Original ResearchFree Access

Intravenous CT Contrast Media and Acute Kidney Injury: A Multicenter Emergency Department–based Study

Published Online:https://doi.org/10.1148/radiol.2021204446

Abstract

Background

Although the historical risk of acute kidney injury (AKI) after intravenous administration of contrast media might be overstated, the risk in patients with impaired kidney function remains a concern.

Purpose

To investigate whether intravenous contrast media administration during CT is associated with a higher risk of AKI and further hemodialysis compared with the risk in patients undergoing unenhanced CT.

Materials and Methods

This retrospective study evaluated patients who underwent contrast-enhanced or unenhanced CT in five Taiwanese emergency departments between 2009 and 2016. The outcomes were AKI within 48–72 hours after CT, AKI within 48 hours to 1 week after CT, or further hemodialysis within 1 month after CT. The associations between contrast media exposure and outcome were estimated by using an overlap propensity score weighted generalized regression model. Subgroup analyses were performed according to the estimated glomerular filtration rate (eGFR).

Results

The study included 68 687 patients (median age, 68 years; interquartile range, 53–74 years; 39 995 men) with (n = 31 103) or without (n = 37 584) exposure to contrast media. After propensity score weighting, contrast media exposure was associated with higher risk of AKI within 48–72 hours after CT (odds ratio [OR], 1.16; 95% CI: 1.04, 1.29; P = .007) but no significant risk at 48 hours to 1 week after CT (OR, 1.00; 95% CI: 0.93, 1.08; P = .90). Among patients with eGFR less than 30 mL/min/1.73 m2, exposure to contrast media was associated with a higher AKI risk (48–72 hours after CT: OR, 1.36; 95% CI: 1.09, 1.70; P = .007) (48 hours–1 week after CT: OR, 1.49; 95% CI: 1.27, 1.74; P < .001) and a higher risk of hemodialysis (OR, 1.36; 95% CI: 1.09, 1.70; P = .008). For patients with eGFR greater than 45 mL/min/1.73.m2, contrast media exposure was not associated with higher AKI risk (P > .05).

Conclusion

Contrast-enhanced CT was associated with higher risk of acute kidney injury and further hemodialysis among Taiwanese patients with an estimated glomerular filtration rate (eGFR) of less than 30 mL/min/1.73 m2 but not those with an eGFR of more than 45 mL/min/1.73 m2.

© RSNA, 2021

Online supplemental material is available for this article.

Summary

Intravenous CT contrast media exposure was associated with a higher risk of acute kidney injury within 48–72 hours after CT, particularly in patients with a pre-CT estimated glomerular filtration rate of less than 30 mL/min/1.73 m2.

Key Results

  • ■ In a retrospective study of 68 687 Taiwanese patients attending an emergency department, patients who underwent intravenous contrast-enhanced CT had a higher risk of acute kidney injury (AKI) within 48–72 hours after exposure (odds ratio [OR], 1.16; P = .007).

  • ■ Among patients with a pre-CT estimated glomerular filtration rate of less than 30 mL/min/1.73 m2, intravenous contrast media exposure was associated with a higher risk of AKI (OR, 1.36–1.49; P < .001 to P = .007) and further hemodialysis within 1 month (OR, 1.36; P = .008).

Introduction

CT with intravenous iodinated contrast media is frequently performed in the emergency department (ED) to improve image quality and diagnostic accuracy. Contrast media administration has been historically considered the third most common cause of iatrogenic acute kidney injury (AKI) (1) and a risk factor for major adverse events, including dialysis, renal failure, and death (2,3). However, previous studies have lacked control groups, and the causality of the relationship between intravenous contrast media exposure and AKI risk has not been established (4). More recent well-designed observational studies have identified an association between contrast media use and the risk of AKI, although disagreements persist (59). For example, Davenport et al and Ellis et al (10,11) noted an association between intravenous contrast media use and the risk of AKI among patients with preexisting kidney dysfunction. However, other studies found no association between intravenous contrast media use and AKI risk, regardless of baseline kidney function (1219). Although recent consensus has downgraded the risk of AKI related to intravenous contrast media exposure (20), ongoing clinical concerns remain regarding the risk of kidney injury with intravenous contrast media exposure, especially among patients with impaired kidney function.

The primary aim of this study was to investigate whether intravenous contrast media administration during CT in the ED was associated with a higher risk of AKI or further dialysis, using patients undergoing unenhanced CT as controls and weighting with the propensity score (PS) overlap method. A secondary aim was to evaluate the association between intravenous contrast media use and AKI risk based on different pre-CT kidney functions.

Materials and Methods

Study Design and Setting

This retrospective study was performed at five Taiwanese hospitals (three tertiary and two secondary hospitals) that have approximately 500 000 ED visits per year. Patient data from 2009 through 2016 were collected from an institutional database (21). The study protocol was approved by the institutional review board (201700961B0C502), and the requirement for informed consent was waived. All patients had provided informed consent for intravenous contrast media administration according to institutional protocol. For contrast-enhanced CT, nonionized iodinated contrast media (iohexol, Omnipaque; GE Healthcare) was used. Before 2012, a standard dose of 100–120 mL per scan was administered, with a maximum dose of 150 mL in patients heavier than 100 kg. Since 2012, the radiology department has implemented a policy to adjust the dose of contrast agent (range, 50–90 mL) in patients with an estimated glomerular filtration rate (eGFR) of less than 60 mL/min/1.73.m2 at the discretion of radiologists.

