Secondary Logo

Journal Logo

Adult Circulatory Support

Associations of Preimplant Red Blood Cell Distribution Width with Clinical Outcomes Among Individuals with Left Ventricular Assist Devices

Miller, P. Elliott*; Houston, Brian A.; Schneider, Andrea L. C.; Bush, Aaron L.§; Whitman, Glenn J.; Stevens, Gerin R.; Tedford, Ryan J.; Russell, Stuart D.

Author Information
doi: 10.1097/MAT.0000000000000431
  • Free

Abstract

Although shown to improve mortality compared with medical therapy,1,2 the use of left ventricular assist devices (LVADs) is still associated with a high frequency of device-related complications including pump thrombosis, infection, stroke, and gastrointestinal (GI) bleeding.2–5 There are a number of risk predictive markers and models proposed to evaluate a patient’s candidacy for an LVAD for mortality,6–9 but fewer have comprehensively studied nonmortality outcomes after LVAD implant.10–13

The red blood cell distribution width (RDW) is a measure of red blood cell size variability.14 Felker et al.15 described RDW elevation as independently associated with cardiovascular death, hospitalization, and all-cause mortality in patients with heart failure (HF). They found the RDW to be more strongly associated with these outcomes than traditional parameters such as ejection fraction, renal function, and New York Heart Association functional class.15 The prognostic value of RDW elevation in HF has subsequently been validated in several other cohorts16,17 and meta-analyses.18,19 However, the relationship between RDW and post-LVAD outcomes remains unexplored.

RDW elevation is associated with oxidative stress, a pro-inflammatory state, and tissue hypoxia.20–23 Because of these mechanisms and the close association between LVAD complications and the hematologic system,24,25 one might postulate that a hematologic marker such as RDW would provide unique insight into prognosis in this population. In this study, we assess the association of preimplant RDW with discrete clinical outcomes, including mortality, GI bleed, stroke, pump thrombosis, and infection.

Materials and Methods

Study Population

The medical records of all subjects aged 18 years or older that underwent care for continuous flow LVAD (HeartMate II [Thoratec, Pleasanton, CA] or HeartWare Framingham, MA]) at the Johns Hopkins Hospital between January 1, 2004 and June 30, 2014 were reviewed. Of the 196 patients identified, eight underwent LVAD implantation at an outside facility or had a preimplant biventricular assist device and were excluded, leaving a total of 188 patients in the final analysis. This study was approved by the Johns Hopkins Hospital Institutional Review Board.

Red Blood Cell Distribution Width Definition

RDW values assigned were all from within 24 hours before LVAD implant and represent the last value before implant for each patient. Complete blood cell counts, including RDW values, were measured using a Sysmex XN series analyzer (Sysmex Corporation, Kobe, Japan). Individuals were separated into tertiles with the preimplant tertile 1 defined as less than 15.7%, tertile 2 defined as 15.7% to 18%, and tertile 3 defined as greater than or equal to 18.1%. The RDW normal reference range at our institution is 11.5% to 14.5%.

Outcome Definitions

All records were reviewed for five discrete clinical outcomes more than 1 year of follow-up with all-cause mortality being our primary outcome and GI bleeding, LVAD-associated infection, stroke, and pump thrombosis our secondary outcomes. Gastrointestinal bleeding was defined as an admission for GI bleed in the medical record with confirmation by endoscopy. Left ventricular assist device-associated infection was defined as a driveline infection, pump pocket infection, or sepsis of unknown source documented in the medical record. Pump thrombosis was defined as a pump thrombosis documented in the medical record requiring enhanced anticoagulation or need for pump exchange. Stroke was defined as a new ischemic or hemorrhagic stroke in the medical record with corroborating imaging.

