記住我
Abstract
Background: As visceral protein expression may influence outcomes in patients with cardiovascular disease, we investigated whether pre-procedural albumin concentration is associated with length of stay (LOS) and 90-day mortality after transcatheter aortic valve repair (TAVR).
Methods: We retrospectively analyzed data from TAVR patients at our institution between January 2013 and December 2017. For all patients, baseline albumin concentration was assessed between one and four weeks before the procedure. To investigate the association between albumin concentration and outcomes, we performed regression analyses, controlling for Society of Thoracic Surgeons, New York Heart Association classification, and Kansas City Cardiomyopathy Questionnaire 12 scores.
Results: Three hundred eighty patients were included in the analyses. Cox-proportional hazards regression showed that patients with albumin concentrations <3.5 g/dL were 80% more likely to have prolonged ICU LOS (HR 1.79; 95%CI 1.04–2.57, P = 0.03) and 70% more likely to have prolonged hospital LOS (HR 1.68; 95%CI 1.01-2.46, P = 0.04) compared to patients with albumin concentrations >3.5 g/dL. Logistic regression showed that patients with albumin concentrations <3.5 g/dL were four times more likely to not survive to 90 days (OR 3.94; 1.13–12.63, P = 0.03) after their TAVR compared to patients with albumin concentrations >3.5 g/dL.
Conclusion: Our data suggest that patients with pre-procedural albumin concentrations <3.5 g/dL are at an increased risk of adverse outcomes after TAVR compared to patients with albumin concentrations ≥3.5 g/dL. Prospective studies are needed to determine whether risk stratification based on pre-procedural albumin can improve outcomes and whether targeted interventions can improve pre-procedural albumin concentrations in potential TAVR candidates.
Keywords: Albumin, aortic stenosis, transcatheter aortic valve replacement
How to cite this article: Introduction 
In the United States, prevalence of aortic stenosis (AS) exceeds 12% in the general population.[1] Without intervention, severe AS carries a 50% mortality at 2 years from diagnosis.[2] For decades, the only interventional approach to the management of severe AS was surgical aortic valve replacement (SAVR). However, SAVR requires sternotomy and cardiac bypass, and has been associated with an intra-operative mortality as high as 10% for patients who have advanced age, left ventricular dysfunction, and chronic renal disease. As such, almost one-third of patients with severe AS were traditionally not offered surgical intervention.[3],[4],[5] Over the last 20 years, percutaneous transcatheter aortic valve replacement (TAVR) was developed and refined in order to reduce peri-procedural morbidity and mortality in high-risk patients.[6] Indeed, 30-day mortality after TAVR has been reported to range between 1% and 9%, compared to 0%–5% for SAVR, whereas 1-year survival after TAVR may range from 76% to 85%, compared to 90% to 95% for SAVR.[7]
TAVR candidates typically have significantly greater comorbidities compared to their surgical counterparts, and therefore higher mortality in this latter cohort of patients may be unavoidable.[8] Moreover, despite it being considered a minimally invasive procedure, in-hospital costs of TAVR are comparable to SAVR.[9] Given that TAVR is an expensive therapy, with relatively high mortality, patient selection may help to improve outcomes. Emerging data suggests that frailty plays an important role in recovery after TAVR.[10] However, to date, there is no universally accepted scoring system or biomarker to assess frailty in TAVR patients.[11] Serum albumin concentrations have been used to assess frailty in various disease states; preliminary data suggest that it may be useful for short (30-day) or long-term (1 year) risk-stratification in TAVR patients.[12],[13] In addition, recent data suggest that although 30-day mortality has been validated and reported most often, 90-day mortality may be a more reliable outcome metric for measuring hospital performance and capturing procedure-related mortality after TAVR.[7] Therefore, our goal was to investigate whether pre-procedural serum albumin concentrations are associated with clinical outcomes within 90-days after TAVR.
Materials and Methods After approval of the study protocol by our Institutional Review Board, we reviewed the medical records of patients who underwent TAVR from January 1, 2013, through December 31, 2017. We only included patients who had serum albumin concentrations measured between one and four weeks before their TAVR. To minimize the impact of systemic inflammation on measured albumin concentrations, we excluded all patients who had been hospitalized within 4 weeks before their measured albumin concentration (IRB approval was obtained on 9/6/19).
