Packed red blood cell transfusion associates with acute kidney injury after transcatheter aortic valve replacement

Aortic stenosis is a common form of degenerative valve disease and its prevalence markedly increases as people age [1, 2]. Many patients with aortic stenosis have comorbidities that place them at higher risk for morbidity and mortality after surgical aortic valve replacement (SAVR). As a consequence, trans-catheter aortic valve replacement (TAVR) has been increasingly utilized as an alternative to SAVR for aortic valve replacement in patients who are intermediate or high risk SAVR candidates. Despite being less invasive than SAVR and not requiring cardiopulmonary bypass (CPB), the risk of TAVR patients developing post-procedure AKI is similar to the risk observed in SAVR patients (~ 10–30%) [3‐6]. The occurrence of AKI is associated with increased morbidity and mortality in both SAVR and TAVR patients [3, 7‐10]. It is therefore important to identify perioperative risk factors that are potentially modifiable for AKI prevention.

Various clinical risk factors have been associated in observational studies with the development of AKI after TAVR [3, 9‐14]. These have included clinical variables such as chronic kidney disease, trans-apical TAVR approach, diabetes, hypertension, peripheral vascular disease, chronic obstructive pulmonary disease, history of myocardial infarction, leukocytosis, bleeding, and blood transfusion [3, 9‐14]. Interestingly, a meta-analysis of observational studies reported that contrast media volume is not significantly associated with development of AKI after TAVR [15].

Transfusion of packed red blood cells (pRBCs) during cardiac surgery or TAVR may be avoided depending upon clinical management of factors such as preoperative anemia, perioperative fluid administration, and utilized transfusion thresholds. In cardiac surgical patients, pRBC transfusion has been associated with development of AKI, with perioperative anemia seeming to present an additional additive risk [16‐18]. Observational cohort studies examining the association between pRBC transfusion and the development of AKI after TAVR have reported conflicting results [11, 12, 19]. To date, the studies that have assessed the association between pRBC transfusion and AKI after TAVR have not included peri-procedure anemia, fluid balance, intra-procedure hypotension, and intra-procedure and post-procedure inotrope/vasopressor use as potential confounders. Since lower nadir hemoglobin (Hgb), hypotension and need for inotropes or vasopressor drugs may occur in conjunction with bleeding and need for blood transfusion, it is important to assess these factors along with pRBC transfusion in order to identify what risk factors might best be targeted to prevent AKI after TAVR.

Given the potential inter-relations between peri-procedure pRBC transfusion, fluid balance, anemia, hypotension and vasoactive drug administration, this study aimed to assess if pRBC transfusion as well as these other potential risk factors associate with the development of AKI after TAVR. The primary hypothesis of this study is that pRBC transfusion associates with AKI after TAVR even after adjusting for other clinical parameters such as peri-procedure anemia and the use of vasoactive drugs.

AKI is known to be associated with increased risk of mortality and renal and non-renal morbidities following TAVR [8, 26]. Despite the less invasive trans-catheter approach versus open SAVR with CPB, 17.2% of the TAVR patients in this study developed post-TAVR AKI. Similar to what has been reported in other TAVR cohorts, most of the AKI observed in this study was Stage 1 AKI [2, 10, 14, 27‐29]. However, even Stage 1 AKI is associated with increased mortality in ambulatory [8], non-cardiac surgery [7], cardiac surgery [30] and TAVR patients [27, 28]. Thus, it is important to identify risk factors that can potentially be modified to mitigate development of AKI after TAVR.

In this study we found that pRBCs transfusion during the TAVR procedure or the first 24 h after TAVR significantly associates with the occurrence of post-TAVR AKI. This was true even after adjusting for other clinical parameters including pre-TAVR eGFR, nadir measured peri-procedure Hgb, and post-procedure inotrope and vasopressor use. Just over one-fourth of our TAVR cohort underwent pRBC transfusion. This transfusion rate is similar to or lower than the rates of pRBC transfusion that have been reported in other TAVR cohort studies [10, 14, 29, 31, 32]. Several prior TAVR cohort studies and a recent meta-analysis have also reported significant associations between pRBC transfusion and development of AKI [11, 12, 31, 32]. What differentiates our study from prior studies is that we concurrently assessed the associations between intra-procedure hypotension, peri-procedure anemia, and inotrope/vasopressor usage with occurrence of post-TAVR AKI. These are additional factors that could potentially affect perfusion and oxygen delivery to the kidneys, predisposing to AKI.

There are multiple biologic characteristics of pRBCs and physiologic responses to transfusion that support the concept of pRBC transfusion as a putative cause of AKI. Stored allogenic pRBCs undergo changes in shape and deformability that can decrease oxygen delivery to tissues such as the kidney [33, 34]. Plasma levels of the inflammatory biomarkers bactericidal permeability increasing protein (BPI) and interleukin-6 have also been reported to be significantly elevated in patients who received pRBCs [35]. An in vivo study by Donadee et al identified storage mediated hemolysis causing impaired vascular function due to endothelial dysfunction and vasoconstriction [36]. RBC hemolysis occurs during pRBC storage and following transfusion can lead to increase in plasma free Hgb and iron which cause dysfunction of microcirculation [18].

pRBC transfusion has been associated with postoperative AKI or reduced postoperative eGFR in several observational studies of patients who underwent cardiac surgery with CPB [16, 17, 37]. These studies found that preoperative anemia and nadir intraoperative Hgb significantly associate with development of AKI [16, 17, 37]. In our study, nadir Hgb did not remain significantly associated with post-TAVR AKI in the multivariable analysis. Additional larger prospective studies may be warranted to assess for interactive influence of peri-procedural anemia and transfusion in TAVR patients.