Selection of Patients

Patients were eligible if they were aged at least 20 years, had undergone contrast-enhanced or unenhanced CT in the ED, and had their serum creatinine (SCr) level measured less than 72 hours before CT and 48 hours to 1 week after CT (22,23). Patients were excluded if they had undergone renal replacement (hemodialysis, peritoneal dialysis, hemofiltration, or renal transplantation) within 1 year or had received contrast media within 3 months before the index ED CT examination. Patients who underwent hemodialysis within 72 hours after the index CT examination were also excluded because they were likely to have undergone hemodialysis before CT. Finally, patients were excluded if they had additional contrast media exposure within 1 week of CT to minimize the additive effects of contrast media exposure. Individual patients could be included more than once if each ED visit fulfilled the inclusion criteria.

Measurements

Data regarding demographic characteristics (age, sex), comorbidities (diabetes, diabetic nephropathy, hypertension, congestive heart failure, and chronic kidney diseases), and medications (angiotensin-converting enzyme inhibitors or angiotensin receptor blockers, nonsteroidal anti-inflammatory drugs, and diuretics) used regularly within 3 months before the index CT examination were retrieved from the hospital electronic health record. Comorbidities were extracted using the International Classification of Disease, Ninth Revision or Tenth Revision, Clinical Modification codes from the outpatient visits and hospitalization records. Comorbidities were considered if they were diagnosed during at least one hospitalization or two outpatient visits within 1 year before the index ED visit. Factors that might alter kidney function, such as in-hospital treatments before outcomes occurred (ie, inotropic agents, dobutamine, angiotensin-converting enzyme inhibitors or angiotensin receptor blockers, nonsteroidal anti-inflammatory drugs, diuretics, and acetylcysteine), fluid challenge, and critical condition at admission, were recorded. Fluid challenge was defined as administration of any intravenous fluid, mostly crystalloid, given as a bolus of more than 200 mL or at a rate of more than 100 mL per hour. Critical condition was identified by whether physicians issued a critically ill notice. It is common practice in Taiwan that if a patient is critically ill, then physicians will issue a notice to their family, which serves as a courtesy and helps initiate discussions about further goals of care. Kidney function was judged based on eGFRs calculated by using the Modification of Diet in Renal Disease 4 equation (24).

Outcomes

The primary outcome was the incidence of AKI during the outcome window, which was defined according to the Acute Kidney Injury Network and Kidney Disease: Improving Global Outcomes criteria (absolute SCr level increase of more than 0.3 mg/dL [26.5 μmol/L] within 48 hours or more than 50% increase from the baseline level within 7 days) (25,26). The outcome window was defined as between 48 and 72 hours and 48 hours to 1 week after CT (22,27,28). These time frames were chosen because patients in the ED usually have their SCr level measured on the same day they undergo CT, and SCr was rarely rechecked within 24–48 hours. Besides, patients were more concerned about the sustained deteriorated kidney function beyond 48 hours. By using an elevated SCr level between 48–72 hours and 48 hours and 1 week, clinically important postcontrast AKI would not be missed. The last SCr measurement was collected in patients with multiple levels during follow-up. The secondary outcome was defined as the risk of hemodialysis within 1 month after CT.

Statistical Analysis

Variables were reported as median and interquartile range or as number and percentage. The generalized regression model with a generalized estimating equation was used to account for self-clustering during comparisons of continuous and categorical variables. Gaussian distribution with identity links or binary distribution with logit links were chosen according to data type. The incidences of outcomes were calculated as the number of events divided by the total number of eligible ED visits. All analyses were performed with SAS software (version 9.4, SAS Institute), and differences were considered significant at P < .05.

The overlap PS weighting method, which weighs the probability of a patient being assigned to the other treatment group, was used to minimize the pre-CT baseline differences between the contrast-enhanced and unenhanced CT groups without reducing the sample size (29). The PS for undergoing contrast-enhanced CT was estimated by using a multivariable logistic regression model, which included age, sex, year of ED visit, hospital branch, comorbidities, previous regular medications within 3 months before the index ED visit, and pre-CT eGFR. All variables in the PS model were preexisting conditions before CT. The covariate balances before and after weighting were compared by using absolute standardized mean differences, with a good balance considered present at a threshold of less than 10% (30). Because overlap PS weighting balances only the pre-CT conditions, the final adjusted models after weighting also included variables such as critical condition, in-hospital treatments that were contemporaneous or after the CT study (ie, inotropic agents, dobutamine, fluid challenge, angiotensin-converting enzyme inhibitors or angiotensin receptor blockers, nonsteroidal anti-inflammatory drugs, diuretics, acetylcysteine), and important independent variables (age, sex, hospital branch, year of ED visit, and pre-CT eGFR) (Table E1 [online]).

Subgroup analyses were performed according to pre-CT kidney function (based on eGFR, classified as greater than or equal to 90, 60–89, 45–59, 30–44, and less than 30 mL/min/1.73 m2), with PS reestimated within each stratum. An interaction term between the eGFR groups and contrast media exposure was used to evaluate the risk of AKI among different eGFR strata.

The robustness of the findings was evaluated using sensitivity analyses. First, one-to-one PS matching was performed through the 9-to-1 digit greedy matching algorithm without replacement (31), and average treatment effects were calculated by counting the clusters in the matched sets. Second, for patients with multiple ED visits, only the first eligible visit was included. Third, Stürmer trimming method (32) was performed because of concern that extreme PS scores might bias the estimates. Fourth, traditional criteria for contrast-induced AKI (33) were used to evaluate the association between contrast media exposure and AKI risk (SCr level increased by more than 0.5 mg/dL [44.2 μmol/L] or more than 25% from baseline).

