Automated Contrast Injection in Contemporary Practice during Cardiac Catheterization and PCI: Effects on Contrast-Induced Nephr

call1.jpg [2]

calltab1.jpg [3]

calltab2.jpg [4]

call2.jpg [5]

calltab3.jpg [6]

Contrast-induced nephropathy (CIN) is a common problem after cardiac catheterization and percutaneous coronary intervention (PCI), and is associated with substantial morbidity and mortality.^1–3 The etiology of CIN has not been clearly delineated. Renal ischemia is the most likely mechanism, although several other mechanisms have been suggested to contribute to this problem.⁴ Preprocedural hydration, N-acetylcysteine and bicarbonate infusion have been used to reduce the incidence of CIN after contrast-related procedures.^5–8 The volume of contrast used during a procedure is also a factor that may correlate with the incidence of CIN.^3,9 Importantly, contrast volume is one of the few procedural variables that can be modified during a coronary catheterization procedure or PCI, and thus, potentially reduce the incidence of CIN. Automated contrast delivery systems are now commercially available and appear to allow lower procedural contrast volume without compromising image quality or safety.^10–12 Whether the use of an automated contrast injection system, by reducing contrast volume, can modify the incidence of CIN is uncertain. Accordingly, we assessed the incidence of CIN after cardiac catheterization and PCI using traditional hand injection techniques, and compared this to the incidence of CIN using an automated contrast delivery system.

Methods
Study population. Between April 2002 and November 2004, 9116 coronary catheterizations and PCI procedures were performed at Wake Forest University Baptist Medical Center (WFUBMC). We evaluated only those patients in whom post-procedure values of renal function were available prior to and >/= 3 days after the procedure. Patients were also excluded if their procedures occurred during the training period with the automated contrast injection system (n = 47), and if they were undergoing hemodialysis (n = 74). The remaining 2,175 patients formed the basis for the study. A total of 1,798 patients underwent diagnostic catheterization or PCI using traditional hand injection techniques. Beginning in January 2004, automated contrast injection was introduced, replacing hand injection, and 377 subsequent patients underwent diagnostic catheterization or PCI using this technique.
Patients were kept n.p.o. after midnight prior to an elective procedure. Preprocedural hydration was generally initiated on all hospitalized patients as an infusion of normal saline, 50–100 ml/hour, from 1–12 hours prior to the procedure. Patients with diabetes mellitus and renal dysfunction with baseline Cr > 1.5 mg/dl were also pretreated with N-acetylcysteine (American Regent, Inc., Shirley, New Jersey) 600 mg orally b.i.d., and in some patients with bicarbonate infusion⁵ at least 1 hour prior to the procedure. Postprocedure hydration was administered to all patients, and N-acetylcysteine was continued if used preprocedurally. Cardiac catheterization and PCI procedures were performed according to standard practice. Ioversol (Optiray 350; Mallinckrodt, Hazelwood, Missouri) is a low-osmolar, nonionic contrast agent that was used in all cases. The ACIST device (ACIST Medical Systems, Eden Prairie, Minnesota) is a self-purging injection syringe connected to an automated manifold that delivers contrast directly to a catheter via a single three-way stopcock. Injection volume, pressure and flow rate are all determined by a software-driven console, with contrast flow controlled by a hand-held actuator.
Outcomes evaluation. The procedural and hospital outcomes of all patients were independently evaluated by research nurses for the occurrence of complications at the time of discharge according to the American College of Cardiology National Cardiovascular Data Registry¹³ and entered into a database following each procedure. Blood urea nitrogen (BUN) and creatinine (Cr) values were obtained from each patient at baseline prior to the procedure, and at least once daily following the procedure. The highest value for both BUN and Cr in the 7 days following the procedure was used in the analysis of CIN. Glomerular filtration rates were estimated using the modification of diet in renal disease (MDRD) 4-variable equation GFR = 186.3 x (SCr)-1.154 x (age, years)^-0.203 x 1.212 (if black) x 0.742 (if female).^14,15
Definitions. CIN was defined as an increase in creatinine > 25% from preprocedural baseline, or an absolute rise in creatinine of more than 0.5 mg/dl.3–5 The development of acute renal failure (ARF) was defined as an increase in baseline creatinine to more than 2 mg/dl; a > 50% increase from an abnormal baseline; or the need for hemodialysis.¹⁶ In-hospital major adverse cardiac events (MACE) included death, myocardial infarction and need for emergent repeat revascularization (by either PCI or CABG).
Statistical analysis. Descriptive statistics (means and SD of continuous factors, frequency counts and relative frequencies of categorical factors) were calculated and compared for statistical significance using the Student’s t-test for continuous factors and Chi-square testing for categorical factors. Pearson correlation coefficients were calculated to examine possible correlation of contrast volume (mL) with percent change in creatinine. Univariate analysis was carried out using the Chi-square test to examine possible associations of covariates in Table 1 with CIN. Standard logistic multivariate regression models were constructed to examine the combined effect of covariates on the association of contrast volume with CIN. Parameter estimates for each covariate were calculated through maximum likelihood estimation. Backward selection was used to remove statistically nonsignificant predictors from the model. Previously known clinical predictors of CIN were retained in the model (i.e., diabetes and history of renal failure), regardless of statistical significance. Multicollinearity was assessed by examining condition indices and variance decomposition proportions for covariates in the logistic regression models. No multicollinearity issues were present in any of the final models. SAS, version 8.02 statistical software package (SAS Institute, Cary, North Carolina) was used for all statistical analysis.

