Longitudinal Risk of Adverse Events in Patients With Acute Kidney Injury After Percutaneous Coronary Intervention: Insights From the National Cardiovascular Data Registry

Acute kidney injury (AKI) is a common complication after percutaneous coronary intervention (PCI)^1,2 and is associated with significant morbidity and mortality. Most studies evaluating the development of post-PCI AKI and its association with adverse outcomes of bleeding, myocardial infarction (MI), and death have focused on in-hospital adverse events, with relatively few focusing on longitudinal, post-discharge outcomes.^1–13 Furthermore, a majority of studies defining these relationships have occurred in single centers and small cohorts, preceding widespread use of guideline-recommended strategies for prevention of AKI,^14–16 using varying definitions of AKI, and without investigation of the longitudinal relationship between post-PCI AKI and recurrent renal injury. Finally, a broader understanding of when these events occur during follow-up is lacking, missing an opportunity to define the highest periods of risk for post-PCI AKI patients.

A recent analysis from the National Cardiovascular Data Registry (NCDR) using the Acute Kidney Injury Network (AKIN) definition of AKI¹⁷ evaluated the incidence, predictors, and in-hospital outcomes of post-PCI patients having AKI.¹ This study represents the largest evaluation of AKI in a contemporary PCI cohort to date and used a standardized definition of AKI that has been adopted by nephrology, critical care, and cardiology communities alike.^4,6,17,18 However, the association between post-PCI AKI and adverse events after hospital discharge remain less well defined, with the risk of recurrent renal injury understudied and the timing of these events unknown. Understanding this relationship in the context of contemporary PCI care may help distinguish those patients who remain at risk for adverse events after discharge and may identify a period of vulnerability where risk is greatest.

Accordingly, we used the NCDR CathPCI registry to evaluate the relationship between post-PCI AKI and the incidence of adverse events after hospital discharge. By leveraging one of the world’s largest PCI registries, we sought to better define the associations of post-PCI AKI and adverse events in a contemporary setting, with standard and accepted definitions of the AKI exposure. Understanding this relationship and the timing of these events in a large, multicenter, real-world population may aid clinicians in the identification of critical periods of vulnerability for patients after PCI.

In this national, multicenter, real-world cohort of Medicare beneficiaries, we found that AKI remains a common complication after PCI, occurring in 8.8% of those patients with periprocedural assessment of renal function. Furthermore, in-hospital AKI is associated with significant morbidity and mortality after hospital discharge. This analysis demonstrates an association between in-hospital AKI after PCI and higher rates of post-discharge mortality, MI, bleeding-related hospitalization, and recurrence of renal insufficiency at 1 year when compared with patients without in-hospital AKI, with the hazard of adverse events increasing in parallel with the severity of in-hospital AKI as defined by AKIN criteria. Finally, these adverse events occur most frequently in the first 30 days after discharge, suggesting a critical period of vulnerability for patients with AKI after PCI.

Previous work demonstrating an association between inpatient AKI and adverse outcomes has been limited by small cohorts, single-center evaluations, restriction to acute MI or clinical trial populations, or the use of nonstandard definitions of AKI.^3–13 Furthermore, the majority of these studies have focused solely on inpatient events. The few that evaluated long-term risks of AKI demonstrated similar findings of increased risk for cardiovascular events, mortality, and bleeding but did not assess long-term effects on recurrent renal injury. In addition, these studies were performed in varied populations, with nonstandardized definitions of AKI, and with many occurring before national societal efforts recognizing and raising awareness of the risks associated with post-PCI AKI. Our analysis evaluated all Medicare beneficiaries undergoing PCI with periprocedural creatinine assessment used a recognized and standard definition of AKI and reflects a national experience in the largest evaluation of the association between post-PCI AKI and post-discharge outcomes to date.

Our findings not only add to the body of literature supporting a relationship between post-PCI AKI and adverse outcomes but also suggest that this relationship persists well after hospital discharge. In addition, our analysis demonstrates a novel association between in-hospital AKI and long-term risk for recurrent renal injury. It is possible that although initial clinically apparent renal injury may be transient, subclinical injury may persist and predispose patients to adverse outcomes beyond the initial event. Finally, our findings demonstrate an early separation in the incidence of adverse events among patients having in-hospital post-PCI AKI, suggesting an early period of vulnerability for these patients within the first 30 days after hospital discharge. It is possible that novel, modifiable risk markers beyond the presence of AKI can be identified for targeted intervention in the immediate post-discharge period.