Results

Patient Characteristics

Among 403 598 CT scans, 90 415 visits had follow-up SCr measurement within 48 hours to 1 week. Of those, 2657 visits were excluded because patients underwent previous renal replacement, 10 354 visits were excluded because patients received contrast media within the previous 3 months, and 3676 visits were excluded because patients underwent dialysis within 72 hours after CT (Fig 1). Finally, 68 687 eligible patients were identified with follow-up SCr measurement available at 48 hours to 1 week, including 39 995 men and 28 692 women (median age, 68 years; IQR, 53–74 years; 22 064 patients were identified with follow-up SCr at 48–72 hours, including 13 003 men and 9061 women (median age, 67 years; IQR, 53–79 years). Patients who underwent contrast-enhanced CT were younger, had fewer comorbidities, and had better pre-CT eGFR (Tables E2, E3 [online]).

Study flowchart. AKI = acute kidney injury, ED = emergency department,                         OR = odds ratio, SCr = serum creatinine.

Figure 1: Study flowchart. AKI = acute kidney injury, ED = emergency department, OR = odds ratio, SCr = serum creatinine.

After overlap PS weighting, baseline characteristics were similar and balanced between the contrast-enhanced and unenhanced CT groups (Table 1). However, the contrast-enhanced CT group received more inotropic agents, more dobutamine, and more fluid challenge. The contrast-enhanced CT group was less likely to receive angiotensin-converting enzyme inhibitors or angiotensin receptor blockers but more likely to receive nonsteroidal anti-inflammatory drugs and diuretics (Table 2).

Table 1: Patients’ Clinical Characteristics after Overlap Propensity Score Weighting

Table 1:

Table 2: Clinical Condition and In-Hospital Treatments after Overlap Propensity Score Weighting

Table 2:

Incidence of AKI and Risk of Further Dialysis: Contrast-enhanced versus Noncontrast CT

The weighted incidence of AKI within 48–72 hours after CT was 9.5% (2086 of 22 064). The contrast media group was more likely to develop AKI within 48–72 hours after CT (10.9% [1105 of 10 143] vs 8.2% [981 of 11 921]; odds ratio [OR], 1.36; 95% CI: 1.24, 1.51; P < .001) (Fig 2, Table E4 [online]), even after adjusting for in-hospital treatments (OR, 1.16; 95% CI: 1.04, 1.30; P = .005).

Association between contrast media exposure and the likelihood of                         acute kidney injury or further dialysis. Subgroup analyses were stratified                         according to the estimated glomerular filtration rate (eGFR). AKI = acute                         kidney injury, OR = odds ratio, Pint = P of interaction. * P <                         .05, ** P < .01, *** P <                         .001. ‡ We reported with the propensity score weighted unadjusted and                         adjusted OR. Variables in the propensity score weighted adjusted model                         included age, sex, pre-CT eGFR, critical condition, in-hospital treatments                         (inotropic agents, dobutamine, fluid challenge, angiotensin-converting                         enzyme inhibitors or angiotensin receptor blockers, nonsteroidal                         anti-inflammatory drugs, diuretics, acetylcysteine), hospital branch, and                         year of index emergency department visit.

Figure 2: Association between contrast media exposure and the likelihood of acute kidney injury or further dialysis. Subgroup analyses were stratified according to the estimated glomerular filtration rate (eGFR). AKI = acute kidney injury, OR = odds ratio, Pint = P of interaction. * P < .05, ** P < .01, *** P < .001. ‡ We reported with the propensity score weighted unadjusted and adjusted OR. Variables in the propensity score weighted adjusted model included age, sex, pre-CT eGFR, critical condition, in-hospital treatments (inotropic agents, dobutamine, fluid challenge, angiotensin-converting enzyme inhibitors or angiotensin receptor blockers, nonsteroidal anti-inflammatory drugs, diuretics, acetylcysteine), hospital branch, and year of index emergency department visit.

The weighted incidence of AKI within 48 hours to 1 week was 6.0% (4092 of 68 687), which was slightly higher among patients who underwent contrast-enhanced CT (6.3% [1966 of 31 103] vs 5.7% [2126 of 37 584]; OR, 1.13; 95% CI: 1.05, 1.21; P < .001). However, no statistical differences were found after adjusting for possible confounding variables (Table E1 [online]) and overlap weighting (OR, 1.00; 95% CI: 0.93, 1.08; P = .99).

The contrast-enhanced CT group had similar risks of hemodialysis within 1 month (contrast-enhanced: 1.6% [500 of 31 103] vs noncontrast: 1.5% [564 of 37 584]; OR, 1.07; 95% CI: 0.94, 1.22; P = .30) and the same after adjusting for in-hospital treatments (OR, 1.01; 95% CI: 0.88, 1.16; P = .91).

Subgroup analyses among different eGFR strata.—The contrast-related risks of AKI and hemodialysis within 1 month varied in the different eGFR strata (P < .05 for all interactions [Fig 2]). Among patients with an eGFR of less than 30 mL/min/1.73.m2, after adjusting for possible confounding factors (Fig 2), contrast media use was associated with higher risks of AKI (48–72 hours: OR, 1.36; 95% CI: 1.09, 1.70; P = .007; 48 hours to 1 week: OR, 1.49; 95% CI: 1.27, 1.74; P < .001) and hemodialysis within 1 month (OR, 1.36; 95% CI: 1.09, 1.70; P = .008).

Among patients with an eGFR of 30–44 mL/min/1.73 m2, contrast media use was associated with a higher risk of AKI within 48–72 hours (OR, 1.35; 95% CI: 1.06, 1.73; P = .02) and risks of hemodialysis (OR, 1.50; 95% CI: 1.05, 2.13; P = .02). However, the association was not found in AKI within 48 hours to 1 week (OR = 1.02; 95% CI: 0.86, 1.21; P = .80).