Covariates

Patient records were assessed for demographic variables, including age (years), sex, race (white; nonwhite), diabetes, body mass index (BMI) at the time of implant, LVAD type (Thoratec HeartMate II; HeartWare), cardiomyopathy etiology (ischemic; nonischemic), preoperative ventilator, and Interagency Registry for Mechanically Assisted Circulatory Support (INTERMACS) profile 1–3 (yes; no). Laboratory values were generally from within 24 hours before LVAD implant and included hemoglobin, mean corpuscular volume (MCV), white blood cell (WBC) count, total bilirubin, estimated glomerular filtration rate (GFR), albumin, and international normalized ratio (INR).

Statistical Analysis

The distributions of demographic and clinical characteristics were compared by RDW tertile using χ2 for categorical variables and t-tests for continuous variables. Laboratory values including hemoglobin, MCV, albumin, INR, and total bilirubin were analyzed as continuous variables, and WBC (<4.5 × 109/L and ≥11 × 109/L vs. 4.5 × 109/L to <11 × 109/L) and GFR (<60 ml/min/1.73 m2vs. ≥60 ml/min/1.73 m2) were analyzed as binary values.

Cox proportional hazard models were used to assess the association between all-cause mortality and secondary clinical outcomes with RDW assessed by two methods, RDW tertiles and continuously. Tertiles were chosen because of similarity of the reference (tertile 1) with normal, physiologic RDW values (as assigned by our institution’s laboratory) as well as to preserve power. Patients that were transplanted before 1 year were censored from analyses. The proportional hazard assumptions were checked with the use of Schoenfeld residuals and graphic methods. We tested for trend across the median of RDW tertiles. Analyses were adjusted for demographic variables and comorbidities shown in previous cohorts to predict adverse outcomes (age, sex, race, diabetes, BMI, and INTERMACS profile) and then further adjusted for laboratory values typically included in LVAD risk scores (hemoglobin, MCV, total bilirubin, estimated GFR, albumin, and INR).7 We included hemoglobin and MCV to assess isolated RDW as an independent hematologic marker of risk.

To tailor our analyses and prevent being overfit, each secondary outcome was adjusted for known preimplant risk factors. Gastrointestinal bleed was adjusted for age, BMI, and cardiomyopathy type.26 Stroke was adjusted for age, race, and diabetes.26 Pump thrombosis was adjusted for age, estimated GFR, and BMI.26,27 Finally, infection was adjusted for age, BMI, diabetes, LVAD type, and INTERMACS profile.28

We additionally modeled the association of continuous RDW with all-cause mortality using a restricted cubic spline model with knots at the 5th, 35th, 65th, and 95th percentiles to assess for linearity. Sensitivity analyses for implant strategy (bridge to transplant vs. destination therapy) and implant era (pre or post 2010) were completed.

All reported p values were based on two-sided tests, and p < 0.05 was considered statistically significant. Statistical analysis was performed using Stata SE 13.1 (StataCorp, College Station, TX).

Results

Characteristics of the study population overall and by RDW tertile are described in Table 1. The average age of individuals was 53.0 years with 22.3% women and 45.7% blacks. The etiology of HF was classified as ischemic in 31.9% and nonischemic in 68.1% of individuals. The HeartMate II device was used in 87.8% of implants with the HeartWare comprising the remaining 12.2% of the population. Roughly, 80% of patients had an INTERMACS profile of 1–3. At the time of LVAD implant, the average BMI was 28.3 kg/m2, 44.7% of patients had diabetes, and 44.2% of individuals had an estimated GFR less than 60 ml/min/1.73 m2. More than 1 year of follow-up, there were a total of 56 deaths, 31 incident GI bleeds, 63 incident LVAD-associated infections, 21 strokes, and 12 episodes of pump thrombosis. The 3 and 12 month mortality in our population was 23.4% and 29.8%, respectively. All baseline characteristics were similar between the tertiles except that individuals in tertile 3 were more likely than tertile 1 to have diabetes, an elevated total bilirubin, a lower albumin, and to be on preoperative inotropic support.