Results
Patient and procedural characteristics are shown in Table 1. Baseline demographics were similar except for a slightly higher percentage of myocardial infarctions in the 7 days prior to their cardiac procedure in the automated injection group, and a slightly higher incidence of vascular disease in the hand injection group. The indication for cardiac catheterization and PCI was an acute coronary syndrome in the vast majority of patients studied in both groups (> 84%). There were slightly fewer diagnostic catheterization cases in the automated injection group compared to the hand injection group (55.4% versus 61.5%; p < 0.05). Sheath size was minimally larger in the automated injection group than the hand injection group (6.3 ± 0.8 Fr versus 6.2 ± 0.9 Fr; p < 0.05). Bicarbonate infusion or acetylcysteine use prior to catheterization or PCI were higher in the automated injection group (36.3%) compared to the hand injection group (24.8%), p < 0.05. Median fluoroscopy times were not statistically different between groups: 11.0 (5.3–19.2 interquartile range) minutes in the automated injection group and 9.7 (5.0–18.6 interquartile range) minutes in the hand injection group (p = NS). In-hospital MACE rates were similar in both groups (7.2%; p = NS), except for the incidence of contrast-induced nephropathy and renal failure noted below (Table 2).
Contrast volume delivered to the patient was significantly lower in the automated injection group (145.6 ± 107.6 ml) than the hand injection group (204.3 ± 147.1 ml), p < 0.05 (Figure 1). Contrast volume administered per case was also significantly lower for the automated injection group compared to the hand injection group for the catheterization, catheterization and PCI, and PCI subgroups.
Contrast-induced nephropathy (CIN) occurred less often in the automated injection group than in the hand injection group (13.3% versus 19.3%; p < 0.05), corresponding to a 31.1% reduction in the incidence of CIN (Table 2). The percent change in the preprocedure to post-procedure creatinine was also significantly less in the automated injection group (11.7% ± 27.9% versus 16.4% ± 40.7%; p < 0.05). Acute renal failure (ARF) occurred in 2.9% of the procedures with the automated injection device, and in 5.7% of the cases with the hand injection, representing a 49.1% reduction in the incidence of ARF (p < 0.05). There was a significantly greater decrease in estimated GFR post-procedure in the hand injection group than in the automated injection group (-9.9 +/- 20.5 ml/minute versus -7.8 ± 17.7 ml/minute; p < 0.05). The incidence of CIN and ARF for patients with diabetes mellitus and for a baseline creatinine > 1.5 mg/dL are also shown in Table 2. The incidence of CIN in these two subgroups was directionally and proportionately similar to the results observed in the overall group, but did not achieve statistical significance.
Univariate predictors of increased CIN included a history of renal failure, PCI procedure, age >/= 64 years, diabetes mellitus, baseline creatinine > 1.5 mg/dl, baseline BUN > 17 mg/dL, and each 100 ml of contrast delivered, while automated contrast injection use and male gender were associated with a decrease in the risk of CIN (Table 3). Stepwise multivariate regression identified history of renal failure, diabetes mellitus and contrast volume as independently predictive of increased CIN, while male gender and automated injection use were independently predictive of reduced CIN (Figure 2). Neither bicarbonate nor N-acetylcysteine had an independent effect on the incidence of CIN.