For all eGFR strata of more than 45 mL/min/1.73 m2, contrast media use was not associated with a higher risk of AKI or further hemodialysis after adjustments for possible confounding (ie, critical condition, in-hospital treatment with a fluid challenge, inotropic agents, dobutamine, angiotensin-converting enzyme inhibitors or angiotensin receptor blockers, nonsteroidal anti-inflammatory drugs, diuretics, acetylcysteine, age, sex, hospital branches, year of visit, and eGFR) (P > .05 for all [Fig 2]).

Sensitivity analyses.—Robustness was evaluated with different sensitivity analyses (SA), including PS matching (SA1), single enrollment (SA2), using the Stürmer trimming to exclude extreme PS values (SA3) (32), and the traditional definition for contrast-induced AKI (SA4: SCr level increased by more than 0.5 mg/dL [44.2 μmol/L] or more than 25%).

Different sensitivity analyses revealed the same patterns as the primary and subgroup analyses (Figs 35). Among patients with eGFR strata of at least 45 mL/min/1.73 m2, contrast media use was not consistently associated with a higher risk of AKI or hemodialysis. Among patients with an eGFR of 30–44 mL/min/1.73.m2, contrast media use was associated with a higher risk of AKI within 48–72 hours and hemodialysis but was not associated with AKI within 48 hours to 1 week. For patients with an eGFR of less than 30 mL/min/1.73.m2, contrast media exposure was consistently associated with a higher risk of AKI within 48–72 hours and within 48 hours to 1 week and further dialysis within 1 month (P < .05 for all; Figs 35).

Association between contrast media administration and the likelihood                         of acute kidney injury (AKI) within 48–72 hours after CT: subgroup                         and sensitivity analyses (SAs). Subgroup analyses were stratified according                         to estimated glomerular filtration rate (eGFR). SA1, propensity score                         matching (PSM): One-to-one PSM was performed through the 9-to-1 digit greedy                         matching algorithm without replacement. SA2, enrolled once: For patients                         with multiple eligible emergency department (ED) visits, we included only                         the first ED visit in the analysis. SA3, Stürmer trimming: Under the                         concern that extreme propensity score (PS) might bias the estimates, we used                         the Stürmer trimming method (ie, trimmed below the 2.5th percentile                         of observed PS in the contrast-enhanced group and above the 97.5th                         percentile of observed PS in the noncontrast group) and reperformed the                         analysis. SA4: The traditional definition of acute kidney                         injury—absolute serum creatinine level increases of more than 0.5                         mg/dL [44.2 μmol/L] or more than 25% over baseline. C+ =                         contrast-enhanced CT, C− = unenhanced CT, OR = odds ratio, Pint = P                         of interaction. * P < .05, ** P < .01,                         *** P < .001. § The numerator and                         denominator of the data are provided in Tables E4 and E5 (online). ‡                         We reported with the PS weighted (or matching) unadjusted and adjusted OR.                         Variables in the adjusted models included age, sex, pre-CT eGFR, critical                         condition, in-hospital treatments (inotropic agents, dobutamine, fluid                         challenge, angiotensin-converting enzyme inhibitors or angiotensin receptor                         blockers, nonsteroidal anti-inflammatory drugs, diuretics, acetylcysteine),                         hospital branch, and year of index ED visit.

Figure 3: Association between contrast media administration and the likelihood of acute kidney injury (AKI) within 48–72 hours after CT: subgroup and sensitivity analyses (SAs). Subgroup analyses were stratified according to estimated glomerular filtration rate (eGFR). SA1, propensity score matching (PSM): One-to-one PSM was performed through the 9-to-1 digit greedy matching algorithm without replacement. SA2, enrolled once: For patients with multiple eligible emergency department (ED) visits, we included only the first ED visit in the analysis. SA3, Stürmer trimming: Under the concern that extreme propensity score (PS) might bias the estimates, we used the Stürmer trimming method (ie, trimmed below the 2.5th percentile of observed PS in the contrast-enhanced group and above the 97.5th percentile of observed PS in the noncontrast group) and reperformed the analysis. SA4: The traditional definition of acute kidney injury—absolute serum creatinine level increases of more than 0.5 mg/dL [44.2 μmol/L] or more than 25% over baseline. C+ = contrast-enhanced CT, C− = unenhanced CT, OR = odds ratio, Pint = P of interaction. * P < .05, ** P < .01, *** P < .001. § The numerator and denominator of the data are provided in Tables E4 and E5 (online). ‡ We reported with the PS weighted (or matching) unadjusted and adjusted OR. Variables in the adjusted models included age, sex, pre-CT eGFR, critical condition, in-hospital treatments (inotropic agents, dobutamine, fluid challenge, angiotensin-converting enzyme inhibitors or angiotensin receptor blockers, nonsteroidal anti-inflammatory drugs, diuretics, acetylcysteine), hospital branch, and year of index ED visit.

Association between contrast media administration and the likelihood                         of acute kidney injury (AKI) within 48 hours to 1 week after CT: subgroup                         and sensitivity analyses (SAs). Subgroup analyses were stratified according                         to estimated glomerular filtration rate (eGFR). SA1, propensity score                         matching (PSM): One-to-one PSM was performed through the 9-to-1 digit greedy                         matching algorithm without replacement. SA2, enrolled once: For patients                         with multiple eligible emergency department (ED) visits, we included only                         the first ED visit in the analysis. SA3, Stürmer trimming: Under the                         concern that extreme propensity score (PS) might bias the estimates, we used                         the Stürmer trimming method (ie, trimmed below the 2.5th percentile                         of observed PS in the contrast-enhanced group and above the 97.5th                         percentile of observed PS in the noncontrast group) and reperformed the                         analysis. SA4: The traditional definition of AKI—absolute serum                         creatinine increases of more than 0.5 mg/dL [44.2 μmol/L] or more                         than 25% over baseline. C+ = contrast-enhanced CT, C− = noncontrast                         CT, OR = odds ratio, Pint = P of interaction, PSM = PS matching. * P                         < .05. ** P < .01. *** P                         < .001. § The numerator and denominator of the data are                         provided in Tables E4 and E5 (online). ‡ We reported with the PS                         weighted (or matching) unadjusted and adjusted OR. Variables in the adjusted                         models included age, sex, pre-CT eGFR, critical condition, in-hospital                         treatments (inotropic agents, dobutamine, fluid challenge,                         angiotensin-converting enzyme inhibitors or angiotensin receptor blockers,                         nonsteroidal anti-inflammatory drugs, diuretics, acetylcysteine), hospital                         branch, and year of index ED visit.