Table 1.
Table 1.:
Patient Characteristics Overall and by RDW Tertiles

Compared with reference (tertile 1), both tertiles 2 (HR 6.95 [2.67–18.10]) and 3 (HR 4.61 [1.74–12.21]) showed a significantly greater risk of all-cause mortality (p < 0.005), which persisted after adjusting for demographic data, cardiovascular, and laboratory values (Table 2). Results in Table 2 are shown with only combined adjustment (demographic, cardiovascular disease, and laboratory), as results were similar throughout. We did not find statistically significant associations with RDW tertiles and the secondary clinical outcomes tested.

Table 2.
Table 2.:
Hazard Ratios (95% Confidence Intervals) for the Associations of Baseline Red Blood Cell Distribution Width Tertile with Outcomes from Adjusted Cox Proportional Hazards Models

Figure 1A shows a Kaplan–Meier survival analysis for mortality by RDW tertile. Tertiles 2 and 3 overlap and separate steeply from tertile 1 early after implant (log-rank p value < 0.001). Individuals in tertile 1 with a preimplant RDW less than 15.7% exhibited approximately 10% mortality at 1 year compared with those in higher RDW tertiles, which demonstrated approximately 40% mortality. Figure 1B–E shows our secondary clinical outcomes, which did not reach statistical significance for infection (Figure 1C), stroke (Figure 1D), or pump thrombosis (Figure 1E). In log-rank analysis, GI bleed (Figure 1B) showed an association (p = 0.02) with RDW. Notably, tertile 2 had the greatest survival from GI bleed, and tertile 3 had the worst. Figure 2 shows the continuous association of RDW with mortality using an adjusted restricted cubic spline model and demonstrates that mortality risk was not linearly associated with RDW.

Figure 1.
Figure 1.:
Kaplan–Meier survival analysis for each outcome by red blood cell distribution width tertile. A: Mortality, B: GI bleed, C: Infection, D: Stroke, E: Pump thrombosis. GI, gastrointestinal; LVAD, left ventricular assist device; RDW, red blood cell distribution width.
Figure 2.
Figure 2.:
Adjusted restricted cubic spline model showing the continuous association of red blood cell distribution width with risk of mortality. The solid line represents the hazard ratios and the dashed lines represent the 95% confidence intervals. Knots at 5th, 35th, 65th, and 95th percentiles of included values of red blood cell distribution width. Restricted cubic spline centered at the 10th percentile of included values of red blood cell distribution width. Restricted cubic spline model truncated at 5th and 95th percentile of overall red blood cell distribution width. Adjusted for age (years; continuous; centered at median), sex (male; female), race (white; black), diabetes (yes; no), body mass index (kg/m2; continuous; centered at median), Interagency Registry for Mechanically Assisted Circulatory Support score 1–3 (yes; no), preoperative ventilator support (yes; no), estimated glomerular filtration rate (binary; <60 ml/min/1.73 m2 vs. ≥60 ml/min/1.73 m2), hemoglobin (g/dl; continuous; centered at median), mean corpuscular volume (fL; continuous; centered at median), white blood cell count (binary; <4.5 × 109/L and ≥11 × 109/L vs. 4.5 × 109/L to <11 × 109/L), total bilirubin (mg/dl; continuous; centered at median), albumin (g/dl; continuous; centered at median), international normalized ratio (continuous; centered at median).

In sensitivity analyses, there were no interactions for both implant strategy and implant era (p > 0.05).

Discussion

In this retrospective study, we found that preimplant RDW was strongly associated with mortality more than follow-up of 1 year after adjustment for demographic, cardiovascular, and laboratory variables. The relationship between RDW and mortality was similar between tertiles 2 and 3, suggesting a threshold effect. We did not find that increasing RDW was associated with any of our secondary outcomes.