To our knowledge no prior study of TAVR patients has assessed the impact of intra-procedure hypotension on development of post-TAVR AKI. One cohort study of 213 TAVR patients did assess the occurrence of any procedural complication that led to severe sustained hypotension and assessed the association between such occurrences and development of AKI. They did not observe an association between such complications and development of post-TAVR AKI, but only 10 patients experienced these complications [32]. We assessed occurrence of MAP < 60 mmHg for greater than or equal to 5 min and also added up the total minutes in all hypotensive episodes that lasted greater than or equal to 5 consecutive minutes. Neither of these intra-procedure hypotension variables significantly associated with AKI in our TAVR cohort. Some have hypothesized that rapid ventricular pacing during TAVR procedure (i.e. period of no blood pressure) may be associated with development of post-TAVR AKI. We did not observe a significant association between rapid pacing and development of post-TAVR AKI, a finding consistent with several other TAVR studies [14, 32].

We assessed hypotension in the intraoperative period but not in the postoperative period, as minute-to-minute post-TAVR blood pressures were not available for our retrospective study cohort. We did find that post-procedure inotrope/vasopressor use was independently associated with development of AKI. Inotrope/vasopressor drug use and its association with AKI has been studied in the cardiac surgical literature with mixed results. A study by Haase et al. did not find an association between vasopressor administration and AKI in patients undergoing on-pump cardiac surgery [37]. A study by Magruder et al. investigating patients who developed AKI following cardiac surgery with CPB showed higher epinephrine dose on ICU arrival and greater total administered dose of epinephrine and norepinephrine in patients who developed postoperative AKI [38]. Porhomayon et al. found a significantly higher incidence of AKI following vasopressin use during coronary artery bypass graft surgery with CPB [39].

It is not uncommon to administer inotropes and vasopressor drugs during the TAVR procedure in order to offset transient effects of anesthesia and/or rapid ventricular pacing. Interestingly, in our study the number of concurrent inotropes and vasopressors administered during the TAVR procedure was not associated with development of post-TAVR AKI. However, maximum number of required concurrent inotropes or vasopressors administered after the TAVR procedure was an independent predictor of AKI. This finding warrants further prospective study of the impact of post-TAVR inotropes and vasopressors with a prospectively defined protocol for how inotropes and vasopressors are dosed and when multiple inotropes and vasopressors are administered. Future study also appears warranted to address whether it is post-TAVR hypotension, low cardiac output, or the vasoconstrictive effects of these drugs (or a combination of all of these factors) that predisposes to AKI.

Iodinated contrast media is administered during TAVR procedures and is known to have properties that can cause intense and prolonged vasoconstriction and direct renal tubular damage [40]. However, awareness of the potential toxicity of contrast has led to common practices of pre-procedure hydration, utilization of low osmolar contrast media and minimizing contrast dose. Contrast dose did not associate with AKI in our cohort of TAVR patients. Contrast dose also did not associate with post-TAVR AKI in several other TAVR cohort studies [11, 14, 15, 19, 27‐29, 31, 32].

Potential limitations of our study should be considered. It should be recognized that this is a single center observational study, and additional studies at other centers should be performed to validate our study’s findings. The estimated odds ratios for our multivariable model could be biased by the number of AKI events observed in our cohort (n = 20), but Vittinghoff and McCulloch (2007) suggest this potential bias is likely minimal [41]. The association between pRBC transfusion and AKI has been observed in some but not all previously reported observational TAVR cohort studies, so our study further corroborates those studies that have reported a significant association between pRBC administration and post-TAVR AKI [12]. However, future studies are needed to work out the pathobiologic mechanism(s) of this observed association. In particular, our findings suggest that anemia may not be a strong modifier of the observed association between pRBC transfusion and post-TAVR AKI. However, our study is an exploratory retrospective assessment, so future larger studies would be useful to validate this observation. Additionally, nadir Hgb was determined from available routine clinical care laboratories. In the setting of bleeding or anticipated bleeding, clinicians might initiate pRBC transfusion before a Hgb level is checked, and, thus, true nadir Hgb might have been missed for some patients. This would possibly underestimate the association between Hgb and AKI. Also, this is a retrospective observational study; therefore a Hgb threshold for initiating pRBCs in clinical practice is not driven by formal study protocol and may have some variability from clinician-to-clinician and patient-to-patient. The significant association between number of inotrope and vasopressor infusions concurrently administered after TAVR and the development of post-procedure AKI is an intriguing finding, but future prospective studies are needed to work out the interactions between patients’ hemodynamic profiles as well as vasoactive drug dosing and the development of AKI in the post-procedure setting. Finally, many of our patients are referred to our medical center for TAVR, but their long-term medical care is continued at outside facilities. Therefore this retrospective study cannot reliably evaluate mid- and long-term post-TAVR outcomes.

Transfusion of pRBCs but not nadir perioperative Hgb independently associated with post-TAVR AKI. Maximum concurrent number of administered inotropes and vasopressors during the post-TAVR period also independently associated with the development of post-TAVR AKI. Future studies are needed to explore the pathobiological mechanisms underlying these associations as well as to assess the impact of approaches such as pre-procedure anemia optimization and restrictive pRBC transfusion thresholds on the development of post-TAVR AKI.