Detailed clinical variables were collected for each eligible patient including: 1) age; 2) sex, 3) body mass index; 4) body surface area; 5) Charlson Comorbidity Index; 6) Society of Thoracic Surgeons (STS) risk score; 7) 6-minute walk test; 8) Kansas City Cardiomyopathy Questionnaire 12 (KCCQ12); 9) pre-procedural albumin concentration; 10) pre-TAVR aortic valve area; 11) pre-TAVR mean gradient across the aortic valve; 12) pre-TAVR left ventricular ejection fraction; and 13) New York Heart Association (NYHA) functional classification. Outcome data collected included intensive care unit (ICU) length-of-stay (LOS), hospital LOS, discharge destination (non-home vs home), and 90-day mortality. Survival data was validated by cross-referencing our clinical records with the Social Security Death Index to ensure that out-of-network patients were not lost to follow-up.
Bivariate tests were used to compare the characteristics of patients who did or did not survive 90-days post-TAVR. Data were reported as either mean ± standard deviation, medians with interquartile ranges (IQRs), or proportions (%), and were compared using t-tests, Mann-Whitney U or log-rank tests, and Chi-square tests, respectively. We then constructed a locally weighted scatterplot smoothing (LOWESS) curve to graphically represent the relationship between pre-procedural serum albumin concentration and risk of 90-day mortality. LOWESS is a type of nonparametric regression that summarizes a relationship between two variables in a fashion that initially relies on limited assumptions about the form or strength of the relationship.[14] The rationale and methods underlying the use of LOWESS for depicting the local relationship between measurements of interest across parts of their ranges have previously been described.[15] Based on the findings of the LOWESS curve, albumin concentration was dichotomized as either “low-risk” or “high-risk” for 90-day mortality after TAVR. Next, Kaplan-Meir curves were constructed to depict any differences in ICU and hospital LOS and compared between “low-risk” and “high-risk” patients who survived to hospital discharge using log-rank tests.
A Cox Proportional-Hazards model, controlling for STS score, NYHA classification, and KCCQ12 was used to investigate the relationship between pre-procedure albumin concentration and ICU as well as hospital LOS. Based on previous studies in our TAVR patients, we assumed that 10% of patients would have pre-procedural albumin concentrations that could be categorized as “high-risk” for 90-day mortality. We also assumed, based on previous studies, that the correlation between albumin concentrations and STS score, NYHA classification, and KCCQ12 was roughly 0.6 (i.e., R2 = 0.36). Therefore, to observe a relative hazard ratio of 2, with alpha set at 0.05 and beta set at 0.2, would require a minimum of 284 events/discharges. On the contrary, logistic regression, adjusting for the same potential confounders, was used to investigate the association of pre-procedure albumin concentration with discharge destination as well as 90-day mortality. Based on previous studies we assumed that the prevalence of non-home discharge was 30% and 90-day mortality was 10%. Therefore, with a logistic regression model including three covariates, alpha set at 0.05, and beta set at 0.2, would require a minimum of 100 and 300 patients, respectively, to detect a relationship between pre-procedure albumin concentration and discharge disposition (non-home vs home) and 90-day mortality. All analyses were performed using Stata v15.1 (StataCorp LLC, College Station, Texas). Two-tailed P < 0.05 and Hazards as well as Odds Ratios (HRs and ORs, respectively) with 95% confidence intervals (CIs) that did not include a value of 1, were considered to be statistically significant.
Results Three hundred eighty patients comprised the analytic cohort and the overall 90-day mortality was 8%. Characteristics of the study cohort are shown in [Table 1]. LOWESS curve analysis showed a notable flattening of the 90-day mortality risk between albumin concentrations of 3 g/dL and 4 g/dL [Figure 1]. As such, we assumed that the “normal” population serum albumin concentration threshold of 3.5 g/dL may be applicable to TAVR patients as well. Kaplan-Meir curves comparing TAVR patients with baseline serum albumin concentrations of <3.5 g/dL to those with concentrations ≥3.5 g/dL, and who survived to hospital discharge, showed a significant difference between groups for ICU [Figure 2] as well as hospital LOS [Figure 3]. Cox-proportional hazards analysis showed that when controlling for STS score, NYHA classification, and KCCQ12, patients with albumin concentrations <3.5 g/dL were almost 80% more likely to have prolonged ICU LOS compared to their counterparts with albumin concentrations ≥3.5 g/dL (HR 1.79; 95%CI 1.04–2.57, P = 0.03). Similarly, when controlling for ICU LOS, NYHA classification, and KCCQ12, patients with albumin concentrations <3.5 g/dL were almost 70% more likely to have prolonged hospital LOS compared to their counterparts with albumin concentrations ≥3.5 g/dL (HR 1.68; 95%CI 1.01–2.46, P = 0.04). Logistic regression analysis, controlling for STS score, NYHA classification, and KCCQ12, showed that patients with pre-procedural albumin concentrations <3.5 g/dL were five times more likely to have a non-home discharge compared to their counterparts with albumin concentrations ≥3.5 g/dL (OR 5.0; 95%CI 1.98–12.66, P < 0.001). Similarly, when controlling for ICU LOS, NYHA classification, and KCCQ12, patients with albumin concentrations <3.5 g/dL were almost 4 times more likely to not survive to 90-days after their TAVR procedure compared to their counterparts with albumin concentrations ≥3.5 g/dL (OR 3.94; 1.13–12.63, P = 0.03).
