Association of Iso-Osmolar vs Low-Osmolar Contrast Media With Major Adverse Renal or Cardiovascular Events in Patients at High Risk for Acute Kidney Injury Undergoing Endovascular Abdominal Aortic Aneurysm Repair

Objective. The purpose of this analysis was to examine the association of iso-osmolar contrast media (IOCM) vs low-osmolar contrast media (LOCM) with major adverse renal or cardiovascular events (MARCE) in patients at high risk of acute kidney injury (AKI), undergoing endovascular abdominal aortic aneurysm repair (EVAR). Methods. Patients at high risk of AKI (defined as age ≥75 years, or one or more of the following comorbidities: diabetes, anemia, chronic kidney disease (CKD stages 1-4) or congestive heart failure), undergoing EVAR from September 2012 to June 2018 were identified using the Premier Hospital Database. We compared the primary endpoint of MARCE (composite of AKI, AKI requiring dialysis, acute myocardial infarction [AMI], stroke/transient ischemic attack [TIA], and death) with IOCM vs LOCM via adjusted multivariable regression analyses. Results. Among 15,777 high-risk patients undergoing EVAR, the occurrence of in-hospital MARCE was 6.8%, including renal events (4.5%), AMI (0.8%), stroke/TIA (0.4%), and death (1.9%), IOCM was used in 7360 patients (47%). Multivariable modeling found IOCM was associated with 1.8% (95% confidence interval [CI], 0.4-3.3; P=.01) lower absolute risk for MARCE (23.9% relative risk reduction; 95% CI, 5.2%-44.2%). Conclusions. Use of IOCM vs LOCM in patients at high risk of AKI undergoing EVAR procedures was associated with a lower risk of MARCE. As prevention of AKI or cardiovascular events after EVAR procedures may lead to reduced morbidity and mortality, this finding may have important clinical implications and should be confirmed through randomized controlled clinical studies. 

J INVASIVE CARDIOL 2021;33(8):E640-E646. Epub 2021 July 16. 

Key words: acute kidney injury (AKI), EVAR, abdominal aortic aneurysms (AAA), chronic kidney disease

Approximately 45,000 abdominal aortic aneurysms (AAA) are diagnosed in the US population each year.¹ As a minimally invasive alternative to traditional open repair, endovascular aneurysm repair (EVAR) for AAA has evolved into first-line treatment, predominately in the elective setting, and the procedure is increasingly used in the contemporary management of AAA patients. Between 2000 and 2010 EVAR increased from 5% to 74% of AAA repairs and is rapidly replacing open repair procedures.¹ Patients undergoing EVAR frequently have comorbidities of diabetes, chronic kidney disease (CKD), congestive heart failure (CHF), and advanced age; therefore, they are at high risk of acute kidney injury (AKI) and adverse cardiac events. The diagnostic and interventional phases of endovascular AAA repair pose a risk for AKI as they require intra-arterial administration of iodinated contrast to facilitate aortography as well as manipulation of the peri-renal aorta increasing the risk for renal embolization or compromise of renal artery flow.^2-5 A prior study found the rate of AKI in EVAR to be 13%,⁵ with some studies suggesting lower risk of AKI with EVAR as compared with open repair, and others studies showing no difference.^3,4 The major adverse cardiovascular event (MACE) free survival rate was favorable in EVAR vs open procedures (30-day post EVAR 97.0% [95% CI, 96.4-97.5] vs open 90.8% [95% CI, 89.9-91.7]; 4 years after repair EVAR 72.9% [95% CI, 71.4-74.4] vs open 69.9% [95% CI, 68.3-71.3]).⁶

Use of iso-osmolar contrast media (IOCM) has been associated with both reduced renal and cardiac events as compared with low-osmolar contrast media (LOCM) in patients undergoing percutaneous coronary procedures. Specifically, IOCM use in several randomized trials and meta-analyses have shown a reduced rate of AKI or dialysis^7-11 while other studies and meta-analyses have suggested a more modest effect or no difference.^12-14 However, when high AKI risk populations were analyzed, there appeared to be a stronger association of IOCM with lower incidence of AKI^9,11 compared with LOCM, which has also been reflected in real world data.¹⁵ The association of contrast media type with adverse renal or cardiovascular outcomes has so far not been studied in at-risk EVAR patients.

The purpose of this analysis was to examine the association of IOCM vs LOCM with major adverse renal or cardiovascular events (MARCE) in patients at high risk of adverse renal events, undergoing EVAR procedures using a large nationally representative real-world dataset from the United States. 