Figure 4: Association between contrast media administration and the likelihood of acute kidney injury (AKI) within 48 hours to 1 week after CT: subgroup and sensitivity analyses (SAs). Subgroup analyses were stratified according to estimated glomerular filtration rate (eGFR). SA1, propensity score matching (PSM): One-to-one PSM was performed through the 9-to-1 digit greedy matching algorithm without replacement. SA2, enrolled once: For patients with multiple eligible emergency department (ED) visits, we included only the first ED visit in the analysis. SA3, Stürmer trimming: Under the concern that extreme propensity score (PS) might bias the estimates, we used the Stürmer trimming method (ie, trimmed below the 2.5th percentile of observed PS in the contrast-enhanced group and above the 97.5th percentile of observed PS in the noncontrast group) and reperformed the analysis. SA4: The traditional definition of AKI—absolute serum creatinine increases of more than 0.5 mg/dL [44.2 μmol/L] or more than 25% over baseline. C+ = contrast-enhanced CT, C− = noncontrast CT, OR = odds ratio, Pint = P of interaction, PSM = PS matching. * P < .05. ** P < .01. *** P < .001. § The numerator and denominator of the data are provided in Tables E4 and E5 (online). ‡ We reported with the PS weighted (or matching) unadjusted and adjusted OR. Variables in the adjusted models included age, sex, pre-CT eGFR, critical condition, in-hospital treatments (inotropic agents, dobutamine, fluid challenge, angiotensin-converting enzyme inhibitors or angiotensin receptor blockers, nonsteroidal anti-inflammatory drugs, diuretics, acetylcysteine), hospital branch, and year of index ED visit.

Association between contrast media administration and the likelihood                         of dialysis within 1 month after CT: subgroup and sensitivity analyses                         (SAs). Subgroup analyses were stratified according to the estimated                         glomerular filtration rate (eGFR). SA1, propensity score matching (PSM):                         One-to-one PSM was performed through the “9-to-1 digit” greedy                         matching algorithm without replacement. SA2, enrolled once: For patients                         with multiple eligible emergency department (ED) visits, we included only                         the first ED visit in the analysis. SA3, Stürmer trimming: Under the                         concern that extreme propensity score (PS) might bias the estimates, we used                         the Stürmer trimming method (ie, trimmed below the 2.5th percentile                         of observed PS in the contrast-enhanced group and above the 97.5th                         percentile of observed PS in the noncontrast group) and reperformed the                         analysis. AKI = acute kidney injury, C+ = contrast-enhanced CT, C− =                         noncontrast CT, OR = odds ratio, Pint = P of interaction. * P                         < .05, ** P < .01, *** P                         < .001. § The numerator and denominator of the data are                         provided in Tables E4 and E5 (online). ‡ We reported with the PS                         weighted (or matching) unadjusted and adjusted OR. Variables in the adjusted                         models included age, sex, pre-CT eGFR, critical condition, in-hospital                         treatments (inotropic agents, dobutamine, fluid challenge,                         angiotensin-converting enzyme inhibitors/angiotensin receptor blockers,                         nonsteroidal anti-inflammatory drugs, diuretics, acetylcysteine), hospital                         branch, and year of index ED visit.

Figure 5: Association between contrast media administration and the likelihood of dialysis within 1 month after CT: subgroup and sensitivity analyses (SAs). Subgroup analyses were stratified according to the estimated glomerular filtration rate (eGFR). SA1, propensity score matching (PSM): One-to-one PSM was performed through the “9-to-1 digit” greedy matching algorithm without replacement. SA2, enrolled once: For patients with multiple eligible emergency department (ED) visits, we included only the first ED visit in the analysis. SA3, Stürmer trimming: Under the concern that extreme propensity score (PS) might bias the estimates, we used the Stürmer trimming method (ie, trimmed below the 2.5th percentile of observed PS in the contrast-enhanced group and above the 97.5th percentile of observed PS in the noncontrast group) and reperformed the analysis. AKI = acute kidney injury, C+ = contrast-enhanced CT, C− = noncontrast CT, OR = odds ratio, Pint = P of interaction. * P < .05, ** P < .01, *** P < .001. § The numerator and denominator of the data are provided in Tables E4 and E5 (online). ‡ We reported with the PS weighted (or matching) unadjusted and adjusted OR. Variables in the adjusted models included age, sex, pre-CT eGFR, critical condition, in-hospital treatments (inotropic agents, dobutamine, fluid challenge, angiotensin-converting enzyme inhibitors/angiotensin receptor blockers, nonsteroidal anti-inflammatory drugs, diuretics, acetylcysteine), hospital branch, and year of index ED visit.