The RDW reflects heterogeneity in the corpuscular volume of a population of red blood cells14 and is a robust prognostic marker in heart failure15–19 as well as several other disease states, including coronary artery disease.22,29 The biologic mechanism explaining these observations remains poorly defined. However, RDW is associated with a pro-inflammatory state and oxidative stress and has been found to be independently associated with other serologic markers of inflammation such as erythrocyte sedimentation rate and C-reactive protein levels.20 It is thought that oxidative stress impairs the membrane fluidity of erythrocytes, reducing the life span of the red blood cells and leading to anisocytosis, marked by elevated RDW.21 Furthermore, an elevated RDW has been shown to be associated with impaired microvascular perfusion and hypoxia even in patients without anemia.22 With other studies noting an association between RDW and inflammation and microvascular disease,20–22 the observed association between RDW and post-LVAD mortality is biologically plausible.

In patients with chronic kidney disease and HF, recent studies have found that RDW is independently associated with altered catabolism of fibroblast growth factor 23 (FGF23),30 which is a direct cardiotropic hormone involved in mineral metabolism and associated with endothelial dysfunction.31,32 Elevated serum concentrations of FGF23 have been associated with heart failure and mortality.33–35 This mechanism is thought to be because of the direct effects of FGF23 on cardiomyocyte cell signaling and activation of fetal growth pathways.31 As a surrogate for FGF23 elevation, an elevated preimplant RDW may signify a population with more myocardial and endothelial dysfunction, and thus higher risk postimplant.

Prior studies have found that changes in RDW over a relatively brief period of time (during a hospital admission) predict outcomes in patients with heart failure. In 229 consecutive patients admitted with acute decompensated heart failure, Uemura et al.36 found that patients whose RDW levels increased during hospitalization had significantly higher all-cause and cardiac mortality than those whose RDW levels decreased. This raises the question of whether changes in RDW levels in LVAD patients (either preimplant or postimplant) could be used to predict adverse events.

Several risk prediction models, such as the HeartMate II Risk Score (HMRS), have tried to assess postimplant LVAD survival. The HMRS incorporated four readily available variables to predict 90 day mortality. It was found to improve risk discrimination compared with the Destination Therapy Risk Score and Model for End-Stage Liver Disease.7 Although we stress that our results are hypothesis generating and must be validated, RDW could offer clinicians an easily obtainable and potent variable for LVAD risk prediction. Further, a larger cohort should assess for improved risk discrimination when RDW is added to currently available risk scores.

Our study needs to be interpreted in light of its limitations. The single-center nature of our study calls for evaluation of RDW in larger, multicenter LVAD populations with a focus not only on mortality but also on discrete postimplant adverse events. Our population may not match the average patient sample in the United States, which limits generalizability. Notably, our analyses were limited to 1 year of follow-up, which represents a relatively short follow-up in a group of patients nearing survival rates of 4 years.37 It may be that preoperative blood transfusions affected the preimplant RDW. However, our retrospective study was unable to accurately capture preoperative transfusion timing and amount for each patient, although it is more likely that patients would have been transfused postoperatively instead of preoperatively. Furthermore, by including hemoglobin and MCV, we attempted to “double” control for anemia as a possible confounder, and our results persisted after adjustment for these factors. We did not adjust for measures of right heart function because of a smaller sample size with this information available but was also notably not included in the HMRS. This study also has several strengths including the novel application of RDW to the LVAD population, robust multivariable adjustment, and a well-described population with 100% outcome ascertainment for patients implanted at our institution.

Conclusions

Preimplant RDW is strongly associated with mortality more than 1 year in patients supported with an LVAD in those with an elevated RDW. This relationship persisted despite adjustment for multiple risk factors. Given the association with inflammation and oxidative stress, an elevated RDW may represent a sicker population of patients. These results require further study as the RDW could offer a readily available adjunctive marker for risk stratification. Future studies should investigate whether short-term changes in RDW signal an LVAD patient population more prone to complications and would therefore necessitate closer monitoring.