We used the Premier Hospital Database (https://www.premierinc.com/), which is a compilation of the administrative claims data representing ~20% of all acute care hospitalizations in the United States for over 15 years. The Premier Hospital Database also contains information on the socio-demographic characteristics, comorbidities, interventional procedures, and medications. The dataset used in this study covered a 5-year period from September 2012 to June 2018 and included a total of 48,977,629 admissions. 

All data used to perform this analysis were de-identified and accessed in compliance with the Health Insurance Portability and Accountability Act. As a retrospective analysis of a deidentified database, the research was exempt from IRB review under 45 CFR 46.101(b) (4).

Study inclusion criteria. Primary inpatient hospital visits for EVAR procedures were included for patients with record of CM usage and meeting criteria for high risk of adverse renal events. Only primary EVAR procedures were included based on the following international classification of procedures coding detail: (1) 39.71 Endovascular implantation of other graft in abdominal aorta; (2) 04V03D6 Restriction of Abdominal Aorta, Bifurcation, with Intraluminal Device, Percutaneous Approach; (3) 04V03DJ Restriction of Abdominal Aorta with Intraluminal Device, Temporary, Percutaneous Approach; or (4) 04V03DZ Restriction of Abdominal Aorta with Intraluminal Device, Percutaneous Approach. By using these codes to select EVAR procedures, patients with AAA procedures via an open approach were not included. By only including patients when the primary reason for hospitalization was an EVAR procedure, we also excluded patients undergoing EVAR procedures incidentally due to other conditions. High-risk patients were defined as those with age ≥75 years, or one or more of the following: diagnosis of diabetes, anemia, chronic kidney disease (CKD) (stage 1-4 or unspecified) or congestive heart failure (CHF). Patients with end-stage renal disease or stage 5 CKD were excluded (see Appendix A for complete code listing).

Predictors and outcome variables. The primary outcome of interest was the composite endpoint of MARCE, defined as the composite of the following: (1) renal events — renal failure with dialysis or acute kidney injury with or without dialysis; (2) acute myocardial infarction (AMI); (3) stroke/transient ischemic attack; or (4) death (see Appendix B for coding detail). In order to qualify as an outcome during the hospital admission, any adverse event was required to be absent on admission and present at discharge. Each component of MARCE was also analyzed separately. 

Using the Premier’s Standard Charge Master (which is a comprehensive table of items billable to a patient or health insurance provider), we identified IOCM (iodixanol) and LOCM (iohexol, ioversol, iopamidol, and other) contrast media. IOCM (vs LOCM) was the main exposure variable of interest. 

Independent variables of interest included patient demographics, comorbid conditions, admission status and CM volume. Patient demographics for this analysis included age, race, sex and year. Comorbid conditions were measured via the Elixhauser Comorbidity Index (ECI) score.¹⁶ The ECI score includes 31 categories of comorbidities that are associated with mortality. These comorbidities were identified using diagnosis codes from the admission for the EVAR procedure. A composite ECI score was calculated from the comorbidity categories (see Appendix C for complete code listing). In addition to the ECI, if a patient had a record of chronic kidney disease (CKD stage 1-4 or unspecified) they were flagged as having CKD. Patients with CKD stage 5 and end-stage renal disease were excluded from this analysis. Hospital admission status was also considered. In Premier, admission status is categorized as follows: emergency (patient requires immediate medical intervention and is typically admitted through the emergency room), urgent (patient requires immediate attention and is generally admitted to the first available/suitable accommodation), trauma center (visit is to a designated trauma center), other or unknown (status is not known), and elective (patient’s condition permits adequate time to schedule the availability of suitable accommodations). For this study, admission status was categorized into two mutually exclusive groups: (1) elective admissions; vs (2) all other categories (emergency, urgent, trauma center, and other), which we labeled as non-elective. Finally, a proxy variable was created to measure contrast media (CM) volume in order to account for the association between contrast volumes and incidence of AKI.¹⁷ The CM volume was a dichotomous variable which indicated whether or not more than 100 mL of CM were utilized. This variable was created based on available formulae using the Premier Standard Charge Master to estimate the amount of CM usage.

