.29-1.97), 2.02 (95%CI: 1.45-2.81), and 2.94 (95%CI: 1.93-4.48). CI-AKI for contrast dose <MACD showed no difference (p=0.5).

Figure 1. MACD Ratio and CI-AKI.

Calculated ratios of contrast volume to the predicted maximum acceptable contrast dose (MACD) are plotted by the crude (grey bars) and risk-adjusted (black diamonds) rates of contrast-induce acute kidney injury (CI-AKI).

Discussion

We have evaluated the clinical usefulness of predicting the MACD prior to PCI and have demonstrated that contrast volumes in excess of the MACD threshold linearly increase the risk for CI-AKI, while contrast volumes below the threshold hold to a lower plateau risk of CI-AKI. We repeated our propensity-matched analysis (all patients: OR: 1.75; 95%CI: 1.49-2.07) among patients with normal baseline renal function (eGFR≥60: OR 1.73; 95%CI: 1.32-2.26) and patients with chronic kidney disease (eGFR < 60: OR: 1.51; 95%CI: 1.21-1.89) and discovered exceeding the MACD threshold is associated with an increased risk of CI-AKI regardless of baseline renal function.

Radio-contrast dye has been postulated to increase vasoconstriction and cause increases in the extracellular calcium influx or decreases in renal blood flow and the loss of nitric oxide production leading to a loss of the outer medullary autoregulation and acute tubular necrosis.^{2, 3, 11} Romano and colleagues have demonstrated low- and iso-osmolar contrast media causes renal cell apoptosis through DNA fragmentation in the renal medulla from medullary hypoxia.¹²

Taliercio et al., showed that repeated exposure to contrast within 72 hours is associated with an increased risk of CI-AKI.¹³ Cigarroa et al., followed by developing the MACD equation currently used today to determine the threshold for safe contrast exposure customized to each patient.⁶ It is important to note that Cigarroa’s equation of MACD was developed during the time that high osmolar contrast media was used; this equation has not been validated in the low and iso-osmolar contrast media use of today. The MACD, or contrast material limit, is calculated by: 5 (mL) of contrast per Kg of body weight / baseline serum creatinine (mg/dL). Cigarroa and colleagues determined that a maximum of 300 mL of contrast should be used regardless of the tailored threshold for each patient.⁶ In terms of risk for CI-AKI, Cigarroa found that CI-AKI developed in only 2% of patients that had contrast exposure under the calculated threshold, but 38% of patients receiving contrast over the MACD developed CI-AKI.⁶ Freeman et al., showed that exceeding the MACD had a significant 6-fold increased risk of developing nephropathy requiring dialysis.¹⁴ Rihal et al., found contrast dose has been shown to be pernicious over customized thresholds. In PCI patients, each 100mL of contrast was associated with a 12% increased risk of CI-AKI.¹⁵ Another method proposed by Laskey et al, in identifying thresholds of contrast is volume-to-creatinine clearance ratio.¹⁶ Laskey proposes that dividing the volume of contrast by the baseline creatinine clearance (CrCl), whereby volume/CrCl ≥3.7 was significantly associated with an early increase in creatinine, or CI-AKI.¹⁶ Most recently, Marenzi and colleagues confirmed that contrast volume and exceeding the MACD is associated with an increased risk of CI-AKI by two-fold in patients with ST-elevation MI undergoing primary PCI.¹⁷ Our PCI cohort including all patients also found 20% of patients undergoing PCI exceeded the MACD. We also demonstrated that the relationship between exceeding the MACD threshold and CI-AKI was linear, whereby a 45% increase risk of CI-AKI was associated for each category (1-1.5, 1.5-2, and > 2) above the threshold. We recommend MACD should be determined prior to the procedure and implemented whenever feasible. Exceptions to this rule include complicated long cases and complete revascularization should be paramount over prevention of CI-AKI; however exceeding the MACD and increasing the potential threat of CI-AKI should be closely monitored and weighed in every case.

There are several limitations to consider. While we were able to confirm results from other institutions, we are reporting on a cohort of patients from a single medical center. Additional work is needed to confirm our results from a multi-centered registry. All patients received IV hydration prior to their PCI, however, our prospective cohort did not capture prophylaxis data and therefore we were unable to accurately capture true prophylaxis. We reported an increased risk of in-hospital events, including dialysis and mortality; the long-term post-discharge association of mortality and end stage renal disease should be investigated. We conducted our analysis on a consecutive cohort of PCI patients with complete pre- and post-PCI serum creatinine values. However, the length of stay ranged from 2.2-2.8 days on average and some patients may not have developed CI-AKI prior to discharge; therefore our capture of CI-AKI is under-representing the true burden of AKI in this population, as has been suggested by Maioli et al.¹⁸ While we conducted a rigorous analysis through multivariable analysis and propensity matching to deal with confounding, there may be residual confounding due to unmeasured covariates. The propensity score matching is not a randomized trial, however, it provides a retrospective method to align similar-risk patients in a cohort analysis. However, on average, patients exceeding the MACD have poorer baseline renal function, more comorbidities, a greater burden of CAD, and procedures that are associated with a higher dose of contrast. We controlled for these differences with adjustment and propensity matching and still found that when patients exceeded their MACD, it was associated with an increased risk of CI-AKI, as well as all other short-term outcomes. It is still possible, though not probable, that these patients did better with their high contrast volume procedures than they would have with lower volume procedures, because they had more complete revascularization and with less complete revascularization (lower contrast volume) they would have had a higher mortality. However, this does not negate the risk of CI-AKI and the subsequent risk of in-hospital and long-term mortality whereby the risk of mortality and subsequent morbidity for incomplete revascularization compared to developing CI-AKI has yet to be determined in the literature. Although we found differences between patients receiving contrast volumes above and below the MACD due to differences in baseline renal function and severity of case mix, we emphasize the necessity of using the MACD among high risk patients to help plan for the procedure through adequate hydration and prophylaxis, minimizing of contrast volume, and consider the risk : benefit ratio of a time-out in the cath-lab when the MACD threshold is approached. Finally, we are validating a twenty-year old equation to calculate the MACD; while this was developed on the last generation of high-osmolality contrast agents it varies in use in today’s cath-labs; a modification of this equation may be required, yet in the mean time we have found this to be a useful tool for guiding tailored care for patients.

Recent work by Nyman and colleagues investigate a modernization of the MACD dose to incorporate the role of grams of iodine in the risk modeling for MACD and the risk of CI-AKI.^{11, 19} Nyman and colleagues demonstrated that CI-AKI was elevated by more than 6-fold among patients receiving more than a 1:1 ratio of grams of iodine to eGFR.¹⁹ More work needs to be done to stream-line simple calculations in our cath-lab programs to inform decision making prior to the case. Our current investigation provides the interventional team with a tool for how much the risk of CI-AKI is increased when the MACD is exceeded. While alternatives to complex cases with long duration and high dye loads need to be considered (staged procedures, medial management, or surgery), the most cost-effective approach to ensure adequate protection against CI-AKI is hydration.

In conclusion, contrast volume does matter and matters most in the patients at highest risk for CI-AKI. Exceeding the MACD for a patient undergoing cardiac catheterization with intervention is associated with an increased risk of CI-AKI by 45% for each 50% in excess of the MACD threshold. Calculating the MACD prior to interventional cardiac catheterization procedures may be a useful tool for minimizing contrast exposure and aiding in decision making for staging interventional procedures. Additional caution should be used for contrast volumes below the MACD as CI-AKI remains an adverse outcome for patients even when the threshold is not exceeded.