References

1. Rose EA, Gelijns AC, Moskowitz AJ, et al.Randomized Evaluation of Mechanical Assistance for the Treatment of Congestive Heart Failure (REMATCH) Study GroupLong-term use of a left ventricular assist device for end-stage heart failure.N Engl J Med200134514351443
2. Jorde UP, Kushwaha SS, Tatooles AJ, et al.HeartMate II Clinical InvestigatorsResults of the destination therapy post-food and drug administration approval study with a continuous flow left ventricular assist device: A prospective study using the INTERMACS registry (Interagency Registry for Mechanically Assisted Circulatory Support).J Am Coll Cardiol20146317511757
3. Slaughter MS, Rogers JG, Milano CA, et al.HeartMate II InvestigatorsAdvanced heart failure treated with continuous-flow left ventricular assist device.N Engl J Med200936122412251
4. Starling RC, Moazami N, Silvestry SC, et al.Unexpected abrupt increase in left ventricular assist device thrombosis.N Engl J Med20143703340
5. Kirklin JK, Naftel DC, Pagani FD, et al.Sixth INTERMACS annual report: A 10,000-patient database.J Heart Lung Transplant201433555564
6. Lietz K, Long JW, Kfoury AG, et al.Outcomes of left ventricular assist device implantation as destination therapy in the post-REMATCH era: Implications for patient selection.Circulation2007116497505
7. Cowger J, Sundareswaran K, Rogers JG, et al.Predicting survival in patients receiving continuous flow left ventricular assist devices: The HeartMate II risk score.J Am Coll Cardiol201361313321
8. Holman WL, Kormos RL, Naftel DC, et al.Predictors of death and transplant in patients with a mechanical circulatory support device: A multi-institutional study.J Heart Lung Transplant2009284450
9. Rao V, Oz MC, Flannery MA, Catanese KA, Argenziano M, Naka YRevised screening scale to predict survival after insertion of a left ventricular assist device.J Thorac Cardiovasc Surg2003125855862
10. Morgan JA, Brewer RJ, Nemeh HW, et al.Stroke while on long-term left ventricular assist device support: Incidence, outcome, and predictors.ASAIO J201460284289
11. Baumwol J, Macdonald PS, Keogh AM, et al.Right heart failure and “failure to thrive” after left ventricular assist device: Clinical predictors and outcomes.J Heart Lung Transplant201130888895
12. Cotts WG, McGee EC Jr, Myers SL, et al.Predictors of hospital length of stay after implantation of a left ventricular assist device: An analysis of the INTERMACS registry.J Heart Lung Transplant201433682688
13. Soleimani B, Haouzi A, Manoskey A, Stephenson ER, El-Banayosy A, Pae WEDevelopment of aortic insufficiency in patients supported with continuous flow left ventricular assist devices.ASAIO J201258326329
14. Karnad A, Poskitt TRThe automated complete blood cell count. Use of the red blood cell volume distribution width and mean platelet volume in evaluating anemia and thrombocytopenia.Arch Intern Med198514512701272
15. Felker GM, Allen LA, Pocock SJ, et al.CHARM InvestigatorsRed cell distribution width as a novel prognostic marker in heart failure: Data from the CHARM Program and the Duke Databank.J Am Coll Cardiol2007504047
16. He W, Jia J, Chen J, et al.Comparison of prognostic value of red cell distribution width and NT-proBNP for short-term clinical outcomes in acute heart failure patients.Int Heart J2014555864
17. Núñez J, Núñez E, Rizopoulos D, et al.Red blood cell distribution width is longitudinally associated with mortality and anemia in heart failure patients.Circ J201478410418
18. Huang YL, Hu ZD, Liu SJ, et al.Prognostic value of red blood cell distribution width for patients with heart failure: A systematic review and meta-analysis of cohort studies.PLoS One20149e104861
19. Shao Q, Li L, Li G, Liu TPrognostic value of red blood cell distribution width in heart failure patients: A meta-analysis.Int J Cardiol2015179495499