Statistical analysis. We examined the association of the IOCM (vs LOCM) with the primary endpoint of MARCE via multivariable regression analysis using the hospital sites as fixed-effects (with hospitals as dummy variables) to control for observable and unobservable differences in severity of patients (and all other hospital factors such as surgical practices, treatments, staffing patterns, physician skill, etc.) across hospitals that may not only be associated with outcomes but also affect choice of contrast medium. The multivariable regression model adjusted for year, patient demographics (age, sex and race), admission status (elective vs non-elective), the ECI, CKD status and volume of CM. The hospital fixed-effect methodology was chosen to control for time invariant within hospital variation such as hospital protocols that specify when to use IOCM or LOCM based on perceived patient risk. The hospital fixed-effect methodology was used to control for variation that is otherwise unobservable but could impact the choice of CM and hence the results of the analyses. All statistical analyses in this study were performed using SAS software, version 9.4 (SAS Institute).

Patient characteristics. A total of 26,234 inpatient visits from 444 hospitals met the inclusion criteria of use of LOCM or IOCM for EVAR. The group of inpatients undergoing EVAR procedures was further refined to patients at high risk of adverse renal events, 60.1% (n = 15,777). This analysis set excluded patients with a record of stage 5 CKD or end-stage renal disease (Figure 1). These 15,777 procedures were grouped by CM (LOCM n = 8417; IOCM n = 7360), with the following CM agents making up the LOCM (n = 8417) cohort: iopamidol (42%), iohexol (38%), ioversol (18%) and other (2%). The CM cohorts had similar sex, race, and age (IOCM 77.5 years and LOCM 77.4 years) (Table 1). IOCM had a slightly higher rate of elective procedures at 83% compared to 79.5% for the LOCM cohort (P<.001). Based on our estimate for CM volume, approximately a third of the patients (35% of the IOCM cohort vs 28% of the LOCM cohort, P<.001) received more than 100 mL of CM. The cohorts had similar burden of comorbidities as shown by the means and standard deviations of the Elixhauser Comorbidity Index (IOCM 3.9 ± 1.8 vs LOCM 4.0 ± 1.9). Common comorbidities were also similar including congestive heart failure (14.0% vs 13.5%), hypertension (83.1% vs 81.5%), and diabetes (33.5% vs 34.0%). CKD was present in about a fourth of the patients, with a higher rate of CKD in the IOCM cohort (26% vs 22%; P<.001).

Outcomes. MARCE occurred in 6.8% of patients, the occurrence for each adverse event that comprise the MARCE composite were as follows: renal events (4.5%), AMI (0.8%), stroke/TIA (0.4%), and death (1.9%). The unadjusted event rates in IOCM vs LOCM patients were as follows: renal events (4.4% vs 4.6%), AMI (0.7% vs 0.9%), stroke/TIA (0.3% vs 0.5%), and death (1.4% vs 2.4%), respectively.

The results of multivariable modeling for the composite outcome of MARCE and its components are shown in Figure 2. After multivariable modeling, use of IOCM was associated with a 1.8% lower absolute risk of MARCE (95% CI, 0.4-3.3; P=.01). The relative risk reduction was 23.9% (95% CI, 5.2-44.2) (Figure 3). When modeled separately, there was a 1.3% lower absolute risk of renal events for patients receiving IOCM. The relative risk reduction for renal events was 28.9%. Mortality was also lower with IOCM, absolute risk reduction 0.9%, relative risk reduction 35.3% (P=.04). When modeled individually, there was no significant difference in the incidence of AMI (0.3%; 95% CI, 0.3-0.8; P=.36) or stroke/TIA (0.2%; 95% CI, -0.2-0.6; P=.36) between IOCM and LOCM. 

There are several important findings observed in this large study of contemporary US practices of contrast use among >15,000 patients undergoing EVAR from 444 hospitals between 2012 and 2018. First, among these high-risk EVAR procedures use of IOCM (vs LOCM) was associated with 24% lower relative risk of MARCE, 29% lower risk of adverse renal outcomes, and 35% lower risk of death, despite adjusting for patient demographics and differences in clinical practice among hospitals, including CM volumes. Second, the absolute differences were large in magnitude: 1.8% for MARCE and 1.3% for renal, translating into NNTs of 57 and 75, respectively. Third, there was an important association of lower all-cause mortality with IOCM vs LOCM with an ~1% absolute lower mortality associated with IOCM, which indicates that perhaps preventing adverse events in this high-risk population is associated with an important reduction in mortality.

Recent publications on real world data have reported both clinical and economic benefits of IOCM in patients undergoing PCI in an inpatient setting. While several studies have explored the role of different contrast media (IOCM vs LOCM) within the context of coronary intervention, our current findings are novel as one of the first attempts to explore this role in EVAR.