Hazard for mortality and adverse events after AKI have not significantly changed from studies performed before our current awareness of the risks associated with AKI, suggesting an ongoing opportunity to develop possible interventions after the development of AKI to mitigate risk of downstream events. Many have pointed to the lack of modifiable risk factors for the development of contrast-induced AKI, outside of avoiding the procedure itself through more discriminant patient selection.^25,26 However, it remains possible that novel, separate, and most importantly, modifiable risk factors for adverse outcomes exist within those that develop AKI that can serve as targets for intervention, such as more judicious and cautious exposure to potentially nephrotoxic agents like angiotensin receptor blockers or angiotensin-converting enzyme inhibitors among those with recent AKI.²⁷ Furthermore, there remains evidence to suggest that a systematic quality improvement intervention to reduce post-PCI AKI can impact not only rates of AKI but also outcomes of mortality and bleeding,²⁸ suggesting that processes outside of the PCI procedure itself can modulate risk. It is also possible that more proximal evaluation of these patients after discharge or through increased scrutiny of their care and post-discharge management can reduce risk of adverse events after discharge.²⁵ Further study is needed to identify novel risk factors and better define possible interventions, but defining this population and their high risk in the immediate post-discharge period is an important first step and offers a time period for clinicians to focus on.

This study should be viewed in the context of its limitations. First, AKI was defined by the change in serum creatinine from the most recent measurement before PCI. It is possible that initial creatinine measurements may not have represented a true baseline and may have been affected by hemodynamic instability or other sources of renal injury, reducing the sensitivity of our analysis to the development of AKI within our cohort. This potential misclassification would bias results toward the null, but our findings nevertheless demonstrated significant association between AKI and adverse events. Second, our study cohort relied on the presence of a complete periprocedural assessment of renal function (both pre- and postprocedure assessment of creatinine) to determine the development of in-hospital AKI, and many patients were excluded because of missing values. We compared the characteristics of those patients excluded from our analysis with those included and found that those patients without a complete periprocedural creatinine assessment had significantly less comorbidities, less acuity of presentation for PCI, less contrast use, and lower rates of multivessel PCI. These patients would be less likely to have post-PCI AKI by validated risk models and less likely to have adverse events post-PCI.^24,29 The inclusion of these patients, by risk model estimation, would reduce the overall rate of both AKI and adverse outcomes, increasing the apparent strength of the relationship between development of AKI and adverse events. Therefore, exclusion of these patients likely further biased our results toward the null hypothesis, which further supports the validity of the demonstrated association. Third, adverse events were obtained via the use of billing and administrative data, with the possibility of missing events. We limited our cohort to only fee-for-service Medicare eligible patients, using validated linkage between CMS and NCDR data to minimize this possibility. However, by limiting our cohort to a Medicate population, the findings of this analysis may not be representative of younger patients having AKI after PCI. Fourth, this is an observational study and the findings represent associations that should be viewed as hypothesis generating and should not imply causality. Although we used robust adjustment methods for predictors of post-PCI mortality and adverse events, residual or unmeasured confounding may remain, impacting the associations demonstrated. Fifth, we are unable to capture medication changes in follow-up that may have occurred as a result of the development of AKI, which may have had impact on our findings. Sixth, increased operator awareness of the risks of AKI may have resulted in increased efforts to reduce the incidence of post-PCI AKI. Although these efforts may impact the rate of AKI seen in our analysis, they do not alter the relationship noted. Finally, the patients and hospitals participating in the NCDR may not accurately reflect all PCI performed in the United States. However, the CathPCI registry captures the majority of PCI procedures performed nationally and is the largest repository of PCI data in the United States, reflecting the majority of PCI performed in this country.

In a contemporary cohort of Medicare beneficiaries undergoing PCI within the United States, in-hospital AKI was significantly associated with increased risk of death, MI, bleeding, and recurrent kidney injury after hospital discharge. The majority of these events occurred within 30 days of hospital discharge. Clinicians should be aware of an increased risk of adverse events after hospital discharge among patients having post-PCI AKI, including recurrent renal injury.