20. Förhécz Z, Gombos T, Borgulya G, Pozsonyi Z, Prohászka Z, Jánoskuti LRed cell distribution width in heart failure: Prediction of clinical events and relationship with markers of ineffective erythropoiesis, inflammation, renal function, and nutritional state.Am Heart J2009158659666
21. Lippi G, Targher G, Montagnana M, Salvagno GL, Zoppini G, Guidi GCRelation between red blood cell distribution width and inflammatory biomarkers in a large cohort of unselected outpatients.Arch Pathol Lab Med2009133628632
22. Tonelli M, Sacks F, Arnold M, Moye L, Davis B, Pfeffer Mfor the Cholesterol and Recurrent Events (CARE) Trial InvestigatorsRelation between red blood cell distribution width and cardiovascular event rate in people with coronary disease.Circulation2008117163168
23. Mozos IMechanisms linking red blood cell disorders and cardiovascular disease.Biomed Res Int20152015682054
24. Fried J, Levin AP, Mody KM, et al.Prior hematologic conditions carry a high morbidity and mortality in patients supported with continuous-flow left ventricular assist devices.J Heart Lung Transplant20143311191125
25. Slaughter MSHematologic effects of continuous flow left ventricular assist devices.J Cardiovasc Transl Res20103618624
26. Boyle AJ, Jorde UP, Sun B, et al.HeartMate II Clinical InvestigatorsPre-operative risk factors of bleeding and stroke during left ventricular assist device support: An analysis of more than 900 HeartMate II outpatients.J Am Coll Cardiol201411880888
27. Kirklin JK, Naftel DC, Kormos RL, et al.Interagency Registry for Mechanically Assisted Circulatory Support (INTERMACS) analysis of pump thrombosis in the HeartMate II left ventricular assist device.J Heart Lung Transplant2014331222
28. Schaffer JM, Allen JG, Weiss ES, et al.Infectious complications after pulsatile-flow and continuous-flow left ventricular assist device implantation.J Heart Lung Transplant201130164174
29. Tonelli M, Sacks F, Arnold M, Moye L, Davis B, Pfeffer Mfor the Cholesterol and Recurrent Events (CARE) Trial InvestigatorsRelation between red blood cell distribution width and cardiovascular event rate in people with coronary disease.Circulation2008117163168
30. van Breda F, Emans ME, van der Putten K, et al.Relation between red cell distribution width and fibroblast growth factor 23 cleaving in patients with chronic kidney disease and heart failure.PLoS One201510e0128994
31. Faul C, Amaral AP, Oskouei B, et al.FGF23 induces left ventricular hypertrophy.J Clin Invest201112143934408
32. Mirza MA, Larsson A, Lind L, Larsson TECirculating fibroblast growth factor-23 is associated with vascular dysfunction in the community.Atherosclerosis2009205385390
33. Seiler S, Reichart B, Roth D, Seibert E, Fliser D, Heine GHFGF-23 and future cardiovascular events in patients with chronic kidney disease before initiation of dialysis treatment.Nephrol Dial Transplant20102539833989
34. Isakova T, Xie H, Yang W, et al.Chronic Renal Insufficiency Cohort (CRIC) Study GroupFibroblast growth factor 23 and risks of mortality and end-stage renal disease in patients with chronic kidney disease.JAMA201130524322439
35. Moe SM, Chertow GM, Parfrey PS, et al.Evaluation of Cinacalcet HCl Therapy to Lower Cardiovascular Events (EVOLVE) Trial Investigators*Cinacalcet, fibroblast growth factor-23, and cardiovascular disease in hemodialysis: The Evaluation of Cinacalcet HCl Therapy to Lower Cardiovascular Events (EVOLVE) Trial.Circulation20151322739
36. Uemura Y, Shibata R, Takemoto K, et al.Elevation of red blood cell distribution width during hospitalization predicts mortality in patients with acute decompensated heart failure.J Cardiol201667268273
37. Kirklin JK, Naftel DC, Pagani FD, et al.Seventh INTERMACS annual report: 15,000 patients and counting.J Heart Lung Transplant20153414951504
Keywords:

left ventricular assist device; red blood cell distribution width; mortality; risk prediction

Copyright © 2016 by the ASAIO