Patients with aortic aneurysmal disease often have multiple comorbidities that increase their risk of adverse renal events. Hypertension, heart failure,¹⁸ diabetes,^1,19 and pre-existing CKD¹⁹ are common. Importantly, renal artery stenosis as an extension of aortic atherosclerosis is also common and an established risk factor for renal ischemia.²⁰ Periprocedural planning for EVAR repair still relies on use of contrast-based computed tomographic angiography (CTA) with the potential for repetitive exposure to contrast agents. 

EVAR patients have a high prevalence of diabetes, CHF and CKD which place them at risk of not only cardiovascular adverse events but also renal events. IOCM vs LOCM have been shown to reduce both cardiovascular and renal events. In a study by Nie et al,²¹ a prospective, double-blind, randomized, controlled trial comparing iodixanol (n=106) and iopromide (n=102) in patients with CKD undergoing coronary angiography with or without PCI, there was a ~7% absolute reduction in the composite CV outcomes of emergent PCI, abrupt vessel closure, stroke, thrombosis, cardiac death, nonfatal MI, CABG; the composite CV outcome was reached more often in the iopromide group compared with the iodixanol group (8.8% vs 1.9%, respectively; P=.025). Similarly, in the NEPHRIC trial,²² there was 15% reduction in AKI with IOCM vs LOCM use. Our study result is consistent with the above randomized trials, that IOCM use (vs LOCM) was associated with reduction in MARCE, predominantly driven by reduction in renal events and mortality.

EVAR is associated with lower risk of AKI in comparison to open repair,⁴ presumably due to lesser hemodynamic insult to the kidneys. Nonetheless the exact etiology of AKI in perioperative EVAR setting is likely multifactorial, with the kidneys being subjected to variety of hemodynamic, mechanical, and pharmacologic insults.²³ Urgent AAA repair, in particular, can be associated with hemodynamic instability – increasing the risk of AKI. Iodinated contrast administration remains the central imaging modality for EVAR. Stent graft placement and dilation of the aorta can liberate atheroembolic debris and result in renal infarction. Compromise of renal blood flow due to graft placement is also a concern during these interventions. Post-procedural contrast administration may also be required when there is concern of an endoleak or inadequate aneurysmal coverage. 

Regardless of the mechanism, acute renal injury post-EVAR has significant clinical consequences; subsequently impacting mortality, cardiovascular morbidity, long-term renal function,²⁴ hospital length of stay, and associated costs.^25,26 In the present study, AKI was lower with the use of IOCM, following similar outcomes in prior studies comparing IOCM with LOCM. Hyperosmolarity of LOCM has been associated with increased intrarenal vasoconstriction, activation of tubuloglomerular feedback, and increases in tubular pressure, all of which may result in decreased glomerular filtration and medullary hypoxemia.¹⁰ In addition, there was a signal of lower mortality with the use of IOCM vs LOCM, that could potentially be explained, at least in part, by the lower incidence of AKI, which in turn has been associated with higher mortality in a variety of clinical scenarios. In addition, the occurrence of AKI may be a surrogate marker of a patient at high risk of subsequent morbidity and mortality.

The Society for Vascular Surgery (SVS) current practice guidelines on the care of patients with AAA acknowledge the potential for uniformly poor outcomes of aortic aneurysm repair in presence of severe renal dysfunction, regardless of the type of repair.²⁷ As such, Chaikof et al suggest comprehensive multiprong strategies to minimize renal injury after endovascular or open repair, including importance of periprocedural hydration to ensure euvolemia, temporarily withholding nephrotoxic medications, and limiting total volume of administered contrast media. Additionally, due to the association of ACE inhibitors and angiotensin receptor antagonists with hypotension on induction of anesthesia, these medications should be held the morning of surgery and restarted after the patient is euvolemic.²⁷ Although recognizing the potential of higher osmolality contrast agents for increased nephrotoxicity, the SVS guidelines do not currently recommend consideration toward different types of contrast media (IOCM vs LOCM). 

Our EVAR-specific assessment complements several meta-analyses^7,8,10 and a systematic review⁸ suggesting favorable reduction of adverse events via utilization of IOCM vs LOCM, most notably within intra-arterial procedures and especially among high-risk cohorts. Given the totality of data and recent supplementary evidence from EVAR-dedicated assessment, IOCM might be considered in high-risk patients undergoing aortic aneurysm repair together with other mitigation strategies such as individual risk assessment, periprocedural hydration, withholding of nephrotoxic medications, and minimization of contrast volume.