Discussion

After intravenous contrast media administration, acute kidney injury (AKI) remains a clinical concern, especially for patients with impaired kidney function. In this retrospective multicenter emergency department (ED)–based study of 68 687 patients, we found that patients with CT contrast media exposure were more likely to develop AKI within 48–72 hours of ED attendance (odds ratio [OR], 1.16; 95% CI: 1.04, 1.30; P = .007). In addition, patients with a pre-CT estimated glomerular filtration rate (eGFR) of less than 30 mL/min/1.73.m2 had a higher risk of AKI (48–72 hours: OR, 1.36; 95% CI: 1.09, 1.70; P = .007; 48 hours to 1 week: OR, 1.49; 95% CI: 1.27, 1.74; P < .001) and dialysis (OR, 1.36; 95% CI: 1.09, 1.70; P = .008). However, contrast media exposure was not associated with a higher risk of AKI (P = .08 to P = .80) among patients with an eGFR of more than 45 mL/min/1.73 m2. For patients with an eGFR of 30–44 mL/min/1.73.m2, the association between contrast media exposure and AKI was observed within 48–72 hours after CT and dialysis within 1 month (AKI: OR, 1.35; 95% CI: 1.06, 1.73; P = .02; dialysis: OR, 1.50; 95% CI: 1.05, 2.13; P = .02) but not in AKI within 48 hours to 1 week (P = .80).

The risk of AKI related to contrast media exposure may be overstated, given the lack of control groups in the previous studies. Whereas animal models and models using renal endothelial and epithelial cells have indicated that contrast media can lead to cell damage (7,34), observational studies and meta-analyses have indicated that AKI after CT is not attributable to contrast media exposure (1216). Our results concur among patients with pre-CT eGFR of more than 45 mL/min/1.73.m2, where contrast media exposure was not associated with a higher risk of AKI or hemodialysis. Nevertheless, among patients with an eGFR of less than 30 mL/min/1.73.m2, contrast media administration was related to a higher likelihood of AKI, in line with findings of previous studies (10,11,35) and the consensus of the American College of Radiology and the National Kidney Foundation (20).

There are various possible explanations for the differences between our results and those of previous studies that reported contrast media administration was unassociated with AKI (1519). We evaluated Taiwanese patients and included a relatively large number of patients with advanced kidney disease, which provided a better estimate of the effects of contrast media. For example, we included 4247–11 198 patients with an eGFR of less than 30 mL/min/1.73.m2, and 650–1550 patients had contrast media exposure. In contrast (Table 3), Hinson et al (15) reported 723 patients had an eGFR of less than 30 mL/min/1.73 m2 (contrast media, 82 patients), McDonald et al (16,17) reported reduced kidney function in 1321 patients (contrast media, 90 patients) and 1486 patients (contrast media, 743 patients), and Davenport et al (10) reported reduced kidney function in 116 patients (contrast media, 44 patients).

Table 3: Comparisons between Different Studies

Table 3:

Our overall incidence of AKI was approximately 9.5% (2086 of 22 064), which was slightly higher than that in the ED-based study by Hinson and colleagues (15). Both studies had similar AKI incidences among patients undergoing unenhanced CT in various eGFR strata. After contrast media exposure, the proportions of AKI were also similar between our study and the studies of Davenport et al and McDonald et al (10,18), especially among patients with eGFR of less than 60 mL/min/1.73.m2.

Among patients with an eGFR of 30–44 mL/min/1.73.m2, the study by Davenport and colleagues and our study had similar AKI incidences and patterns (10). However, the difference in the sample size may reduce the power to detect a difference. In contrast, compared with the study by McDonald and colleagues in eGFR strata of 30–59 mL/min/1.73.m2 (17), we observed similar proportions of AKI only in the contrast media group but not in patients who underwent unenhanced CT (Table 3). Possible explanations include the difference in baseline characteristics. Besides patient race and patient source (solely from the ED vs a mixture of patients from the ED, hospitalized patients, or outpatients), the comorbidities were different (eg, diabetes [18% vs 40%], hypertension [26% vs 80%], and heart failure [5% vs 34%]). Among patients with eGFR of 30–44 mL/min/1.73 m2, we observed that the risks of AKI or hemodialysis were generally associated with contrast media exposure, although not all sensitivity analyses revealed significant results. Nonetheless, all these studies consistently indicated higher AKI risks in patients with poorer kidney function, regardless of contrast media exposure, indicating that impaired kidney function is a risk factor for AKI (16).

Nevertheless, contrast-enhanced CT may be indicated in emergency scenarios, regardless of kidney function. Given contrast media administration may be associated with a higher AKI risk among patients with less-preserved kidney function, contrast media use should be based on the expected diagnostic yield and whether alternate modalities are available.

Our study had several limitations. First, we considered Taiwanese patients who visited the ED, which provided only nonprotocolized physician-dependent AKI prophylaxis (ie, fluid hydration, oral acetylcysteine). Although we tried to adjust for different clinical treatments in the final model, our results may not be generalizable to other populations or settings. For example, AKI risk may be lower in the outpatient setting, as better patient preparation may be possible (vs that in the ED). Second, we used International Classification of Disease, Ninth Revision or Tenth Revision, Clinical Modification codes and prescriptions to identify comorbidities. Such records are prone to coding errors. We attempted to minimize the influence of miscoding by requiring diagnoses from at least one hospitalization or two outpatient visits. However, patients presenting de novo to the ED lacked this information. Besides, patients may not have been classified as having chronic kidney disease until it reached the advanced stage in our database. Nevertheless, miscoding would be a random event unlikely to cause systematic error or bias. Third, treatment assignment was not randomized, and there were differences between the contrast-enhanced and unenhanced CT groups. Although we attempted to minimize this issue by using a PS-weighted model to eliminate disparities, unmeasured confounding was a possibility. Fourth, we did not collect data regarding the stability of kidney function. Patients underwent contrast-enhanced or unenhanced CT based on the ED SCr, not previous SCr or kidney function stability. Thus, kidney function stability was not a confounding factor. Fifth, we excluded patients who underwent dialysis within 72 hours after CT, which might have excluded patients who developed severe kidney injury after contrast media exposure, although the number of affected patients should be negligible. Sixth, we considered only AKI risks within 48–72 hours and 48 hours to 1 week and dialysis within 1 month after the CT scan. Therefore, our results may not reflect the long-term effects of contrast media exposure.