Figure 2: Intensive care unit length of stay stratified by albumin concentration (n = 380). Kaplan-Meir curves show that intensive care unit length of stay (in hours) is significantly longer in patients with albumin concentration <3.5 g/dL compared to patients with concentrations ≥3.5 g/dL. All patients survived to hospital discharge
Figure 3: Hospital length of stay stratified by albumin concentration (n = 380). Kaplan-Meir curves show that hospital length of stay (in days) is significantly longer in patients with albumin concentration <3.5 g/dL compared to patients with concentrations >3.5 g/dL. All patients survived to hospital discharge Discussion In this single-center, retrospective study, we show that TAVR patients with low pre-procedural albumin concentrations (<3.5 g/dL) have a higher risk of prolonged ICU and hospital LOS, fivefold risk of non-home discharge, as well as a fourfold risk of 90-day mortality when compared to patients with normal albumin concentrations (≥3.5 g/dL). Although the results of our study are observational in nature, the biological plausibility of these findings is undeniable.
Albumin is the most abundant protein in the bloodstream, making up approximately 60% of total circulating proteins.[16] In adults, the normal range for albumin concentration varies between 3.5 and 5.0 g/dL.[16] Albumin is synthesized in the liver and functions as a transporter of exogenous as well as endogenous ligands and is a modulator of plasma oncotic pressure.[17] Hypoalbuminemia can occur due to states of decreased protein synthesis, such as in chronic liver disease or malnutrition.[18],[19] It can also result from acute loss of serum albumin (e.g., renal or gastrointestinal disease) or sudden fluid shifts (e.g., sepsis or acute systemic inflammation).[20] Furthermore, low albumin concentrations may serve as a biomarker for cachexia in patients with longstanding illnesses.[18] Cardiac cachexia is a phenomenon that occurs in patients with advanced cardiovascular disease and is especially prevalent in patients with heart failure.[21] It has been suggested that inflammatory cytokine activation, up-regulation of the renin–angiotensin axis, neurohormonal abnormalities, anorexia, as well as an imbalance between anabolic and catabolic pathways, play a crucial role in its pathogenesis.[22],[23],[24],[25],[26] As such, pre-procedural albumin concentrations may help to identify potential TAVR candidates at increased risk for malnutrition and/or cachexia who may benefit from enhanced preventive measures.