Study limitations. The limitations of this study include those which are inherent to observational studies in general. While we would have liked to capture all EVAR procedures in the US, the data source for this study was the Premier Hospital Database that represents 20% of all inpatient discharges in the US, thus considered fairly representative of the US population.²⁸ Outcomes were coded using ICD-9 and ICD-10 codes however, these have been considered to have a high degree of accuracy for comparative analysis.²⁹ Furthermore, as this study used a conservative approach by defining renal events using specific codes intended to identify postprocedural renal events related to CM usage, in reality, the event rate may be higher. In other studies researchers have taken a much broader definition of AKI that included tubular necrosis, lesion of renal cortical necrosis, lesion of renal medullary necrosis, and other specified pathological lesions in the kidney.^30-32 Another limitation is that we made certain assumptions in estimating CM volume. While we acknowledge the likely presence of a bias in the CM volume estimate, the direction of the bias is expected to be the same for IOCM and LOCM procedures alike and not expected to be differential by CM type. Another limitation of this data source is that it does not track patients longitudinally. Thus, all MARCE events were captured during the EVAR hospitalization only. Patients experiencing a MARCE event after being discharged from their primary hospitalization could not be captured. Our specific definition of renal events combined with events occurring only during the primary hospitalization lends itself to a very conservative estimate and biases us against finding a significant difference in MARCE between IOCM and LOCM. However, in general, patients experiencing in-hospital renal and cardiac events remain at greater risk of subsequent events and their inclusion would perhaps continue to show long-term risk reduction associated with IOCM as well. Another limitation is that we are unable to distinguish between technical procedural difficulty. For example, low-risk EVAR procedures were not distinguished from more complex interventions or higher risk procedures due to aneurysm location (infra-renal vs juxta-renal/supra-renal repair) and stenting grafts utilized were not categorized (fenestrated, branched, chimney stents, etc.). Both of these modifications would help identify patients at highest risk of MARCE. Lastly, due to the lack of laboratory values, we could not define CI-AKI by serum creatinine levels, rather, the outcome was defined by coding detail which may underestimate the occurrence of this event.

In conclusion, use of IOCM vs LOCM during EVAR procedures in patients at high risk of adverse renal events was associated with a lower incidence of MARCE with renal events driving this association. The absolute risk differences indicate that the magnitude of the association is large and clinically important with low NNTs. This finding may have significant clinical implications as development or worsening of renal function after EVAR procedures can lead to increased morbidity and mortality and use of IOCM may be a way to reduce morbidity and mortality in these challenging patients. Accordingly, randomized controlled clinical studies are recommended to confirm this finding.

From ¹Interventional Cardiology, Dartmouth-Hitchcock Medical Center, Geisel School of Medicine at Dartmouth, Lebanon, New Hampshire; ²Interventional Cardiology, Barnes Jewish Hospital, ³Cardiac Catheterization Laboratory, University of Texas Health Sciences Center at San Antonio, San Antonio, Texas; ⁴Biostatistics, CTI Clinical Trial & Consulting Services, Covington, Kentucky; ⁵Real World Evidence, CTI Clinical Trial & Consulting Services, Covington, Kentucky; ⁶Center for Complex Coronary Interventions, Minneapolis Heart Institute, Minneapolis, Minnesota; and ⁷Center for Coronary Artery Disease, Minneapolis Heart Institute Foundation, Minneapolis, Minnesota.

This study was presented in the moderated posters session at the 2019 Transcatheter Cardiovascular Therapeutics Conference (TCT) in San Francisco, California, September 2019.

Funding: This study was sponsored by GE Healthcare.

Disclosures: The authors have completed and returned the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Amin reports consulting income from GE Healthcare and Terumo. Dr Prasad reports research funding from ACIST, Osprey Medical, Recor Medical, and Freeman Heart Association; speaker’s bureau for AstraZeneca; consulting for GE Healthcare and Osprey Medical. MP Ryan was an employee of, and Dr Gunnarsson is a consultant to CTI Clinical Trial & Consulting Services, a consulting company for GE Healthcare, the study sponsor. Dr Brilakis reports consulting/speaker honoraria from Abbott Vascular, American Heart Association (associate editor, Circulation), Biotronik, Boston Scientific, Cardiovascular Innovations Foundation (Board of Directors), CSI, Elsevier, GE Healthcare, InfraRedx, Medtronic, Siemens, and Teleflex; research support from Regeneron and Siemens; shareholde in MHI Ventures. 

Manuscript accepted December 10, 2020.

Address for correspondence: Amit P. Amin, MD, MSc, MBA, Interventional Cardiology, Dartmouth-Hitchcock Medical Center, Associate Professor of Medicine, Geisel School of Medicine at Dartmouth, Lebanon, NH 03756. Email: Amit.P.Amin@hitchcock.org