Discussion
In this retrospective observational study, the use of automated contrast injection in conjunction with contemporary strategies of hydration and N-acetylcysteine use for diagnostic catheterization and PCI procedures was associated with a 30% lower use in contrast volume, a 30% lower incidence of CIN, and a 50% lower incidence of acute renal failure. To the best of our knowledge, this is the first study to demonstrate that methods to reduce contrast use and strategies to minimize CIN are associated with a lower incidence of CIN and ARF after diagnostic catheterization and PCI procedures.
Automated contrast injection systems have been evaluated in prior studies and have demonstrated that their use results in lower contrast volume per case.^10–12 Kern and colleagues¹¹ found that the use of any automated contrast injection system with 4 Fr diagnostic catheters enhanced coronary artery visualization compared to manual injection at a comparable level of contrast volume per case. The results from our study confirm the contrast-sparing benefits of automated contrast injection use. Although not specifically measured, the lower contrast volumes with automated injection occurred as a result of lower contrast volume per injection, with a smaller reduction in the total number of injections per case. Also, although not specifically assessed in this study, coronary artery visualization was as good, if not better, with automated contrast use compared to hand injection.
Our observation that the incidence of CIN was lower in the automated contrast injection group is consistent with its effect on lowering contrast volume. Several studies have identified contrast volume as an independent predictor of CIN,^3,9 although this has not been a consistent finding. Data supporting the adverse effects of contrast have indicated that there appears to be a threshold for developing CIN at contrast volume >/= 200 ml,9 although lower volumes have also been associated with an increased incidence of CIN. In our study, the average contrast use in the patients developing CIN was 290 ml, consistent with these prior data. Rihal et al3 observed a 1.1 relative risk of CIN for each 100 ml of contrast used, suggesting that the total contrast volume used also increased the risk of CIN. In support of this, we also observed a 1.1 relative risk of CIN for each 100 ml of contrast used, independent of the method of delivering the contrast. While we observed a detrimental effect of contrast volume in this study, we did not assess the role of the type of contrast agent on the incidence of CIN. Since prior studies have indicated that the osmolality of contrast agents may affect the incidence of CIN, especially in high-risk patients, our findings cannot be extrapolated to contrast agents other than the one used in this study.⁶
The effects of strategies to prevent CIN during invasive procedures have been extensively studied.⁶ Preprocedural hydration with saline was shown to reduce the incidence of CIN, and was adopted universally. A more recent study⁵ indicated that a bicarbonate infusion might be better than saline in reducing the incidence of CIN, but this finding has not been confirmed in additional studies. Finally, N-acetylcysteine use has been examined in multiple studies. A recent meta-analysis indicated a generally favorable effect of N-acetylcysteine on the incidence of CIN,^7,8 although substantial heterogeneity in study results was noted. The greater use of N-acetylcysteine or bicarbonate infusion in the automated injection group reflects the growing awareness of the benefit of these strategies in reducing the incidence of CIN. Although multivariate analysis of the incidence of CIN in this study did not indicate that these strategies had an independent effect on the incidence of CIN, a beneficial effect cannot be excluded.
Study limitations. There are several factors not accounted for in this study, which may have influenced the results. First, this is an observational study and subject to biases in selection of patients that may have influenced the results. We limited the study to patients hospitalized for 3 or more days after their procedure in order to assess the incidence of CIN >/= 48 hours after their procedure, and did not assess outcomes in all patients undergoing procedures. The study cohort represented 24% of all patients undergoing cardiac catheterization or PCI procedures during the time period of the study. However, almost all of these patients underwent diagnostic catheterization and PCI after presenting with an acute coronary syndrome; both groups had similar lengths of stay, and the in-hospital MACE rates were identical. Thus, it is unlikely that there were significant baseline differences in the two study groups that may have affected the observed outcomes. Second, changes in procedural techniques may have occurred that could have influenced the study results, including shorter procedures requiring less contrast use. However, fluoroscopy times per case were unchanged using the two contrast injection techniques, suggesting that shorter cases per se were unlikely to have accounted for the reduction in contrast volume observed in this study. Third, as noted above, pharmacologic therapies used to limit CIN including the use of N-acetylcysteine⁷ and sodium bicarbonate⁵ infusion may have contributed to the decrease in CIN observed in this study, since there was an increase in use of these agents during the time period the automated injection device was introduced. However, multivariate logistic regression analysis did not identify either bicarbonate infusion or N-acetylcysteine use as predictive of CIN in this study. Further studies will need to be performed to assess the relative benefit of each of these factors on the incidence of CIN.

Conclusion
In conclusion, the use of an automated contrast injection system in conjunction with contemporary strategies of hydration and N-acetylcysteine use during diagnostic catheterization and PCI was associated with a significant reduction in use of contrast volume, as well as in the incidence of CIN. These data suggest that contrast-sparing strategies as part of routine diagnostic catheterization and PCI procedures may be extremely useful in reducing complications from these procedures.