In conclusion, in Taiwanese patients, CT contrast media exposure in the emergency department was associated with the likelihood of acute kidney injury, which was related to the pre-CT kidney function. Patients with an estimated glomerular filtration rate of less than 30 mL/min/1.73 m2 had a higher likelihood of acute kidney injury and further dialysis within 1 month.

Disclosures of Conflicts of Interest: T.H.S. disclosed no relevant relationships. C.H.H. disclosed no relevant relationships. Y.L.C. disclosed no relevant relationships. Y.C.W. disclosed no relevant relationships. C.F.K. disclosed no relevant relationships. C.H.L. disclosed no relevant relationships. C.C.L. disclosed no relevant relationships. H.Y.C. disclosed no relevant relationships.

Acknowledgments

The authors appreciate and acknowledge the support of the Center for Big Data Analytics and Statistics at Linkou Chang Gung Memorial Hospital for statistical consultation and data analysis.

Author Contributions

Author contributions: Guarantors of integrity of entire study, T.H.S., C.H.H.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; approval of final version of submitted manuscript, all authors; agrees to ensure any questions related to the work are appropriately resolved, all authors; literature research, T.H.S., C.H.H., Y.L.C.; clinical studies, T.H.S., C.H.H., Y.L.C., C.F.K., C.H.L.; statistical analysis, T.H.S., C.H.H., Y.L.C., H.Y.C.; and manuscript editing, T.H.S., C.H.H., Y.L.C., Y.C.W., C.H.L., C.C.L., H.Y.C.

* T.H.S. and C.H.H. contributed equally to this work.