Indeed, Bogdan et al.[27] investigated the relationship between pre-procedural serum albumin concentrations and post-TAVR mortality (n = 150) at 2-years. Mortality risk was found to be higher in patients with albumin concentration <4 g/dL compared to patients with albumin concentrations ≥4 g/dL (HR 2.28; 95% CI 1.17–4.44, P = 0.01). In addition, Koifman et al.[28] investigated the relationship between serum albumin concentrations and in-hospital, 30-day, as well as 1-year mortality after TAVR (n = 476). Similar to our study, the investigators defined low serum albumin concentrations as <3.5 g/dL and normal serum albumin concentrations as ≥3.5 g/dL. The study results showed that patients with low baseline serum albumin concentrations had higher 30-day mortality (12% vs 6%, P = 0.01) and 1-year mortality (29% vs 19%, P = 0.01) compared to patients with normal baseline albumin concentrations. Furthermore, Shimura et al.[29] investigated the link between potential frailty markers (serum albumin concentration, grip strength, gait speed, and Clinical Frailty Score) and post-TAVR outcomes. The study cohort (n = 1,215) was divided into 2 two groups: TAVR patients with hypoalbuminemia (defined as <3.5 g/dL) and those with normal serum albumin (defined as ≥3.5 g/dL). According to the authors, after propensity score matching, they observed higher all-cause and non-cardiovascular mortality in the hypoalbuminemia group compared to the normal serum albumin group. Unfortunately, the details of the methodology used for propensity score matching and for comparing mortality between the two groups were not provided. Hebeler et al.[13] also investigated frailty markers (serum albumin concentration, gait speed, grip strength, and Katz Index of Independence in Activities of Daily Living) and their influence on 1-year mortality after TAVR. Although serum albumin concentration was reported as the only potential frailty marker associated with an increased risk of 1-year mortality, the effect size was not reported. And finally, Kiani et al.[30] conducted a large multicenter retrospective study (n = 36,242) to investigate the association of frailty markers (serum albumin concentration, hemoglobin level, performance on 5-meter walk test) with outcomes post-TAVR. Patients with low serum albumin concentration (<3.4 g/dL) had prolonged hospital LOS (6 days; IQR 3-9 vs 4 days; IQR 3-6, P < 0.001), higher rate of readmission for heart failure at 30-days as well as 1-year (5% vs 3%, P < 0.001 and 16% vs 11%, P < 0.001), higher rate of readmission for bleeding at 30-days as well as 1-year (13% vs 11%, P < 0.001 and 26% vs 20%, P < 0.001), and greater risk of mortality at 30-days as well as 1-year (HR 1.29; 95% CI 1.12-1.48, P < 0.001 and HR: 1.52; 95% CI 1.41-1.64, P < 0.001), compared to patients with normal albumin concentration (≥3.4 g/dL). Our study adds to this body of knowledge by showing a relationship between pre-procedural albumin concentrations and more specific procedure-related clinical outcomes.
Although our findings are interesting, it is important to be mindful of the limitations of our study. Due to the observational nature of our work, a causal relationship between low pre-procedural albumin concentrations and 90-day mortality after TAVR cannot be established. Despite efforts to control for various important variables, there may be residual confounding which may have influenced our findings. Notably, we were unable to fully control for pre-procedural nutritional status or quantity/quality of food intake. Moreover, our study was conducted at a single, teaching hospital, which is a large referral center for high-risk TAVR patients from in and around New England. These and other factors may limit the generalizability of our findings. Furthermore, data regarding our primary outcome of 90-day mortality was robustly available in 95% of patients (i.e., in-network patients or out-of-network patients with follow-up communications beyond 90-days post-TAVR). Despite efforts to cross-reference mortality records with the Social Security Death Index for all patients, it may have been possible to miss patients who did not survive to 90-days in this small subset of patients (i.e., out-of-network with limited follow-up). And finally, we assumed that albumin concentrations did not change significantly in patients from the time of assessment (one to four weeks before procedure) to the day of their TAVR. Several factors, including even mild inflammation, can dramatically alter albumin concentrations and as such, the timing of albumin concentration measurements may influence risk-stratification. These and other issues will be to be expanded upon and addressed in future studies.
Conclusion Our results suggest that low pre-procedural albumin concentration is associated with prolonged ICU as well as hospital LOS, higher likelihood of a non-home discharge, and an increased risk of mortality within 90-days after TAVR compared to patients with normal pre-procedural albumin concentrations. Prospective studies are needed to determine whether risk stratification based on pre-procedural albumin concentration can improve outcomes in potential TAVR patients. In addition, further studies are needed to determine whether targeted interventions can improve visceral protein expression (i.e., optimize anabolic metabolism) in potential TAVR candidates with low baseline circulating albumin concentrations.
Financial support and sponsorship
Department of Anesthesiology and Perioperative Medicine, Tufts Medical Center.
Conflicts of interest
SAQ received consulting fees from Abbott Nutrition, Fresenius Kabi, and Alcresta Therapeutics unrelated to the content of this manuscript. All other authors have no conflicts to declare.
References 
Correspondence Address:
Sadeq A Quraishi
Department of Anesthesiology and Perioperative Medicine; Tufts Medical Center; 800 Washington Street, Ziskind 6038; Boston, MA - 02118
USA
Source of Support: None, Conflict of Interest: None
CheckDOI: 10.4103/aca.aca_114_21
留言 (0)