REBEL Cast Ep98: Intravenous Contrast and Long-Term Kidney Impairment

Background: It has long been thought that intravenous contrast can lead to acute kidney injury. Recent data, however, has called this dogmatic teaching into question.  Unfortunately, the data arguing against the association of contrast with AKI comes from observational trials and, thus, carry with it numerous biases.  One potential bias is baseline differences in the risk between exposure groups with patients not receiving contrast perceived to be at higher risk and those receiving contrast at lower risk of PC-AKI. Another example is selection bias due to requiring subsequent renal function testing in patients deemed to be higher risk and not those at lower risk.  Both of these can form a control group at high risk of kidney injury which creates a bias in favor of contrast and potentially masking harm.

REBEL Cast Ep98: Intravenous Contrast and Long-Term Kidney Impairment

Paper: Goulden R et al. Association of Intravenous Radiocontrast With Kidney Function: A Regression Discontinuity Analysis. JAMA Intern Med 2021. [Link is HERE]

Clinical Question: Is intravenous contrast exposure associated with clinically significant long-term kidney impairment?

What They Did:

• Quasi-experimental, retrospective, cohort study including all ED patients undergoing d-dimer testing between 2013 to 2018 in a single province in Canada

Fuzzy Regression Discontinuity Design:

• Quasi-experimental pretest-posttest design
• Elicits causal effects of interventions by assigning a cutoff or threshold above or below which an intervention is assigned
• By comparing observations lying closely on either side of the threshold, it is possible to estimate the average treatment effect when randomization is not possible
• This model is effective at replicating results from randomized experiments while other observational study designs are not
• Sharp vs Fuzzy Design
  • Sharp: Variable cut point perfectly predicts treatment status (Yes or No)
  • Fuzzy: Variable cut point does not perfectly predict treatment status (Some treatment group members do not receive treatment and some comparison group members do receive treatment (i.e. cross overs)

Outcomes:

• Primary: Estimated GFR (eGFR) up to 6 months following index ED visit
• Secondary:
  • Receipt of renal replacement therapy (dialysis or kidney transplantation) in the 6 months after index ED visit
    • AKI = increase in Cr level of 50% or 0.3mg/dL within 7d
  • All-cause mortality at 6 months

Inclusion:

• ≥18 years of age
• D-dimer measured during an ED visit

Exclusion:

• No baseline eGFR within 2 hours of d-dimer result
• Receiving dialysis
• Renal transplantation in preceding 6 months

Results:

• 156,028 patients received a d-dimer test
  • Mean eGFR = 86mL/min/1.73m²
  • Patients just above and below the CTPA eligibility cutoff were similar in terms of measured confounders
• NO EVIDENCE for an association of contrast with eGFR up to 6 months later associated with CTPA exposure:
  • Data available for 84,624 patients (54%)
  • Mean change in eGFR = -0.4mL/min/1.73m²
  • 95% CI -4.9 to 4.0
• No association with need for renal replacement therapy
  • 165 (0.11%) of patients required renal replacement therapy
    • 161 dialysis
    • 4 kidney transplants
  • Risk Difference 0.07%
  • 95% CI -0.47% to 0.61%
• No association with mortality
  • 6,656 (4.3%) patients died in the 6 months following index ED visit
  • Risk Difference 0.3%
  • 95% CI -2.9% to 3.2%
• No association with acute kidney injury within 7d
  • Available for 42,691 patients (27%)
  • 4,147 (9.7%) developed AKI
  • Risk Difference 4.3%
  • 95% CI -2.7% to 12.9%
• Subgroup Analyses:
  • Potentially consistent signal of harm among patients with diabetes (mean eGFR change -6.4 mL/min/1.73m²; 95% CI -15.4 to 0.2) although not statistically significant
  • eGFR <45mL/min/1.73m2 only had a difference of 0.5mL/min/1.73m2 (95% CI 012.1 to 33.8; p = 0.74) but again not statistically significant

Strengths:

• Large sample size from a population-wide based data set
• Used a well validated regression discontinuity design to avoid potential confounders
• Groups were well balanced at baseline in terms of age, baseline eGFR, diabetes, hypertension, CAD, cancer and Charlson comorbidity index score
• Performed 8 sensitivity analyses of the primary outcome. 7 of them found no evidence of association
• Inclusion of patients was not based on the ability to repeat creatinine and based solely on baseline characteristics which helps remove selection bias
• No evidence of discontinuities in any of the measured confounders which is similar to a well done randomized clinical trial

Limitations:

• AKI secondary outcomes limited by missing data
• Subgroup analyses were relatively underpowered
• It is possible in patients at higher risk of AKI, were less likely to have CTPA determined by d-dimer results and therefore potentially underrepresented
• Data only available for 50% of patients
• Only 1 type of CT and one group of patients. Results could be different with different contrast loads/indications

Discussion:

• The authors chose the primary outcome of measured eGFR up to 6 months after the index ED visit as they felt it was more patient-centered than AKI (i.e. need for permanent kidney replacement therapy) and it also addressed the issue of selection bias that could arise with using short term AKI as an outcome (indications for testing eGFR months after ED visit much less likely to be affected by variables associated with initial probability of CTPA exposure)

Author Conclusion: “Using a cohort study design and analysis that permits stronger causal interpretation than existing observational research, we found no evidence for a harmful effect on kidney function of intravenous contrast administered for CTPA in an emergency setting.”

Clinical Take Home Point: In this methodologically novel trial, there was no association between IV contrast from CTPA and worsening eGFR up to 6 months after index ED visit.  Although a randomized clinical trial would be great, this may be the highest-level evidence we achieve.  This trial plus other observational trials on this topic all point to a change in protocols to allow for IV contrast in patients regardless of kidney function.

References:

1. Goulden R et al. Association of Intravenous Radiocontrast With Kidney Function: A Regression Discontinuity Analysis. JAMA Intern Med 2021. [Link is HERE]
2. Guduguntla V et al. Exploiting Clinical Decision-Making Threholds to Recover Causal Effects From Observational Data: Randomization Without Trials. JAMA Intern Med 2021. [Link is HERE]

For More Thoughts on This Topic Checkout:

• REBEL EM: The Kompas Trial – Sodium Bicarb Prehydration in Adults with CKD Prior to IV Contrast-Enhanced CT
• REBEL EM: Post Contrast Acute Kidney Injury (PC-AKI) – We Don’t Need Another Retrospective Observational Study
• REBEL EM: Contrast Induced Nephropathy – A Modern-Day Medical Myth
• REBEL EM: Is Contrast Induced Nephropathy Really Not a Thing?
• REBEL EM: The AMACING Trial – Prehydration to Prevent Contrast Induced Nephropathy (CIN)?
• REBEL EM: Contrast Induced Nephropathy – Fact or Myth?

Post Peer Reviewed By: Anand Swaminathan, MD (Twitter: @EMSwami)