References

  • 1. Nash K, Hafeez A, Hou S. Hospital-acquired renal insufficiency. Am J Kidney Dis 2002;39(5):930–936. Crossref, MedlineGoogle Scholar
  • 2. Mitchell AM, Kline JA, Jones AE, Tumlin JA. Major adverse events one year after acute kidney injury after contrast-enhanced computed tomography. Ann Emerg Med 2015;66(3):267–274.e4. Crossref, MedlineGoogle Scholar
  • 3. Coca SG, Peixoto AJ, Garg AX, Krumholz HM, Parikh CR. The prognostic importance of a small acute decrement in kidney function in hospitalized patients: a systematic review and meta-analysis. Am J Kidney Dis 2007;50(5):712–720. Crossref, MedlineGoogle Scholar
  • 4. Rao QA, Newhouse JH. Risk of nephropathy after intravenous administration of contrast material: a critical literature analysis. Radiology 2006;239(2):392–397. LinkGoogle Scholar
  • 5. Ehrmann S, Aronson D, Hinson JS. Contrast-associated acute kidney injury is a myth: yes. Intensive Care Med 2018;44(1):104–106. Crossref, MedlineGoogle Scholar
  • 6. Kashani K, Levin A, Schetz M. Contrast-associated acute kidney injury is a myth: we are not sure. Intensive Care Med 2018;44(1):110–114. Crossref, MedlineGoogle Scholar
  • 7. Weisbord SD, du Cheryon D. Contrast-associated acute kidney injury is a myth: no. Intensive Care Med 2018;44(1):107–109. Crossref, MedlineGoogle Scholar
  • 8. Rudnick MR, Leonberg-Yoo AK, Litt HI, Cohen RM, Hilton S, Reese PP. The controversy of contrast-induced nephropathy with intravenous contrast: what is the risk? Am J Kidney Dis 2020;75(1):105–113. Crossref, MedlineGoogle Scholar
  • 9. Hinson JS, Ehmann MR, Klein EY. Evidence and patient safety prevail over myth and dogma: consensus guidelines on the use of intravenous contrast media. Ann Emerg Med 2020;76(2):149–152. Crossref, MedlineGoogle Scholar
  • 10. Davenport MS, Khalatbari S, Cohan RH, Dillman JR, Myles JD, Ellis JH. Contrast material–induced nephrotoxicity and intravenous low-osmolality iodinated contrast material: risk stratification by using estimated glomerular filtration rate. Radiology 2013;268(3):719–728. LinkGoogle Scholar
  • 11. Ellis JH, Khalatbari S, Yosef M, Cohan RH, Davenport MS. Influence of clinical factors on risk of contrast-induced nephrotoxicity from IV iodinated low-osmolality contrast material in patients with a low estimated glomerular filtration rate. AJR Am J Roentgenol 2019;213(5):W188–W193. Crossref, MedlineGoogle Scholar
  • 12. Aycock RD, Westafer LM, Boxen JL, Majlesi N, Schoenfeld EM, Bannuru RR. Acute kidney injury after computed tomography: a meta-analysis. Ann Emerg Med 2018;71(1):44–53.e4. Crossref, MedlineGoogle Scholar
  • 13. Dekkers IA, van der Molen AJ. Propensity score matching as a substitute for randomized controlled trials on acute kidney injury after contrast media administration: a systematic review. AJR Am J Roentgenol 2018;211(4):822–826. Crossref, MedlineGoogle Scholar
  • 14. McDonald JS, McDonald RJ, Comin J, et al. Frequency of acute kidney injury following intravenous contrast medium administration: a systematic review and meta-analysis. Radiology 2013;267(1):119–128. LinkGoogle Scholar
  • 15. Hinson JS, Ehmann MR, Fine DM, et al. Risk of acute kidney injury after intravenous contrast media administration. Ann Emerg Med 2017;69(5):577–586.e4. Crossref, MedlineGoogle Scholar
  • 16. McDonald JS, McDonald RJ, Carter RE, Katzberg RW, Kallmes DF, Williamson EE. Risk of intravenous contrast material–mediated acute kidney injury: a propensity score–matched study stratified by baseline-estimated glomerular filtration rate. Radiology 2014;271(1):65–73. LinkGoogle Scholar
  • 17. McDonald JS, McDonald RJ, Williamson EE, Kallmes DF. Is intravenous administration of iodixanol associated with increased risk of acute kidney injury, dialysis, or mortality? A propensity score–adjusted study. Radiology 2017;285(2):414–424. LinkGoogle Scholar
  • 18. McDonald JS, McDonald RJ, Lieske JC, et al. Risk of acute kidney injury, dialysis, and mortality in patients with chronic kidney disease after intravenous contrast material exposure. Mayo Clin Proc 2015;90(8):1046–1053. Crossref, MedlineGoogle Scholar
  • 19. McDonald RJ, McDonald JS, Carter RE, et al. Intravenous contrast material exposure is not an independent risk factor for dialysis or mortality. Radiology 2014;273(3):714–725. LinkGoogle Scholar
  • 20. Davenport MS, Perazella MA, Yee J, et al. Use of intravenous iodinated contrast media in patients with kidney disease: consensus statements from the American College of Radiology and the National Kidney Foundation. Radiology 2020;294(3):660–668. LinkGoogle Scholar
  • 21. Shao SC, Chan YY, Kao Yang YH, et al. The Chang Gung Research Database—a multi-institutional electronic medical records database for real-world epidemiological studies in Taiwan. Pharmacoepidemiol Drug Saf 2019;28(5):593–600. Crossref, MedlineGoogle Scholar
  • 22. STARSurg Collaborative. Perioperative intravenous contrast administration and the incidence of acute kidney injury after major gastrointestinal surgery: prospective, multicentre cohort study. Br J Surg 2020;107(8):1023–1032. Crossref, MedlineGoogle Scholar
  • 23. Mehran R, Dangas GD, Weisbord SD. Contrast-associated acute kidney injury. N Engl J Med 2019;380(22):2146–2155. Crossref, MedlineGoogle Scholar
  • 24. Levey AS, Coresh J, Greene T, et al. Expressing the modification of diet in renal disease study equation for estimating glomerular filtration rate with standardized serum creatinine values. Clin Chem 2007;53(4):766–772. Crossref, MedlineGoogle Scholar
  • 25. James MT, Tonelli M, Ghali WA, et al. Renal outcomes associated with invasive versus conservative management of acute coronary syndrome: propensity matched cohort study. BMJ 2013;347f4151. Crossref, MedlineGoogle Scholar
  • 26. Kidney Disease: Improving Global Outcomes. KDIGO Clinical Practice Guideline for Acute Kidney Injury. Kidney Int Suppl 2012;2(1):1–138. https://kdigo.org/guidelines/acute-kidney-injury/. Google Scholar
  • 27. Heller M, Krieger P, Finefrock D, Nguyen T, Akhtar S. Contrast CT scans in the emergency department do not increase risk of adverse renal outcomes. West J Emerg Med 2016;17(4):404–408. Crossref, MedlineGoogle Scholar
  • 28. Haveman JW, Gansevoort RT, Bongaerts AH, Nijsten MW. Low incidence of nephropathy in surgical ICU patients receiving intravenous contrast: a retrospective analysis. Intensive Care Med 2006;32(8):1199–1205. Crossref, MedlineGoogle Scholar
  • 29. Thomas LE, Li F, Pencina MJ. Overlap weighting: a propensity score method that mimics attributes of a randomized clinical trial. JAMA 2020;323(23):2417–2418. Crossref, MedlineGoogle Scholar
  • 30. Stuart EA, Lee BK, Leacy FP. Prognostic score–based balance measures can be a useful diagnostic for propensity score methods in comparative effectiveness research. J Clin Epidemiol 2013;66(8 Suppl):S84–S90.e1 https://doi.org/10.1016/j.jclinepi.2013.01.013. Crossref, MedlineGoogle Scholar
  • 31. Austin PC. A comparison of 12 algorithms for matching on the propensity score. Stat Med 2014;33(6):1057–1069. Crossref, MedlineGoogle Scholar
  • 32. Stürmer T, Webster-Clark M, Lund JL, et al. Propensity score weighting and trimming strategies for reducing variance and bias of treatment effect estimates: a simulation study. Am J Epidemiol 2021;190(8):1659–1670. Crossref, MedlineGoogle Scholar
  • 33. Solomon R. Contrast-medium-induced acute renal failure. Kidney Int 1998;53(1):230–242. Crossref, MedlineGoogle Scholar
  • 34. Zager RA, Johnson AC, Hanson SY. Radiographic contrast media–induced tubular injury: evaluation of oxidant stress and plasma membrane integrity. Kidney Int 2003;64(1):128–139. Crossref, MedlineGoogle Scholar
  • 35. Davenport MS, Khalatbari S, Dillman JR, Cohan RH, Caoili EM, Ellis JH. Contrast material–induced nephrotoxicity and intravenous low-osmolality iodinated contrast material. Radiology 2013;267(1):94–105. LinkGoogle Scholar

Article History

Received: Dec 2 2020
Revision requested: Mar 2 2021
Revision received: July 22 2021
Accepted: Aug 5 2021
Published online: Oct 12 2021
Published in print: Dec 2021